# CELEBRATING TWENTY YEARS OF THE BRAZILIAN SYMPOSIUM ON CARDIOVASCULAR PHYSIOLOGY

EDITED BY: Camille M. Balarini and Valdir A. Braga PUBLISHED IN: Frontiers in Physiology

### *Frontiers Copyright Statement*

*© Copyright 2007-2017 Frontiers Media SA. All rights reserved. All content included on this site, such as text, graphics, logos, button icons, images, video/audio clips, downloads, data compilations and software, is the property of or is licensed to Frontiers Media SA ("Frontiers") or its licensees and/or subcontractors. The copyright in the text of individual articles is the property of their respective authors, subject to a license granted to Frontiers.*

*The compilation of articles constituting this e-book, wherever published, as well as the compilation of all other content on this site, is the exclusive property of Frontiers. For the conditions for downloading and copying of e-books from Frontiers' website, please see the Terms for Website Use. If purchasing Frontiers e-books from other websites or sources, the conditions of the website concerned apply.*

*Images and graphics not forming part of user-contributed materials may not be downloaded or copied without permission.*

*Individual articles may be downloaded and reproduced in accordance with the principles of the CC-BY licence subject to any copyright or other notices. They may not be re-sold as an e-book.*

*As author or other contributor you grant a CC-BY licence to others to reproduce your articles, including any graphics and third-party materials supplied by you, in accordance with the Conditions for Website Use and subject to any copyright notices which you include in connection with your articles and materials.*

> *All copyright, and all rights therein, are protected by national and international copyright laws.*

*The above represents a summary only. For the full conditions see the Conditions for Authors and the Conditions for Website Use.*

ISSN 1664-8714 ISBN 978-2-88945-176-0 DOI 10.3389/978-2-88945-176-0

### About Frontiers

Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.

### Frontiers Journal Series

The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing. All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.

### Dedication to Quality

Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.

Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view. By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.

### What are Frontiers Research Topics?

Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area! Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org

# **CELEBRATING TWENTY YEARS OF THE BRAZILIAN SYMPOSIUM ON CARDIOVASCULAR PHYSIOLOGY**

Topic Editors: **Camille M. Balarini,** Federal University of Paraiba, Brazil **Valdir A. Braga,** Federal University of Paraiba, Brazil

This e-book is dedicated to the celebration of 20 years of the Brazilian Symposium on Cardiovascular Physiology. In 1996 groups from the School of Medicine of Ribeirao Preto, University of Sao Paulo (FMRP-USP) and from the Federal University of Sao Paulo (UNIFESP) joined together to discuss cardiovascular physiology. In subsequent editions of the meeting, the participation of other groups from all over the country has grown and acquired the status of a national symposium. The participants now agree that the symposium should be itinerant and that the chair group is responsible for its organization. In 2016, we proudly reached the 20th edition of the Brazilian Symposium on Cardiovascular Physiology. It is certainly a memorable date and a great opportunity to share the accomplishments of Brazilian groups in the field of cardiovascular physiology.

**Citation:** Balarini, C. M., Braga, V. A., eds. (2017). Celebrating Twenty Years of the Brazilian Symposium on Cardiovascular Physiology. Lausanne: Frontiers Media. doi: 10.3389/978-2-88945-176-0

# Table of Contents

### **1. Introduction**

*06 Editorial: Celebrating Twenty Years of the Brazilian Symposium on Cardiovascular Physiology*

Camille M. Balarini and Valdir A. Braga

### **2. Historical Perspective**

*09 The Origin and Advancement of Cardiovascular Physiology in Brazil: The Contribution of Eduardo Krieger to Research Groups* Elisardo C. Vasquez

### **3. Experimental models**


### **4. Autonomic nervous system**


Phileno Pinge-Filho, Lisete C. Michelini and Marli C. Martins-Pinge

*62 Blockade of Rostral Ventrolateral Medulla (RVLM) Bombesin Receptor Type 1 Decreases Blood Pressure and Sympathetic Activity in Anesthetized Spontaneously Hypertensive Rats*

Izabella S. Pinto, Aline A. Mourão, Elaine F. da Silva, Amanda S. Camargo, Stefanne M. Marques, Karina P. Gomes, James O. Fajemiroye, Angela A. da Silva Reis, Ana C. S. Rebelo, Marcos L. Ferreira-Neto, Daniel A. Rosa, André H. Freiria-Oliveira, Carlos H. Castro, Eduardo Colombari, Diego B. Colugnati and Gustavo R. Pedrino

### *73 Chronic Treatment with Ivabradine Does Not Affect Cardiovascular Autonomic Control in Rats*

Fernanda C. Silva, Franciny A. Paiva, Flávia C. Müller-Ribeiro, Henrique M. A. Caldeira, Marco A. P. Fontes, Rodrigo C. A. de Menezes, Karina R. Casali, Gláucia H. Fortes, Eleonora Tobaldini, Monica Solbiati, Nicola Montano, Valdo J. Dias Da Silva and Deoclécio A. Chianca Jr.

*83 Ovarian Hormone Deprivation Reduces Oxytocin Expression in Paraventricular Nucleus Preautonomic Neurons and Correlates with Baroreflex Impairment in Rats* Vitor U. De Melo, Rayssa R. M. Saldanha, Carla R. Dos Santos, Josiane De Campos Cruz, Vitor A. Lira, Valter J. Santana-Filho and Lisete C. Michelini

### **5. Renal function**

### *92 Increased Blood Pressure Variability Prior to Chronic Kidney Disease Exacerbates Renal Dysfunction in Rats*

Frederico F. C. T. Freitas, Gilberto Araujo, Marcella L. Porto, Flavia P. S. Freitas, Jones B. Graceli, Camille M. Balarini, Elisardo C. Vasquez, Silvana S. Meyrelles and Agata L. Gava

### *104 Dipeptidyl Peptidase IV Inhibition Exerts Renoprotective Effects in Rats with Established Heart Failure*

Daniel F. Arruda-Junior, Flavia L. Martins, Rafael Dariolli, Leonardo Jensen, Ednei L. Antonio, Leonardo dos Santos, Paulo J. F. Tucci and Adriana C. C. Girardi

### **6. Gut microbiota**

### *120 New Insights on the Use of Dietary Polyphenols or Probiotics for the Management of Arterial Hypertension*

José L. de Brito Alves, Vanessa P. de Sousa, Marinaldo P. Cavalcanti Neto, Marciane Magnani, Valdir de Andrade Braga, João H. da Costa-Silva, Carol G. Leandro, Hubert Vidal and Luciano Pirola

### *128 Effects of Kefir on the Cardiac Autonomic Tones and Baroreflex Sensitivity in Spontaneously Hypertensive Rats*

Brunella F. Klippel, Licia B. Duemke, Marcos A. Leal, Andreia G. F. Friques, Eduardo M. Dantas, Rodolfo F. Dalvi, Agata L. Gava, Thiago M. C. Pereira, Tadeu U. Andrade, Silvana S. Meyrelles, Bianca P. Campagnaro and Elisardo C. Vasquez

### **7. Vascular function**

*140 C-Type Natriuretic Peptide Induces Anti-contractile Effect Dependent on Nitric Oxide, Oxidative Stress, and NPR-B Activation in Sepsis*

Laena Pernomian, Alejandro F. Prado, Bruno R. Silva, Aline Azevedo, Lucas C. Pinheiro, José E. Tanus-Santos and Lusiane M. Bendhack

### *153 Neuronal Nitric Oxide Synthase in Vascular Physiology and Diseases* Eduardo D. Costa, Bruno A. Rezende, Steyner F. Cortes and Virginia S. Lemos

*161 Increased Nitric Oxide Bioavailability and Decreased Sympathetic Modulation Are Involved in Vascular Adjustments Induced by Low-Intensity Resistance Training*

Fabrício N. Macedo, Thassio R. R. Mesquita, Vitor U. Melo, Marcelo M. Mota, Tharciano L. T. B. Silva, Michael N. Santana, Larissa R. Oliveira, Robervan V. Santos, Rodrigo Miguel dos Santos, Sandra Lauton-Santos, Marcio R. V. Santos, Andre S. Barreto and Valter J. Santana-Filho

*172 Different Anti-Contractile Function and Nitric Oxide Production of Thoracic and Abdominal Perivascular Adipose Tissues*

Jamaira A. Victorio, Milene T. Fontes, Luciana V. Rossoni and Ana P. Davel

### **8. Cardiac function**

*182 Comparative mRNA and MicroRNA Profiling during Acute Myocardial Infarction Induced by Coronary Occlusion and Ablation Radio-Frequency Currents*

Eduardo T. Santana, Regiane dos Santos Feliciano, Andrey J. Serra, Eduardo Brigidio, Ednei L. Antonio, Paulo J. F. Tucci, Lubov Nathanson, Mariana Morris and José A. Silva Jr.

*196 Low-Level Laser Application in the Early Myocardial Infarction Stage Has No Beneficial Role in Heart Failure*

Martha T. Manchini, Ednei L. Antônio, José Antônio Silva Junior, Paulo de Tarso C. de Carvalho, Regiane Albertini, Fernando C. Pereira, Regiane Feliciano, Jairo Montemor, Stella S. Vieira, Vanessa Grandinetti, Amanda Yoshizaki, Marcio Chaves, Móises P. da Silva, Rafael do Nascimento de Lima, Danilo S. Bocalini, Bruno L. de Melo, Paulo J. F. Tucci and Andrey J. Serra

*204 Exercise Training Attenuates Right Ventricular Remodeling in Rats with Pulmonary Arterial Stenosis*

Brunno Lemes de Melo, Stella S. Vieira, Ednei L. Antônio, Luís F. N. dos Santos, Leslie A. Portes, Regiane S. Feliciano, Helenita A. de Oliveira, José A. Silva Jr., Paulo de Tarso C. de Carvalho, Paulo J. F. Tucci and Andrey J. Serra

*216 Simulated Microgravity and Recovery-Induced Remodeling of the Left and Right Ventricle*

Guohui Zhong, Yuheng Li, Hongxing Li, Weijia Sun, Dengchao Cao, Jianwei Li, Dingsheng Zhao, Jinping Song, Xiaoyan Jin, Hailin Song, Xinxin Yuan, Xiaorui Wu, Qi Li, Qing Xu, Guanghan Kan, Hongqing Cao, Shukuan Ling and Yingxian Li

### **9. New Insights**

### *228 A Disintegrin and Metalloprotease 17 in the Cardiovascular and Central Nervous Systems*

Jiaxi Xu, Snigdha Mukerjee, Cristiane R. A. Silva-Alves, Alynne Carvalho-Galvão, Josiane C. Cruz, Camille M. Balarini, Valdir A. Braga, Eric Lazartigues and Maria S. França-Silva

# Editorial: Celebrating Twenty Years of the Brazilian Symposium on Cardiovascular Physiology

Camille M. Balarini 1, 2 \* and Valdir A. Braga<sup>2</sup>

*<sup>1</sup> Department of Physiology and Pathology, Health Sciences Center, Federal University of Paraiba, Joao Pessoa, Brazil, <sup>2</sup> Biotechnology Center, Federal University of Paraiba, Joao Pessoa, Brazil*

Keywords: cardiovascular physiology, cardiovascular diseases, hypertension, vascular function

**Editorial on the Research Topic**

### **Celebrating Twenty Years of the Brazilian Symposium on Cardiovascular Physiology**

### Edited by:

*Geoffrey A. Head, Baker IDI Heart and Diabetes Institute, Australia*

### Reviewed by:

*Geoffrey A. Head, Baker IDI Heart and Diabetes Institute, Australia Fiona D. McBryde, University of Auckland, New Zealand*

> \*Correspondence: *Camille M. Balarini camille.balarini@gmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *16 February 2017* Accepted: *07 March 2017* Published: *21 March 2017*

### Citation:

*Balarini CM and Braga VA (2017) Editorial: Celebrating Twenty Years of the Brazilian Symposium on Cardiovascular Physiology. Front. Physiol. 8:166. doi: 10.3389/fphys.2017.00166* This research topic is dedicated to the celebration of 20 years of the Brazilian Symposium on Cardiovascular Physiology. In 1996 groups from the School of Medicine of Ribeirao Preto, University of Sao Paulo (FMRP-USP) and from the Federal University of Sao Paulo (UNIFESP) joined together to discuss cardiovascular physiology. In subsequent editions of the meeting, the participation of other groups from all over the country has grown and acquired the status of a national symposium. The participants now agree that the symposium should be itinerant and that the chair group is responsible for its organization. In 2016, we proudly reached the 20th edition of the Brazilian Symposium on Cardiovascular Physiology. It is certainly a memorable date and a great opportunity to share the accomplishments of Brazilian groups in the field of cardiovascular physiology.

The groups devoted to investigate cardiovascular physiology in Brazil descended, in a direct or indirect manner, from the Argentinian physiologist and Nobel Prize winner Bernardo Houssay. One of his disciples, Miguel Covian, was invited to be the chair of the Department of Physiology in the School of Medicine of Ribeirao Preto when it was funded in the 1950's. At the same time, the physicians at Federal University of Rio Grande do Sul started a collaborative work with Bernardo Houssay's and Braun Menendez's groups. That historical perspective is presented in this issue in the article by Vasquez. In his paper, he also shares the contributions of Professor Eduardo M. Krieger to the development of cardiovascular physiology in Brazil.

The etiology of cardiovascular and metabolic diseases reveals the involvement of different genetic, environmental, nutritional and behavioral aspects. In this issue, Costa-Silva et al. discuss the role of maternal diet on the development of cardiometabolic diseases. The authors point out that epigenetic alterations can be, at least in part, responsible for the increased risk of developing cardiometabolic problems. The authors discuss in detail the manner in which maternal protein undernutrition or overnutrition during the perinatal period can increase the risk of cardiovascular and metabolic diseases.

It is important to highlight that the success of the experimental design for investigating cardiovascular physiology depends on the availability of suitable experimental models. In this context, Crestani reviews the effects of acute and chronic emotional stress in cardiovascular function. The author focused on the cardiovascular responses observed in different animal models of emotional stress. Considering the impact of stress in the cardiovascular system in humans, this is certainly a promising area of research.

The influence of the autonomic nervous system in cardiovascular function is remarkable. Thus, different research groups in Brazil are devoted to investigating such influence in health and disease. In the present issue, Accorsi-Mendonça et al. present a historical retrospective on the characteristics of rostral ventrolateral medulla (RVLM) presympathetic neurons and discuss the concept that those cells work as pacemakers for the generation of the sympathetic activity. Interestingly, physical activity can modulate central areas involved in autonomic control of the cardiovascular system. In this context, Raquel et al. show that swimming modulates nitric oxide (NO) availability and glutamatergic neurotransmission in the RVLM, contributing to a decrease in sympathetic activity and an increase in baroreflex control of blood pressure (BP). This provides insight and support to the idea that physical activity should be included when treating hypertensive patients.

Pharmacological modulation of the autonomic nervous system is an important target to treat cardiovascular diseases (CVD). Alternative experimental approaches in this field include the search for new natural and synthetic compounds. In this regard, Pinto et al. observed that bombesin, a peptide isolated from frog skin, increases blood pressure and renal sympathetic nerve activity (RSNA) when administered into the RVLM of normotensive and spontaneously hypertensive rats (SHR). In addition, Silva et al. show that ivabradine reduced resting heart rate and blood pressure, with no effects on cardiovascular reflexes or RSNA.

Other important group of neurons involved in autonomic cardiovascular control are present in the paraventricular nucleus of the hypothalamus (PVN). This region possesses reciprocal communications with the nucleus tractus solitarii (NTS) and synapses with both the RVLM itself and the subfornical organ (SFO), which receives information from directly from the blood stream since it lacks a blood brain barrier. In this context, De Melo et al. hypothesize that the decrease in ovarian hormones during menopause blunts oxytocin expression and signaling in pre-autonomic PVN neurons, leading to baroreflex impairment, autonomic imbalance and arterial hypertension. This reinforces clinical findings that women are more prone to develop CVD after menopause than men of the same age and provides experimental data to further support hormonal reposition therapies.

Considering that an increase in blood pressure variability (BPV) is an indicative of poor prognosis in cardiovascular outcomes, Freitas et al. demonstrate that increased BPV prior to the onset of chronic kidney disease can reduce renal blood flow, increase renal vascular resistance and increase uraemia and glomerulosclerosis, exacerbating renal dysfunction. In this scenario, the authors suggest that increased BPV may be considered as a marker for target-organ damage.

Recently, the gut microbiota has gained attention as it can be involved in the onset of diverse pathological states, including the development of CVD. In this context, de Brito-Alves et al. show evidence, from both experimental and clinical approaches, that the use of polyphenols and probiotics reduces blood pressure and improves cardiovascular function. In agreement with this, Klippel et al. present an experimental study in which the administration of kefir (a probiotic composed by different bacteria such as Lactobacillus kefiranofaciens, Lactobacillus kefir, and Candida kefir) to SHR resulted in the amelioration of vagal and sympathetic imbalance, improvement of baroreflex sensitivity and a reduction in blood pressure. Taken together, these reports emphasize the potential of probiotics as adjuvants on CVD treatment.

Despite the fundamental role played by the central nervous system (CNS) in cardiovascular control, the peripheral control of blood flow to target organs in response to specific organ demands presents an important role in BP homeostasis. This is particularly evident during sepsis and septic shock, when a massive vasodilation causes a remarkable reduction in blood pressure, with a high mortality rate. During this state, there is also a vascular hyporesponsivity to vasoconstrictors, limiting therapeutic options. Thus, Pernomian et al. used an experimental model of sepsis to evaluate the participation of C-type natriuretic peptide on this response. The authors suggest that this peptide is involved in hyporeactivity to vasoconstrictors in aorta, revealing a novel potential target for septic shock.

In the present issue, Costa et al. provide a relevant review on the vascular effects of neuronal nitric oxide synthase (nNOS). It is important to highlight that, although endothelial NOS (eNOS) is considered the main NOS isoform in vessels, nNOS is an important source of both NO and hydrogen peroxide (H2O2), two endothelium-dependent vasodilators. Imbalances in nNOS expression and/or activity have been described in arterial hypertension and atherosclerosis, corroborating the idea that this isoform is particularly relevant for proper vessel function. Nitric oxide signaling can be modulated by physical activity not only in the CNS as previously discussed but also in the periphery. In this context, Macedo et al. evaluated the participation of NO in vascular tone adjustments in response to low-intensity resistance training. They observed increased expression of eNOS and nNOS, culminating in an improvement of vascular function, increase in baroreflex sensitivity and reduction in blood pressure.

Additionally, a further guardian of vascular homeostasis is perivascular adipose tissue (PVAT), which can secrete vasoactive substances. In this context, Victorio et al. comparatively evaluated the effects of PVAT from abdominal and thoracic aorta. The authors observed functional regional differences along the aorta, with a greater production of vasodilators in thoracic vs. abdominal PVAT.

Despite the efforts of diverse research groups in Brazil and abroad, CVD still is the leading cause of death worldwide. The efficient function of the cardiovascular system depends to a large degree on good cardiac function, which may be impaired after myocardial infarction (MI), for example. In this issue, Santana et al. compare two different techniques to induce MI in rats. Patterns of gene expression were seen to differ between the two methods. This study certainly contributes to the standardization of suitable experimental models that will allow new approaches to prevent and treat MI. For instance, Manchini et al. used lowlevel laser therapy to improve left ventricular systolic function. Although the results are promising, the beneficial effects seem to be transient and long-term studies are needed. Heart failure is commonly observed after MI and this compromises not only the function of the heart per se but also leads to damage in target organs. It was observed by Arruda-Junior et al. that rats with heart failure presented impaired renal function (fluid retention, reduction in glomerular filtration rate, increase in protein excretion), which was preserved by the inhibition of dipeptidyl peptidase IV. de Melo et al. provide us with another example of beneficial outcomes from physical exercise. The authors evaluated the effects of exercise training in a model of right ventricular remodeling and found that it attenuated myocardial remodeling and improved right ventricular function.

Different strategies to deal with CVD often come from the discovery of new physiological pathways. In this context, Xu et al. reviewed the role played by ADAM17 (A disintegrin A metalloprotease 17) in the cardiovascular system as well as in CNS areas involved in cardiovascular control. Although it is a very promising target molecule, the diversity of its substrates (this enzyme is involved with more than 70 different substrates) including inflammatory substances, have slowed the progression of translational studies.

As highlighted here, based on the quality of the different studies presented, the initiative of creating the Brazilian Symposium on Cardiovascular Physiology 20 years ago has contributed enormously to the development of research on cardiovascular physiology in Brazil. The country now holds more than 20 different research groups established from the north to the south of the country producing worldwide recognized science. Most laboratories are well-equipped with cutting-edge technology allowing in-deep investigations into cardiovascular phenomena. The main limitations currently faced by Brazilian researchers are the strict rules governing the importation of research goods (such as chemicals, reagents, live animals, viruses for gene transfer and general lab equipment) and the uneven distribution of research funding across the country. In conclusion, we would like to thank very much all the Brazilian research groups who attended to the 20th Brazilian Symposium on Cardiovascular Physiology and the reviewers who generously agreed to review the manuscripts presented in this issue.

## AUTHOR CONTRIBUTIONS

CB and VB participated in all stages during the elaboration of this manuscript. Authors read and approved the final version.

### ACKNOWLEDGMENTS

Authors thank reviewers who generously agreed to review the manuscripts presented in this research topic.

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Balarini and Braga. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# The Origin and Advancement of Cardiovascular Physiology in Brazil: The Contribution of Eduardo Krieger to Research Groups

### Elisardo C. Vasquez 1, 2 \*

*<sup>1</sup> Laboratory of Translational Physiology, Federal University of Espírito Santo, Vitória, Brazil, <sup>2</sup> Pharmaceutical Sciences Graduate Program, Vila Velha University, Vila Velha, Brazil*

Since 1996, symposia devoted to the discussion of advances in cardiovascular physiology have been alternately organized by Brazilian research groups, most of which were created or joined by Ph.D. trainees of Eduardo M Krieger. Therefore, as *Frontiers in Physiology* is publishing a topic devoted to the celebration of the 20th edition of the Brazilian Symposium of Cardiovascular Physiology, it is a great opportunity to talk about the contributions of Eduardo Krieger to the development of cardiovascular physiology. In this historical mini-review, first, the influence of the Argentinian group of Bernardo Houssay and Braun Menéndez on cardiovascular physiology in Brazil is discussed. Second, the contribution of Eduardo Krieger to the creation of several of those groups and to the development of science and technology is reviewed. Finally, the origin and consolidation of the group of Vitoria is highlighted as an example of a research group that was influenced by the University of Sao Paulo-Faculty of Medicine of Ribeirao Preto and has trained hundreds of Master and Ph.D. students in the area of cardiovascular research.

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Luciana Venturini Rossoni, University of São Paulo, Brazil Mark Chapleau, University of Iowa, USA Luis Carlos Reis, Federal Rural University of Rio de Janeiro, Brazil*

### \*Correspondence:

*Elisardo C. Vasquez evasquez@pq.cnpq.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *04 March 2016* Accepted: *28 March 2016* Published: *19 April 2016*

### Citation:

*Vasquez EC (2016) The Origin and Advancement of Cardiovascular Physiology in Brazil: The Contribution of Eduardo Krieger to Research Groups. Front. Physiol. 7:135. doi: 10.3389/fphys.2016.00135*

Keywords: Eduardo Krieger, Braun Menéndez, Bernardo Houssay, cardiovascular physiology, angiotensin

## INTRODUCTION

The scientific career of Eduardo Krieger has been described through interviews that have been conducted in Portuguese and translated to English (Moura and Zorzetto, 2012) and was recently commented on in a book about the career of his master Braun Menéndez (Krieger, 2015). Now, thanks to the editors of this special topic devoted to celebrate the 20th Brazilian Symposium of Cardiovascular Physiology, there is an opportunity to demonstrate in Frontiers in Physiology our gratitude to those who have immensely contributed to the creation and consolidation of research groups in cardiovascular physiology in Brazil.

This mini-review encompasses the point of view of a senior investigator in cardiovascular physiology who received his Ph.D. research training from Eduardo Krieger (**Figures 1e,f**) at the University of Sao Paulo-Faculty of Medicine of Ribeirao Preto (FMRP-USP), focused on the contribution of his master and the FMRP-USP to the creation of various research groups in cardiovascular physiology. Moreover, our knowledge of the beginning of cardiovascular physiology research in Brazil comes from the chair of the Department of Physiology of the FMRP-USP, Miguel R. Covian, who created a group of eight Ph.D. students interested in philosophy and to talk about science and scientists, including the fascinating history of Bernardo Houssay's leadership. This mini-review is consistent with a philosophical phrase I use, "in our life and science, as much as we know where we came from, as much as we will be able to decide where to go."

## ARGENTINIAN MASTERS WHO CONTRIBUTED TO THE DEVELOPMENT OF CARDIOVASCULAR PHYSIOLOGY IN BRAZIL

The beginning of cardiovascular physiology research in Brazil occurred in the middle of the past century, with the convergence of several factors. In 1947, Bernardo Houssay, the leader of a group of physiologists at the Faculty of Medicine of the University of Buenos Aires, was awarded the Nobel Prize in Physiology or Medicine. His disciples included important investigators such as the physiologist Braun Menéndez, who was one of the discoverers of angiotensin in 1940 and the cardiologist Alberto Taquini (**Figures 1a–d**). In 1951, two important public agencies in Brazil were created: The National Council for the Development of Science and Technology (CNPq), with the mission of supporting science and technology, and the Coordination for the Improvement of Higher Education (CAPES), with the mission of supporting science at universities.

In 1953, the physician Rubens Maciel, working at the School of Medicine of the Federal University of Rio Grande do Sul (UFRGS) in the city of Porto Alegre, acquired financial support from CAPES to establish an agreement between the School of Medicine and the Institute of Biology and Medicine Research recently created by Bernardo Houssay's group, with the goal of collaboratively training new Brazilian physiologists. Additionally, in 1953, the FMRP-USP was created and 2 years later Miguel R. Covian, from Bernardo Houssay's group, was hired to chair the Department of Physiology at that institution.

## THE MASTERS OF EDUARDO KRIEGER

In 1953, Eduardo M. Krieger, native of the State of Rio Grande do Sul, who during the last term of his medical course in Porto

FIGURE 1 | The Argentinian group of cardiovascular physiology (1940). Taquini (a), Braun Menéndez, co-discoverer of angiotensin (b), Houssay (c), and Leloir (d), the two Nobel Prize laureates. The master of Brazilian cardiovascular physiology and hypertension, Eduardo Krieger (e), and his disciple Elisardo Vasquez (f) (2014).

Alegre decided to become a cardiologist and was working at the School of Medicine with the cardiologist Rubens Maciel, met the famous Braun Menéndez and became fascinated by his enthusiasm for science and his facility for problem solving. Reciprocally, Rubens Maciel and Braun Menéndez noted the potential and scientific ability of the young Eduardo Krieger and proposed him to work on the renin-angiotensin system at the Braun Menéndez and Bernardo Houssay laboratory (1955), which was located in a large house belonging to the Braun Menéndez family. The group had left the university in protest against the military dictatorship. Eduardo Krieger described that period of his work in the laboratory of Braun Menéndez as a fascinating opportunity to interact with other famous scientists, such as Luis Leloir, who had a lab in a smaller house close to the Menéndez house and who was awarded the Nobel Prize in 1970 for his discoveries in chemistry (**Figures 1a–d**). In Porto Alegre, Eduardo Krieger was dedicated to work in cardiovascular physiology and experimental hypertension in collaboration with Rubens Maciel and Braun Menéndez. Then, having been recommended by Bernardo Houssay and with a grant from the Rockefeller Foundation, in 1956, he decided to pursue scientific training in the USA. There, he worked with William Hamilton at the University of Georgia, where he also had the opportunity to meet Raymond Ahlquist, who discovered the α- and β-adrenoceptors. While there and with plans to return to Porto Alegre to work in cardiology, Krieger was invited by the head of the Dept. Physiology of the FMRP-USP, Miguel Covian, to assume the position of professor and to create a cardiovascular physiology division.

## EDUARDO KRIEGER'S CAREER AT THE FMRP-USP

Eduardo Krieger, now 88 years old, initiated his career as professor of cardiovascular physiology in 1957 at the FMRP-USP upon graduating from the first medical class at the institution. He has noted that he found a scientific atmosphere at the FMRP-USP that resembled what he had observed at Bernardo Houssay's and Braun Menéndez's labs. The FMRP-USP had excellent research laboratories, substantial financial support from the Rockefeller Foundation for new investigators and houses on-campus for the residence of full-time investigators. All of these favorable conditions were designed to develop research and were attractive not only to Eduardo Krieger but also to other investigators who had been invited to fill open positions in the departments of biochemistry, physiology, and pharmacology at FMRP-USP. Two years later, Krieger planned a visit by his mentor, Braun Menéndez, to develop collaborative research projects, but it was canceled due to the premature and unexpected death of the still young Braun Menéndez.

## THE CARDIOVASCULAR DISCOVERIES OF EDUARDO KRIEGER

A couple of years after arriving at the FMRP-USP, while waiting for the equipment he had applied for, Krieger decided to study the mechanisms by which the peripheral and central nervous systems regulate blood pressure (BP) in conditions of hypothermia. In that study, he stimulated the nerves of the cervical region in rats and casually observed that the stimulation of the central part of the vagus nerve caused both increases and decreases in BP. Then, under a higher magnification microscope, he noted that the vagus nerve was not a single nerve but included a second nerve. He observed that the stimulation of one nerve caused an increase in BP and that the stimulation of the other nerve caused a decrease in BP. After that observation, Krieger conducted systematic experiments to determine how those nerves controlled BP and concluded that it was possible to isolate the sympathetic nerve from the vagus nerve in the rat species and that the sympathetic nerve contained aortic baroreceptor fibers. Those studies led him to work on another project, which resulted in the creation of the sino-aortic denervation (SAD) model, which became known and used worldwide after its publication in Circulation Research in 1964 (Krieger, 1964); the article has accumulated more than 600 citations. Already training postgraduate students, Eduardo Krieger subsequently focused his studies on the baroreflex control of BP and the conditions of sustained hypertension. He and his Ph.D. students demonstrated that the baroreceptors are reset to operate at higher BP levels in hypertension and that a complete resetting occurs when the increase in the pressure threshold equals the increase in BP, which is observed after 48 h of sustained hypertension in rats (Krieger, 1989).

In 1985, after he had retired from the FMRP-USP, Krieger accepted an invitation to be a full professor and investigator at the Heart Institute (InCor) of the USP Hospital with the mission of developing integrative investigations in cardiovascular physiology and in experimental and clinical hypertension, transferring the knowledge from the bench to the bedside.

## THE CONTRIBUTION OF EDUARDO KRIEGER TO SCIENTIFIC ORGANIZATIONS

The career of the tireless Eduardo Krieger was not restricted to publishing papers and supervising new investigators. He was one of three editors who, in 1981, created the important internationalized Brazilian Journal of Medical and Biological Research. After being president of the Brazilian Society of Physiology (1979–1985), he was the primary contributor to the creation of the Federation of Experimental Societies of Biology (FeSBE) and its first president (1985–1991). Many famous investigators working in cardiovascular physiology attended FeSBE meetings during Professor Krieger's tenure as President. Furthermore, the organization provided hundreds of opportunities for young physiologists to train at prominent universities around the world. Krieger was one of the creators and presidents of the Brazilian Society of Hypertension (1992–1994) and the Inter-American Society of Hypertension. He served as chairman of the Brazilian Academy of Sciences (1992–2007), an organization linking science to society and government; and has served as vice-chairman of the Foundation for the Development of Science of Sao Paulo (Fapesp) since 2010. Thus, he continues to play a major role in the development of science and technology throughout Brazil.

## PH.D. STUDENTS TRAINED IN CARDIOVASCULAR PHYSIOLOGY BY EDUARDO KRIEGER

By the 1970s, when CAPES created the national post-graduation system, the universities were undergoing an expansion and many recently graduated individuals and hired professors were waiting for an opportunity to begin a Ph.D. degree. Unfortunately, at that time few labs were able to offer Ph.D. training in cardiovascular research. The laboratory of Eduardo Krieger was one of the most sought after and active laboratories and within a few years he had trained several Ph.D. students. Some of the graduates were hired by the Departments of Physiology and Pharmacology of the FMRP-USP and created new labs and new possibilities for training Ph.D. students in cardiovascular physiology, in collaboration with Krieger's lab. Many of the Ph.D. graduates returned to their previous positions in other universities or were rapidly hired and joined or created new cardiovascular research groups at various places. Overall, Eduardo Krieger trained more than 30 Ph.D. students and one-third of them became full professors who have been internationally recognized as important investigators and who have created or joined other groups of cardiovascular physiology throughout Brazil.

Considering the students who were supervised by Krieger at the FMRP-USP and those supervised by his disciples, more than 200 Ph.D.'s were trained by the group of Eduardo Krieger in the investigation of cardiovascular physiology. This number is markedly greater if one considers the period after his retirement, when he worked as the head of the Hypertension Unit at InCor-USP, where he trained Ph.D. students in basic science areas and clinicians to work in new fields of research, such as BP regulation during sleep and exercise and the recording of sympathetic nerve activity in physiological and clinical conditions.

## THE CONTRIBUTION OF THE FMRP-USP TO THE CREATION OF THE VITORIA GROUP: THE ORIGIN OF AN IMPORTANT RESEARCH GROUP IN CARDIOVASCULAR PHYSIOLOGY

The success of the cardiovascular physiology group at the Federal University of Espirito Santo (Ufes) in Vitoria could be considered unexpected. In 1981, Ufes had only one post-graduation program and did not have a department of physiology or physiological sciences; I brought the first Wistar rats to the department for research from the FMRP-USP. On the other hand, the group became feasible because there was a resident and very active physiologist, Dalton Vassallo, who welcomed Ph.D. trainees in biochemistry (one), physiology (myself), and pharmacology (four) from the FMRP-USP and one who had a Ph.D. degree from the University of Birmingham in the UK. Additionally, the new members of the group started sharing space, material, equipment, and technicians and working together to understand the control of cardiovascular function using the rat model of SAD and rat models of hypertension. In 1989, the group of eight researchers was consolidated and was authorized by CAPES to create a Master's degree course, which was inaugurated with a Lecture from Eduardo Krieger.

At that time, my master suggested that I should go to the USA to work with Michael J. Brody at the University of Iowa, who gave me the position of visiting associate professor (1989– 1991). I opted to work on the central control of BP and the rostral ventrolateral medulla (Vasquez et al., 1992). However, similar to Braun Menéndez, Brody had an abrupt death and we did not have the opportunity to plan a collaborative work with him at Ufes.

In 1991, when attending The Council for High Blood Pressure Research of the American Heart Association, in Baltimore, Eduardo Krieger told me that the Vitoria group had reached maturity and that we should attempt to create the first postgraduate program for Ph.D.'s at Ufes. We did so in 1993, and in a couple of years, that program achieved the maximum score (A) from CAPES.

In addition to his direct contribution to the creation of the Vitoria group, Krieger also had important influences on the training of members of this group by many famous cardiovascular physiologists. We highlight the collaboration between the Vitoria group and the group of Michael Brody, François Abboud, Mark Chapleau, and Alan Kim Johnson in Iowa City on two occasions; the group of Kurt Varner, Daniel Kapusta, and Louis Barker at Louisiana State University; the group of Michel Safar at Paris Descartes University; the group of Bernard Fleury at Hôpital Saint-Antoine in Paris; and the group of Virend Somers at the Mayo Clinic College of Medicine, all of which had been planned when I was working in Iowa City in two occasions (1989–1991 and 1998–2000). Other important and long-lasting scientific influences came from Michael Spyer at The University College of London Medical School in collaboration with Henrique Futuro-Neto, from Mercedes Salaices at the Universidad Autonoma de Madrid in collaboration with Dalton Vassallo and from José Krieger at InCor-USP in collaboration with José Geraldo Mill.

### REFERENCES


Interestingly, the collaborative work of Silvana Meyrelles in Chapleau's lab and of myself in Kim Johnson's lab (1998– 2000) enabled us to improve our studies with addition of molecular biology techniques, gene therapy, and knockout mouse models at Ufes. Thus, we began translational research using a combination of in vivo and in vitro approaches, which culminated with the creation of the Lab of Translational Physiology, which has substantially contributed to the progress of our group. In the past 20 years, this laboratory has focused on the characterization of cardiovascular dysfunction in mouse models of atherosclerosis and hypertension, the identification of new therapies, and decreasing the time that a scientific discovery takes to reach clinical practice (Vasquez et al., 2012), as has been proposed by the societies of physiology (Seals, 2013). To date, the Vitoria group has provided outstanding training to 275 Masters and 130 Ph.D. students in cardiovascular physiology. Those Ph.D.'s have successfully filled positions in academia, either joining established groups or creating new groups.

### CONCLUSION

Eduardo Krieger, who was trained by Braun Menéndez, Houssay, and Hamilton, has contributed enormously to the advancement of science through the creation and leadership of scientific organizations and molding of young investigators to work in cardiovascular physiology. He was an exemplary master and taught us that as investigators, we should maximize our research efforts to contribute to better life conditions for the population.

## AUTHOR CONTRIBUTIONS

EV contributed with the conception, review design, and contents of the manuscript.

### ACKNOWLEDGMENTS

EV is supported by CNPq (Ref. 476525/2012-8) and Fapes (Grant Universal 2014 Proc 67597482).


**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Vasquez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Developmental Origins of Cardiometabolic Diseases: Role of the Maternal Diet

### João H. Costa-Silva\*, Aiany C. Simões-Alves and Mariana P. Fernandes

*Departamento de Educação Física e Ciências do Esporte, Centro Acadêmico de Vitória, Universidade Federal de Pernambuco, Pernambuco, Brazil*

Developmental origins of cardiometabolic diseases have been related to maternal nutritional conditions. In this context, the rising incidence of arterial hypertension, diabetes type II, and dyslipidemia has been attributed to genetic programming. Besides, environmental conditions during perinatal development such as maternal undernutrition or overnutrition can program changes in the integration among physiological systems leading to cardiometabolic diseases. This phenomenon can be understood in the context of the phenotypic plasticity and refers to the adjustment of a phenotype in response to environmental input without genetic change, following a novel, or unusual input during development. Experimental studies indicate that fetal exposure to an adverse maternal environment may alter the morphology and physiology that contribute to the development of cardiometabolic diseases. It has been shown that both maternal protein restriction and overnutrition alter the central and peripheral control of arterial pressure and metabolism. This review will address the new concepts on the maternal diet induced-cardiometabolic diseases that include the potential role of the perinatal malnutrition.

### Edited by:

*Camille M. Balarini, Federal University of Paraíba, Brazil*

### Reviewed by:

*James Todd Pearson, National Cerebral and Cardiovascular Center, Japan Ana Paula Davel, State University of Campinas, Brazil*

> \*Correspondence: *João H. Costa-Silva joao.hcsilva@ufpe.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *21 August 2016* Accepted: *14 October 2016* Published: *16 November 2016*

### Citation:

*Costa-Silva JH, Simões-Alves AC and Fernandes MP (2016) Developmental Origins of Cardiometabolic Diseases: Role of the Maternal Diet. Front. Physiol. 7:504. doi: 10.3389/fphys.2016.00504* Keywords: developmental plasticity, perinatal nutrition, cardiometabolic control, protein restriction

## INTRODUCTION

Cardiovascular and metabolic diseases, such as hypertension, type II diabetes, and dyslipidemia are highly prevalent in the world and have important effects on the public health, increasing risk factors for the development of other diseases, including coronary heart disease, stroke, and heart failure (Landsberg et al., 2013). The etiology of these cardiometabolic diseases includes a complex phenotype that arises from numerous genetic, environmental, nutritional, behavioral, and ethnic origins (Landsberg et al., 2013; Ng et al., 2014). In this regard, it has been observed that the eating habits and behaviors and nutritional condition in early phases of life may play a key role on the etiology of these diseases by inducing physiological dysfunctions (Lucas, 1998; Victora et al., 2008; Wells, 2012). This phenomenon can be understood in the context of phenotypic plasticity and it refers to the ability of an organism to react to both an internal and external environmental inputs with a change in the form, state, physiology, or rate of activity without genetic changes (West-Eberhard, 2005b). Indeed the nutritional factors rise as important element in this theme and it has been highlighted since Barker (Barker, 1990, 1994, 1995, 1998, 1999a,b, 2000; Barker and Martyn, 1992; Fall and Barker, 1997; Osmond and Barker, 2000). In this context, new evidence from epidemiological and clinical studies have showed the association of the maternal under- and overnutrition with development of cardiometabolic dysfuntions (Ashton, 2000; Hemachandra et al., 2006; Antony and Laxmaiah, 2008; Conde and Monteiro, 2014; Costa-Silva et al., 2015; Parra et al., 2015). Thus, this review will address the new concepts about the involvement of the maternal protein malnutrition and overnutrition on the development of the cardiometabolic diseases.

## PERINATAL ORIGIN OF CARDIOMETABOLIC DISEASES: THE ROLE OF PHENOTYPIC PLASTICITY

Biological and medical consequences of perinatal nutritional factors have been extensively studied in the field of the "developmental origins of health and diseases" proposed by Barker and colleagues since 1986 (Barker and Osmond, 1986; Barker et al., 1989, 1993; Barker, 2007). This field of research proposes that cardiometabolic diseases can be "programmed" by the "adaptative" effects of both under- and overnutrition during early phases of growth and development on the cell physiology (Barker and Osmond, 1986; Hales and Barker, 1992; Alfaradhi and Ozanne, 2011; Chavatte-Palmer et al., 2016). As stated before, it aims to study how an organism reacts to a different environmental input, such as malnutrition, and induces changes in the phenotype, but without altering the genotype (Barker et al., 2005; West-Eberhard, 2005a; Labayen et al., 2006; Andersen et al., 2009; Biosca et al., 2011). In this context, epigenetic alterations, such as DNA methylation, histone acetylation, and microRNA expression are considered the molecular basis of the phenotypic plasticity (Wells, 2011). These modifications termed as "epigenetic" were firstly described by Conrad Waddigton in 1940 and it studies the relationship between cause and effect in the genes to produce a phenotype (Jablonka and Lamb, 2002). Nowadays, this concept is employed to describe the process of the gene expression and its linking to modifications in the cromatin structure without altering DNA sequence (Chong and Whitelaw, 2004; Egger et al., 2004). Among all epigenetic modifications, the DNA methylation is one that has been best studied and is related to addition of methyl groups on DNA cytosine residues, normally on the cytosine followed by guanine residue (CpG dinucleotides), which can produce inhibition of the gene expression by impairing transcriptional factor binding (Waterland and Michels, 2007; Mansego et al., 2013; Chango and Pogribny, 2015; Mitchell et al., 2016). In this context, it has been investigated how nutritional aspect may induce these epigenetic modifications.

Macro- and micro-nutrient compositions have been identified as important nutritional factors inducing epigenetic processes, such as DNA methylation (Mazzio and Soliman, 2014; Szarc vel Szic et al., 2015). It is considered at least three ways by which nutrients can induce DNA methylation, alter gene expression, and modify cellular phenotype: (i) by providing methyl group supply for inducing S- adenosyl-L-methionine formation (genomic DNA methylation), modifying the methyltransferase activity, or impairing DNA demethylation process; (ii) by modifying chromatin remodeling, or lysine and arginine residues in the N-terminal histone tails; and (iii) by altering microRNA expression (Chong and Whitelaw, 2004; Egger et al., 2004; Hardy and Tollefsbol, 2011; Stone et al., 2011). In this context, altered contents of amino acids, such as methionine and cysteine, as well as reduced choline and folate diet amount can modify the process of the DNA methylation leading to both DNA hyper- and hypomethylation (Fiorito et al., 2014). For example, deficiency of choline can precipitate DNA hypermethylation associated with organ dysfunction, mainly in liver metabolism (Karlic and Varga, 2011; Wei, 2013).

High fat diet (HFD) during perinatal period has been identified as risk factor to predispose and induce epigenetic processes in the parents and their offspring (Mazzio and Soliman, 2014; Szarc vel Szic et al., 2015). Both hypo- and hypermethylation processes participate in this dysregulation attributed to HFD consumption (Ng et al., 2010; Milagro et al., 2013). In adipose tissue, for example, it was observed that gene promoter of the fatty acid synthase enzyme suffered methylation (Lomba et al., 2010) and that important obesity-related genes such as leptin have disruption on their methylation status (Milagro et al., 2009).

### MATERNAL PROTEIN UNDERNUTRITION: EARLY- AND LONG-TERM OUTCOMES

Maternal malnutrition is associated with the risk of developing cardiovascular disease and co-morbidities in offspring's later life including hypertension, metabolic syndrome, and type-II diabetes (Barker et al., 2007; Nuyt, 2008; Nuyt and Alexander, 2009). In humans, studies have provided support for the positive association between low birth weight and increased incidence of hypertension (Ravelli et al., 1976; Hales et al., 1991; Sawaya and Roberts, 2003; Sawaya et al., 2004).

Maternal low-protein diet model during both gestation and lactation is one of the most extensively studied animal models of phenotypic plasticity (Ozanne and Hales, 2004; Costa-Silva et al., 2009; Falcão-Tebas et al., 2012; Fidalgo et al., 2013; de Brito Alves et al., 2014; Barros et al., 2015). Feeding a low-protein diet (8% protein) during gestation and lactation is associated with growth restriction, asymmetric reduction in organ growth, elevated systolic blood pressure, dyslipidemia, and increased fasting plasma insulin concentrations in the most of studies in rodents (Ozanne and Hales, 2004; Costa-Silva et al., 2009; Falcão-Tebas et al., 2012; Fidalgo et al., 2013; Leandro et al., 2012; de Brito Alves et al., 2014, 2016; Ferreira et al., 2015; Paulino-Silva and Costa-Silva, 2016). However, it is known that the magnitude of the cardiovascular and metabolic outcomes are dependent on the both time exposure to protein restricted-diet (Zohdi et al., 2012, 2015) and growth trajectory throughout the postnatal period (Wells, 2007, 2011). A rapid and increased catch-up

**Abbreviations:** AKT/PKB, Protein kinase B; CB, Carotid body; CNS, Central nervous system; CRP, C-reactive protein; ERK, Extracellular signal-regulated kinase; GSH, Glutathione reduced; HFD, High fat diet; HIF-1α, Hypoxic inducible factor 1 alpha; IGF2, Insulin-like growth factor 2; IL-6, Interleukin-6; IR, Insulin receptor; IRS, Insulin receptor substrate; mTOR, Mammalian target of rapamycin; PI3K, Phosphatidylinositol 3-kinase; RAS, Renin-angiotensin system; ROS, Reactive oxygen species; TNF-α, Tumor necrosis factor alpha.

growth and childhood weight gain appear to augment metabolic disruption in end organs, for example liver (Tarry-Adkins et al., 2016; Wang et al., 2016).

Although, the relationship between maternal protein restriction, sympathetic overactivity and hypertension have been suggested (Johansson et al., 2007; Franco et al., 2008; Barros et al., 2015), few studies have described the physiological dysfunctions responsible for producing these effects. Nowadays, it is well accepted that perinatal protein malnutrition raise risks of hypertension by mechanisms that include abnormal vascular function (Franco Mdo et al., 2002; Brawley et al., 2003; Franco et al., 2008), altered nephron morphology and function, and stimulation of the renin-angiotensin system (RAS) (Nuyt and Alexander, 2009; Siddique et al., 2014). Recently, studies have highlighted contribution of the sympathetic overactivity associated to enhanced respiratory rhythm and O2/CO<sup>2</sup> sensitivity on the development of the maternal low-protein diet-induced hypertension by mechanisms independent of the baroreflex function (Chen et al., 2010; Barros et al., 2015; Costa-Silva et al., 2015; de Brito Alves et al., 2015; Paulino-Silva and Costa-Silva, 2016). Offspring from dams subjected to perinatal protein restriction had relevant short-term effects on the carotid body (CB) sensitivity and respiratory control. With enhanced baseline sympathetic activity and amplified ventilatory and sympathetic responses to peripheral chemoreflex activation, prior to the establishment of hypertension (de Brito Alves et al., 2014, 2015). The underlying mechanism involved in these effects seems to be linked with up-regulation of hypoxic inducible factor (HIF-1α) in CB peripheral chemoreceptors (Ito et al., 2011, 2012; de Brito Alves et al., 2015). However, the epigenetic mechanisms in these effects are still unclear. It is hypothesized that epigenetic mechanism produced by DNA methylation could be involved (Altobelli et al., 2013; Prabhakar, 2013; Nanduri and Prabhakar, 2015).

The central nervous system (CNS) compared to other organ systems has increased vulnerability to reactive oxygen species (ROS). ROS are known to modulate the sympathetic activity and their increased production in key brainstem sites is involved in the etiology of several cardiovascular diseases, for example, diseases caused by sympathetic overexcitation, such as neurogenic hypertension (Chan et al., 2006; Essick and Sam, 2010). Ferreira and colleagues showed that perinatal protein undernutrition increased lipid peroxidation and decreased the activity of several antioxidant enzymes (superoxide dismutase, catalase, glutathione peroxidase, and glutathione reductase activities) as well as elements of the GSH system, in adult brainstem. Dysfunction in the brainstem oxidative metabolism, using the same experimental model, were observed in rats immediately after weaning associated to the increase in ROS production, with a decrease in antioxidant defense and redox status (Ferreira et al., 2015, 2016). Related to the metabolic effects on the heart, it was observed that these animals showed decreased mitochondrial oxidative phosphorylation capacity and increased ROS in the myocardium. In addition, maternal low-protein diet induced a significant decrease in enzymatic antioxidant capacity (superoxide dismutase, catalase, glutathione-S-transferase, and glutathione reductase activities) and glutathione level when compared with normoprotein group (Nascimento et al., 2014).

Regarding hepatic metabolism, studies showed that protein restricted rats had suppressed gluconeogenesis by a mechanism primarily mediated by decrease on the mRNA level of hepatic phosphoenolpyruvate carboxykinase, a key gluconeogenic enzyme, and enhancement of the insulin signals through the insulin receptor (IR)/IR substrate (IRS)/phosphatidylinositol 3-kinase (PI3K)/mammalian target of rapamycin complex 1 (mTOR) pathway in the liver (Toyoshima et al., 2010). In relation to lipid metabolism, there was decreased liver triglyceride content in adult rats exposed to protein restriction during gestation and lactation. It was suggested that this effect could be due to increased fatty-acid transport into the mitochondrial matrix or alterations in triglyceride biosynthesis (Qasem et al., 2015). A maternal protein restriction was shown to reduce the lean and increase the fat contents of 6-month old offspring with a tendency for reduced number of muscle myofibers associated with reduced expression of mRNA of Insulin-like growth factor 2 gene (IGF2 mRNA) in pigs (Chavatte-Palmer et al., 2016).

## MATERNAL OVERNUTRITION AND RISK FACTOR FOR THE CARDIOMETABOLIC DYSFUNTIONS

Nutritional transition is a phenomenon well documented in developing countries in the twentieth and twenty-first centuries, and has induced high incidence of the chronic diseases and high prevalence of the obesity (Batista Filho and Rissin, 2003; Batista Filho and Batista, 2010; Ribeiro et al., 2015). It is evident that protein malnutrition was an health problem in the first half of the twentieth century. Now, it was replaced by a diet enriched in saturated fat or other HFDs, predisposing to overweight, and obesity (Batista et al., 2013). Nowadays, it suggested that two billion people in the world are overweight and obese individuals, with major prevalence is related to diet induced-obesity, which have been associated to cardiovascular and endocrine dysfunctions (Hotamisligil, 2006; Aubin et al., 2008; Zhang et al., 2012; Ng et al., 2014; Wensveen et al., 2015).

Recently, the obesity has been considered a physiological state of chronic inflammation, characterized by elevated levels of inflammatory markers including C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor alpha (TNFα) (Wensveen et al., 2015; Erikci Ertunc and Hotamisligil, 2016; Lyons et al., 2016). Maternal HFD chronic consumption enhances the circulating free fatty acids and induce the activation of inflammatory pathways, enhancing chronic inflammation in offspring (Gruber et al., 2015). Studies of Roberts et al. (2015) found that cardiometabolic dysfunction was associated with changes such as elevated serum triglycerides, elevated oxidative stress levels, insulin resistance, vascular disorders, and development of hypertension (Roberts et al., 2015).

In animals on a HFD the hormone leptin has been considered one of the most important physiological mediators of the cardiometabolic dysfunction (Correia and Rahmouni, 2006; Harlan et al., 2013; Harlan and Rahmouni, 2013). Since hyperleptinemia, common in overweight and obesity conditions, produce a misbalance in autonomic system, with sympathetic overactivation (Machleidt et al., 2013; Kurajoh et al., 2015; Manna and Jain, 2015), and reduced sensitivity of vagal afferent neurons (de Lartigue, 2016). This disorder of vagal afferent signaling can activate orexigenic pathways in the CNS and drive hyperphagia, obesity, and cardiometabolic diseases at long-term (de Lartigue, 2016). Some authors have described that, at least in part, cardiovascular dysfuntion elicited by HFD or obesity may be due to changes in the neural control of respiratory and autonomic systems (Bassi et al., 2012, 2015; Hall et al., 2015; Chaar et al., 2016). Part of these effects were suggested to be influenced by atrial natriuretric peptide and renin-angiotensin pathways (Bassi et al., 2012; Gusmão, 2012).

Interestingly, it has been shown that offspring from mothers fed HFD have high risk to develop pathologic cardiac hypertrophy. This condition would be linked to re-expression of cardiac fetal genes, systolic, and diastolic dysfunction and sympathetic overactivity on the heart. These effects lead to reduced cardioprotective signaling that would predispose them to cardiac dysfunctions in adulthood (Taylor et al., 2005; Wang et al., 2010; Fernandez-Twinn et al., 2012; Blackmore et al., 2014). Regarding arterial blood pressure control, it has been described that maternal HFD induces early and persistent alterations in offspring renal and adipose RAS components (Armitage et al., 2005). These changes seem to be dependent upon the period of exposure to the maternal HFD, and contribute to increased adiposity and hypertension in offspring (Samuelsson et al., 2008; Elahi et al., 2009; Guberman et al., 2013; Mazzio and Soliman, 2014; Tan et al., 2015). Studies in baboons subjected to HFD showed that microRNA expression and putative gene targets involved in developmental disorders and cardiovascular diseases were up-regulated and others were down-regulated. The authors suggested that the epigenetic modifications caused by HFD may be involved in the developmental origins of cardiometabolic diseases (Maloyan et al., 2013).

Other metabolic outcomes induced by HFD have been pointed out in the last years and it has demonstrated that HFD displayed a drastic modification on metabolic control of the glucose metabolism and lead to increased insulin level in serum (Fan et al., 2013) and enhanced insulin action through AKT/PKB (protein kinase B) and ERK (extracellular signal-regulated kinase), and activation of mammalian target of rapamycin (mTOR) pathways in cardiac tissue (Fernandez-Twinn et al., 2012; Fan et al., 2013). Offspring from HFD mothers showed alterations in blood glucose and insulin levels, with high predisposition to insulin resistance and cardiac dysfunction (Taylor et al., 2005; Wang et al., 2010). Part of these effects are associated with enhanced production of ROS and reduction in the levels of the anti-oxidant enzymes, such as superoxide dismutase, suggesting a misbalance in the control of the oxidative stress (Fernandez-Twinn et al., 2012).

Altogether, this review addressed the new concept on the maternal diet induced-cardiometabolic diseases that include the potential role of the perinatal malnutrition. It showed that the etiology of these diseases is multifactorial involving genetic and environmental influences and their physiological integration. It is well recognized that both perinatal undernutrition and overnutrition are related with the risk of developing metabolic syndrome and hypertension in adult life (**Figure 1**). The underlying mechanism can be explained in the context of phenotypic plasticity during development that includes adaptive change on the CNS, heart, kidney, liver, muscle, and adipose tissue metabolisms with consequent physiology dysfunction and

FIGURE 1 | Schematic drawing showing the physiological effects induced by maternal and fetus exposure to under- or overnutrition through DNA methylation and their consequences on the organ physiology and increased risk of the cardiometabolic diseases in the offspring.

with subsequent cardiometabolic diseases. Moreover, maternal undernutrition or overnutrition may predispose epigenetic modifications in dams and their offspring, with predominance of DNA methylation, leading to altered gene expression during development and growth. Further, it can provide a different physiological condition which may contribute to the developmental origins of the cardiometabolic diseases. These physiological dysfunctions seem to be linked to the impaired central and peripheral control of both metabolic and cardiovascular functions by mechanisms that include enhanced

### REFERENCES


sympathetic-respiratory activities and disruption in metabolism of end organs at early life. It is suggested that those effects could be associated to inflammatory conditions and impaired oxidative balance, which may contribute to adult cardiometabolic diseases.

### AUTHOR CONTRIBUTIONS

JC, AS, and MF drafted and revised critically the work for important intellectual content and final review of the manuscript.


response to maternal protein restriction. Brain Res. 1642, 553–561. doi: 10.1016/j.brainres.2016.04.049


Burden of Disease Study 2013. Lancet 384, 766–781. doi: 10.1016/S0140- 6736(14)60460-8


hepatic fibrosis, inflammation, and oxidative stress in a male rat model of poor maternal nutrition and accelerated postnatal growth. Am. J. Clin. Nutr. 103, 579–588. doi: 10.3945/ajcn.115.119834


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Costa-Silva, Simões-Alves and Fernandes. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Emotional Stress and Cardiovascular Complications in Animal Models: A Review of the Influence of Stress Type

### Carlos C. Crestani\*

*Faculdade de Ciências Farmacêuticas, UNESP - Univ Estadual Paulista, Araraquara, Brasil*

Emotional stress has been recognized as a modifiable risk factor for cardiovascular diseases. The impact of stress on physiological and psychological processes is determined by characteristics of the stress stimulus. For example, distinct responses are induced by acute vs. chronic aversive stimuli. Additionally, the magnitude of stress responses has been reported to be inversely related to the degree of predictability of the aversive stimulus. Therefore, the purpose of the present review was to discuss experimental research in animal models describing the influence of stressor stimulus characteristics, such as chronicity and predictability, in cardiovascular dysfunctions induced by emotional stress. Regarding chronicity, the importance of cardiovascular and autonomic adjustments during acute stress sessions and cardiovascular consequences of frequent stress response activation during repeated exposure to aversive threats (i.e., chronic stress) is discussed. Evidence of the cardiovascular and autonomic changes induced by chronic stressors involving daily exposure to the same stressor (predictable) vs. different stressors (unpredictable) is reviewed and discussed in terms of the impact of predictability in cardiovascular dysfunctions induced by stress.

### Edited by:

*Valdir Andrade Braga, Federal University of Paraiba, Brazil*

### Reviewed by:

*Eugene Nalivaiko, University of Newcastle, Australia Angela J. Grippo, Northern Illinois University, USA*

> \*Correspondence: *Carlos C. Crestani cccrestani@yahoo.com.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *13 May 2016* Accepted: *09 June 2016* Published: *24 June 2016*

### Citation:

*Crestani CC (2016) Emotional Stress and Cardiovascular Complications in Animal Models: A Review of the Influence of Stress Type. Front. Physiol. 7:251. doi: 10.3389/fphys.2016.00251* Keywords: psychological stress, restraint stress, chronic variable stress, social isolation, social defeat, crowding stress, baroreflex, autonomic activity

## INTRODUCTION

Convergent clinical and preclinical studies have provided evidence on the important role of psychosocial factors in the etiology and progression of cardiovascular diseases (Friedman and Rosenman, 1959; Rosenman et al., 1975; Kawachi et al., 1994; Ford et al., 1998; Rozanski et al., 1999; Rugulies, 2002; Smith et al., 2004; Grippo and Johnson, 2009; Roest et al., 2010; Carnevali et al., 2013a; Sgoifo et al., 2014). Among several psychosocial factors, psychological stress has been recognized as a modifiable risk factor for several cardiovascular dysfunctions (Steptoe and Kivimaki, 2012; von Känel, 2012; Inoue, 2014). Indeed, epidemiological and experimental results on humans and animals have demonstrated the influence of psychosocial stress on cardiovascular health (Rosengren et al., 2004; Kivimaki et al., 2006b; Grippo and Johnson, 2009; Steptoe and Kivimaki, 2012; Jarczok et al., 2013; Sgoifo et al., 2014). The association between stress and cardiovascular diseases has been shown to be independent of traditional cardiovascular risk factors (e.g., age, sex, smoking, diabetes mellitus, and obesity; Rosengren et al., 2004; Kivimaki et al., 2006b; Steptoe and Kivimaki, 2012).

The impact of stress on physiological and psychological processes is determined by stressor characteristics, such as chronicity, predictability, and severity (Nalivaiko, 2011; Steptoe and Kivimaki, 2012; Herman, 2013). Indeed, distinct cardiovascular and neuroendocrine changes have been reported following acute vs. chronic stress regimens (Nalivaiko, 2011; Herman, 2013). Furthermore, studies in animals comparing the impact of predictable vs. unpredictable stressor stimuli demonstrated that the latter induces more severe behavioral and physiological consequences (Bassett et al., 1973; Magariños and McEwen, 1995; Marin et al., 2007; Flak et al., 2012; Kopp et al., 2013; Smith et al., 2013; Duarte et al., 2015a). A dose–response pattern of stress-induced cardiovascular dysfunctions has also been reported, and stress severity may be modulated by the duration and/or frequency of the exposure to stress stimulus (Marmot et al., 1991; Kivimaki et al., 2006a; Chandola et al., 2008; Steptoe and Kivimaki, 2012).

The purpose of the present review is to discuss experimental research in animal models describing the impact of emotional stress on cardiovascular function. The primary focus of this review is to discuss the influence of stress type in stress-induced cardiovascular changes. Therefore, the first section discusses the influence of chronicity by summarizing the cardiovascular responses during acute stress sessions and the cardiovascular dysfunctions following repeated exposure to aversive stimuli (i.e., chronic stress). The second section discusses the impact of the predictability of stressor stimulus in cardiovascular and autonomic changes induced by chronic stress.

### CARDIOVASCULAR RESPONSES TO STRESS: INFLUENCE OF CHRONICITY

### Cardiovascular Responses during Acute Stress Sessions

A coordinated and complex set of physiological changes is generated during acute stress sessions. Changes in the autonomic nervous system activity promote immediate responses during aversive threats, which are mainly characterized by changes in the cardiovascular function. Cardiovascular responses typically consist of increase in blood pressure, heart rate (HR), and cardiac output (Hubbard et al., 1986; Schadt and Hasser, 1998; Dos Reis et al., 2014). However, these responses are consequences of more complex changes in cardiovascular function. Below is discussed the blood flow redistribution, changes of cardiac autonomic activity, and modulation of baroreflex function observed during an acute stress session. Additionally, the recovery of cardiovascular function after ending the stress session is described.

### Blood Flow Redistribution

Blood flow redistribution occurs from the visceral and cutaneous beds toward the skeletal muscle vasculature during aversive threats (Kirby et al., 1987; Kapusta et al., 1989; Knardahl and Hendley, 1990; Viken et al., 1991; Zhang et al., 1996; Schadt and Hasser, 1998; Blessing, 2003). Blood flow redistribution is mediated mainly by vasoconstriction of the splanchnic, renal, cutaneous, and celiac vascular beds and by vasodilatation of the skeletal muscle vasculature (Iriuchijima et al., 1982; Kirby et al., 1987; Kapusta et al., 1989; Zhang et al., 1992, 1996; Anderson and Overton, 1994; Schadt and Hasser, 1998; Blessing, 2003; Mohammed et al., 2014). The increase in skeletal muscle blood flow is primarily due to an increase in the level of circulating catecholamines and activation of β2-adrenoceptors in the skeletal muscle beds, whereas vasoconstriction in the visceral vasculature is mediated by catecholamine released from both the adrenal medulla and sympathetic nerve fibers and activation of αadrenoceptors (Iriuchijima et al., 1982; Knardahl and Hendley, 1990; Zhang et al., 1992, 1996).

An increase in the total peripheral resistance during stress was reported in borderline hypertensive rats (Hatton et al., 1997). However, studies in normotensive animals documented little or no change in this parameter (Hubbard et al., 1986; Martin et al., 1996; Zhang et al., 1996; Schadt and Hasser, 1998), indicating that vasodilation mostly balanced vasoconstriction. Therefore, the increase in cardiac output is possibly the main mechanism mediating the moderate increase in arterial pressure during aversive threats in rodents. In this regard, hemodynamic changes during stress also include an increase in venomotor tone via activation of α2-adrenoceptors (Martin et al., 1996; Schadt and Hasser, 1998), which increase venous return and contribute to stress-induced increase in the cardiac output.

### Cardiac Autonomic Activity

The tachycardic response during aversive threats is mediated by an increase in the sympathetic tone to the heart (Iwata and LeDoux, 1988; Baudrie et al., 1997; Sgoifo et al., 1999; van den Buuse et al., 2001; Carrive, 2006; Crestani et al., 2010b; Dos Reis et al., 2014). However, stress-induced tachycardic response was enhanced following systemic treatment with muscarinic cholinergic receptor antagonists (Iwata and LeDoux, 1988; Baudrie et al., 1997; Nijsen et al., 1998; Carrive, 2006; Crestani et al., 2009; Dos Reis et al., 2014), indicating an increase in the parasympathetic tone to the heart. Indeed, coactivation of cardiac sympathetic and parasympathetic activities has been documented during aversive situations (Iwata and LeDoux, 1988; Baudrie et al., 1997; Carrive, 2006; Dos Reis et al., 2014). The increase in the parasympathetic tone to the heart allows precise response control, reducing the amplitude of the response, and counteracting excessive cardiac activation associated with cardiac sympathetic system activation alone. The parasympathetic activation and the consequent reduction of cardiac response possibly contributes to a functional state stabilization of the heart (Paton et al., 2005).

### Modulation of Baroreflex Function

Change in baroreflex function is a relevant mechanism controlling cardiovascular and autonomic responses during aversive threats. The baroreflex is an important mechanism for arterial pressure regulation. For example, an increase in arterial pressure activates baroreceptors within the wall of the carotid arteries and aorta that in turn increase cardiac parasympathetic activity and decrease sympathetic activity, leading to decreased HR, vascular resistance, and venous return (Michelini, 1994; Sved and Gordon, 1994). However, as stated above, the increase in arterial pressure during stress is followed by a parallel enhancement in sympathetic activity and HR. Hilton (1963) addressed first the question regarding how parallel increase in blood pressure, HR, and sympathetic activity occurs during aversive situations. Based on the responses obtained by stimulation of the hypothalamic defense area, which mimics the cardiovascular responses induced by stress, baroreflex function was first claimed to be suppressed during defensive behavior (Hilton, 1963; Coote et al., 1979). However, further studies consistently demonstrated that pressor, tachycardic, sympathetic activation, and regional vasoconstrictor responses observed during emotional stress are enhanced in sinoaortic baroreceptordenervated rats (Norman et al., 1981; Buchholz et al., 1986; Zhang et al., 1996; Dos Reis et al., 2014), indicating an active role of the baroreflex counteracting stress-induced autonomic and cardiovascular responses during defensive responses. Indeed, baroreflex is clearly functional during aversive threats, but it curves for both renal sympathetic nerve activity and HR shifts to the right and upward (Shammas et al., 1988; Hatton et al., 1997; Schadt and Hasser, 1998; Kanbar et al., 2007; Burke and Head, 2009; Crestani et al., 2010b; Miki and Yoshimoto, 2013), allowing a parallel increase in arterial pressure, HR, and sympathetic activity. A schematic representation of the right and upward shift of the baroreflex curve during aversive situations is presented in **Figure 1**.

The rightward shift represents a resetting of the operating point of the baroreflex toward higher arterial pressure values (**Figure 1**). Therefore, the baroreflex operates under higher arterial pressure values during stress. This idea is supported by evidence that a decrease in the increase in arterial pressure due to treatment with α-adrenoceptor antagonists (i.e., maintaining arterial pressure values below the new operating point of the

resetting). For details, see text in the section "*Cardiovascular responses during*

baroreflex) enhanced the tachycardic response during aversive situations (Baudrie et al., 1997; Carrive, 2002; Dos Reis et al., 2014), possibly to increase arterial pressure to some preset value. The baroreflex resetting seems to be a mechanism allowing the simultaneous increase in arterial pressure and sympathetic activity/HR rather than generating the cardiovascular responses to stress, once sinoaortic baroreceptor denervation facilitated stress-induced cardiovascular and autonomic adjustments, as stated above (Norman et al., 1981; Buchholz et al., 1986; Zhang et al., 1996; Dos Reis et al., 2014). However, Zhang et al. (1996) reported that vasodilatation in the skeletal muscle vasculature was reduced in sinoaortic baroreceptor-denervated rats, indicating that baroreflex activity resetting may contribute to blood flow redistribution during aversive threats.

The abovementioned evidence indicates that autonomic and cardiovascular responses to stress are mediated by mechanisms acting independent of the changes in baroreflex function. However, the increase in sympathetic activity and HR is responsible for the upward shift of the baroreflex curves (Schadt and Hasser, 1998; Crestani et al., 2010b; Miki and Yoshimoto, 2013; **Figure 1**). Indeed, the organism operates at higher values of HR and sympathetic activity during stress, inducing a vertical change in the baroreflex curves (Schadt and Hasser, 1998; Crestani et al., 2010b; Miki and Yoshimoto, 2013; **Figure 1**).

### Recovery of Cardiovascular Function after Ending Stress Session

Arterial pressure and HR progressively decrease toward baseline values during recovery from an aversive stimulus, returning to pre-stress levels ∼30–60 min after ending stress (McDougall et al., 2000; Vianna and Carrive, 2005; Igosheva et al., 2007; Porter et al., 2007; Crestani et al., 2010b; Krause et al., 2011). Nevertheless, we have previously reported that besides arterial pressure and HR had already returned to pre-stress values 30 min after ending an acute session of restraint stress, the baroreflex activity remained similar to that observed during stress, returning to control values only 60 min after ending the stress session (Crestani et al., 2010b). These findings indicate non-persistent effects of stress on cardiovascular function, but suggest a period of vulnerability during the recovery period, once the baroreflex function is reset to higher arterial pressure values. Nevertheless, persistence of the stress effects on baroreflex activity may be important to counteract excessive activation of depressor mechanisms during the recovery period, thus allowing adequate return of blood pressure to normal values without abrupt falls. Indeed, central and peripheral depressor mechanisms are activated during the recovery period and mediate the return of arterial pressure to resting values (Yip and Krukoff, 2002; D'Angelo et al., 2006).

### Cardiovascular Responses to Chronic Stressors

The physiological and behavioral responses during acute stress sessions constitute important adaptive responses, maintaining homeostasis and ensuring survival (Sterling and Eyer, 1988; Sterling, 2012). However, the frequent/prolonged activation (i.e., over-exposure to stress changes) or generation of inadequate

*acute sessions of stress.*"

responses (i.e., insufficient responses to individual needs) during stressful events can result in disease development (McEwen and Stellar, 1993; McEwen, 1998; Danese and McEwen, 2012). A situation of over-exposure to stress responses occurs during repeated exposure to aversive threats (i.e., chronic stress; McEwen, 1998). Regarding the cardiovascular consequences of stress, besides reports of acute cardiovascular events such as ventricular arrhythmias, myocardial infarction, and cardiomyopathy (potentially fatal) after exposure to acute stressful situations (Cannon, 1942; Lown et al., 1973; Natelson and Cagin, 1979; Sgoifo et al., 1997, 1999; Ueyama et al., 2002; Mostofsky et al., 2014; Lagraauw et al., 2015), enduring cardiovascular changes (days to years) are mainly provoked by long-term exposure to stressful events (Nalivaiko, 2011). Therefore, the frequency of exposure to stress (chronicity) is an important factor in determining the pattern of stressinduced cardiovascular diseases. Below, we discuss the evidence of enduring changes in arterial pressure, HR, cardiac function, autonomic activity, and baroreflex function that lasts beyond the duration of chronic stress protocols.

### Arterial Pressure

Despite of pronounced neuroendocrine (e.g., increased hypothalamic–pituitary–adrenal, HPA, axis activity), somatic (e.g., reduced body weight gain, adrenal hypertrophy, and thymus atrophy), and behavioral (e.g., anxiety- and depressionlike behavior) changes following long-term exposure to stressful events, preclinical studies have failed to consistently reproduce epidemiological findings demonstrating a relationship between chronically stressful life events and hypertension in animals (Timio et al., 1988; Markovitz et al., 2004). Indeed, inconsistent findings have been reported regarding the effects of chronic stress on arterial pressure in animal models. Nalivaiko (2011) reviewed in details publications that investigated the effect of different chronic stress models on arterial pressure in several rat strains. An important consideration raised was the influence of the method used to assess arterial pressure. For example, most studies reviewed by Nalivaiko (2011) that identified a significant hypertensive effect following exposure to chronic stressors used the tail-cuff method (indirect) to measure the arterial pressure. Even in studies that validated the findings obtained through the tail-cuff method by direct methods (through implantation of a catheter into the femoral or carotid artery at the end of the stress protocol), the arterial pressure increase obtained in the direct method was lower than that in the tail-cuff method (Nalivaiko, 2011). Similar analysis of studies published after 2007 (date of the more recent study included in the Nalivaiko's review; **Table 1**) further reinforces observation regarding the influence of the method used to assess arterial pressure in the identification of stress-induced hypertension.

Animals need to be restrained (similar to the restraint stress model) in the tail-cuff method, and marked pressor and tachycardic responses are observed, even following recommended habituation (Grundt et al., 2009). Based on this, Nalivaiko (2011) proposed that the identification of hypertensive effect preferentially by the tail-cuff method would be due to a more vigorous reaction of the animals stressed to the tail-cuff procedure. Hence, it has been reported that exposure to chronic stressors increases pressor response during acute aversive stimuli (Grippo et al., 2002, 2006; Cudnoch-Jedrzejewska et al., 2010, 2014; Maslova et al., 2010).

Another factor mentioned by Nalivaiko (2011) could be the possible influence of coping strategies in models providing actual or perceived control over a stressful situation, which may occur, for example, in highly predictable protocols (e.g., during repeated exposure to the same stressor). We recently addressed this issue by comparing the impact on arterial pressure (measured by the direct method) of a predictable (repeated restraint stress, RRS) vs. an unpredictable (chronic variable stress, CVS) chronic stress protocol (Duarte et al., 2015a) (see the next section for details). We observed a similar small pressor effect (+10–15 mm Hg) following exposure to either RRS or CVS for 10 days (Duarte et al., 2015a). Other studies using longer CVS protocols (4–8weeks) reported either similar effect (∼+8 mm Hg) (Grippo et al., 2008) or absence of changes on arterial pressure measured by either direct methods or telemetry (Grippo et al., 2002, 2006; Cudnoch-Jedrzejewska et al., 2010; Demirtas et al., 2014; Stanley et al., 2014). In addition, Bobrovskaya et al. (2013) did not identify changes in mean arterial pressure measured by telemetry (only a small increase in systolic arterial pressure was identified) following repeated exposure to footshock (more intense aversive stimulus) for 6 weeks in an unpredictable schedule. Thus, the small hypertensive effect obtained by direct measurement of arterial pressure seems to be independent of intensity, predictability, and duration of chronic stress protocol.

An additional factor considered in our review (**Table 1**) was an elevated baseline arterial pressure in the control group, which would indicate some degree of stress induced by the experimental procedure (independent of chronic stressor) that in turn could buffer effects of chronic stress. However, independently of the chronic stress paradigm, a similar small hypertensive effect (∼10 mm Hg) was observed following long-term stress exposure in studies wherein baseline mean arterial pressure in control groups ranged from 95 to 120 mmHg (Grippo et al., 2008; Daubert et al., 2012; Duarte et al., 2015a; Cruz et al., 2016) (**Table 1**). It has also been documented that cardiovascular dysfunctions induced by chronic stressors may be related to animals age (Crestani, 2016); hence, an influence of age (e.g., young vs. adult) was considered in this review (**Table 1**). However, the influence of age seems to be related to the type of chronic stressor. For example, social isolation for 3 weeks increased arterial pressure in adolescent rats, without affecting this parameter in adult animals (Cruz et al., 2016). In contrast, RRS induced a mild hypertensive effect in adult animals alone (Duarte et al., 2015a).

### Heart Rate and Cardiac Dysfunctions

More consistent results have been reported for HR. **Table 2** summarizes the main findings of changes in baseline HR following exposure to chronic stressors. Increase in baseline HR has been reported in many studies, but this effect seems to be related to the chronic stressor type. Indeed, resting tachycardia has been consistently documented following exposure to two


 (MAP).

TABLE


TABLE

1


Continued

*repeated restraint stress; SHR, spontaneously*

*Bold indicates studies that identified effect of chronic emotional stress in baseline MAP.*

\**Cardiovascular*

 *recording in anesthetized*

 *animals.*

 *hypertensive*

 *rats; SHR/Y, borderline hypertensive*

 *consomic rat strain; WKY, Wistar-Kyoto*

 *rat.*


rate

(HR).

TABLE

2


*(Continued)*


TABLE

2


Continued

*(Continued)*


TABLE

2


Continued

*Bold indicates studies that identified effect of chronic stress in baseline HR.*

\**Cardiovascular*

 *recording in anesthetized*

 *animals.*

animal models of chronic stress: the CVS and chronic social isolation in prairie voles (a highly social rodent species; **Table 2**). Resting tachycardia induced by both these models was independent of the protocol length. For example, increased baseline HR was observed following exposure to protocols ranging from 10 days to 8 weeks of CVS (Grippo et al., 2002, 2008; Mercanoglu et al., 2008; Bouzinova et al., 2012; Duarte et al., 2015a) and 5 days to 8 weeks of social isolation in prairie voles (Grippo et al., 2007, 2011; McNeal et al., 2014). Moreover, the increase in HR was not affected by baseline HR in control animals (indicative of some degree of stress of the experimental procedure, see discussion above) and by animal age (**Table 2**).

Although, most studies investigating the effect of chronic social isolation in prairie voles have used female animals (**Table 2**), resting tachycardia was also reported in male prairie voles following a period of social isolation (McNeal et al., 2014). CVS-induced increase in HR was evaluated only in males animals, but was documented in both Wistar and Sprague– Dawley rats (Grippo et al., 2002, 2006, 2008; Bouzinova et al., 2012; Duarte et al., 2015a; Matchkov et al., 2015), indicating that the effect is independent of the rat strain. As discussed for the effect of stress on arterial pressure, we note higher responses in studies that measured HR by the tail-cuff method compared with data obtained by other methods (**Table 2**). As discussed above, it can be related to a more vigorous reaction of the stressed animals to the tail-cuff procedure (Nalivaiko, 2011).

Despite the absence of changes in baseline HR, some models induce profound changes in cardiac function. For example, repeated social defeat episodes caused accumulation of fibrous tissue in the left ventricular myocardium, maladaptive cardiac hypertrophy, changes in electrical conduction system of the heart (e.g., reduced myocardial refractoriness and impaired conduction), and increased susceptibility to cardiac arrhythmias (Gelsema et al., 1994; Costoli et al., 2004; Carnevali et al., 2013b, 2015; Sgoifo et al., 2014). Disruption of the circadian rhythm for HR has also been reported following repeated exposure to social defeat (Tornatzky and Miczek, 1993; Sgoifo et al., 2002; Carnevali et al., 2015).

RRS is another chronic stressor that does not induce significant changes in baseline HR (**Table 2**), but cardiac hypertrophy (Nagaraja and Jeganathan, 1999; Bruder-Nascimento et al., 2012; Duarte et al., 2015a), electrocardiogram (ECG) abnormality, myocardium injury, and cardiac dysfunction (Zhao et al., 2007; Roth et al., 2015) have been reported in animals exposed to this experimental model of stress. Moreover, RRS increased the size of infarction and the incidence of potentially fatal arrhythmias induced by myocardial ischemia (Scheuer and Mifflin, 1998), increased plaque formation in the coronary arteries and occurrence of myocardial infarctions in mice susceptible to atherosclerosis (Roth et al., 2015), and exacerbated hypertension and left ventricular hypertrophy, fibrosis, and diastolic dysfunction related to metabolic syndrome (Habib et al., 2015; Matsuura et al., 2015). These results indicate that RRS may affect the outcome of cardiac diseases.

Consistent findings have also indicated that crowding stress (i.e., social stress that causes competition for resources such as space, food, and water) does not affect baseline HR (**Table 2**). Nevertheless, analysis of responses to myocardial ischemia indicated an impairment of postischemic recovery of cardiac mechanical function and aggravation of tachyarrhythmia and ventricular fibrillation in animals subjected to crowding stress (Ravingerova et al., 2011; Ledvenyiova-Farkasova et al., 2015), suggesting that exposure to this social stressor reduces the tolerance to cardiac ischemia. Maladaptive cardiac hypertrophy, ECG abnormality, and impairment of coronary perfusion of the myocardium have also been reported following exposure to crowding stress (Nagaraja and Jeganathan, 1999; Ravingerova et al., 2011).

In addition to causing tachycardia, cardiac dysfunctions have also been reported following exposure to either CVS or chronic social isolation in prairie voles. Indeed, increased susceptibility to experimentally and stress-induced cardiac arrhythmias was reported in animals subjected to either stressor (Grippo et al., 2004, 2010, 2012a; Liang et al., 2015), and CVS increased the infarcted area after myocardial ischemia (Mercanoglu et al., 2008). These findings indicate that these experimental models of stress may increase the vulnerability and severity of cardiac complications. In addition, cardiac contractile dysfunction was reported in animals subjected to CVS protocols (Grippo et al., 2006; Xie et al., 2012).

Studies in the literature also investigated a possible impact of chronic stressors in the pacemaker activity of the sinoatrial node, referred as intrinsic HR. It can be assessed experimentally by dual blockade of sympathetic and parasympathetic activities of the heart via combined systemic treatment with β-adrenoceptor antagonists (i.e., sympathetic blocker; e.g., propranolol or atenolol) and muscarinic cholinergic receptors (i.e., parasympathetic blocker; e.g., methylatropine). However, intrinsic HR was not affected after exposure to either CVS (Grippo et al., 2002; Almeida et al., 2015; Duarte et al., 2015a), chronic social isolation in prairie voles (Grippo et al., 2007, 2009; McNeal et al., 2014), or RRS (Duarte et al., 2015a). This finding seems to be independent of the duration of chronic stress protocol. For example, intrinsic HR was investigated in CVS protocols ranged from 10 days to 4 weeks (Grippo et al., 2002; Almeida et al., 2015; Duarte et al., 2015a), whereas social isolation in prairie voles ranged from 5 days to 4 weeks (Grippo et al., 2007; McNeal et al., 2014). To the best of my knowledge, a possible impact of other animal models in cardiac pacemaker activity has never been evaluated.

### Autonomic Activity

Reduction in HR variability following exposure to different animal models of chronic stress has been described (Grippo et al., 2002, 2009; Wood et al., 2012; Sévoz-Couche et al., 2013), which indicates changes in cardiac autonomic activity. Indeed, analysis of cardiac autonomic activity by either power spectral analysis of oscillatory components of HR or pharmacological blockade of cardiac sympathetic (e.g., treatment with propranolol) and parasympathetic (e.g., treatment with methylatropine) activities indicated significant stress-induced changes in cardiac autonomic activity. Nevertheless, these alterations seem to be specific to the chronic stressor type. For example, pharmacological blockade of cardiac autonomic activity revealed that CVS increased the sympathetic tone to the heart, without significantly affecting cardiac parasympathetic activity (Grippo et al., 2002; Duarte et al., 2015a). It was also evidenced by demonstration of an increase in LF/HF ratio of HR variability (Bundzikova-Osacka et al., 2015), which indicates a change in cardiac sympathovagal balance toward a sympathetic predominance. In contrast, the changes in cardiac autonomic activity induced by chronic social isolation in prairie voles, evidenced by pharmacological autonomic blockade, were characterized by both a decrease in parasympathetic and an increase in sympathetic tone to the heart (Grippo et al., 2007, 2009; McNeal et al., 2014). Despite specific changes in sympathetic and parasympathetic activities, the changes induced by either CVS or social isolation in prairie voles were characterized by a change in cardiac sympathovagal balance toward a sympathetic predominance, which is in line with the resting tachycardia induced by these experimental models. Moreover, the increase in sympathetic contribution of cardiac autonomic balance corroborates the increased susceptibility to cardiac arrhythmias induced by both models (Grippo et al., 2004, 2006, 2012a), as well as with cardiac contractile dysfunction (Grippo et al., 2006; Xie et al., 2012) and increased severity of myocardial ischemia (Mercanoglu et al., 2008) observed after exposure to CVS. Studies have not reported changes in cardiac sympathovagal balance following exposure to RRS protocols (Daubert et al., 2012; Duarte et al., 2015a,b), which is in line with the absence of changes in baseline HR.

Cardiac autonomic imbalance has been reported in Sprague– Dawley rats subjected to repeated sessions of social defeat (Wood et al., 2012; Sévoz-Couche et al., 2013). By analyzing HR variability, results indicated a shift in sympathovagal balance toward sympathetic predominance (Wood et al., 2012; Sévoz-Couche et al., 2013), which was likely mediated by both a decrease in parasympathetic activity and increase in sympathetic tone to the heart (Sévoz-Couche et al., 2013). This finding is in line with the increase in baseline HR observed in Sprague–Dawley rats following repeated exposure to social defeat (Sévoz-Couche et al., 2013). However, Sprague–Dawley rats seem to be selectively susceptible to resting tachycardia induced by repeated social defeat once studies did not identify changes in baseline HR in either mice (Bartolomucci et al., 2003; Costoli et al., 2004; Carnevali et al., 2012) or wide-type Groningen (Sgoifo et al., 2001; Carnevali et al., 2013b), Long-Evans (Tornatzky and Miczek, 1993), Wistar-Kyoto (Carnevali et al., 2015), Lister hooded (Chung et al., 1999), and Dahl-s (Adams and Blizard, 1986; Adams et al., 1987) rats. Nevertheless, differences in experimental protocol may also account for the differences in findings of autonomic/cardiovascular changes because exposures to social defeat in different studies ranged from 5 to 25 days (**Table 2**). Moreover, the defeated animal cohabited with its aggressor in some protocols, being subjected to intermittent episodes of aggressive interaction but with continuous sensorial contact (Bartolomucci et al., 2003; Costoli et al., 2004), whereas animals were exposed only to intermittent episodes of social defeat in other studies (Sgoifo et al., 2001; Carnevali et al., 2013b), including those that identified changes in HR (Wood et al., 2012; Sévoz-Couche et al., 2013).

The increase in sympathetic activity induced by chronic stressors seems not to be restricted to the heart. For instance, Grippo et al. (2008) reported that a 4-week CVS protocol increased lumbar sympathetic nerve activity. In addition, an increase in tyrosine hydroxylase expression and activity has been reported in the sympathetic ganglia and adrenal medulla following exposure to chronic stressors (Nankova et al., 1994, 1996; Bobrovskaya et al., 2013). The relevance of the response in the adrenal medulla is unclear since changes in plasma concentration of adrenaline and noradrenaline were not identified following exposure to several chronic stress protocols, including RRS, footshock, social isolation, crowding, and CVS (Konarska et al., 1989; Dronjak et al., 2004; Gavrilovic et al., 2005; Spasojevic et al., 2009). Therefore, sympathoexcitation induced by chronic stressors seems to be mediated by an increase in activity of sympathetic nerves rather than changes in the sympatho-adrenomedullary system.

Long-term potentiation was reported in the sympathetic ganglia of animals subjected to cage-switch stress for 4 weeks (Alkadhi et al., 2005). It is an activity-dependent sustained increase in ganglionic transmission, which is similar to that described in the brain. Therefore, this sustained enhancement of synaptic efficacy may constitute an important mechanism underlying the increase in sympathetic tone following long-term exposure to aversive stimuli (Alkadhi et al., 2005).

### Baroreflex Function

Changes in the baroreflex control of HR were reported following exposure to several chronic stressors, including CVS, RRS, repeated social defeat stress, and chronic social isolation (Conti et al., 2001; Porter et al., 2004; Daubert et al., 2012; Xie et al., 2012; Sévoz-Couche et al., 2013; Almeida et al., 2015; Duarte et al., 2015a; Cruz et al., 2016). However, the effects in the baroreflex activity seem to be dependent on the duration of chronic stress. For example, exposure to CVS protocols of either 10 (Duarte et al., 2015a) or 14 days (Xie et al., 2012; Almeida et al., 2015) induced changes in the baroreflex control of HR, but a 4-week protocol did not affect baroreflex HR responses (Grippo et al., 2008). Nevertheless, the reflex increase in lumbar sympathetic nerve activity induced by hypotension was decreased in animals following exposure to a CVS protocol for 4 weeks (Grippo et al., 2008), indicating that lumbar sympathetic nerve and cardiac autonomic innervation are affected differently by CVS. In addition, different effects of CVS in reflex bradycardia induced by increased blood pressure were obtained when baroreflex function was evaluated in animals anesthetized (facilitation; Xie et al., 2012) and unanesthetized rats (impairment; Almeida et al., 2015; Duarte et al., 2015a), indicating that anesthesia may affect the analysis of cardiovascular changes induced by chronic stressors.

Most studies describing stress effects evaluated the baroreflex activity using the classical pharmacological approach (i.e., arterial pressure changes were induced by intravenous infusion of vasoactive agents). However, analysis of the baroreflex responses over the physiological range of fluctuations in arterial pressure without any pharmacological manipulations (i.e., spontaneous baroreflex) has also provided evidence of an impact of chronic stressors, but the effects are stress-specific type. For example, changes in spontaneous baroreflex activity were reported in animals subjected to repeated exposure to social defeat (Sévoz-Couche et al., 2013) and CVS (Almeida et al., 2015), but not following RRS (Daubert et al., 2012; Duarte et al., 2015b). Therefore, an impact of RRS in baroreflex activity was evidenced by the classical pharmacological approach alone (Conti et al., 2001; Porter et al., 2004; Duarte et al., 2015a). Acute ablation of specific central nervous system regions has different effects on the baroreflex responses assessed by the classical pharmacological approach and the sequence analysis technique (Crestani et al., 2010a; de Andrade et al., 2014), indicating that differences in the neural circuitry of reflex responses within the narrow range of physiological variations and during more pronounced arterial pressure changes could explain the specific influence of RRS on the baroreflex responses over the full range of arterial pressure changes. Nevertheless, further studies are necessary to clarify this issue.

The studies that evaluated cardiovascular function in unanesthetized animals provided evidence that both CVS (Grippo et al., 2008; Almeida et al., 2015; Duarte et al., 2015a) and repeated social defeat stress (Sévoz-Couche et al., 2013) impaired the baroreflex function. Impairment of the baroreflex function is proposed to be involved in the physiopathology of hypertension (Grassi et al., 2006; Honzikova and Fiser, 2009), and is associated with overactivity of sympathetic activity (Grassi et al., 2004). Therefore, sympathoexcitation and mild hypertension (reported in some studies, **Table 1**) induced by these chronic stressors may be mediated by changes in the baroreflex activity. In contrast, an increase in baroreflex sensitivity was observed following exposure to RRS (Conti et al., 2001; Duarte et al., 2015a). Moreover, a recent study did not identify an impact of chronic social isolation in baroreflex function in adult rats (Cruz et al., 2016). Therefore, an involvement of baroreflex changes in the physiopathology of stress-induced cardiovascular complications is dependent on the chronic stress type.

## CARDIOVASCULAR RESPONSES TO STRESS: INFLUENCE OF PREDICTABILITY

Most studies that characterized the influence of predictability of stressor in its responses evaluated alterations in neuroendocrine function, behavioral responses, somatic parameters, and brain morphology/function. Indeed, a possible difference in the impact of predictable vs. unpredictable stressors in cardiovascular function and autonomic activity was addressed only recently (Duarte et al., 2015a,b). Therefore, before discussing the influence of predictability in cardiovascular/autonomic responses to stress, a summary of the impact of predictable vs. unpredictable stressors in the HPA axis, anxiety- and depression-like behaviors, and somatic parameters is presented in order to discuss experimental results in rodents that characterized the influence of predictability of stressor stimulus and its consequences.

## Characterization of Influence of Predictability

The impact of predictability of stressor stimulus was initially examined by comparing responses to signaled (e.g., aversive stimulus preceded by light) vs. unsignaled footshock. Some of these studies indicated that signaling minimized stress reactions (e.g., corticosterone release) and stress-induced pathology (e.g., stomach ulceration; Perkins, 1955; Seligman, 1968; Weiss, 1970), which supported the hypothesis that a reliable predictor of an aversive stimulus minimizes its responses (Seligman, 1968). However, these results were not consistent (Brady et al., 1962; Paré, 1971; Bassett et al., 1973), and the definition of predictability as identical with the presence or absence of a signal preceding a regularly occurring aversive stimulus was criticized (Bassett et al., 1973). Indeed, the predictability of a stressor in terms of time has been proposed as an important dimension of its consequences. For example, repeated stress applied varying the interval between each exposure (i.e., irregular) induced greater and more persistent increase in plasma corticosterone, fatty acid, and glucose levels than regular exposure (Bassett et al., 1973; Quirce et al., 1981; Smith et al., 2013). Increase in anxiety- and depression-like behaviors and somatic changes such as adrenal hypertrophy and reduction in body weight gain were also rather observed following stressor applied irregularly than regularly (Martí and Armario, 1997; Smith et al., 2013), besides some of these effects may be related to stressor intensity (Martí and Armario, 1997). Studies have also used protocols that vary the duration of each stress session, but with constant interval between the sessions (Rockman et al., 1987; Ortiz et al., 2015). Nevertheless, a comparison of the responses induced by this stress paradigm with those induced by a stressor applied regularly has never been investigated.

Studies have also evaluated the predictability by varying the type of stressor. These studies have compared the impact of daily exposure to the same stressor type (i.e., homotypic) vs. the exposure to different aversive stimuli (i.e., heterotypic). Typically, these studies have compared the effects of the RRS applied in a predictable schedule vs. the CVS, which is a widely used paradigm that involves daily exposure of rodents to different stressors at unpredictable times (Willnér, 2005; Grippo, 2009; Frisbee et al., 2015). In this regard, the CVS has been demonstrated to induce more severe somatic changes such as adrenal hypertrophy and thymus involution (Magariños and McEwen, 1995; Zucchi et al., 2009; Flak et al., 2012; Kopp et al., 2013; Duarte et al., 2015a), which is possibly related to an increase in baseline HPA axis activity observed mainly following exposure to heterotypic stressors (Magariños and McEwen, 1995; Marin et al., 2007). Increase in anxiety- and depression-like behaviors was also more severe following exposure to CVS than to RRS (Haile et al., 2001; Pastor-Ciurana et al., 2014; Yoon et al., 2014; Zhu et al., 2014; Gao et al., 2016). Although some studies report that CVS induces increased reduction in body weight gain (Marin et al., 2007; Gao et al., 2016), several studies have demonstrated that homotypic and heterotypic stressors similarly affect this parameter (Magariños and McEwen, 1995; Vyas et al., 2002; Flak et al., 2012; Yoon et al., 2014; Duarte et al., 2015a).

The lesser impact of homotypic stressors in neuroendocrine, behavioral, and some somatic parameters is possibly related to a habituation process upon repeated exposure to the same stressor, which is reduced during exposure to heterotypic stressors (Herman, 2013). This habituation process is mainly evidenced by a progressive reduction in HPA axis activation (Grissom and Bhatnagar, 2009). Indeed, adaptation to chronic stress determined by habituation of responses upon repeated stress exposure has been proposed, which limits the long-term impact of stress (Grissom and Bhatnagar, 2009; Herman, 2013). In this regard, the habituation process has been demonstrated to be impaired when homotypic stressors are applied irregularly (De Boer et al., 1989; Martí and Armario, 1997; Smith et al., 2013), which is in line with evidence discussed above wherein stressors applied irregularly induce more severe responses.

## Influence of Predictability in Cardiovascular Responses to Stress

Cardiovascular and autonomic changes have been reported following exposure to both predictable and unpredictable stressors. However, few studies to date have investigated the influence of predictability by directly comparing cardiovascular and autonomic changes induced by predictable vs. unpredictable stress protocols. De Boer et al. (1989) initially demonstrated that noise applied regularly, but not irregularly, increased plasma noradrenaline concentration, whereas noradrenaline response to noise was decreased in animals subjected to irregular protocol. Plasma adrenaline response to noise was reduced by both regular and irregular protocols (De Boer et al., 1989). These results do not support an influence of predictability in the sympatho-adrenomedullary response to stress. However, a possible influence of regular vs. irregular aversive stimuli on cardiovascular parameters, such as blood pressure and HR, has never been evaluated. Indeed, an impact of predictable vs. unpredictable stressors in cardiovascular function was addressed only recently. In this regard, Duarte et al. (2015a) compared the effect of the homotypic stressor RRS (predictable) vs. the heterotypic stressor CVS (unpredictable) on the baseline values of arterial pressure and HR, baroreflex function, and cardiac autonomic activity. They observed that both chronic stressors increased the baseline arterial pressure values (Duarte et al., 2015a). However, increase in baseline HR values and cardiac sympathetic activity as well as impaired baroreflex function was observed only in animals subjected to CVS (Duarte et al., 2015a). Although, these results provide evidence of a more severe impact of unpredictable vs. predictable stressors in cardiovascular function and autonomic activity, relevant cardiovascular changes were also detected following exposure to the predictable stressor RRS.

As stated above, adaptation to chronic stress determined by habituation of physiological responses upon repeated exposure to the same stressor has been proposed, which limits its long-term impact (Grissom and Bhatnagar, 2009; Herman, 2013). This idea has been supported mainly by consistent findings of habituation of HPA axis activation (Grissom and Bhatnagar, 2009). **Table 3** summarizes the studies that compared cardiovascular and autonomic responses during an acute stress session and after repeated exposure to the stressor. We note that studies have not consistently demonstrated a habituation of cardiovascular responses upon repeated exposure to a stressor (**Table 3**). For example, some findings indicated a decrease in pressor and tachycardiac response upon repeated exposure to restraint stress (Chen and Herbert, 1995; Bechtold et al., 2009), but several other studies reported similar cardiovascular responses to this stressor during both acute and repeated exposures (McDougall et al., 2000, 2005; Conti et al., 2001; Daubert et al., 2012). Analysis of cardiovascular and autonomic responses during repeated exposure to social defeat has also identified inconsistent results. Indeed, some results evidenced habituation (Adams et al., 1987; Chung et al., 1999; Costoli et al., 2004), whereas several other studies reported similar responses during repeated exposure to social defeat (Adams and Blizard, 1986; Meehan et al., 1995; Sgoifo et al., 2001, 2002; Carnevali et al., 2012, 2013b). Regardless of inconsistency of the results, evidence that cardiovascular responses do not readily habituate indicates a reduced process of adaptation (Grissom and Bhatnagar, 2009; Herman, 2013), which supports the findings of Duarte et al. (2015a) and other researchers (see above sections) demonstrating significant cardiovascular dysfunctions following exposure to predictable stressors (e.g., RRS and repeated social defeat).

Reasons for the discrepancy in findings of habituation of cardiovascular responses are unclear, but may be the result of methodological differences. For example, the protocol of one of the studies that identified habituation included some stress-free days (i.e., animals were subjected to restraint stress 5–7 days/week for 21–25 days; Bechtold et al., 2009), which is different from the studies that did not identify habituation to restraint stress wherein animals were continuously subjected to stress. In this regard, stress-free periods between periods of repeated stress exposure has been proposed to facilitate the habituation process (Grissom and Bhatnagar, 2009). Differences between species and strains may also contribute. For example, more consistent evidence of habituation to social defeat was observed in mice (Bartolomucci et al., 2003; Costoli et al., 2004), whereas similar responses upon repeated exposure to social defeat was observed in rats (Adams and Blizard, 1986; Meehan et al., 1995; Sgoifo et al., 2001, 2002; Carnevali et al., 2013b). A study using rats reported habituation upon repeated exposure to social defeat (Chung et al., 1999), whereas another report in mice did not identify habituation (Carnevali et al., 2012). However, strains of mice and rats used in these studies were different from those wherein a habituation process was observed in mice rather than in rats (**Table 3**), indicating that the strain may also affect the habituation process. Nevertheless, further studies are necessary to better address the habituation of the cardiovascular responses upon repeated exposure to the same stressor.

### CONCLUDING REMARKS

A significant increase in research using animal models on the impact of emotional stress in cardiovascular function and autonomic activity has occurred in last years. These studies


TABLE

3

*Bold indicates studies that identified reduction of cardiovascular*

\**Tail-cuff method of cardiovascular*

 *recording.*

 *and/or autonomic responses upon repeated exposure to the same stressor (i.e., habituation).*

to

the

stressor. have provided evidence that the impact of stress on the cardiovascular function is determined by stressor stimulus characteristics, such as chronicity and predictability. Regarding chronicity, changes in cardiovascular function and autonomic activity are well-documented during acute stress sessions. These responses constitute important short-term adaptive mechanisms, maintaining homeostasis and ensuring survival. However, enduring autonomic imbalance and cardiovascular dysfunctions are provoked mainly by long-term exposure to stressful events (i.e., chronic stress). In this sense, studies have provided consistent results regarding the effect of chronic emotional stress in baseline HR values, autonomic activity, and baroreflex function. Nevertheless, the chronic stress paradigm seems to be an important factor that determines the pattern of changes in these parameters. A detailed analysis of the effects of chronic stress in arterial pressure reveals inconsistent results; hence, an animal model of stress-induced hypertension is still missing.

Analysis of neuroendocrine and behavioral responses to stress has clearly demonstrated an influence of predictability of stressor in its effects. Nevertheless, only a limited number of studies compared the cardiovascular and autonomic changes following exposure to predictable vs. unpredictable stressors. These studies have not demonstrated a clear influence of predictability in cardiovascular dysfunctions induced by chronic stress. However, a possible influence of predictability in stress-induced cardiovascular complications is only beginning to be investigated, and a number of issues need to be addressed before more conclusive analysis can be performed. For example, adaptation to chronic stress determined by

### REFERENCES


habituation of physiological responses upon repeated exposure to the same stressor has been proposed (Grissom and Bhatnagar, 2009; Herman, 2013), which explains the more severe impact of homotypic vs. heterotypic stressors in HPA axis activity, behavioral responses, and somatic parameters. However, habituation of the cardiovascular and autonomic responses is still a matter of debate. Furthermore, evaluation of cardiovascular changes following exposure to homotypic stressors applied regularly vs. irregularly is a relevant unvalued aspect in determining the influence of predictability, because this approach compares responses with an aversive stimulus of similar intensity applied at predictable vs. unpredictable schedules. Therefore, further studies are necessary to elucidate the influence of predictability of stressor in its cardiovascular responses.

### AUTHOR CONTRIBUTIONS

CC designed, drafted, and revised the manuscript; and prepared the figures and tables.

### ACKNOWLEDGMENTS

The author wishes to acknowledge the financial support from FAPESP (grants # 2012/14376-0 and 2015/05922-9), CNPq (grant #456405/2014-3), and Programa de Apoio ao Desenvolvimento Científico da Faculdade de Ciências Farmacêuticas da UNESP – PADC. CC is a CNPq research fellow (process #305583/2015-8).

in an unpredictable chronic, mild stress model of depression in rats. Eur. J. Pharmacol. 710, 67–72. doi: 10.1016/j .ejphar.2013.04.007


Cannon, W. B. (1942). "Voodoo" death. Am. Anthropol. 44, 169–181.


role of corticotropin-releasing factor. Neuropeptides 35, 71–81. doi: 10.1054/npep.2001.0847


stressors in adolescent and adult rats. Psychosom. Med. 77, 186–199. doi: 10.1097/PSY.0000000000000141


behaviors. Biol. Psychiatry 62, 1162–1170. doi: 10.1016/j.biopsych.2007.0 4.011


**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Crestani. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Pacemaking Property of RVLM Presympathetic Neurons

Daniela Accorsi-Mendonça † , Melina P. da Silva † , George M. P. R. Souza, Ludmila Lima-Silveira, Marlusa Karlen-Amarante, Mateus R. Amorim, Carlos E. L. Almado, Davi J. A. Moraes and Benedito H. Machado\*

*Department of Physiology, School of Medicine of Ribeirão Preto, University of São Paulo, São Paulo, Brazil*

Despite several studies describing the electrophysiological properties of RVLM presympathetic neurons, there is no consensus in the literature about their pacemaking property, mainly due to different experimental approaches used for recordings of neuronal intrinsic properties. In this review we are presenting a historical retrospective about the pioneering studies and their controversies on the intrinsic electrophysiological property of auto-depolarization of these cells in conjunction with recent studies from our laboratory documenting that RVLM presympathetic neurons present pacemaking capacity. We also discuss whether increased sympathetic activity observed in animal models of neurogenic hypertension (CIH and SHR) are dependent on changes in the intrinsic electrophysiological properties of these cells or due to changes in modulatory inputs from neurons of the respiratory network. We also highlight the key role of INaP as the major current contributing to the pacemaking property of RVLM presympathetic neurons.

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Simon McMullan, Macquarie University, Australia Julie Yu-Hwa Chan, Kaohsiung Chang Gung Memorial Hospital, Taiwan*

### \*Correspondence:

*Benedito H. Machado bhmachad@fmrp.usp.br*

> *† Joint first authors.*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *25 June 2016* Accepted: *07 September 2016* Published: *22 September 2016*

### Citation:

*Accorsi-Mendonça D, da Silva MP, Souza GMPR, Lima-Silveira L, Karlen-Amarante M, Amorim MR, Almado CEL, Moraes DJA and Machado BH (2016) Pacemaking Property of RVLM Presympathetic Neurons. Front. Physiol. 7:424. doi: 10.3389/fphys.2016.00424* Keywords: neurogenic hypertension, sympathetic activity, presympathetic neurons

## RVLM AND SYMPATHETIC OUTFLOW

Presympathetic neurons located in the rostral ventrolateral medulla (RVLM) are responsible for generating the sympathetic drive to the cardiovascular system and ultimately determine cardiac output and vascular resistance (Dampney, 1994). Original studies by Owsjannikow and Dittmar, from Carl Ludwig's laboratory, suggested the presence of a vasomotor center in the medulla (apud Seller, 1996). These authors performed controlled lesions in the brainstem and simultaneously recorded arterial pressure. After several precise anteroposterior transections in the brain axis, they observed that a small area in the ventrolateral medulla was critical to keep the baseline arterial pressure and identified this region as a vasomotor area (apud Seller, 1996). These findings by the Germans physiologists, in the second half of the 19th century, were the first description of the pressor area and contributed to the identification of spinally projecting sympatho-excitatory neurons.

Additional evidence about the relevance of RVLM in the maintenance of baseline arterial pressure was provided in a study by Guertzenstein and Silver (1974), in which they demonstrated that bilateral inhibition of specific areas in the ventral medulla, using inhibitory amino acid glycine, produced a large fall in the arterial blood pressure, similar to that described by Dittmar after medullo-spinal transections. The role of RVLM in controlling the cardiovascular function was also described in a study by Granata et al. (1983), which reinforced the concept of a key region in the medullary surface for the maintenance of arterial blood pressure. Moreover, RVLM activation by either electrical stimulation or application of excitatory amino acid (glutamate) or even RVLM disinhibition by application of GABA receptor antagonist (bicuculline), in anesthetized or conscious animals, elicited an increase in sympathetic activity and arterial blood pressure (Willette et al., 1983; Reis et al., 1984; Ross et al., 1984a; de Paula and Machado, 2000; Sakima et al., 2000; Moraes et al., 2011), while bilateral electrolytic lesions, microinjection of GABA or administration of tetrodotoxin, leads to a large fall in the arterial pressure to levels comparable to those observed after transection below brainstem (Dampney and Moon, 1980; Willette et al., 1983; Reis et al., 1984; Benarroch et al., 1986).

Fine anatomical studies by Amendt et al. (1979) and Ross et al. (1981, 1984a) using retrograde transport and immunocytochemical technique demonstrated that RVLM neurons project directly to the thoracic spinal cord, where preganglionic sympathetic neurons are located. Studies by Ross et al. (1984a,b) also demonstrated that terminals of RVLM neurons contain phenylethanolamine-N-methyl transferase (PNMT), characterizing these neurons as C1 adrenalinesynthesizing neurons. Photostimulation of this neuronal phenotype, using lentivirus that expresses channelrhodopsin-2, increased the sympathetic nerve activity and arterial blood pressure in rats in vivo confirming the involvement of these cells in the cardiovascular regulation (Abbott et al., 2009).

The direct neural projection from RVLM neurons to spinal cord was also electrophysiologically characterized using antidromic stimulation (Barman and Gebber, 1985; McAllen, 1986; Morrison et al., 1988). In addition to their spinally projection, RVLM neurons are also characterized by the reduction in their firing frequency or silence completely in face of baroreflex stimulation (Schreihofer and Guyenet, 1997). Therefore, it is very well documented that there are presympathetic neurons in RVLM and that their integrity is essential to maintain the level of sympathetic activity and, consequently, the baseline levels of arterial blood pressure.

### ELECTROPHYSIOLOGICAL CHARACTERIZATION OF PRESYMPATHETIC NEURONS

After anatomical and functional evidence that cell bodies of presympathetic neurons were located at RVLM (Amendt et al., 1979; Ross et al., 1981, 1984a; Barman and Gebber, 1985; Morrison et al., 1988) several studies were performed to evaluate their pacemaker activity. Initially, studies using anesthetized animals described that RVLM neurons presented tonic activity, a state of continuous discharge of action potential, and their firing was highly synchronized with the sympathetic nerve discharge, the arterial pulse and respiration (Barman and Gebber, 1985; Haselton and Guyenet, 1989; Granata and Kitai, 1992). There is also experimental evidence that the firing frequency of RVLM presympathetic neurons is modulated by the afferents inputs from the arterial baroreceptors (Barman and Gebber, 1985; McAllen, 1986; Granata and Kitai, 1992).

A very important study by Sun et al. (1988a) considered two theories to explain the tonic activity of RVLM presympathetic neurons observed in anesthetized animals: (1) the pacemaker theory, suggesting that these neurons have intrinsic capacity to generate rhythmic activity and (2) the network theory suggesting that the activity of these neurons is dependent on the balance of tonic excitatory and inhibitory synaptic inputs arising from other brain regions. Although, different studies have documented the presence of excitatory and inhibitory inputs to RVLM neurons (Brown and Guyenet, 1985; Cravo and Morrison, 1993; Dampney, 1994; Schreihofer et al., 2000; Schreihofer and Guyenet, 2002; Gao and Derbenev, 2013), the main issue about these cells was related to their pacemaking capacity. More recently, it was documented that glial cells are also involved in the control of arterial pressure, since selective stimulation of RVLM astrocytes, using optogenetic approach, induced ATP release, depolarization of the presympathetic neurons with consequent increase in the sympathetic nerve activity and arterial pressure (Marina et al., 2013). Therefore, the controversy about the capacity the RVLM neurons generate spontaneous and rhythmic activity persisted by several years.

In this context, Sun et al. (1988a) provided evidence supporting the concept that RVLM neurons, under experimental conditions in which synaptic activity is low, are pacemakers. These authors using anesthetized adult rats and in vitro experiments (bloc of vascularly perfused bulb), reduced the excitatory neurotransmission using glutamate-receptor antagonist (kynurenic acid) and recorded the firing frequency of RVLM neurons using extracellular recordings. Intracisternal injection of kynurenic acid increased the firing frequency of functionally identified barosensitive neurons. On the other hand, several studies documented that microinjections of kynurenic acid into RVLM produced no major changes in the sympathetic nerve activity (Sun and Guyenet, 1987; Kiely and Gordon, 1994; Araujo et al., 1999). In a subsequent study, Sun et al. (1988a) suggested that glutamatergic receptor antagonist may also reduce the neuronal activity in CVLM, which sends inhibitory inputs to RVLM neurons, as demonstrated previously by Willette et al. (1984). Studies performed under the effect of kynurenic acid, Sun et al. (1988a) showed that the majority of synaptic inputs to RVLM presympathetic neurons are reduced and that rhythmic firing pattern observed in these cells using extracellular recordings was due to pacemaker activity. Therefore, based on these experiments Sun et al. (1988a) suggested that presympathetic RVLM neurons have intrinsic pacemaker properties. In their study they stated: "The final proof of the pacemaker theory will have to await the result of intracellular recording experiments."

In order to explore in further detail the possible pacemaker activity of RVLM presympathetic neurons, Sun et al. (1988b) performed intracellular recordings in brainstem slices of young adult rats. They described that RVLM neurons display a typical pacemaker membrane potential trajectory with no evidence of excitatory synaptic inputs. However, in their study it was not used any pharmacological tool to exclude possible inputs from neuronal network. They ruled out neuronal network involvement in the generation of the regular firing frequency of these neurons because no excitatory post-synaptic potentials (EPSPs) were observed in response to intracellular hyperpolarizing currents. Moreover, an important methodological advancement of this study was the use of a dye to identify spinal cord-projecting RVLM neurons, since the RVLM region is functionally, anatomically as well as chemically, heterogeneous. For this purpose, Sun et al. (1988b) performed injection of rhodamine microbeads into the spinal cord (T3 level) in anesthetized animal and few days later they visualized labeled cells on the slices of the ventral medulla. Taking the advantage of retrogradely identified cells and intracellular recordings, Sun et al. (1988b) confirmed that the RVLM presympathetic neurons present electrophysiological properties of auto-depolarization, i.e., characteristics of pacemaker neurons in accordance with the following criteria: (1) pacemaker firing frequency, (2) tonic discharge of at least 4 spikes per second, (3) loss of pacemaker activity during hyperpolarization around −80 mV, (4) absence of detectable EPSPs even during hyperpolarization. However, it is important to mention that the intracellular recordings may damage the neuronal membrane during the penetration of pipette into the cell, producing a leak current, depolarized resting potential (Li et al., 2004) and inactivation of voltage-dependent sodium channels with a consequent decrease in the frequency discharge (Staley et al., 1992).

After these important studies by Sun et al. (1988a,b), several others from different laboratories tried to identify the presence of pacemaker activity in RVLM neurons using different experimental approaches, such as whole-cell patch clamp technique. This approach is more appropriate for recording neuronal activity since it produces less damage to the membrane of recorded cell and provides more information about the intrinsic properties of neurons, such as ionic conductances related to the firing frequency, which is not feasible using extracellular records. In this context, Kangrga and Loewy (1995) performed experiments to analyze the membrane potential of these cells, using brainstem slices from neonatal rats, retrogradely labeled neurons and whole-cell patch clamp. They identified two types of labeled RVLM neurons: pacemaker and non-pacemaker neurons. The pacemaker cells were classified according to the following criteria: the regenerative spontaneous firing frequency at a constant rate and membrane potential trajectory presenting gradual depolarizing interspike ramps. Kangrga and Loewy (1995) also suggested that the intrinsic tonic firing frequency of RVLM neurons was due to the pacemaker activity and not due to the synaptic inputs. However, it is important to note that two factors make the interpretation of these experiments difficult: (1) identification of the intrinsic properties, since no pharmacological antagonism was used to isolate the recorded cell from the neuronal network and, (2) the results obtained using neonatal rats may be different from those observed in adult animals, since the density distribution of ionic channels in neurons is established in the brain development period between P17 and P19 (Beckh et al., 1989; Straka et al., 2005), and the expression and functional properties of several receptors involved in the synaptic transmission may also change during the development (Ben-Ari, 2002; Luján et al., 2005).

Studies by Lipski et al. (1996) also attempted to shed light on the controversy about the pacemaker activity of RVLM neurons and the possible role of the neural network, studying the activity of RVLM neurons using intracellular recordings in anesthetized adult rats. In their work, Lipski et al. (1996) identified RVLM presympathetic neurons by 2 criteria: (1) inhibition of neuronal activity after stimulation of the aortic depressor nerve and (2) antidromic responses evoked by stimulation of RVLM bulbospinal axons. In contrast to the intrinsic pacemaker properties as previously suggested by Sun et al. (1988a,b) and Kangrga and Loewy (1995), the findings by Lipski et al. (1996) indicated that the spontaneous firing frequency in RVLM presympathetic neurons results from synaptic inputs based on the following evidence: (1) action potentials were normally preceded by depolarizing potentials showing features of fast EPSPs and (2) there was no evidence of regular, ramp-like depolarization between action potentials. Therefore, the results by Lipski et al. (1996) raised again new questions about the "pacemaker" activity of RVLM presympathetic neurons and brought for discussion the network theory related to the generation of action potential in RVLM neurons.

In another study, Lipski et al. (1998) evaluated spontaneous firing frequency in acutely dissociated retrogradely labeled RVLM neurons from neonatal rats (13- to 19-days old), in which all cell-to-cell interactions were eliminated. Using wholecell patch clamp they verified that these cells presented no pacemaker activity and suggested that the depolarization of these neurons observed in the whole animal was dependent on the influence of inhibitory and excitatory tonic projections from different neuronal networks in the brainstem, such as excitatory projections from the nucleus of tractus solitarius to CVLM neurons, which in turn send monosynaptic inhibitory projections to RVLM presympathetic neurons (Agarwal and Calaresu, 1991).

Although, the pacemaker property of RVLM presympathetic neurons were evaluated in several studies (Sun et al., 1988a,b; Kangrga and Loewy, 1995; Lipski et al., 1996, 1998) there was no consensus about their spontaneous activity, probably due to different experimental conditions, such as the age of animal, presence of anesthesia and different electrophysiological methods used to recordings the neuronal activity (**Figure 1**). In addition, it is important to note that the electrophysiological record of RVLM neurons from juvenile and adult rats is not a simple task mainly due to the high degree of technical difficulties in recording neurons in the ventral medulla, a brainstem area presenting high density of myelin.

In our laboratory, using an in situ preparations of juvenile rats (P30-P31), we observed that all RVLM presympathetic neurons fire spontaneously and the frequency of these cells were heterogeneous ranging from 8 to 22 Hz, with depolarized (−52 mV) and hyperpolarized (−63 mV) values of membrane potential (Moraes et al., 2013). Moreover, we observed that respiratory network modulates the activity of RVLM presympathetic neurons, which allow us to classify them into four types, being three of them modulated by respiratory activity. These neurons receive multiple synaptic inputs, observed by a high level of synaptic "noise," and their baseline firing frequency was probably determined by the balance of excitatory and inhibitory inputs, as observed previously in anesthetized rats by Lipski et al. (1996). However, after pharmacological blockade of fast synaptic transmission in the in situ preparation, we clearly verified the intrinsic properties of auto-depolarization in RVLM neurons, confirming their pacemaker properties (**Figure 2**,

Moraes et al., 2013). Due to this discrepancy in the firing rate and resting membrane potential, we combined single cell RT-qPCR and immunohistochemistry to characterize the neurochemical profile of these neurons and we observed that all respiratorymodulated RVLM presympathetic neurons investigated were glutamatergic neurons. However, the expression of tyrosine hydroxylase was detected in the inspiratory-modulated and nonrespiratory modulated RVLM presympathetic neurons, but not in the post-inspiratory modulated neurons, pointing out to the existence of different subpopulations of RVLM presympathetic pacemaker neurons (Moraes et al., 2013).

We also performed experiments designed to record retrogradely labeled RVLM presympathetic neurons in brainstem slices preparations from juvenile-adult animals (P35, Almado et al., 2014). Ten days after the surgical procedures to retrogradely label these cells, whole-cell recordings in brainstem slices revealed that RVLM neurons are under synaptic modulation and presented a regular and spontaneous firing frequency, which are in agreement with our findings in the in situ preparation. Moreover, after blockade of fast synaptic transmission, the firing frequency of RVLM neurons decreased significantly but their activity was not abolished, indicating that these cells have intrinsic properties required to auto-depolarization. Thus, RVLM presympathetic neurons in slices from juvenile/adults rats also behave as pacemakers under our experimental condition. Therefore, our studies performed in the in situ preparation, as well as in the brainstem slices from juvenile-adults rats, support the concept that RVLM presympathetic neurons are indeed pacemakers (**Figure 2**).

## RVLM NEURONS AND NEUROGENIC HYPERTENSION

Several studies have suggested that changes in the intrinsic properties of RVLM neurons are the main cause of cardiovascular disorders, such as neurogenic hypertension, which is characterized by the chronic increase of the arterial blood pressure mediated by sympathetic overactivity rather than vascular and renal dysfunctions (Han et al., 1998; Guyenet, 2006; Toney et al., 2010; Kumagai et al., 2012).

Considering that experimental models of neurogenic hypertension, such as rats submitted to chronic intermittent hypoxia (CIH) and spontaneously hypertensive (SH) rats, show a significant increase in sympathetic tone, more recently we became directly involved with this important issue. Our studies were designed to analyze whether RVLM neurons, from juvenile-adult animals, present an enhancement in their spontaneous firing frequency and whether this enhancement is

(2014, Copyright License Number: 3894220907506).

responsible for the sympathetic overactivity and hypertension observed in CIH and SH rats. To reach these goals, we used in situ, as well as in vitro preparations. Firstly, we performed blind whole cell patch clamp recordings of RVLM neurons, using in situ preparations of juvenile rats (Paton, 1996). This preparation has the advantage of being anesthesia-free with intact brainstem circuits, while the lack of pulsatility makes the brain amenable to whole-cell recordings (Moraes et al., 2013). Although, we have observed an increase in the firing frequency of RVLM presympathetic neurons from CIH rats in the late-expiratory phase of the respiratory cycle (late-E), the blockade of fast synaptic transmission revealed similar intrinsic firing frequency, membrane potential, input resistance as well as intrinsic excitability when compared with RVLM presympathetic neurons from control rats. These important findings show that the sympathetic overactivity observed in this model of neurogenic hypertension is not due to changes in the intrinsic properties of RVLM presympathetic neurons. Therefore, these cells are not in charge of sympathetic overactivity observed in CIH rats. Furthermore, in the intact respiratory and sympathetic brainstem networks, respiratorymodulated RVLM presympathetic neurons from SH rats revealed an increase in their activity, also in the late-expiratory phase, when compared with those neurons from normotensive rats. It is important to highlight that after synaptic blockade, the pacemaking capacity of RVLM presympathetic neurons was similar in either control, CIH or SH rats, indicating clearly that their increased firing frequency during the lateexpiratory phase was driven by excitatory synaptic inputs

from neurons of the respiratory network (Moraes et al., 2013, 2014).

In a series of experiments performed in slices, we also analyzed the effects of CIH on the electrophysiological properties of RVLM presympathetic neurons. As described for the in situ approach, in the presence of synaptic blockade RVLM neurons from CIH rats presented no changes in their resting membrane potential, firing frequency and input resistance. Thus, the intrinsic properties of RVLM presympathetic neurons in brainstem slices from juvenile rats exposed to CIH were similar to those observed in neurons from control rats, as we observed in in situ preparations of CIH and SH rats. Although, our findings indicate that RVLM presympathetic neurons are pacemakers, our studies using two experimental models of neurogenic hypertension (CIH and SH) also indicate that the increased firing frequency of these cells is not due to changes in their intrinsic properties (**Figure 3**), but is associated with changes in their modulation by synaptic inputs from the respiratory network.

Taken together, our data, from in situ and in vitro preparations, allow us to consider that: (1) RVLM presympathetic neurons have intrinsic mechanism that allow them to behave as pacemaker neurons; and (2) the increased respiratory modulation to these neurons is mainly due to excitatory drives (Moraes et al., 2013, 2014), supporting the hypothesis that sympathetic overactivity, present in different models of neurogenic hypertension, might involve changes in the neurons from the respiratory network; the observed changes in these respiratory network neurons seems to increase the excitatory inputs to the RVLM presympathetic neurons (Czyzyk-Krzeska and Trzebski, 1990; Moraes et al., 2013, 2014).

## RVLM PRESYMPATHETIC NEURONS: CONDUCTANCE AND ROLES IN THE AUTO-DEPOLARIZATION

Although, changes in the intrinsic electrophysiological properties of RVLM presympathetic neurons are not the cause of neurogenic hypertension in CIH and SH rats, a considerable amount of effort has been committed to understanding the mechanisms underlying the ability of these neurons to autodepolarize. Studies by Lipski et al. (1998) using retrograde labeled isolated RVLM neurons demonstrated that these cells express high and low voltage-activated calcium channels. These type of channels presents several functionalities including: (1) neurotransmission, (2) activation of calcium-dependent potassium channels, and (3) neuronal excitability control (Llinás, 1988, 2014; Catterall, 2011). Furthermore, low voltage-activated Ca2<sup>+</sup> channels have been implicated in the auto-depolarization of RVLM presympathetic neurons during short periods of hypoxia (Sun and Reis, 1994a). However, Kangrga and Loewy (1995), using brainstem slices observed that only in 2 out 13 RVLM presympathetic neurons tested, the application of CdCl2, a broad-spectrum calcium channel blocker, abolished the spontaneous firing frequency of cells, while the majority showed a significant enhancement. These results by Kangrga and Loewy (1995) suggested that some RVLM neurons may require Ca <sup>2</sup><sup>+</sup> influx and/or synaptic drive for regenerative firing and are in agreement with studies by Sun and Reis (1994a) demonstrating that increases in the firing frequency of RVLM neurons are synchronized with the rapid increase in Ca2<sup>+</sup> channel conductance.

In our laboratory, we also investigated the role of calcium channels in the auto-depolarization behavior of RVLM presympathetic neurons. Using in situ preparations, we observed that Ni2+, a blocker of type T calcium channels, did not eliminate the activity of neurons, but increased their firing frequency (Moraes et al., 2013). However, we believe that type T calcium channels may play an indirect role in the RVLM neurons activity by stimulating calcium-activated potassium channel (BKCa channels), which in turn may influence the firing frequency of neurons as suggested by Pierrefiche et al. (1995). This suggestion is supported by a slight increase in the action potential duration and the marked decrease in the amplitude and duration of after-hyperpolarization observed previously by Kangrga and Loewy (1995). All together these studies revealed that calcium currents seems to be involved, but are not the main conductance in the intrinsic auto-depolarization observed in the majority of RVLM presympathetic neurons.

A second conductance that seems to be involved in the auto-depolarization of RVLM presympathetic neurons is related to voltage-dependent potassium channels. Previously, it was demonstrated in neonate rats that RVLM presympathetic neurons express a variety of voltage-dependent K<sup>+</sup> channels (Kangrga and Loewy, 1995; Li et al., 1995). Therefore, this previous information leads us to investigate whether these channels could drive the intrinsic activity of RVLM presympathetic neurons. We documented that RVLM presympathetic neurons show a delay during the depolarizing phase of action potentials generation. In this case, the current that underlies this delayed excitation seems to be similar to transient potassium current. However, when we blocked this conductance using 4-aminopyridine, it resulted in a decrease in after-hyperpolarization amplitude and an increase in the firing frequency. These findings highlight the contribution of this conductance to the action potential kinetics, but not for the auto-depolarization characteristic of RVLM presympathetic neurons (Moraes et al., 2013).

Considering auto-depolarization as a possible summation of a set of smaller conductances operating at membrane potentials just below the spike threshold, we also explored the presence and contribution of such conductances to RVLM presympathetic neurons, mainly those involved in pacemaker activity. The first to be investigated was a current originated by hyperpolarizationactivated cyclic-nucleotide-gated channels or HCN channels (McCormick and Pape, 1990; Wahl-Schott and Biel, 2009) that is responsible for keeping the resting potential near to the threshold value. Although, HCN channels strongly modulates spontaneous discharge of several cells (Gu et al., 2005; Rodrigues and Oertel, 2006; Kase and Imoto, 2012), it does not seem to be essential for auto-depolarization of RVLM presympathetic neurons and consequently for blood pressure control, since ZD7288, a specific blocker of these channels, did not change the excitability of RVLM presympathetic neurons, sympathetic activity or mean arterial pressure (Miyawaki et al., 2003; Moraes et al., 2013; Tallapragada et al., 2016).

Another conductance potentially involved in intrinsic autodepolarization of RVLM neurons is the Na<sup>+</sup> conductance resistant to TTX, or the INaP currents. INaP has been implicated in the regulation of subthreshold excitability in a variety of excitable cells (French and Gage, 1985; Stafstrom et al., 1985). Computational modeling has shown that INaP is involved with spontaneous excitability due to its contribution for afterhyperpolarization phase increasing the cellular excitability by reducing threshold, and also by increasing the discharge frequency in response to depolarizing current (Vervaeke et al., 2006).

Studies by Kangrga and Loewy (1995) demonstrated that the spontaneous firing of RVLM presympathetic neurons is due to a cellular mechanism that is fully dependent on INaP. In a recent study from our laboratory, we obtained similar findings, since a blocker of sodium channels responsible for INaP (riluzole) abolished the spontaneous activity of these cells (Moraes et al., 2013) revealing the conductance responsible for the intrinsic auto-depolarization of the RVLM presympathetic neurons. Therefore, the INaP current is essential for the spontaneous activity of these neurons (**Figure 4**).

Although, our previous studies documented the absence of changes in the intrinsic properties of RVLM presympathetic neurons in neurogenic hypertensive models, we cannot ignore that under some conditions, their intrinsic properties may be altered. There are some studies reporting the

of calcium and potassium channels increase the intrinsic firing frequency of RVLM presympathetic neurons while the blockade of TTX-resistant sodium channel abolished the spontaneous activity of these neurons.

possibility that these neurons present chemosensitivity, especially to hypoxia (Sun et al., 1992; Sun and Reis, 1993, 1994a,b; Wang et al., 2001; Koganezawa and Terui, 2007; Koganezawa and Paton, 2014). Therefore, it is possible that RVLM presympathetic neurons present a detection system of brainstem hypoperfusion/ischaemia through specific membrane conductances. It has been emphasized that hypertension can be produced in response to brain hypoperfusion, but the mechanism for detecting this brain condition remains poorly understood (Paton et al., 2009; Cates et al., 2011). We suggest that RVLM presympathetic neurons can switch from synaptically modulated firing frequency to almost pure pacemaking-driven discharge, in a reversible way, during severe hypercapnic/hypoxia, such as gasping and it may represent the last physiological strategic response to increase sympathetic activity and to the survival of the animals under these challenges. It is also important to note that previous studies demonstrated the involvement of intrinsic membrane conductance such as potassium and

### REFERENCES

Abbott, S. B., Stornetta, R. L., Socolovsky, C. S., and West, G. H., Guyenet, P. G. (2009). Photostimulation of channelrhodopsin-2 expressing ventrolateral medullary neurons increases sympathetic nerve activity and blood pressure in rats. J. Physiol. 587(Pt 23), 5613–5631. doi: 10.1113/jphysiol.2009.1 77535

calcium currents as well as persistent sodium currents for the intrinsic response of RVLM presympathetic neurons exposed to hypoxia (Sun and Reis, 1994a; Koganezawa and Paton, 2014). The contribution of different intrinsic membrane conductances for generating changes in pacemaker RVLM presympathetic neuronal activity in hypoperfusion/ischemia awaits additional experiments and it is a critical step for our understanding of the electrophysiological complexity of the RVLM presympathetic neurons under physiological challenges.

## CONCLUSION

In this review we described the functional characteristics of RVLM presympathetic neurons and discussed their intrinsic capacity to auto-depolarize and work as pacemakers, a controversial concept in the recent past. Experimental evidence that these neurons are not responsible for sympathetic overactivity observed in models of neurogenic hypertension, such as CIH and SH rats were also discussed. In addition, we highlighted that the main cause of the increased frequency discharge of RVLM presympathetic neurons in these experimental models of neurogenic hypertension is likely related to changes in synaptic inputs from the respiratory network.

## AUTHOR CONTRIBUTIONS

BM, coordinate the group of Ph.D. students, post-docs and young faculty to write this review, wrote several parts of the manuscript and revised the final version; DM, wrote several parts of the manuscript and revised the final version; DA, revised the literature, wrote the manuscript, organized the figures and the revised the final version; MS, revised the literature, wrote the manuscript, organized the figures and the revised the final version; GS, wrote some sections of the manuscript and revised the final version; LS, wrote some sections of the manuscript and revised the final version; MK, wrote some sections of the manuscript and revised the final version; MA, wrote some sections of the manuscript and revised the final version; CA, wrote some sections of the manuscript and revised the final version.

## FUNDING

The experiments and results from our laboratory presented in this review were part of the Thematic Project funded by FAPESP (2013/06077-5).


network in rats. J. Neurosci. 33, 19233–19237. doi: 10.1523/JNEUROSCI.3041- 13.2013


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Accorsi-Mendonça, da Silva, Souza, Lima-Silveira, Karlen-Amarante, Amorim, Almado, Moraes and Machado. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Corrigendum: Pacemaking Property of RVLM Presympathetic Neurons

Daniela Accorsi-Mendonça, Melina P. da Silva, George M. P. R. Souza, Ludmila Lima-Silveira, Marlusa Karlen-Amarante, Mateus R. Amorim, Carlos E. L. Almado, Davi J. A. Moraes and Benedito H. Machado\*

*Department of Physiology, School of Medicine of Ribeirão Preto, University of São Paulo, São Paulo, Brazil*

Keywords: neurogenic hypertension, sympathetic activity, presympathetic neurons

### **A corrigendum on**

this article in any way.

### **Pacemaking Property of RVLM Presympathetic Neurons**

### Edited and reviewed by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

> \*Correspondence: *Benedito H. Machado bhmachad@fmrp.usp.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *09 November 2016* Accepted: *10 November 2016* Published: *23 November 2016*

### Citation:

*Accorsi-Mendonça D, da Silva MP, Souza GMPR, Lima-Silveira L, Karlen-Amarante M, Amorim MR, Almado CEL, Moraes DJA and Machado BH (2016) Corrigendum: Pacemaking Property of RVLM Presympathetic Neurons. Front. Physiol. 7:575. doi: 10.3389/fphys.2016.00575* M., Amorim, M. R., et al. (2016). Front. Physiol. 7:424. doi: 10.3389/fphys.2016.00424 Due to an oversight, the authors did not properly cite two important publications by Roger A. Dampney. In section "RVLM and sympathetic outflow," the second paragraph should read as

by Accorsi-Mendonça, D., da Silva, M. P., Souza, G. M. P. R., Lima-Silveira, L., Karlen-Amarante,

follows:

Additional evidence about the relevance of RVLM in the maintenance of baseline arterial pressure was provided in a study by Guertzenstein and Silver (1974), in which they demonstrated that bilateral inhibition of specific areas in the ventral medulla, using inhibitory amino acid glycine, produced a large fall in the arterial blood pressure, similar to that described by Dittmar after medullo-spinal transections. Equally important were the contributions by Dampney (1981) and Dampney et al. (1982), which original studies documented that microinjections of L-glutamate into the ventral medulla increased arterial pressure in anesthetized rabbits. The role of RVLM in controlling the cardiovascular function was also described in a study by Granata et al. (1983), which reinforced the concept of a key region in the medullary surface for the maintenance of arterial blood pressure. Moreover, RVLM activation by either electrical stimulation or application of excitatory amino acid (glutamate) or even RVLM disinhibition by application of GABA receptor antagonist (bicuculline), in anesthetized or conscious animals, elicited an increase in sympathetic activity and arterial blood pressure (Willette et al., 1983; Reis et al., 1984; Ross et al., 1984a; de Paula and Machado, 2000; Sakima et al., 2000; Moraes et al., 2011), while bilateral electrolytic lesions, microinjection of GABA or administration of tetrodotoxin, leads to a large fall in the arterial pressure to levels comparable to those observed after transection below brainstem (Dampney and Moon, 1980; Willette et al., 1983; Reis et al., 1984; Benarroch et al., 1986). The authors apologize for this oversight. This error does not affect the scientific conclusions of

## REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Accorsi-Mendonça, da Silva, Souza, Lima-Silveira, Karlen-Amarante, Amorim, Almado, Moraes and Machado. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Swimming Training Modulates Nitric Oxide-Glutamate Interaction in the Rostral Ventrolateral Medulla in Normotensive Conscious Rats

Hiviny de A. Raquel <sup>1</sup> , Gustavo S. Masson<sup>2</sup> , Barbara Falquetto Barna<sup>2</sup> , Nágela G. Zanluqui <sup>3</sup> , Phileno Pinge-Filho<sup>3</sup> , Lisete C. Michelini <sup>2</sup> and Marli C. Martins-Pinge<sup>1</sup> \*

*<sup>1</sup> Department of Physiological Sciences, Center of Biological Sciences, State University of Londrina, Londrina, Brazil, <sup>2</sup> Department of Physiology & Biophysics, Institute of Biomedical Sciences, University of São Paulo, São Paulo, Brazil, <sup>3</sup> Department of Pathological Sciences, Center of Biological Sciences, State University of Londrina, Londrina, Brazil*

We evaluated the effects of swimming training on nitric oxide (NO) modulation to glutamate microinjection within the rostral ventrolateral medulla (RVLM) in conscious freely moving rats. Male Wistar rats were submitted to exercise training (Tr) by swimming or kept sedentary (Sed) for 4 weeks. After the last training session, RVLM guide cannulas and arterial/venous catheters were chronically implanted. Arterial pressure (AP), heart rate (HR), and baroreflex control of HR (loading/unloading of baroreceptors) were recorded in conscious rats at rest. Pressor response to L-glutamate in the RVLM was compared before and after blockade of local nitric oxide (NO) production. In other Tr and Sed groups, brain was harvested for gene (qRT-PCR) and protein (immunohistochemistry) expression of NO synthase (NOS) isoforms and measurement of NO content (nitrite assay) within the RVLM. Trained rats exhibited resting bradycardia (average reduction of 9%), increased baroreflex gain (Tr: −4.41 ± 0.5 vs. Sed: −2.42 ± 0.31 b/min/mmHg), and unchanged resting MAP. The pressor response to glutamate was smaller in the Tr group (32 ± 4 vs. 53 ± 2 mmHg, *p* < 0.05); this difference disappeared after RVLM pretreatment with carboxy-PTIO (NO scavenger), Nw-Propyl-L-Arginine and L-NAME (NOS inhibitors). eNOS immunoreactivity observed mainly in RVLM capillaries was higher in Tr, but eNOS gene expression was reduced. nNOS gene and protein expression was slightly reduced (−29 and −9%, respectively, *P* > 0.05). Also, RVLM NO levels were significantly reduced in Tr (−63% vs. Sed). After microinjection of a NO-donor, the attenuated pressor response of L-glutamate in Tr group was restored. Data indicate that swimming training by decreasing RVLM NO availability and glutamatergic neurotransmission to locally administered glutamate may contribute to decreased sympathetic activity in trained subjects.

Keywords: baroreflex, RVLM, nitric oxide synthase, heart rate, arterial pressure

### INTRODUCTION

Aerobic training induces adaptations in central autonomic areas involved in the control of the cardiovascular system (Ichiyama et al., 2002; Martins-Pinge, 2011). Such changes modify the parasympathetic and sympathetic outflow to heart and vessels, with robust changes in the peripheral sympathetic activity (Mitchell and Victor, 1996). In this sense, the rostral ventrolateral

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Zhihong Yang, University of Fribourg, Switzerland Luciana A. Campos, Universidade Camilo Castelo Branco, Brazil*

\*Correspondence:

*Marli C. Martins-Pinge martinspinge@uel.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *11 April 2016* Accepted: *27 May 2016* Published: *13 June 2016*

### Citation:

*Raquel HA, Masson GS, Barna BF, Zanluqui NG, Pinge-Filho P, Michelini LC and Martins-Pinge MC (2016) Swimming Training Modulates Nitric Oxide-Glutamate Interaction in the Rostral Ventrolateral Medulla in Normotensive Conscious Rats. Front. Physiol. 7:221. doi: 10.3389/fphys.2016.00221* medulla (RVLM), the main sympathetic output for heart and blood vessels (Dampney, 1994), emerges as a potent target for modulatory action of exercise on autonomic control. However, only few studies have evaluated functional RVLM changes in animals previously submitted to exercise training (Becker et al., 2005; Martins-Pinge et al., 2005; Mueller, 2007; Ogihara et al., 2014).

RVLM premotor neurons driving the excitatory tone to the sympathetic preganglionic neurons are mainly glutamatergic (Guertzenstein and Silver, 1974; Dampney, 1994). Indeed microinjections of L-glutamate in the RVLM produce pressor response in both anesthetized (Willette et al., 1987) and awake rats (Bachelard et al., 1990; Martins-Pinge et al., 2005). In addition, studies have shown increased release of glutamate within the RVLM during static muscle contractions (Caringi et al., 1998; Lillaney et al., 1999; Ishide et al., 2003), suggesting this neurotransmitter is directly involved in the exercise pressor reflex (Ally, 1998).

Interestingly, we previously observed smaller pressor responses to RVLM glutamate administration in conscious rats previously submitted to a protocol of swimming training (Martins-Pinge et al., 2005). Experimental evidence indicated that exercise, through flow-induced shear stress, increases nitric oxide (NO) production to cause local vasodilation (Green et al., 2004; McAllister and Laughlin, 2006). It was also demonstrated the presence of NO synthase (NOS) isoforms in the RVLM (Chan et al., 2001), which, under physiological conditions are able to activate the local synthesis of NO. It has been proposed that NO is produced during the activation of NMDA receptors, suggesting its involvement in the activation of glutamatergic pathways (Wu et al., 2001). Indeed the involvement of NO in glutamatergic neurotransmission within the RVLM was previously reported (Martins-Pinge et al., 1999). However, the production of NO in the RVLM as well as its potential effects on the autonomic control of the circulation following exercise training has not been investigated yet.

Knowing that RVLM neurons express NOS isoforms and that NO has a functional role in glutamatergic neurons (Dampney, 1994), we hypothesized that exercise would change RVLM NO availability, thus contributing to the small pressor response observed in swimming-trained rats. Therefore, in the present study we analyzed the pressor response to glutamate microinjection in the RVLM in sedentary and swimming-trained rats before and after pretreatment with NOS blockers and NO scavenger within the RVLM. In addition, we compared gene and protein expression of eNOS and nNOS as well as NO availability in the RVLM of sedentary and trained rats. We also performed L-glutamate microinjections in RVLM of sedentary and trained rats previously treated with a NO-donor.

## MATERIALS AND METHODS

### Animals

Adult male Wistar rats, weighing 220–240 g at the beginning of protocols were used. They were housed at the Central Animal Facility of the State University of Londrina, Brazil at controlled room temperature (22 ± 1 ◦C) with a 12-h darklight cycle and free access to standard chow and water. All surgical and experimental protocols were in accordance and recommendations of Brazilian National Council for Animal Experimentation Control (CONCEA) and approved by the Ethics Committe of the State University of Londrina, Brazil (process number: 35247.2011.45).

## Exercise Training Protocol

The animals were randomly allocated to two groups: trained group (Tr) submitted to swimming training and sedentary group (Sed), which was not submitted to the exercise protocol. The swimming training, according to Martins-Pinge et al. (2005), was conducted between 11:00 AM and 1:00 PM in a glass tank (4000 cm<sup>2</sup> of surface area, 60 cm deep) with water heated to 31 ± 1 ◦C. The training protocol consisted of 4 weeks of swimming being carried out 60 min per day, 5 times a week. During the first week the animals swam 15 min on the 1st day, 30 min the 2nd day, 45 min on the 3rd day and 60 min from the 4th day on.

## Guide Cannula Implantation in the RVLM

One day after the last exercise session the rats were anesthetized with sodium pentobarbital (40 mg/kg, ip) and underwent stereotaxic surgery for implantation of guide cannulas directed to RVLM. Rats were placed in the stereotaxic apparatus (David Kopf) with the incisor bar 5 mm below the interaural line, according to Martins-Pinge et al. (1997). At the end of this procedure, the animals received a prophylactic dose of antibiotic (40.000 IU) and returned to the Animal Facility for 3 days for surgical recovery.

### Artery and Vein Catheterization and Cardiovascular Recordings

Twenty-four hours before the experiments, the rats were again anesthetized (tribromoethanol, 250 mg/kg, ip) for implantation of catheters in the femoral artery and vein for arterial pressure (AP) and heart rate (HR) recordings and drugs administration, respectively. The arterial cannula was attached to a pressure transducer (Model MLT0380, Powerlab) connected to a computerized system (Powerlab, AD Instruments) and ∼30 min were allowed for adaptation to the environment (individual cage in a quiet room). Baseline AP and HR were continuously recorded in conscious freely-moving rats for 30 min.

### Baroreflex Testing

Baroreflex function was analyzed by loading/unloading of baroreceptors with intravenous bolus injections (100µL) of phenylephrine (0.1 up to 12.8µg/kg) and sodium nitroprusside (0.2 up to 25.6µg/kg). Subsequent injections were not made until the returning of MAP and HR to basal values. MAP and HR values were measured immediately before (control) and at the peak of each response. Baroreceptor reflex control of HR, determined for each rat, was estimated by the sigmoidal logistic equation fitted to data points, as described previously (Kent et al., 1972; Head and Mccarty, 1987). The equation linking HR responses to pressure changes was: HR = P1+P2/[1+eP3(BP– P4)], where P1 = lower HR plateau, P2 = HR range, P3 = the curvature coefficient and P4 = BP50 (the value of MAP at half of the HR range). The average gain of baroreflex function (BrS) was also calculated (Kent et al., 1972; Head and Mccarty, 1987). Baroreflex testing was performed in groups of trained and sedentary rats without guide cannula implantation.

### RVLM Microinjections

The protocol consisted of initial unilateral microinjection of L-glutamate (5 nmol/100 nL). After microinjection, MAP and HR responses were followed for 5–10 min recovery interval for returning of cardiovascular parameters to baseline values. Then, RVLM was treated with one of the following drugs: saline 0.9% (Sed: n = 4; Tr: n = 5) or NO scavenger, Carboxy-PTIO (1 nmol/100 nL) (Sed: n = 8; Tr: n = 10) or the nNOS inhibitor Nw-Propyl-L-Arginine (4 nmol/100 nL) (Sed: n = 8; Tr: n = 10) or the unspecific NOS inhibitor L-NAME (15 nm/100 nL) (Sed: n = 8; Tr: n = 7). RVLM L-glutamate (5 nmol/100 nL) microinjection was then repeated and MAP and HR responses were followed for 5–10 min up to the return of cardiovascular parameters to baseline values.

Another protocol consisted of previous treatment of RVLM with DeaNonoate (an NO-donor, 50 nmol/100 nL) followed by microinjection of L-glutamate in the RVLM of sedentary and trained animals (Sed: n = 4; Tr: n = 4).

The concentrations of drug used were based on the following literatures: L-glutamate, and L-NAME:. Martins-Pinge et al. (2007); carboxy-PTIO and N-Propyl-L-arginine: Busnardo et al. (2010). Dea-NONOate: Yao et al. (2007). At the end of each experimental protocol, the animals were euthanized with an extra dose of anesthetic and then held marking procedures of microinjection sites and removal of the brain for subsequent histological analysis.

### Confirmation of RVLM Microinjections

At the end of the experimental protocols, the rats were euthanized with an overdose of sodium pentobarbital. Sites of RVLM administrations were marked by microinjection of Evans blue dye (2%/100 nL). Brain was removed and stored in 10% formaldehyde for subsequent histological analysis. Sequential slices (40µm) of brainstem were cut in a cryostat, placed in gelatinized slides and stained with 1% neutral red. The sections were examined microscopically with the aid of a rat brain atlas (Paxinos, 1998). Only rats with confirmed RVLM microinjection were included in experimental groups (see **Figure 2F**).

## Tissue Harvesting for qPCR and Immunohistochemistry Assays

Gene and protein expression in the RVLM were analyzed in other groups of sedentary and trained rats not submitted to RVLM cannulation. At the end of experimental protocols, rats were deeply anesthetized (60 mg/kg pentobarbital i.p) and the brains perfused with phosphate-buffered saline (PBS 0.1 M, pH 7.4, ∼30 mL/min for 4–5 min, via a left ventricle cannula) immediately after the respiratory arrest (Cavalleri et al., 2011). In 8–10 rats/group, fresh brains were rapidly removed and frozen in a dry ice box. Bilateral punches of RVLM were obtained from frozen brain stem sections (rostral to the Obex, 1000– 1200µm of thickness) and stored in a deep freezer in individual eppendorfs with 1 mL Trizol <sup>R</sup> until processing. In the remaining 3–4 rats/group, after the initial perfusion with PBS, brains were fixed with 4% paraformaldehyde (PFA, 30 mL/min for 20– 30 min). Brains were removed from the skull, post-fixed in 4% PFA for 24–48 h. Series of coronal sections (40µm) from brain stem were cut using the Leica-CM3050 cryostat and stored in a cryoprotectant solution (20% glycerol plus 30% ethylene glycol in 50 mM phosphate buffer, pH 7.4, −20◦C) for up to 2 weeks until histological processing (Schreihofer and Guyenet, 1997).

### Real-Time qPCR

RVLM mRNA expression was estimated in Tr and Sed by the real-time qPCR. The total RNA was extracted using the Trizol <sup>R</sup> , dissolved in 10µL of DEPC water and stored at -80◦C. Then, the samples were treated with DNAse I for cDNA synthesis by reverse transcription (Revert Aid TMM-MuLV Reverse Transcriptase) according to the protocol supplied by the manufacturer. The cDNA obtained was then stored at −20◦C. The samples were subjected to amplification by Real Time qPCR method using Platinum SYBRGreen qPCR Supermix-UDG (Cavalleri et al., 2011) and specific primers for the two NOS isoforms: eNOS (Gene Bank: NM\_021838.2/Fragment Size: 94pb, sense primer: GCCAAACAGGCCTGGCGCAA, antisense primer: GTGCTGTCCTGCAGTCCCGA) and nNOS (Gene Bank: NM\_0522799.1/Fragment Size: 118pb, sense primer: CGCTACGCGGGCTACAAGCA, antisense primer: GCACGTCGAAGCGGCCTCTT). mRNA expression was estimated by semi-quantitative real time PCR (7500 Real-Time PCR System). The specificity of the SYBRGreen assays was confirmed by analysis of the melting points of the curves. The endogenous gene was the hypoxanthine guanine phosphoribosyl transferase—HPRT (Gene bank: NM\_012583.2/Fragment Size: 125pb), which is continuously expressed in all cells of the body and not altered by physical training (Cavalleri et al., 2011). Analysis of gene expression was made by the geNorm software VBA applet for Microsoft Excel, considering the values of threshold cycle (Ct) and the 11Ct method (Cavalleri et al., 2011). The results were expressed as fold increase. All reagents and primers were purchased from Invitrogen (San Diego, CA, USA).

### Immunohistochemistry

Endothelial nitric oxide syntase (eNOS) and neuronal nitric oxide syntase (nNOS) immunoreactivities were detected in sequential brain stem slices using mouse anti-eNOS/NOS Type III antibody (1:200, BD Transduction Laboratories) and mouse anti-nNOS (1:200, BD Transduction Laboratories), respectively, as previously described (Llewellyn-Smith et al., 2005; Barna et al., 2012). Biotin-SP-conjugated AffiniPure Donkey Anti-Mouse IgG (H+L) (1:500, Jackson Immuno Research Laboratories immunoperoxidase assay) was used as the secondary antibody. Brain stem slices were mounted in sequential rostrocaudal order; slides were dried and covered with Krystalon (EMD Chemicals Inc, NJ). Brain sections were analyzed in a Zeiss Axioskop 2 microscope (Oberkochen, Germany) to check the location of neurons and vessels marked. RVLM neurons immunoreactive to nNOS and vessels immunoreactive to eNOS were identified and quantified by a blind investigator. The images from both experimental groups were digitized with identical acquisition settings. Image analysis was performed with Image J software (Wright Cell Imaging Facility—Toronto Western Research Institute, ON). An automated tracing procedure that incorporated the threshold paradigm was applied the acquired images. The background intensity was calculated from random adjacent areas in the RVLM. ROIs of predetermined sizes were used to determine the density of eNOS threshold signal. Values for each area per rat were averaged to obtain the mean value for each experimental group.

### Nitrite Levels in RVLM

Indirect NO concentration in RVLM was estimated in punches obtained from sedentary (n = 10) and trained (n = 8) rats through the measurement of nitrite as described previously (Navarro-Gonzálvez et al., 1998; Panis et al., 2011). Group Samples (Sed: 3.6 ± 0.3 and Tr: 3.5 ± 0.4 mg), keeping the concentration of 100 mg of wet weight tissue per milliliter of PBS, were used. All reagents for the nitrite assay were obtained from Sigma Chemical Co. The results were expressed in uM of nitrite/mg of RVLM tissue.

## Statistical Analysis

All data are reported as mean ± SEM. Nitrite concentration, gene and protein expression in Gtr and Gsed, baroreflex sensitivity and MAP and HR responses determined by RVLM drugs microinjections in both groups were compared by Student T test or paired "T" test as appropriate. Differences between groups were analyzed by one-way ANOVA followed by Newman-Keuls as the post hoc test. Differences were considered significant when P<0.05.

## RESULTS

In all groups of rats studied, swimming training was accompanied by resting bradycardia (average reduction of ∼9%, when compared to respective sedentary groups, **Table 1**). In addition we observed improved baroreceptor reflex control of HR (**Figure 1A**) and increased baroreflex gain (**Figure 1B**) in the trained animals compared to sedentary controls (Gtr: −4.41 ± 0.5 vs. Gsed: −2.42 ± 0.31 b/min/mmHg, P < 0.05). These responses confirmed the efficacy of exercise training to improve cardiovascular control. Also, as observed in **Table 1**, swimming training did not change baseline MAP in the normotensive groups of rats.

Accordingly with previous data (Martins-Pinge et al., 2005), the L-glutamate administrations in RVLM elicited

TABLE 1 | Resting values of mean arterial pressure (MAP) and heart rate (HR) and basal values of MAP before the first and the second Glutamate microinjections in sedentary and swimming trained groups treated with Saline, Carboxi-PTIO, Nw-Propyl-L-Arginine, and L-NAME within the RVLM.


*(*\**p* ≤ *0.05 compare to corresponding sedentary group; Student t-test).*

marked pressure increases that were significantly reduced by swimming training, and the same was observed in all groups analyzed here before the different treatments (**Figure 2**). In sedentary rats, the pressor response to L-Glu in the RVLM before and after local microinjections of saline were, respectively, 53 ± 2 mmHg (**Figure 2A**, open bar) and 52 ± 6 mmHg (**Figure 2A**, dark bar) and, in the trained rats were 35 ± 2 mmHg (**Figure 2A**, open bar) and 32 ± 2 mmHg (**Figure 2A**, dark bar). However, previous microinjection of Carboxi-PTIO canceled the differences in the pressor responses to L-glutamate between sedentary and trained animals (**Figure 2B**). The RVLM previous treatment with Nw-Propyl-L-Arginine, a nNOS inhibitor, caused the same results (**Figure 2C**). A similar observation was made after L-NAME (a nonselective NOS inhibitor) administration in the RVLM, in which, the pressor responses to L-Glu in the RVLM were not different between sedentary and trained rats (**Figure 2D**).

As depicted in the map of the ventral surface of the medulla (**Figure 2E**) and in a coronal section of the brain stem (**Figure 2F**), dye injection at the end of experiments confirmed that all microinjections were directed to the RVLM.

The gene expression of NOS isoforms within the RVLM was also analyzed in both groups. There was a marked reduction of RVLM eNOS expression in the trained group (from 1.20 ± 0.25 in Sed to 0.41 ± 0.08 in Tr, P < 0.05, **Figure 3C**) and a small decrease in nNOS mRNA expression that did not attain significance (Sed: 1.46 ± 0.42 vs. Tr: 1.04 ± 0.27, P > 0.05, **Figure 3B**). Immunohistochemistry for eNOS in the RVLM revealed a "blood vessels pattern," confirming the presence of eNOS mainly in the endothelium of capillaries within the RVLM (**Figures 4B,B1,C,C1**). In agreement with the increased capillary supply observed in brain areas of trained animals (Dunn et al., 2012; Huang et al., 2013) quantitative analysis showed that trained rats exhibited increased eNOS immunoreactivity when compared to sedentary controls (Sed:

2.3 ± 0.2 vs. Tr: 11.7 ± 0.2, P < 0.05, **Figure 4A**). On the other hand, nNOS immunoreactivity in the RVLM was present essentially in neuronal cell bodies (**Figures 4E,F**). There was a slight training-induced reduction in RVLM nNOS positive neurons (-9.4%), but values did not attain significance (Sed: 3.2 ± 0.2 vs. Tr: 2.9 ± 0.1, P > 0.05, **Figure 4D**). Interestingly, the comparison of nitrite concentration indicated lower NO availability within the RVLM of trained rats (Sed: 9.1 ± 2.0 vs. Tr: 3.4 ± 0.4µM of nitrite/mg of tissue, a reduction of 63%, P < 0.05, **Figure 3A**).

Considering that NO seems to be decreased in RVLM of trained rats, we performed L-glutamate microinjection in TR and Sed rats after treatment with DeaNonoate (**Figure 5A**). After the NO-donor, L-glutamate pressor responses were increased in trained rats (Sed: 45.29 ± 5.76 vs. Tr: 65.03 ± 5.24, P > 0.05, **Figure 5B**).

### DISCUSSION

First of all, since several studies in the literature described increased baroreflex sensitivity after aerobic training (Brum et al., 2000; Medeiros et al., 2004; Ceroni et al., 2009; Cavalleri et al., 2011; Masson et al., 2014), we also investigated the ability of swimming training to alter baroreceptor reflex control of HR. We observed that 4 weeks of swimming training markedly improved baroreflex gain and increased the operational range of the reflex. Training-induced improvement of baroreflex sensitivity by providing a better ability to correct instantaneous pressure oscillations reduces pressure variability and consequent capillary lesions thus improving cardiovascular homeostasis. This study also confirmed previous observations in the literature that swimming training did not change pressure levels of normotensive rats, but induced resting bradycardia (Medeiros et al., 2004; Mehanna et al., 2007; Mastelari et al., 2011; Sant'Ana et al., 2011).

In the present study, some new observations were obtained with the swimming training protocol of 4 weeks: (1) similar to other training protocols, swimming training was also able to improve the baroreceptor reflex control of HR; (2) the lower training-induced RVLM NO content significantly contributes to the lower pressor response to glutamate since the difference between trained and sedentary rats disappear after the local administration of NO scavenger or NOS inhibitors; (3) reduced NO availability in trained rats may decrease RVLM glutamatergic activity, thus reducing glutamate-induced sympathoexcitation; (4) both isoforms are able to release NO in the RVLM, nNOS and eNOS seems to be involved in the reduced NO content during RVLM glutamate administration. (5) After adding a NO-donor, the attenuated pressor response to L-glutamate in Tr group was restored. Together these data indicate that swimming training, although not changing basal pressor levels, is able to refrain glutamate-stimulated pressure increases by reducing RVLM NO modulation of sympathetic activity.

In trained rats, the lower pressor response to glutamate administration in the RVLM confirmed previous data from our laboratory (Martins-Pinge et al., 2005). In addition, the present results by comparing in the same rats the pressure responsiveness to glutamate before and after the blockade of NO availability (by either NO scavenger and blockade of NOS isoforms), showed that the smaller pressor response exhibited by trained rats was due to a low NO content in the RVLM. Indeed the measurement of nitrite content within this area confirmed the reduced NO availability after swimming training. On the other hand, pressor responsiveness was not changed in sedentary rats submitted to the same RVLM treatments. Since the difference in the pressor response was only observed in trained rats before treatments, was not present after NO withdrawal in the trained group and was not significantly affect by NO removal in sedentary rats, we suggest that training abrogated pressor responsiveness to glutamate by decreasing NO release and its excitatory effects on glutamatergic neurons. We also confirm this hypothesis when after adding a NO-donor, the increase in MAP by L-glutamate was restored.

Regarding the effects of NO in the RVLM, there are controversies in the literature. Some investigators reported blood

pressure increases after administration of L-arginine and NO donors (Hirooka et al., 1996; Martins-Pinge et al., 1999), while others observed a significant decrease (Shapoval et al., 1991; Zanzinger et al., 1995; Tseng et al., 1996; Kagiyama et al., 1997). It is possible that these contradictory effects are grounded in different concentrations used by researchers: while high NO doses in the RVLM lead to decreases, lower doses produce arterial pressure increases (Morimoto et al., 2000). It is important to note that pharmacological studies in the central nervous system of different species showed that NO may interact with both the glutamatergic excitatory and GABAergic inhibitory neurons thus being able to cause neuronal excitation or neuronal inhibition (Tseng et al., 1996; Chen et al., 2001; Ishide et al., 2003; Martins-Pinge et al., 2013). However, considering the predominance of sympathetic premotor neurons in the RVLM that are glutamatergic (Dampney, 1994; Mischel et al., 2015), and the observations that blockade of endogenous NO release by NO scavengers in the RLVM was accompanied by hypotension and bradycardia (Chan et al., 2001) and reduced pressor response (present set of data) we may suggest that NO within the RVLM modulates preferentially the sympathoexcitation mediated by glutamatergic neurons. In addition, RVLM blockade of NOS isoforms by both Nw-Propyl-L-Arginine and L-NAME was accompanied by smaller pressor response to locally administered glutamate. These data together with previous studies showing MAP, HR, and renal sympathetic nerve activity reductions after NOS blockade in the RVLM (Hirooka et al., 1996; Chan et al., 2003; Martins-Pinge et al., 2007) indicate an excitatory effect of locally released NO that is blunted by swimming training.

Both isoforms are able to synthesize NO when properly stimulated by the increased neuronal activity (nNOS) or by the augmented shear stress (eNOS). Our data showed that trained rats exhibited a huge increase in eNOS protein expression and a possible compensatory downregulation of eNOS gene expression due to the increased capillary profile observed after training (Dunn et al., 2012; Huang et al., 2013). Notice that RVLM glutamate injection was made when trained and sedentary rats are resting in their home cages, therefore not exhibiting a hyperkinetic circulation to stimulate eNOS via increased shear stress. On the other hand the slight reduction in nNOS expression (gene and protein) observed in the RVLM of trained rats may account for the reduced NO availability upon neuronal activation by glutamate administration. This does not preclude additional activation of eNOS and endothelial NO production during acute bouts of exercise. Indeed Ishide et al. (2003, 2005) showed that

both, nNOS and eNOS are involved in RVLM NO synthesis during muscle contractions. Importantly, our data showed that 4 weeks of swimming training caused a marked reduction of NO availability in the RVLM, as measured by nitrite concentration. The ability of swimming training to reduce both the increased NOS expression and the elevated NO synthesis observed in hypertensive rats was recently demonstrated in the RVLM of 2K-1C trained rats (Sousa et al., 2015).

In summary, our data indicate that swimming training decreases RVLM NO availability, therefore reducing glutamatergic activity in sympathetic premotor neurons and the stimulated pressor response. The present set of data demonstrates an important modulatory role of RVLM glutamatergic neurons by locally released NO in exercise trained subjects.

## AUTHOR NOTE

The present study evaluated glutamate and nitric oxide interactions in the RVLM of normotensive rats previously submitted to swimming training. The functional studies were performed in conscious rats, avoiding anesthesia influence on neural function, focusing on glutamate effects in RVLM and the contributions of eNOS and nNOS isoforms and its cardiovascular control. We observed that exercise training decreased nitric oxide production in RVLM, collaborating to decrease sympathetic activity in trained subjects. Also, no studies have been evaluating those aspects in conscious rats.

## AUTHOR CONTRIBUTIONS

Conception and design: HR, LM, MM. Acquisition, analysis and interpretation of data: HR, GM, BB, NZ,. Analyzed the data: HR, GM, BB, NZ, PP, LM, MM. Materials and reagents: PP, LM, MM. Drafting or revising and final approval: HR, GM, BB, NZ, PP, LM, MM.

## ACKNOWLEDGMENTS

The authors thank the Fundo de Auxílio ao Ensino, Pesquisa e Extensão da Universidade Estadual de Londrina (FAEPE), Brazil; Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES); Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP proc. 2011/51410-9); and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for PP, LM, and MM research fellowship.

### REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Raquel, Masson, Barna, Zanluqui, Pinge-Filho, Michelini and Martins-Pinge. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Blockade of Rostral Ventrolateral Medulla (RVLM) Bombesin Receptor Type 1 Decreases Blood Pressure and Sympathetic Activity in Anesthetized Spontaneously Hypertensive Rats

Edited by: *Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Simon McMullan, Macquarie University, Australia Josiane Campos Cruz, Federal University of Paraíba, Brazil*

### \*Correspondence:

*Gustavo R. Pedrino pedrino@pq.cnpq.br; gpedrino@gmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *11 February 2016* Accepted: *20 May 2016* Published: *02 June 2016*

### Citation:

*Pinto IS, Mourão AA, da Silva EF, Camargo AS, Marques SM, Gomes KP, Fajemiroye JO, da Silva Reis AA, Rebelo ACS, Ferreira-Neto ML, Rosa DA, Freiria-Oliveira AH, Castro CH, Colombari E, Colugnati DB and Pedrino GR (2016) Blockade of Rostral Ventrolateral Medulla (RVLM) Bombesin Receptor Type 1 Decreases Blood Pressure and Sympathetic Activity in Anesthetized Spontaneously Hypertensive Rats. Front. Physiol. 7:205. doi: 10.3389/fphys.2016.00205* Izabella S. Pinto<sup>1</sup> , Aline A. Mourão<sup>1</sup> , Elaine F. da Silva<sup>1</sup> , Amanda S. Camargo<sup>1</sup> , Stefanne M. Marques <sup>1</sup> , Karina P. Gomes <sup>1</sup> , James O. Fajemiroye<sup>2</sup> , Angela A. da Silva Reis <sup>3</sup> , Ana C. S. Rebelo<sup>4</sup> , Marcos L. Ferreira-Neto<sup>5</sup> , Daniel A. Rosa<sup>1</sup> , André H. Freiria-Oliveira<sup>1</sup> , Carlos H. Castro<sup>1</sup> , Eduardo Colombari <sup>6</sup> , Diego B. Colugnati <sup>1</sup> and Gustavo R. Pedrino<sup>1</sup> \*

*<sup>1</sup> Department of Physiological Sciences, Center for Neuroscience and Cardiovascular Research, Biological Sciences Institute, Federal University of Goiás, Goiânia, Brazil, <sup>2</sup> Postgraduate Programme in Pharmaceutical Sciences, Faculty of Pharmacy, Federal University of Goiás, Goiânia, Brazil, <sup>3</sup> Department of Biochemistry and Molecular Biology, Biological Sciences Institute, Federal University of Goiás, Goiânia, Brazil, <sup>4</sup> Department of Morphology, Biological Sciences Institute, Federal University of Goiás, Goiânia, Brazil, <sup>5</sup> Department of Physiology, College of Physical Education, Federal University of Uberlândia, Uberlândia, Brazil, <sup>6</sup> Department of Physiology and Pathology, School of Dentistry, São Paulo State University, Araraquara, Brazil*

Intrathecal injection of bombesin (BBS) promoted hypertensive and sympathoexcitatory effects in normotensive (NT) rats. However, the involvement of rostral ventrolateral medulla (RVLM) in these responses is still unclear. In the present study, we investigated: (1) the effects of BBS injected bilaterally into RVLM on cardiorespiratory and sympathetic activity in NT and spontaneously hypertensive rats (SHR); (2) the contribution of RVLM BBS type 1 receptors (BB1) to the maintenance of hypertension in SHR. Urethane-anesthetized rats (1.2 g · kg−<sup>1</sup> , i.v.) were instrumented to record mean arterial pressure (MAP), diaphragm (DIA) motor, and renal sympathetic nerve activity (RSNA). In NT rats and SHR, BBS (0.3 mM) nanoinjected into RVLM increased MAP (33.9 ± 6.6 and 37.1 ± 4.5 mmHg, respectively; *p* < 0.05) and RSNA (97.8 ± 12.9 and 84.5 ± 18.1%, respectively; *p* < 0.05). In SHR, BBS also increased DIA burst amplitude (115.3 ± 22.7%; *p* < 0.05). BB<sup>1</sup> receptors antagonist (BIM-23127; 3 mM) reduced MAP (–19.9 ± 4.4 mmHg; *p* < 0.05) and RSNA (−17.7 ± 3.8%; *p* < 0.05) in SHR, but not in NT rats (−2.5 ± 2.8 mmHg; −2.7 ± 5.6%, respectively). These results show that BBS can evoke sympathoexcitatory and pressor responses by activating RVLM BB<sup>1</sup> receptors. This pathway might be involved in the maintenance of high levels of arterial blood pressure in SHR.

Keywords: rostral ventrolateral medulla, bombesin, BB<sup>1</sup> receptors, BIM-23127, SHR

## INTRODUCTION

Bombesin (BBS), a tetradecapeptide isolated from the skin of the frog Bombina bombina (Anastasi et al., 1971), have shown broad spectrum of biological activities (Brown, 1983; Gonzalez et al., 2008; Jensen et al., 2008). The BBS activates three G protein-coupled receptors: bombesin receptor 1 (BB1), bombesin receptor 2 (BB2), and bombesin receptor 3 (BB3). BBS-like peptides—Neuromedin B (NB) and gastrin releasing peptide (GRP) are natural ligand of the BB<sup>1</sup> and BB<sup>2</sup> receptors, respectively (Jensen et al., 2008). Natural agonist of the BB<sup>3</sup> receptor still remains unknown. However, it seems that BB<sup>3</sup> receptor plays an important physiological role, since BB<sup>3</sup> receptor knockout mice developed obesity associated with hypertension and impairment of glucose metabolism (Ohki-Hamazaki et al., 1997). In human, the BB<sup>1</sup> receptor gene is at chromosome 6p21-pter (Jensen et al., 2008). The BB<sup>1</sup> receptor signal occurs primarily through phospholipase-C-mediated cascades, that involve activation of Gqα protein and consequent stimulation of protein kinase C (Jensen et al., 2008).

In mammals, BBS receptors and BBS-like peptides are distributed in the Central Nervous System (CNS) (Woodruff et al., 1996; Jensen et al., 2008) including regions involved in the cardiorespiratory control (Chung et al., 1989; Lynn et al., 1996; Li et al., 2016). The administration of BBS has been reported to enhancebreathing (Holtman et al., 1983; Glazkova and Inyushkin, 2006), raise plasma concentration of catecholamine (Brown and Fisher, 1984), tachycardia (Zogovic and Pilowsky, 2011), increase blood pressure (Brown, 1983; Zogovic and Pilowsky, 2011), and sympathetic tone (Zogovic and Pilowsky, 2011) in normotensive (NT) rats. Zogovic and Pilowsky (2011) showed that intrathecal injection of BBS is associated with sympathoexcitatory and pressor responses. In the same study, the authors also reported that the administration of an antagonist of BBS receptor 2 attenuated the effects of BBS on blood pressure of NT rats. However, the involvement of the rostral ventral medulla (RVLM) in the BBS-induced cardiorespiratory and autonomic responses as well as the contribution of the BBS receptor type 1 to the maintenance of blood pressure in NT rats and spontaneously hypertensive rats (SHR) is still unclear.

The RVLM contains neurons that regulate peripheral sympathetic vasomotor tone and blood pressure (Guertzenstein, 1973; Guertzenstein and Silver, 1974; Guyenet et al., 1989; Guyenet, 2006; Toney and Stocker, 2010). The RVLM is localized ventral to the rostral part of the nucleus ambigus (NA), caudal to the facial nucleus and ventral to the Bötzinger complex. The RVLM neurons project to the sympathetic preganglionic neurons located in the intermediolateral (IML) cell column of the spinal cord (Loewy, 1981; Millhorn and Eldridge, 1986; Guyenet, 2006). The neurons projecting from RVLM could modulate peripheral sympathetic activity to the kidneys, vessels, heart, and adrenal gland.

The hyperactivity of RVLM neurons has been implicated in the maintenance of hypertension in different experimental models (Yang et al., 1995; Fink, 1997; Ito et al., 2000, 2003; Matsuura et al., 2002; Adams et al., 2007; Stocker et al., 2007; Toney and Stocker, 2010). Previous studies have shown that the injection of excitatory amino acid (EAA) antagonist into the RVLM reduced arterial pressure in SHR but not in NT rats (Ito et al., 2000). In addition, the electrophysiological studies have shown that firing rate of RVLM neurons is significantly faster in neonatal and adult SHR than NT rats (Chan et al., 1991; Matsuura et al., 2002). These findings indicate that hyperactivity of the RVLM neurons could contribute to the development and maintenance of hypertension in SHR.

Hence, we hypothesized whether or not the injection of BBS into RVLM affects cardiorespiratory and sympathetic activities in NT rats and SHR. In order to test this hypothesis, the recording of the mean arterial pressure (MAP), renal sympathetic nerve activity (RSNA) and diaphragm (DIA) motor activity were carried out to evaluate changes induced by unilateral and bilateral injection of BBS into RVLM of urethane-anesthetized NT rats and SHR. In addition, the contribution of tonic activation of BBS receptor 1 in the maintenance of high levels of arterial pressure in SHR was also evaluated.

## METHODS

## Animals and Ethical Approval

Male Wistar NT rats and SHR weighing 250 to 330 g were used. Animals were housed in a temperature-controlled room (22– 24◦C) with a 12:12-h light-dark cycle (lights on at 07:00), free access to food, and tap water. All experimental procedures were designed in strict adherence to the National Health Institute Guidelines for Care and Use of Laboratory Animals, and approved by Ethics Committee for Animal Care and Use (CEUA) of the Federal University of Goiás (number of ethical committee: 025/12).

### Surgical Procedures

Rats were anesthetized with halothane (2–3% in O2; Tanohalo; Cristália, SP, Brazil). The right femoral vein and artery were catheterized for drug administration and blood pressure recording, respectively. After catheterization, anesthesia was maintained by intravenous administration of urethane (1.2 g · kg−<sup>1</sup> b.wt.; Sigma–Aldrich, St. Louis, MO, USA). The trachea was cannulated to reduce airway resistance. Bipolar stainless steel electrodes were implanted in the DIA muscle for electromyography (EMG) recording of inspiratory motor activity. Rats were later mounted prone in a stereotaxic apparatus for craniotomy and instrumented for the recording of RSNA. The body temperature of rats was maintained at 37 ± 0.5◦C with a thermostatically controlled heated table.

### Recording of Cardiorespiratory Parameters

In order to record the arterial pressure, the arterial catheter was connected to a pressure transducer which is coupled to an amplifier (Bridge Amp FE221; ADInstruments, Colorado Springs, CO, USA). The pulsatile pressure was recorded continuously with a data acquisition system (PowerLab; ADInstruments, Colorado Springs, CO, USA). The MAP was calculated from the pulsatile signal using the LabChart software (v.7.3.7, ADInstruments, Colorado Springs, CO, USA). Analogical signals of the electrocardiogram (ECG), obtained through electrodes positioned in the forelimbs, were amplified 1000 times and filtered between 100 and 1000 Hz (Bridge Amp; ADInstruments, Colorado Springs, CO, USA). The heart rate (HR) was calculated as instantaneous frequency of the ECG signal (LabChart v.7.3.7, ADInstruments, Colorado Springs, CO, USA). The DIA motor activity signals was amplified 10,000 times (Bridge Amp; ADInstruments, Colorado Springs, CO, USA) and band-pass filtered (100–2000 Hz). The signal were rectified and integrated in 50 ms intervals using LabChart software (v.7.3.7; ADInstruments, Colorado Springs, CO, USA). The DIA motor activity was evaluated by burst amplitude (expressed as percentage difference from baseline) and frequency (considered as respiratory frequency, fR, and expressed in cycles per minute, cpm).

## Recording of Renal Sympathetic Nerve Activity

RSNA was recorded through the left renal nerve with bipolar silver electrodes. The renal nerve was located, dissected and covered with mineral oil prior to the placement of electrodes for recording (Nujol - Schering-Plough, São Paulo, SP, Brazil). The signals were obtained using a high-impedance probe connected to the amplifier (P511; Grass Instruments, Quincy, MA, USA). The signal was amplified 20,000 times, digitized and band-pass filtered (30–1000 Hz). The nerve signal was recorded continuously (with a PowerLab System-ADInstruments; Colorado Springs, CO, USA), rectified and integrated at 1 s intervals using LabChart software (v.7.3.7.; ADInstruments; Colorado Springs, CO, USA). At the end of each experiment, ganglionic blocker hexamethonium (30 mg · kg−<sup>1</sup> , b.wt., i.v.; Sigma–Aldrich, St. Louis, MO, USA) was administered to determine the background noise. The level of RSNA was expressed as a percentage of baseline after subtraction of the noise.

## Respiratory Synchronization

Analyses of the respiratory synchronization were made offline using Spike2 software (version 8; Cambridge Electronic Design Limited, Cambridge, CAM, England). In order to analyze respiratory modulation, RSNA was rectified and signals were smoothed using a time constant of 50 ms. The DIA-triggered averages of RSNA were generated after nanoinjection of vehicle and bilateral BBS into RVLM. Averages of RSNA were made using 15 DIA burst as trigger events. The time of inspiration was determined based on the duration of the inspiratory DIA burst while expiratory time was determined between consecutive DIA burst. The RSNA post-inspiratory peak was later evaluated.

## Nanoinjections into RVLM

Animals were mounted prone in a stereotaxic apparatus (David Kopf Instruments, Tujunga, CA, USA) with incisor bar 11 mm below the interaural line. After partial removal of the occipital bone, the meninges covering the dorsal surface of the brainstem was opened up surgically to visualize the calamus scriptorius. In order to nanoinject into the RVLM, a glass micropipette was positioned as follows: 2.5 mm rostral from the calamus scriptorius, ±2.0 mm lateral from the midline and 2.5 mm ventral from the dorsal surface.

Firstly, RVLM was localized through unilateral nanoinjection (50 nl) of L-glutamate (10 mM; Sigma-Aldrich, St. Louis, MO, USA) in all experimental groups. RVLM was considered previously identified when L-glutamate nanoinjection increased MAP in approximately 20 mmHg. After the return of cardiorespiratory and sympathetic activities to basaline, the bilateral nanoinjection (50 nl each) of vehicle (150 mM NaCl; Sigma-Aldrich, St. Louis, MO, USA), BBS (0.3 mM; Bachem AG, Bubendorf, Switzerland) or bombesin receptor 1 (BB1) antagonist (3 mM; [D-2- NaI<sup>5</sup> -Cys6,11-Tyr<sup>7</sup> ,D-Trp<sup>8</sup> ,Val10,2- NaI12-Somatostatin-14 (5–12) amide trifluoroacetate salt]; BIM-23127; Bachem AG, Bubendorf, Switzerland) was injected into RVLM. In order to confirm the integrity of RVML, L-glutamate (10 mM; 50 nl) was nanoinjected again. This nanoinjection produced similar increase in MAP (∼20 mmHg) as previously observed in the initial L-glutamate nanoinjection.

## Histology

At the end of the experiment, 2% Evans Blue solution (50 nl; Sigma-Aldrich, St. Louis, MO, USA) was nanoinjected bilaterally into RVLM for histological analyses with the aim of confirming the accuracy of injection sites. Rats were perfused transcardially with saline (150 mM NaCl; 300 mL), followed by 10% formaldehyde (300 mL; Synth Ltda, Diadema, SP, Brazil). The brains were later removed and fixed in 10% formaldehyde. Frozen brains were cut into 40 µm coronal sections and stained with 1% neutral red to determine the nanoinjection sites.

## Experimental Protocols

The MAP, DIA motor activity and RSNA were recorded (NT, n = 7; SHR, n = 7). After a period of stabilization, bilateral nanoinjections of BBS (0.3 mM in 50 nl each) or equivalent volume of vehicle (150 mM NaCl in 50 nl each) was nanoinjected into RVLM. In another group of NT rats (n = 6) and SHR (n = 6), BIM-23127 (BB<sup>1</sup> receptor antagonist; 3 mM in 50 nl each) was nanoinjected bilaterally into RVLM. In order to confirm the blockade of BB<sup>1</sup> receptors, in separated groups of NT rats (n = 3) and SHR (n = 6), unilateral nanoinjection of BBS (0.3 mM in 50 nl each) 15 min before and 15 min after BIM-23127 (3 mM in 50 nl each) was carried out.

## Statistical Analysis

The GraphPad Prism software (v.6; GraphPad Software, Inc., La Jolla, CA, USA) was used for the statistical analysis of experimental data. The basal values and changes in the respiratory synchronization induced by BBS nanoinjection were compare between the groups using an unpaired and paired Student's t-test. The autonomic and cardiovascular effects induced by nanoinjection of vehicle, BBS, and BIM into the RVLM were analyzed using two-way ANOVA prior to the Newman–Keuls test. The value of p < 0.05 was considered statistically significant.

## RESULTS

### Histological Analysis

**Figure 1A** show a representative photomicrograph of the brainstem section that indicates the accuracy of nanoinjection site—RVLM. The center of the nanoinjection site distributions at rostral and caudal levels of RVLM are shown in **Figure 1B**. Only animal with RVLM confined nanoinjections were analyzed.

### Cardiorespiratory and Sympathetic Changes Produced by the Injections of BBS into RVLM

**Table 1** shows the body weight and basal values of MAP, RSNA, and fR of NT rats and SHR. There were no changes in RSNA and body weight among groups. However, higher MAP and lower fR were observed in SHR when compared with NT rats (p < 0.05).

Unilateral nanoinjection of L-glutamate into RVLM increased MAP in both NT rats (1MAP: 21.6 ± 5.6 mmHg vs. vehicle −0.7 ± 0.5 mmHg; p < 0.05) and SHR (26.3 ± 6.3 mmHg vs. vehicle 0.4 ± 0.5 mmHg; p < 0.05; **Figures 2**, **3A**). The injection of L-glutamate did not elicit significant alteration in RSNA, DIA burst amplitude and fR in NT rats (1RSNA: 28.8 ± 5.2% vs. vehicle −0.3 ± 1.1%; 1DIA Burst Amp: −19.9 ± 15.3% vs. vehicle −0.5 ± 1.7%; 1fR: −6.7 ± 8,2 cpm vs. vehicle −2,5 ± 1.8 cpm; **Figures 2**, **3C,D**) and SHR (1RSNA: 40.0 ± 3.2% vs. vehicle 1.4 ± 0.6%; 1DIA Burst Amp: 10.7 ± 12% vs. vehicle −3.5 ± 1.6%; 1fR: −3.5 ± 2.4 cpm vs. vehicle −0.03 ± 1.2 cpm; **Figures 2**, **3B–D**).

Unilateral nanoinjection of BBS increased MAP in NT rats (17.1 ± 1.7 mmHg vs. vehicle −0.7 ± 0.3 mmHg; p < 0.05; **Figures 2A**, **3A**) without significant change in RSNA (45.5 ± 6.9% vs. vehicle −0.3 ± 1.1%; **Figures 2A**, **3B**), DIA burst amplitude (21.3 ± 6.1% vs. vehicle −0.5 ± 1.7%; **Figures 2A**, **3C**) and fR (−0.1 ± 3.7 cpm vs. vehicle −2.5 ± 1.8 cpm; **Figures 2A**, **3D**). In SHR, unilateral nanoinjection of BBS increased MAP (28.8 ± 3.2 mmHg vs. vehicle 0.3 ± 0.4 mmHg; p < 0.05; **Figures 2B**, **3A**) without significant alteration in RSNA (31.8 ± 3.7% vs. vehicle 1.4 ± 0.6%; **Figures 2B**, **3B**) and fR (11.9 ± 3.5 cpm vs. vehicle 0.0 ± 1.2 cpm; **Figures 2B**, **3D**). Unilateral nanoinjection of BBS enhanced DIA burst amplitude (81.2 ± 37.9% vs. vehicle −3.5 ± 1.5%; p < 0.05; **Figures 2B**, **3C**).

In NT rats, bilateral nanoinjection of BBS increased MAP (33.9 ± 6.6; 27.8 ± 2.9; 16.8 ± 3.4 mmHg vs. vehicle −0.7 ± 0.3 mmHg at 1, 5, and 15 min, respectively; p < 0.05; **Figures 2A**, **3A**) and RSNA (97.8 ± 12.9; 73.1 ± 23.2; 66.9 ± 30.2% vs. vehicle −0.3 ± 1.1%, at 1, 5, and 15 min, respectively; p < 0.05; **Figures 2A**, **3B**). An increase in MAP (37.1 ± 4.6; 26.2 ± 5.2; 19.5 ± 5.5 mmHg vs. vehicle 0.3 ± 0.4 mmHg, at 1, 5, and 15 min, respectively; p < 0.05; **Figures 2B**, **3A**) and RSNA (87.9 ± 18.1; 68.7 ± 16.8% vs. vehicle 1.4 ± 0.6%, at 1 and 5 min, respectively; p < 0.05; **Figures 2B**, **3B**) were observed in SHR.

BIM-23127. NA, nucleus ambiguous.


TABLE 1 | Body weight (b.w) and basal values of mean arterial pressure (MAP), renal sympathetic nerve activity (RSNA), and respiratory frequency (fR) of normotensive (NT) and spontaneously hypertensive rats (SHR) that received nanoinjections of bombesin (BBS; 0.3 mM) or BIM-23127 (3 mM) into RVLM.

*Values are expressed as mean* ± *SEM.* \**different from normotensive groups; p* < *0.05. n* = *3–7.*

bombesin (BBS; 0.3 mM; 1, 5, and 15 min) into RVLM of normotensive (A) and hypertensive rats (B). Pulsatile arterial pressure (PAP), renal sympathetic nerve activity (RSNA), integrate of renal sympathetic nerve activity (R RSNA), integrate of diaphragm motor activity (R DIAEMG) and hexam (hexamethonium).

The DIA burst amplitude was not altered by bilateral nanoinjections of BBS in NT rats (23.2 ± 3.3; 19.3 ± 6.0; 15.3 ± 17.3% vs. vehicle −0.5 ± 1.7%; at 1, 5, and 15 min, respectively; **Figures 2A**, **3C**). On the other hand, bilateral nanoinjections of BBS increased DIA burst amplitude in SHR (111.0 ± 25.3; 94.5 ± 23.1; 90.9 ± 27.7% vs. vehicle −3.5 ± 1.5%; at 1, 5, and 15 min, respectively; p < 0.05; **Figures 2B**, **3C**). The bilateral nanoinjection of BBS did not induce significant changes in fR of NT rats (−5.1 ± 9.3; −2.6 ± 7.4 cpm; −2.4 ± 6.4 vs. vehicle −2.5 ± 1.8 cpm; at 1, 5, and 15 min, respectively; **Figures 2A**, **3D**) and SHR (10.5 ± 9.1; 8.7 ± 8.9; 15.4 ± 9.2 cpm vs. vehicle 0.0 ± 1.2 cpm; at 1, 5, and 15 min, respectively; **Figures 2B**, **3D**).

The unilateral nanoinjection of BBS did not elicit significant alteration in HR of NT rats (−14.5 ± 4.9 bpm vs. vehicle −3.2 ± 1.6 bpm). The bilateral nanoinjection of BBS decreased HR (−30.1 ± 7.3 bpm; −27.0 ± 6.8 bpm; −23.9 ± 6.8 bpm vs. vehicle −3.2 ± 1.6 bpm; at 1, 5, and 15 min, respectively; p <

### mM NaCl), bombesin (I BBS, 0.3 mM), and bilateral nanoinjection of bombesin (II BBS; 0.3 mM; 1, 5, and 15 min) into RVLM on mean arterial pressure (MAP; A), renal sympathetic nerve activity (RSNA; B), diaphragm burst amplitude (DIA Burst Amp; C) and respiratory frequency (fR; D) in normotensive and hypertensive rats. \*different from vehicle; †different from normotensive rats. *p* < 0.05; *n* = 7.

0.05). In SHR, the unilateral (18.4 ± 3.8 bpm vs. vehicle −8.5 ± 1.9 bpm; p < 0.05) and bilateral nanoinjections of BBS (26.3 ± 7.0 bpm; 27.4 ± 7.7 bpm; 25.8 ± 8.2 bpm vs. vehicle −8.5 ± 1.9 bpm; at 1, 5, and 15 min; respectively; p < 0.05) increased HR.

### Synchronization of Sympathetic Discharge During the Respiratory Cycle after Bilateral Nanoinjection of BBS into RVLM

The BBS nanoinjected into RVLM elicited an increase in postinspiratory RSNA peak in SHR (0.309 ± 0.07 a.u vs. vehicle 0.180 ± 0.05 a.u.; p < 0.05; **Figure 4B**) but not in NT rats (0.227 ± 0.06 a.u vs. vehicle 0.103 ± 0.05 a.u.; **Figure 4A**).

## Cardiorespiratory and Sympathetic Changes Produced by BB<sup>1</sup> Receptor Blockade in the RVLM Neurons in NT rats and SHR

The bilateral nanoinjection of BIM-23127 into RVLM did not produce significant changes in MAP (−2.2 ± 2.8; −1.4 ± 2.7; −2.4 ± 1.9 mmHg vs. vehicle 1.0 ± 1.0 mmHg, at 1, 5, and 15 min, respectively; **Figures 5A**, **6A**), RSNA (−2.9 ± 2.5; −3.4 ± 2.7; −3.3 ± 2.3% vs. vehicle 0.0 ± 0.9% at 1, 5, and 15 min, respectively; **Figures 5A**, **6B**), DIA burst amplitude (−0.7 ± 3.5; −0.5 ± 3.1; −3.7 ± 3.5% vs. vehicle 2.1 ± 3.5% at 1, 5, and 15 min, respectively; **Figures 5A**, **6C**) and fR (−0.1 ± 1.9; −4.0 ± 4.2; −7.8 ± 4.1 cpm vs. vehicle −1.1 ± 1.1 cpm at 1, 5, and 15 min, respectively; **Figures 5A**, **6D**) in NT rats.

In SHR, bilateral nanoinjection of BIM-23127 reduced MAP (−19.9 ± 4.4; −14.8 ± 4.6; −15.6 ± 4.1 mmHg vs. vehicle −0.0 ± 1.2 mmHg, at 1, 5, and 15 min respectively; p < 0.05; **Figures 5B**, **6A**) and RSNA (−17.7 ± 3.8; −11.3 ± 2.4; −12.4 ± 3.5% vs. vehicle 0.7 ± 1.0% at 1, 5, and 15 min, respectively; p < 0.05; **Figures 5B**, **6B**). No significant changes were observed in the DIA burst amplitude (−10.5 ± 5.9; –14.0 ± 5.4; −11.8 ± 5.9% vs. vehicle 1.5 ± 2.9% at 1, 5, and 15 min, respectively; **Figures 5B**, **6C**) and fR (−2.3 ± 3.5; −2.2 ± 3.5; −2.7 ± 3.0 cpm vs. vehicle −4.6 ± 2.2 cpm at 1, 5, and 15 min, respectively; **Figures 5B**, **6D**) in SHR.

The blockade of BB<sup>1</sup> receptor did not alter HR in both NT rats (−3.8 ±3.0 bpm; −4.9 ± 2.4 bpm; −7.8 ± 2.6 bpm vs. vehicle 1.9 ± 1.6 bpm, at 1, 5, and 15 min, respectively) and SHR (−11.5 ±6.7 bpm; −6.1 ± 5.0 bpm; −5.7 ± 6.2 bpm vs. vehicle 0.4 ± 1.2 bpm, at 1, 5, and 15 min, respectively).

## Cardiorespiratory and Sympathetic Responses to BBS Injected after Blockade of RVLM BB<sup>1</sup> Receptors

**Table 2** shows the cardiorespiratory and sympathetic changes induced by unilateral nanoinjection of BBS into RVLM before and after injection of BIM-23127 (BB<sup>1</sup> receptor antagonist) in NT rats and SHR. In both groups, the BBS-induced increase in MAP and RSNA was abolished by BIM-23127. The significant changes in BBS-induced DIA burst amplitude was abolished by the blockade of BB<sup>1</sup> receptors SHR.

## DISCUSSION

Previous studies have shown that BBS increases blood pressure (Brown and Guyenet, 1985; Zogovic and Pilowsky, 2011) and sympathetic tone (Zogovic and Pilowsky, 2011). However, these

TABLE 2 | Changes in the mean arterial pressure (MAP), renal sympathetic nerve activity (RSNA) and diaphragm burst amplitude (DIA burst Amp) induced by unilateral injection of bombesin (BBS; 0.3 mM) into RVLM before and after BB<sup>1</sup> receptor blockade with BIM-23127 (3 mM) in normotensive (NT) and spontaneously hypertensive (SHR) rats.


*Values are expressed as mean* ± *SEM.* \**different from BBS before BIM-23127; p* < *0.05. n* = *3–6.*

studies did not established the effects of BBS in the brainstem and their role in the maintenance of hypertension. In the present study, we provided the first evidence that BBS acting in RVLM elicits a significant increase in MAP and RSNA of both NT rats and SHR. In addition, the administration of BBS into RVLM increased DIA burst amplitude and postinspiratory RSNA burst in SHR. The blockade of BB<sup>1</sup> receptors in the RVLM reduces MAP and RSNA in SHR but not in NT rats. These results strongly indicate that the sympathoexcitation associated with pressor response to BBS administration is mediated by BB<sup>1</sup> receptors that are located on RVLM neurons. It has been suggested that tonic activation of BB<sup>1</sup> receptors is involved in the maintenance of high arterial pressure in SHR.

Zogovic and Pilowsky (2011), showed a long-lasting increase in the splanchnic SNA (sSNA), blood pressure and phrenic nerve amplitude by the intrathecal injection of BBS. Glazkova and Inyushkin (2006) reported that microinjection of BBS in the solitary tract nucleus (NTS) stimulated respiration, and increased the level of pulmonary ventilation, respiratory volume, and bioelectrical activity of the inspiratory muscles. Our results showed that BBS injected into the RVLM increased arterial blood pressure and sympathetic vasomotor tone in both NT and SHR. This result suggests a modulatory action of BBS in RVLM neurons.

Recently, Li et al. (2016) showed that small neural subpopulation is involved in the control of breathing. The retrotrapezoid nucleus/parafacial respiratory group

(RTN/pFRG) expresses neuromedin B (BBS-like peptide genes) and GRP. This neural subpopulation project to preBötzinger Complex (preBötC, the respiratory rhythm generator that expresses neuromedin B and GRP receptors). The vasomotor presympathetic neurons in all anteroposterior extension of RVLM are intercalated with ventral respiratory column neurons (Smith et al., 2007; Alheid and McCrimmon, 2008; Wang et al., 2009). As a result of the proximity with respiratory neurons, the BBS injection into the RVLM could activate neuromedin B or GRP receptors and contribute to the increase in respiratory drive that was observed in SHR.

The RVLM neurons play essential role in the generation of sympathetic outflow (Guertzenstein, 1973; Guyenet, 2006; Wang et al., 2009) and regulation of peripheral chemosensitive and barosensitive sympathetic efferents (Sun and Reis, 1995; Miyawaki et al., 1996; Dampney et al., 2002; Alheid and McCrimmon, 2008). The neuronal activity of the RVLM is determined by the action of excitatory and inhibitory synapses that involve glutamate and γ-aminobutyric acid (GABA) neurotransmitters, respectively (Sun and Reis, 1995; Miyawaki et al., 1996; Ito et al., 2000; Schreihofer et al., 2000; Alheid and McCrimmon, 2008). In addition, some neuropeptides have been reported to play important modulatory role in the integration cardiovascular responses (Ito et al., 2000; Alheid and McCrimmon, 2008; Abbott and Pilowsky, 2009).

Several studies have shown that the activation or blockade of neuropeptide receptors could cause a long-term response (Abbott and Pilowsky, 2009; Zogovic and Pilowsky, 2011). The nature of response can be explained (partly) by the receptor-ligand type. For instance, the sensitization of G-protein coupled receptors are related to a wide range of intracellular event cascades such as changes in ion channel permeability, activation of kinases, and protein phosphorylation (Springell et al., 2005a,b).

Zogovic and Pilowsky (2011) reported that intrathecal injection of BBS elicited gradual increase in MAP and splanchnic sympathetic nerve activity (sSNA) within 5 min. This increase returned to control level after approximately 35 min of BBS injection. In our study, we showed that BBS injection into RVLM caused a rapid increase in blood pressure. This increase persisted during 15 min of bilateral nanoinjection of BBS. The injection of BIM induced rapid decrease in the MAP and RSNA (the decrease was maintained during 15 min). Some authors have reported instantaneous response to BBS administration (Erspamer et al., 1972; Chahl and Walker, 1981; Bayorh and Feuerstein, 1985; Kaczynska and Szereda-Przestaszewska, 2011). Intravenous administration of BBS caused immediate increase in

MAP of anesthetized animals (Erspamer et al., 1972; Kaczynska and Szereda-Przestaszewska, 2011). According to the author, this cardiovascular response appeared to be mediated via αadrenergic receptors (Kaczynska and Szereda-Przestaszewska, 2011).

Experimental evidences have demonstrated that an increase in RVLM neuronal activity could contribute to the development and maintenance of hypertension in SHR (Smith and Barron, 1990; Yang et al., 1995; Matsuura et al., 2002; Ito et al., 2003). The bicuculline (GABAa receptor antagonist) injection into RVLM slightly increased the arterial blood pressure in SHR when compared to NT rats. This result suggest excessive excitatory drive of RVLM pre-sympathetic neurons in hypertensive rats (Smith and Barron, 1990). The inhibition of glutamatergic neurotransmission by kynurenic acid (KYN) injection into RVLM of SHR decreased the arterial pressure (Ito et al., 2000). Matsuura et al., (2002), showed that basal membrane potential in irregularly firing RVLM neurons is less negative in neonatal SHR. In consequence, the RVLM neurons in these animals are more easily excitable. The firing rate is faster in neonatal SHR when compared with NT rats. In our study, we demonstrated that the blockade of RVLM BB<sup>1</sup> receptors decreased MAP and RSNA in SHR, but not in NT rats. These findings suggest that the activation of BB<sup>1</sup> receptors in the RVLM could have contributed to the maintenance of high arterial pressure in SHR.

Our findings are consistent with previous studies, which showed a decrease blood pressure of anesthetized hypertensive rats as a result of pharmacological blockade of RVLM neurons (Bergamaschi et al., 1999, 2014; Ito et al., 2000, 2003; Suhaimi et al., 2010; Du et al., 2013). The BBS injection into RVLM, in the present study, induced similar increase in the MAP and RSNA of SHR when compared with NT rats. The anesthesia reduced MAP in hypertensive rats more than in NT controls. Unlike non-anesthetized rats, the subjection of rats to anesthesia is assumed to have prevented the some responses to BBS.

Our results showed that BBS injection into RVLM induced sympathoexcitation with an increase in the blood pressure of NT rats and SHR. This effect indicates that RVLM could integrate neuronal pathway that are involved in BBS induced cardiorespiratory and sympathetic effects. The blockade of BB<sup>1</sup> receptors in the RVLM of SHR, but not NT rats, decreased MAP and RSNA. This result suggests that the tonic activation of BB<sup>1</sup> receptors is involved in the maintenance of high blood pressure in anesthetized SHR. The changes in the cardiorespiratory and sympathetic activity as a result of an activation or inhibition of RVLM BBS receptors suggests an important cardiorespiratory role of BBS and related peptides.

### AUTHOR CONTRIBUTIONS

Conceived and designed the experiments: GP, CD, AF, AD, AR, DC. Performed the experiments: ID, AM, SM, AC, KG. Analyzed the data: ID, AM, ED, MF, JF, DC, EC, GP. Contributed reagents/materials/analysis tools: DC, CD, DR, EC, AF, AD, AR, GP. Wrote the paper: ID, AM, ED, MF, JF, DR, EC, AF, AR, GP.

### ACKNOWLEDGMENTS

This work was supported by Fundação de Amparo a Pesquisa do Estado de Goiás (FAPEG) grants 2012/0055431086 (GP) and 2009/10267000352 (GP) and by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) grants 477832/2010-5 (GP), 483411/2012-4 (GP), and 447496/2014-0 (GP). The funders had no role in study design, data collection, analysis, decision to publish, or preparation of the manuscript. Moreover, all authors have contributed sufficiently in this study to be included as authors.

### REFERENCES


a hierarchy of three oscillatory mechanisms. J. Neurophysiol. 98, 3370–3387. doi: 10.1152/jn.00985.2007


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Pinto, Mourão, da Silva, Camargo, Marques, Gomes, Fajemiroye, da Silva Reis, Rebelo, Ferreira-Neto, Rosa, Freiria-Oliveira, Castro, Colombari, Colugnati and Pedrino. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Chronic Treatment with Ivabradine Does Not Affect Cardiovascular Autonomic Control in Rats

Fernanda C. Silva1, <sup>2</sup> , Franciny A. Paiva1, 2, Flávia C. Müller-Ribeiro<sup>3</sup> , Henrique M. A. Caldeira<sup>1</sup> , Marco A. P. Fontes <sup>3</sup> , Rodrigo C. A. de Menezes 1, 2 , Karina R. Casali <sup>4</sup> , Gláucia H. Fortes <sup>5</sup> , Eleonora Tobaldini <sup>6</sup> , Monica Solbiati <sup>6</sup> , Nicola Montano<sup>6</sup> , Valdo J. Dias Da Silva<sup>7</sup> and Deoclécio A. Chianca Jr. 1, 2 \*

*<sup>1</sup> Laboratory of Cardiovascular Physiology, Department of Biological Sciences, Institute of Exact and Biological Sciences, Federal University of Ouro Preto, Ouro Preto, Brazil, <sup>2</sup> Graduate Program in Biological Sciences – CBIOL/NUPEB, Federal University of Ouro Preto, Ouro Preto, Brazil, <sup>3</sup> Laboratory of Hypertension, Institute of Biological Sciences, Department of Physiology and Biophysics, Federal University of Minas Gerais, Belo Horizonte, Brazil, <sup>4</sup> Laboratory of Biomedical Engineering, Institute of Science and Technology, Federal University of São Paulo, São José dos Campos, Brazil, <sup>5</sup> Department of Physiology, University of Uberaba, Uberaba, Brazil, <sup>6</sup> Department of Clinical Sciences and Community Health, IRCCS Ca' Granda Foundation, Ospedale Maggiore Policlinico, University of Milan, Milan, Italy, <sup>7</sup> Department of Physiology, Institute of Biological and Natural Sciences, Federal University of Triângulo Mineiro, Uberaba, Brazil*

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Débora Simões Almeida Colombari, School of Denstistry, Brazil Maria Socorro França-Silva, Federal University of Paraíba, Brazil*

\*Correspondence:

*Deoclécio A. Chianca Jr. deochianca@gmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *13 May 2016* Accepted: *06 July 2016* Published: *26 July 2016*

### Citation:

*Silva FC, Paiva FA, Müller-Ribeiro FC, Caldeira HMA, Fontes MAP, de Menezes RCA, Casali KR, Fortes GH, Tobaldini E, Solbiati M, Montano N, Dias Da Silva VJ and Chianca DA Jr. (2016) Chronic Treatment with Ivabradine Does Not Affect Cardiovascular Autonomic Control in Rats. Front. Physiol. 7:305. doi: 10.3389/fphys.2016.00305* A low resting heart rate (HR) would be of great benefit in cardiovascular diseases. Ivabradine—a novel selective inhibitor of hyperpolarization-activated cyclic nucleotide gated (HCN) channels- has emerged as a promising HR lowering drug. Its effects on the autonomic HR control are little known. This study assessed the effects of chronic treatment with ivabradine on the modulatory, reflex and tonic cardiovascular autonomic control and on the renal sympathetic nerve activity (RSNA). Male Wistar rats were divided in 2 groups, receiving intraperitoneal injections of vehicle (VEH) or ivabradine (IVA) during 7 or 8 consecutive days. Rats were submitted to vessels cannulation to perform arterial blood pressure (AP) and HR recordings in freely moving rats. Time series of resting pulse interval and systolic AP were used to measure cardiovascular variability parameters. We also assessed the baroreflex, chemoreflex and the Bezold-Jarish reflex sensitivities. To better evaluate the effects of ivabradine on the autonomic control of the heart, we performed sympathetic and vagal autonomic blockade. As expected, ivabradine-treated rats showed a lower resting (VEH: 362 ± 16 bpm vs. IVA: 260 ± 14 bpm, *p* = 0.0005) and intrinsic HR (VEH: 369 ± 9 bpm vs. IVA: 326 ± 11 bpm, *p* = 0.0146). However, the chronic treatment with ivabradine did not change normalized HR spectral parameters LF (nu) (VEH: 24.2 ± 4.6 vs. IVA: 29.8 ± 6.4; *p* > 0.05); HF (nu) (VEH: 75.1 ± 3.7 vs. IVA: 69.2 ± 5.8; *p* > 0.05), any cardiovascular reflexes, neither the tonic autonomic control of the HR (tonic sympathovagal index; VEH: 0.91± 0.02 vs. IVA: 0.88 ± 0.03, *p* = 0.3494). We performed the AP, HR and RSNA recordings in urethane-anesthetized rats. The chronic treatment with ivabradine reduced the resting HR (VEH: 364 ± 12 bpm vs. IVA: 207 ± 11 bpm, *p* < 0.0001), without affecting RSNA (VEH: 117 ± 16 vs. IVA: 120 ± 9 spikes/s, *p* = 0.9100) and mean arterial pressure (VEH: 70 ± 4 vs. IVA: 77 ± 6 mmHg, *p* = 0.3293). Our results suggest that, in health rats, the long-term treatment with ivabradine directly reduces the HR without changing the RSNA modulation and the reflex and tonic autonomic control of the heart.

Keywords: ivabradine, HCN channels, renal sympathetic nerve activity, cardiovascular reflexes, tonic control, vagal activity, sympathetic activity

## INTRODUCTION

There is a clear association between increased resting heart rate (rHR) and mortality rate, especially in patients suffering from cardiovascular disease (Fox et al., 2008; Verrier and Tan, 2009). It is noteworthy that this association is true not only for very high values of rHR, since rHR upper than 83 bpm has been associated with an increased risk for all-cause and cardiovascular mortality (Diaz et al., 2005). From a physiological point of view, rHR reduction promotes a bigger and better coronary perfusion leading to a greater oxygen balance and cardiac performance. Thus, a low rHR would be of great benefit for patients with cardiovascular disease (i.e., heart failure and angina pectoris) (Hall and Palmer, 2008; Gent et al., 2015).

Beta-adrenoceptors antagonists have been used for reducing heart rate. They are indicated as first-line therapy in patients with myocardial ischemia and heart failure. However, their usage could exacerbate cardiovascular (through a negative inotropic effect) and respiratory complications, mainly in elderly patients (Rochon et al., 1999; Chaudhary et al., 2015). Consequently, a pure bradycardic agent might be useful in conditions in which beta-adrenergic blocker cannot be used due to their side effects. In this regard, ivabradine—a novel selective inhibitor of HCN channels—has emerged as a promising "pure" heart rate (HR) lowering drug (DiFrancesco and Camm, 2004; Bucchi et al., 2006).

To date, the mammalian genome encodes four HCN isoforms (HCN1 to HCN4), which relate to ion conductivity, mainly in the central nervous system and the heart (Notomi and Shigemoto, 2004; Harzheim et al., 2008). HCN4, which has been described as the main HCN channel of sinoatrial node, and the prevalent isoform expressed in the heart conduction system (Brioschi et al., 2009; Furst and D'Avanzo, 2015), is the preferential target of ivabradine (Chaudhary et al., 2015). In vitro studies have demonstrated that ivabradine specifically blocks the HCN pore in a low-moderate concentration range (Bucchi et al., 2002; DiFrancesco, 2010). It greatly inhibits the hyperpolarization-active current (If) and reduces the firing rate of the sinoatrial node cells at small concentrations, without influencing other ion currents (calcium and potassium) (Bois et al., 1996). Moreover, experimental and clinical studies have corroborated the use of ivabradine as a favorable therapeutic strategy, since they revealed no unwanted cardiovascular outcome (negative inotropic or lusitropic effects), preserving ventricular contractility (DiFrancesco and Camm, 2004; Sulfi and Timmis, 2006). The lack of cardiovascular side effects and the specificity of ivabradine on lowering heart rate provide relevant advantages for its clinical usage. However, little is known about the effects of ivabradine on the cardiovascular autonomic control.

HR, whose control is achieved through intrinsic and extrinsic mechanisms, respectively, by If pacemaker current of HCN channels and the autonomic nervous system (Verrier and Tan, 2009), are determinants of myocardial oxygen demand and may also affect myocardial perfusion. The autonomic system, which exerts reflex and tonic control over the cardiovascular homeostasis, influences the HCN voltage-dependence, changing the diastolic depolarization and, consequently, the HR (DiFrancesco, 1993). In this context, knowledge concerning the effects of chronic treatment with ivabradine on cardiovascular autonomic control has become essential. A few reports have described some autonomic effects of If blockers, including ivabradine, during acute endovenous treatment (Barzilai and Jacob, 2015; Dias da Silva et al., 2015). However, to our knowledge, no study has attempted to investigate the chronic effects of ivabradine on reflex autonomic control of HR and sympathetic nerve activity. Therefore, in the present study we assessed, in rats, the effects of chronic treatment with ivabradine on the cardiovascular autonomic modulation, the cardiovascular reflexes regulation (baroreflex, Bezold-Jarish reflex and chemoreflex) and tonic autonomic HR control, as well as on the RSNA.

Featuring the effects of chronic treatment with ivabradine on the HR, may provide a considerable in vivo understanding of its effects on the cardiovascular autonomic system. Therefore, improving our knowledge on the pharmacodynamics of this drug and providing substantial outcomes for the clinical usage of ivabradine as a cardiac medication.

## MATERIALS AND METHODS

## Experimental Model

Experiments were performed on male Wistar rats (300 ± 10 g), supplied by the Center of Animal Science of the Federal University of Ouro Preto (UFOP). They were kept in grouped cages (n = 3) on a 12 h light/dark cycle, at a controlled room temperature (23◦C), with free access to commercial chow and filtered water. Efforts were made to avoid any unnecessary distress to the rats, in accordance to the Brazilian Council for Animal Experimentation. All procedures were approved by the institutional ethics committee for animal research of UFOP (CEUA 2014/66; 2015/59), and were performed according to the regulations set forth by the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals.

### Experimental Preparation

Rats were divided in two treatment groups, respectively, vehicle and ivabradine receiving intraperitoneal (i.p.) injections of PBS (1mL/kg/day; single doses) or ivabradine (10 mg/kg/day or 2 mg/kg/day; single doses) during 7 or 8 consecutive days according to experimental design (detailed in Experimental Design Section). On the 5th day of treatment, the animals were submitted to surgical procedures, as described below.

### Surgical Procedures Related to Experiment 1

Rats were anesthetized with ketamine and xylazine solution (80 mg/kg; 7 mg/kg; i.m.). Polyethylene catheters were placed into the femoral artery and vein, respectively, for cardiovascular recordings and drugs infusion, which have been described in detail elsewhere (Martins et al., 2011). Prophylactic treatment with antibiotics [Veterinary Pentabiotic—penicillin (benzatinbenzilpenicillin, procain benzilpenicillin and potassicbenzilpenicillin), streptomicyn and dihydrostreptomycin: 1 mL/kg; i.m.] and anti-inflammatory (Ketoprofen: 4 mg/kg; i.m.) drugs were performed in order to prevent postsurgical infections and inflammation, respectively (Silva et al., 2013). After 48 h of recovery from the anesthesia and surgery, rats were conducted to experimental protocol (Rodrigues-Barbosa et al., 2012).

### Surgical Procedures Related to Experiment 2

Rats were anesthetized with urethane (1.2–1.4 g/kg, i.p. with supplementary doses of 0.1 g/kg i.v., if required). The adequacy of anesthesia was verified by the absence of the corneal reflex and a withdrawal response to nociceptive stimulation of a hind paw. A tracheotomy was performed to maintain and unobstructed airway, and all the animals were allowed to breathe freely. Polyethylene catheters were placed into the femoral artery and vein, respectively, for pulsatile arterial pressure recording and drug injections. The rat was mounted in a stereotaxic apparatus and the left renal nerve was exposed and prepared as described previously (Muller-Ribeiro et al., 2012; Xavier et al., 2014). Briefly, the renal nerve was exposed from a retroperitoneal approach. It was carefully separated from surrounding tissue and placed on a silver bipolar recording electrode and immersed in mineral oil for sympathetic nerve activity recording. The signals from the recording electrode were amplified and filtered (bandwidth 100–2000 Hz). The signals were then digitized (1000 samples/s) and recorded using the PowerLab system. Chart software was used to rectify and integrate the RSNA signals.

## Experimental Design

As mentioned in the experimental preparation section, rats were submitted to chronic treatment during 7 and 8 days, in accordance to the previous literature (da Silva et al., 1994). All experiments started only after stabilization of physiological parameters for at least 30 min.

### Experiment 1: The Effects of Chronic Treatment with Ivabradine (8 days) on Modulatory, Reflex and Tonic Autonomic Control of Heart Rate (n = 14)

This experiment was conducted in non-anesthetized freely moving rats (vehicle group: PBS 1 mL/kg/day i.p.—single doses; n = 6 and ivabradine group: 2 mg/kg/day i.p.—single doses; n = 8; both during 8 days), in order to test the effects of ivabradine on the cardiovascular variability, reflex sensitivities and tonic control of HR. In this experiment, we used 2 mg/kg of ivabradine, since higher doses evoked a severe bradycardia. The adopted dose was based on previous studies (Dias da Silva et al., 2015). On the 7th day of treatment, rats were submitted to 30 min of basal period recording, which was used to quantify HR and systolic arterial pressure (SAP) variability parameters in the time—and the frequency-domain (spectral analysis), followed by the assessment of cardiovascular reflexes and cardiac autonomic tone. On the 8th day of treatment, rats were re-submitted to 30 min of basal period, followed by the measurement of cardiac autonomic tone, as will be detailed later.

In order to evaluate cardiovascular variability parameters, mean values of SAP, mean arterial pressure (MAP), diastolic arterial pressure (DAP), and pulse interval (PI) were calculated for each 20–30-min period of recording. For the cardiovascular variability study, the signals of arterial pressure (AP) were processed using a software (PRE24 software, kindly provided by Dr. Alberto Porta, University of Milan, Italy) to generate beatto-beat time series of PI, DAP, and SAP. The variance of these values in each period was considered a variability index in the time-domain. In order to minimize the potential influence of the absolute numerical values on the calculation of variance, we have also performed the calculation of the normalized variance, which consisted in dividing each single PI, SAP, or DAP values by the mean value of the entire respective time series and multiplying the result for 100% (Sacha, 2014). Therefore, the normalized variance (calculated from the adjusted time series and expressed in %) is corrected for the mean value of the time series. The variability of PI, DAP and SAP was also evaluated in the frequency domain using an autoregressive spectral analysis method. The theoretical and analytical proceedings are described in previous studies (Malliani et al., 1991; TFESCNASPE, 1996). Briefly, beat by beat time series of PI, DAP, and SAP were divided into serial segments of 300 beats, wherein all successive segments were overlapped by 50% (150 beats) on the previous segments (Welch's method). Using stationary time series segments, autoregressive parameters were estimated using Levinson-Durbin's method, and the model order was chosen according to Akaike's criteria (Rubini et al., 1993; TFESCNASPE, 1996). Then, on each individual stationary segment of 300 beats, spectral decomposition was performed using appropriate autoregressive software. The normalization procedure, applied only to the variability of the PI, was performed by dividing the power of the low frequency component (low frequency—LF, 0.20–0.750 Hz) or high frequency (high frequency—HF, 0.75– 3.00 Hz) by total spectral power, which is subtracted from the power of the very low frequency band (very low frequency— VLF, 0.01 to 0.20 Hz), and multiplying the result by 100 (Rubini et al., 1993). The spectral parameters obtained for each individual stationary segment of 300 beats were averaged, and the average values resulting from 30 min of recording were calculated for each animal.

In order to assess the baroreflex sensitivity, we administered intravenous bolus injection (i.v.) of phenylephrine (2, 4, and 8µg/kg) and sodium nitroprusside (4, 8, and 16µg/kg, i.v.) for calculating the 1HR/1MAP index (Oliveira et al., 2005), which was obtained through the mean of the three doses. We also administered phenylbiguanide (1.25; 2.5 and 5µg/kg, i.v.) and potassium cyanide (80 and 160µg/kg, i.v.) to evaluate, respectively, the Bezold-Jarish reflex and chemoreflex sensitivities (Penitente et al., 2007; Bezerra et al., 2011). All aforementioned i.v. injections were performed every 5 min. Subsequently, to evaluate the ivabradine influence on the tonic autonomic control of the heart, we also performed the sympathetic and vagal autonomic blockade after propranolol (5 mg/kg, i.v.) and methylatropine (4 mg/kg, i.v.) injections, respectively, to calculate the sympathetic and vagal effects, as well as the intrinsic HR and tonic sympathovagal index (Goldberger, 1999). The autonomic blockers were administered in a random sequence with a 15 min interval between them, on the 7th and 8th days of treatment. After double blockade, the cardiovascular recordings lasted for 15 min. Briefly, the sympathetic effect was considered as the difference between the HR after sympathetic blockade and resting HR (Sympathetic effect = HR after sympathetic blockade - resting HR). Vagal effect was calculated as the difference between HR after vagal blockade and resting HR (Vagal effect = HR after vagal blockade - resting HR). The tonic sympathovagal index was obtained as the ratio between resting and intrinsic HR, considering that the intrinsic HR (iHR) was the HR obtained after double autonomic blockade (Goldberger, 1999).

### Experiment 2: The Effects of Chronic Treatment with Ivabradine (7 days) on Heart Rate (HR), Mean Arterial Pressure (MAP) and Renal Sympathetic Nerve Activity (RSNA) (n = 10)

This experiment was conducted in urethane-anesthetized rats (vehicle group: PBS 1 mL/kg/day i.p.; single doses; n = 5 and ivabradine group: 10 mg/kg/day i.p.; single doses; n = 5, both during 7 days), in order to test the effects of ivabradine on HR, MAP, and RSNA. The ivabradine dose was chosen based on earlier studies (Du et al., 2004; Luszczki et al., 2013). For this purpose, such parameters were recorded during 60 min. The RSNA signal was amplified (10 K), filtered (100–1000 Hz), displayed on an oscilloscope and monitored via an audio-amplifier. The filtered nerve activity signal was rectified, integrated (resetting every second), displayed online and acquired using Powerlab 4/20 LabChart 7.1 (ADInstruments, Sydney, Australia). All data were digitized at 1 kHz. The noise of the recording system was determined post mortem (urethane: 0.5 mL; i.v) and subtracted from the RSNA values obtained during the experiment (Muller-Ribeiro et al., 2012; Xavier et al., 2014). The quantitation of spikes/second was conducted using a previously described methodology (Gomes da Silva et al., 2012). Body temperature was monitored since the beginning of surgical procedure using a rectal thermometer and maintained in the range of 37–37.5◦C using a heating pad (Xavier et al., 2014).

### Statistical Analysis

Regarding Experiment 1, baseline values of HR and SAP variability parameters were compared. In addition, for reflexes and autonomic tone studies, baseline values of HR and MAP were obtained by averaging the 1 min-period that preceded drugs injections. Maximum changes (mean ± SEM) were calculated using the peak response after drugs injections (for cardiovascular reflexes analysis) or using the last 1 min-period corresponding to each autonomic blockade recording (for autonomic tone analysis). The effects between groups (ivabradine 2 mg/kg vs. vehicle) were compared using Student's unpaired t-test.

In relation to the Experiment 2, data were obtained by averaging the values of the whole recording. They were expressed as absolute values and reported as mean ± SEM. The effects between groups (ivabradine 10 mg/kg vs. vehicle) were compared using Student's unpaired t-test.

Prism 5.0 (GraphPad Software, La Jolla, CA, USA) was used to analyze all data. The significance level was set at p < 0.05.

### RESULTS

## Experiment 1: Chronic Treatment with Ivabradine (8 Days) Did Not Change Cardiovascular Variability as Well as the Reflex and Tonic Autonomic Control of Heart Rate in Non-anesthetized Rats

As expected, compared with vehicle-treated group (VEH), the ivabradine-treated rats (IVA: 2 mg/kg/day; i.p.) presented lower resting (VEH: 362 ± 16 bpm vs. IVA: 260 ± 14 bpm, p = 0.0005; **Figure 1A**) and intrinsic HR (VEH: 369 ± 9 bpm vs. IVA: 326 ± 11 bpm, p = 0.0146; **Figure 3D**). Ivabradine treatment also decreased resting MAP (VEH: 115 ± 3 mmHg vs. IVA: 102 ± 2 mmHg, p = 0.0020; **Figure 1B**).

Heart rate variability analysis by means of spectral autoregressive decomposition of pulse interval (PI) timeseries in all experimental groups is summarized in **Table 1**. Treatment with ivabradine significantly increased variance and absolute values of LF and HF spectral components, without any changes in normalized LF and HF parameters and LF/HF ratio in ivabradine-treated rats when compared to vehicle-treated



*PI, Pulse interval; VLF, very low frequency component; LF, low frequency component; LF (nu), LF in normalized units; HF, high frequency component; HF (nu), HF in normalized units.* \**p* < *0.05 vs. vehicle.*

TABLE 2 | Mean values (±SEM) of systolic arterial pressure (SAP), variance and VLF, LF and HF spectral components of arterial pressure variability in non-anesthetized Wistar rats from vehicle or ivabradine-treated rats.


*SAP, Systolic arterial pressure; VLF, very low frequency component; LF, low frequency component; HF, high frequency component.* \**p* < *0.05 vs. vehicle.*

animals (**Table 1**). Systolic arterial pressure variability analysis is summarized in **Table 2**. Despite of the significantly lower values of SAP, no changes in SAP variability parameters were observed in ivabradine-treated rats when compared to vehicle-treated animals (**Table 2**). The behavior of diastolic arterial pressure variability was similar to the SAP variability (data not shown).

Regarding the possible influence of the chronic treatment with ivabradine on HR baroreflex control, we administered intravenous injection (i.v.) of phenylephrine (PHE: 2, 4, and 8µg/kg) and sodium nitroprusside (SNP: 4, 8, and 16µg/kg, i.v.) to assess the baroreflex sensitivity. The baroreflex bradycardic and tachycardic gains were calculated through 1HR/1MAP index. No changes were observed in bradycardic (VEH:−2.7 ± 0.4 bpm/mmHg vs. IVA:−2.9 ± 0.8 bpm/mmHg, p = 0.8025; **Figure 2A**) and tachycardic gains (VEH:−6.9 ± 0.6 bpm/mmHg vs. IVA:−7.7 ± 1.1 bpm/mmHg, p = 0.6432; **Figure 2B**) between groups. We also administered phenylbiguanide (PBG: 1.25; 2.5 and 5µg/kg, i.v.) and potassium cyanide (KCN: 80 and 160µg/kg, i.v.) to assess the Bezold-Jarish reflex and chemoreflex regulation, respectively. The treatment with ivabradine during 8 days did not modify the cardiovascular responses to Bezold-Jarish reflex activation, since all PBG injected doses induced similar hypotension and bradycardia in VEH and IVA groups (1**MAP: PBG 1.25**−VEH:−4 ± 1 mmHg vs. IVA:−7 ± 2 mmHg, p = 0.2525; **PBG 2.5**−VEH:−5.5 ± 2 mmHg vs. IVA:−10 ± 3 mmHg, p = 0.2303; **PBG 5**−VEH:−5 ± 1 mmHg vs. IVA:−10 ± 3 mmHg, p = 0.1318; **Figure 2C**); (1**HR: PBG 1.25**−VEH:−45 ± 14 bpm vs. IVA: −63 ± 20 bpm, p = 0.5; **PBG 2.5**−VEH:−143 ± 35 bpm vs. IVA:−124 ± 25 bpm, p = 0.2848; **PBG 5**−VEH:−139 ± 25 bpm vs. IVA:−122 ± 20 bpm, p = 0.3199; **Figure 2D**). Similarly, the ivabradine treatment did not change the chemoreflex responsiveness. The amplitude of hypertensive and bradycardic responses evoked by KCN injections were similar in both groups (1**MAP: KCN 80**−VEH: 14 ± 6 mmHg vs. IVA: 16 ± 5 mmHg, p = 0.4259; **KCN 160**−VEH: 21 ± 4 mmHg vs. IVA: 17 ± 4 mmHg, p = 0.5162; **Figure 2E**); (1**HR: KCN 80**−VEH:−66 ± 23 bpm vs. IVA:−92 ± 19 bpm, p = 0.2018; **KCN 160**−VEH:−87 ± 22 bpm vs. IVA:−121 ± 15 bpm, p = 0.1087; **Figure 2F**).

In order to evaluate the influence of ivabradine chronic treatment on the tonic autonomic control of the heart, we performed the vagal and sympathetic autonomic blockade through methylatropine (4 mg/kg, i.v.) and propranolol (5 mg/kg, i.v.) injections, respectively, to calculate the vagal and sympathetic effects, as well as the tonic sympathovagal index. No differences on vagal (1**HR;** VEH: 96 ± 18 bpm vs. IVA: 86 ± 22 bpm, p = 0.3571; **Figure 3A**) neither on sympathetic effects (1**HR;** VEH: −47 ± 11 bpm vs. IVA: −36 ± 13 bpm, p = 0.2724; **Figure 3B**) were observed between VEH and IVA groups. Additionally, the ivabradine treatment did not alter the sympathovagal balance, expressed by tonic sympathovagal index (VEH: 0.91 ± 0.02 vs. IVA: 0.88 ± 0.03, p = 0.3494; **Figure 3C**).

### Experiment 2: Chronic treatment with Ivabradine (7 days) Reduced Heart Rate (HR), Without Significant Effects on Mean Arterial Pressure (MAP) and Renal Sympathetic Nerve Activity (RSNA) in Urethane-anesthetized Rats

Representative records **(A,B)** and changes in HR **(C)**, MAP **(D)** and RSNA **(E)** induced by the chronic treatment with ivabradine were shown in **Figure 4**. Compared with vehicle treatment, the ivabradine administration (10 mg/kg/day; i.p.) during 7 consecutive days markedly reduced resting HR (VEH: 364 ± 12 bpm vs. IVA: 207 ± 11 bpm, p < 0.0001; **Figure 4C**). However, such ivabradine treatment did not change MAP (VEH: 70 ± 4 vs. IVA: 77 ± 6 mmHg, p = 0.3293; **Figure 4D**) and RSNA (VEH: 117 ± 16 vs. IVA: 120 ± 9 spikes/sec, p = 0.9100; **Figure 4E**).

### DISCUSSION

In this study, we evaluated the effects of chronic treatment with ivabradine—a "pure" HR lowering drug which selectively inhibits the pacemaker HCN channels—on the autonomic control of the HR in rats. Our results showed that, in healthy animals, a

long-term ivabradine administration reduced resting HR without promoting any change on the cardiovascular reflexes and tonic autonomic regulation of the heart, as well as on RSNA. As no study has investigated this set of variables to date, our data provides new in vivo insights about the chronic effects of ivabradine on the cardiovascular autonomic system.

Exploring whether chronic treatment with ivabradine could influence the autonomic regulation of the HR, we assessed its effects on the reflex and on the tonic autonomic control. In the present investigation, performed in freely moving rats, ivabradine treatment (2 mg/kg/day) during 8 days induced a significant reduction in resting and intrinsic HR. Furthermore, it also significantly decreased resting MAP, which could be ascribed to a cardiac output fall due to a marked ivabradineinduced bradycardia. It is plausible to infer that, since we did not observe any ivabradine-induced RSNA and autonomic changes, which indirectly suggest no effect on the peripheral resistance. The ivabradine-induced bradycardia (28%) was greater than that achieved in other experimental (Dias da Silva et al., 2015) and clinical studies (15–20%) (Fox et al., 2008; Swedberg et al., 2010). The discrepancy between the bradycardia magnitudes may be explained, at least, by two reasons. Firstly, the mentioned experimental report was performed on anesthetized rats, which were submitted to acute ivabradine injection (Dias da Silva et al., 2015). Secondly, there is a plausible difference between the pharmacodynamic of ivabradine in rats and humans, which exhibit lower HR values. The ivabradine blockade is currentdependent and favored by depolarization, when the drug molecules are impelled to their intracellular binding site by the inward flow of permeating cations (sodium and potassium). In addition, ivabradine binds to HCN channels in open state, whose configuration depends on the attachment of the cyclic adenosine monophosphate (cAMP). Consequently, ivabradine presents upper activity when cAMP levels are high, such as in adrenergic stimulation that results in higher HR. It is a particular property, which provides its strong use-dependent action and suggests that its rate-reducing efficiency might be enhanced at high rates (Bucchi et al., 2002; DiFrancesco, 2005), as observed in non-anesthetized rats. Although, ivabradineinduced bradycardia observed in our study was not comparable to those considered appropriated to therapeutic action, it was crucial for our objective to obtain a greater bradycardia, since it assured a worst scenario for its potential influence on autonomic system.

We also analyzed the effects of chronic treatment with ivabradine on cardiovascular reflex control (baroreflex, Bezold-Jarish reflex and chemoreflex). The ivabradine treatment did not affect the bradycardic and tachycardic gains of baroreflex,

FIGURE 3 | Effects of chronic treatment with ivabradine on the tonic autonomic control of the heart rate in non-anesthetized Wistar rats. (A) Vagal and (B) sympathetic effects were obtained, respectively, by the difference between vagal blockade (by methylatropine: 4 mg/kg, i.v.) or sympathetic blockade (by propranolol: 5 mg/kg, i.v.) and resting HR. (C) Sympathovagal balance was expressed by the tonic sympathovagal index, which is the ratio between resting and intrinsic HR. (D) Intrinsic HR (iHR, bpm) —the HR was obtained after autonomic double blockade. All aforementioned parameters were evaluated in non-anesthetized Wistar rats submitted to vehicle (PBS 1 mL/kg/day i.p.; single dose; *n* = 6) or ivabradine (2 mg/kg/day i.p.; single dose; *n* = 8) treatment during 8 days. Bars represent mean ± SEM of vehicle (VEH—black) and ivabradine (IVA—white) groups. \**p* < 0.05, VEH vs. IVA (Student's unpaired *t*-test).

neither the hypotensive and bradycardic responses evoked by Bezold-Jarish reflex activation, as well as the hypertensive and bradycardic responses induced by peripheric chemoreceptors stimulation. To our knowledge, it was the first time that these reflexes were studied in non-anesthetized rats submitted to long-term use of ivabradine. A few reports leading with this subject investigated the acute intravenous injection effects of HCN blockers in rats (Kruger et al., 2000; Dias da Silva et al., 2015). One study, in which acute zatebradine i.v. injection was conducted in non-anesthetized sham-rats, revealed an increase in arterial baroreflex sensitivity (Kruger et al., 2000). However, a recent study from our group performed in anesthetized rats, that received acute ivabradine i.v. injection, showed that ivabradine did not alter the baroreflex sensitivity (Dias da Silva et al., 2015). As mentioned above, chronic treatment, during 7–8 days, could permit a resetting of cardiovascular autonomic reflexes to a new hemodynamic context.

In order to deepen the understanding about the influence of chronic use of ivabradine on autonomic regulation of the HR, we also evaluated its effects on modulatory and tonic autonomic control. For that, we performed HR variability analysis, and vagal and sympathetic autonomic blockade, respectively. We also evaluated the tonic sympathovagal index, which is a validated methodology to assess the vagal and sympathetic cardiac tonic balance (Goldberger, 1999). Regarding to HR variability analysis, we observed in ivabradine-treated animals a markedly higher value of variance, a time–domain index, which was accompanied to higher values in LF and HF spectral components, without any changes in normalized LF and HF components, as well as in LF/HF ratio. Even though, at the first glance, these findings could suggest an increased sympathetic (higher LF component) and parasympathetic modulation (higher variance and HF component). The normalization procedure and LF/HF ratio show a balanced autonomic modulation. As previously reported by our group (Dias da Silva et al., 2015) and others (Rocchetti et al., 2000; Zaza and Lombardi, 2001; Monfredi et al., 2014), the increase in all spectral components (expressed in absolute units) and in the variance (sum of the individual spectral components) can be considered, since these parameters are intrinsically dependent on mean interbeat interval (pulse interval), which was directly increased by ivabradine more than possible changes in the autonomic influences toward sinus node. In fact, our results of autonomic blockade with atropine and propranolol seem to reinforce this idea, since no differences were observed on vagal and sympathetic effects. Additionally, the tonic sympathovagal index was not affected by ivabradine treatment. It was <1 in vehicle and ivabradine groups, indicating a vagal dominance in both (Goldberger, 1999). Taken together, these data suggest that, in rats, the sustained use of ivabradine did not alter the autonomic balance to the heart. Despite of that, as showed by Mangin et al. (1998), the intrinsic dependency of variance and spectral parameters on mean HR does not rule out the possibility that these parameters may be also modulated by neural influences, which probably could be the case under several conditions of real autonomic imbalance, as observed in many cardiovascular disease conditions (Malliani et al., 1991; TFESCNASPE, 1996; Schwartz and De Ferrari, 2011). Despite of a small decrease in arterial pressure, SAP variability data did not show any difference between both vehicle and ivabradine groups. LF component of SAP variability has been considered an indirect marker of vascular sympathetic modulation, suggesting that chronic treatment with ivabradine does not change sympathetic activity not only to the heart but also to the peripheral vessels. However, given the large standard errors of means observed on SAP variability parameters, the lack of differences between vehicle and ivabradine treatment should be interpreted cautiously and not as a definitive statement that ivabradine does not change arterial pressure variability. More experiments with a higher number of cases in a near future should be performed to clarify this issue.

In order to further understanding ivabradine actions on sympathetic control, we assessed the sympathetic nervous system responsiveness to long-term treatment with ivabradine by a direct methodology. Thus, we analyzed the effects of chronic injections of ivabradine (10 mg/kg/day; during 7 days) on RSNA, as well as in HR and MAP in urethaneanesthetized rats. Recordings of RSNA are technically more amenable in anesthetized than in freely moving animals (Xavier et al., 2014), however the cardiovascular autonomic control is affected by anesthetics. Urethane presents minor interference with autonomic activity, but reduces the sympathetic tone whereas it seems to leave the parasympathetic tone relatively intact (Shimokawa et al., 1998; Bencze et al., 2013).Taking this into account, as Experiment 2 was performed

under urethane-anesthesia, we used a ivabradine dose 5-fold higher to that administered in Experiment 1 because it was substantial to assure a enough severe scenario to evaluate the potential influence of ivabradine on autonomic control. Ivabradine treatment reduced resting HR, without affecting MAP, corroborating previous studies (Du et al., 2004; Verrier et al., 2014; Gent et al., 2015). Furthermore, it did not change RSNA. RSNA is closely correlated with arterial pressure and is believed to be a reliable indicator of overall sympathetic vasomotor activity (Burgess et al., 1997). This finding is consistent with the SAP and HR variabilities parameters observed in nonanesthetized rats, strongly suggesting that long-term ivabradine administration does not compromise the sympathetic autonomic regulation. A recent study by our group showed that, in rats, ivabradine-induced bradycardia was associated with increased cardiac sympathetic nerve activity, resulting from baroreceptor unloading (Dias da Silva et al., 2015). Additionally, another study reported that, in sham-rats, zatebradine (another HCN blocker) increases the heart rate variability—a marker of autonomic modulation of HR (Kruger et al., 2000). However, such data methodologically differs from ours in some aspects: (i) they refer to acute i.v. injections while in our study the ivabradine effects were assessed 7–8 days after continuous treatment, a time frame long enough to induce baroreceptor resetting and, (ii) the second report used a HCN blocker less specific than ivabradine. Ivabradine, which is an open channel-required blocker, reaches its binding site by entering in the HCN pore from an intracellular side, being able to specifically block the pore channel in a lowmoderate concentration range (Bucchi et al., 2002).

Corroborating the widespread idea that ivabradine is a promissor cardiac medication, three current clinical studies have shown that acute treatment with ivabradine did not alter the HR and blood pressure variabilities and baroreflex sensitivity, suggesting no implications on the sympatho-vagal balance in healthy men (Heusser et al., 2016; Schroeder et al., 2016) and in postural tachycardia syndrome patients (Barzilai and Jacob, 2015). In addition, some studies have reported that ivabradine exerts no side effects on the: (i) conductivity and refractoriness (in atrium, atrioventricular node, His–Purkinje system and ventricles), (ii) left ventricular ejection fraction, (iii) stroke volume and (iv) some ECG measurements (corrected QT, PR, and QRS intervals) (DiFrancesco and Camm, 2004), reinforcing the advantage and viability of ivabradine use to the detriment of others reducing HR drugs, such as beta-blockers. Blockade of beta receptors, which are found throughout the heart, can promote a beneficial reduction of HR, though it could trigger side effects due to its presence in several organs, including in the cardiovascular and respiratory systems (Tattersfield, 1991; Bois et al., 1996; Sulfi and Timmis, 2006). Beta-blockers also slow the If current via the reduction in sympathetic activity and cAMP

formation, while ivabradine acts specifically inhibiting If current (Sulfi and Timmis, 2006).

In accordance to all aforementioned, our data indicates that long-term treatment with ivabradine, in healthy rats, reduces the resting and intrisic heart rate, without compromising: (i) baroreflex sensitivity, (ii) Bezold-Jarish reflex control, (iii) chemoreflex resposiveness, (iv) BP and HR variabilities, (v) vagal and sympathetic tones, (vi) tonic sympathovagal index and (vii) RSNA. Characterizing the actions of chronic treatment with ivabradine on the cardiovascular autonomic control in rats can provide an additional understanding of its effects on cardiovascular autonomic system, and then significantly improve our knowledge to support the ivabradine clinical use.

### AUTHOR CONTRIBUTIONS

FS drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted the all data. FP acquired the data and also substantially contributed to work design and to analysis and interpretation of the all data. FM substantially contributed to acquisition and analysis of the RSNA data, HC substantially contributed to Bezold Jarish reflex data acquisition. MF, RD, KC, GF, ET, and MS

### REFERENCES


substantially contributed to all data interpretation, NM, VD, and DC designed the work, and substantially contributed to analysis and interpretation of the all data. All authors revised the work critically, approved the final version to be published and declared accountable for all aspects of the work.

### FUNDING

This study was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, grant #400851/2014-8 for VD), Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Fundação de Amparo à Pesquisa de Minas Gerais (FAPEMIG), Universidade Federal de Ouro Preto (UFOP) and Universidade Federal do Triângulo Mineiro (UFTM), Brazil.

### ACKNOWLEDGMENTS

The authors are grateful to Center of Science Animal/UFOP for supplying the rats, to Milton Alexandre de Paula and Marly de Lourdes Ferreira Lessa for technical assistance.


attenuated in transgenic rats with low brain angiotensinogen. Brain Res. 1448, 101–110. doi: 10.1016/j.brainres.2012.02.021


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer MF and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review. The reviewer DC declared a shared affiliation, though no other collaboration, with one of the author KC to the handling Editor, who ensured that the process nevertheless met the standards of a fair and objective review.

Copyright © 2016 Silva, Paiva, Müller-Ribeiro, Caldeira, Fontes, de Menezes, Casali, Fortes, Tobaldini, Solbiati, Montano, Dias Da Silva and Chianca. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Ovarian Hormone Deprivation Reduces Oxytocin Expression in Paraventricular Nucleus Preautonomic Neurons and Correlates with Baroreflex Impairment in Rats

Vitor U. De Melo1, 2, Rayssa R. M. Saldanha<sup>1</sup> , Carla R. Dos Santos <sup>3</sup> , Josiane De Campos Cruz <sup>4</sup> , Vitor A. Lira<sup>2</sup> , Valter J. Santana-Filho1 † and Lisete C. Michelini 3 † \*

### Edited by:

*Ovidiu Constantin Baltatu, Universidade Camilo Castelo Branco, Brazil*

### Reviewed by:

*Rohit Ramchandra, University of Auckland, New Zealand Cara Hildreth, Macquarie University, Australia*

\*Correspondence:

*Lisete C. Michelini michelin@usp.br*

*† These authors have contributed equally to this work.*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *06 July 2016* Accepted: *26 September 2016* Published: *13 October 2016*

### Citation:

*De Melo VU, Saldanha RRM, Dos Santos CR, Cruz JC, Lira VA, Santana-Filho VJ and Michelini LC (2016) Ovarian Hormone Deprivation Reduces Oxytocin Expression in Paraventricular Nucleus Preautonomic Neurons and Correlates with Baroreflex Impairment in Rats. Front. Physiol. 7:461. doi: 10.3389/fphys.2016.00461* *<sup>1</sup> Department of Physiology, Federal University of Sergipe, São Cristóvão, Brazil, <sup>2</sup> Department of Health and Human Physiology, Obesity Research and Education Initiative, Fraternal Order of Eagles Diabetes Research Center, Abboud Cardiovascular Research Center, Pappajohn Biomedical Institute, University of Iowa, Iowa, IA, USA, <sup>3</sup> Department of Physiology and Biophysics, Institute of Biomedical Sciences, University of São Paulo, São Paulo, Brazil, <sup>4</sup> Department of Biotechnology, Federal University of Paraiba, João Pessoa, Brazil*

The prevalence of cardiovascular diseases including hypertension increases dramatically in women after menopause, however the mechanisms involved remain incompletely understood. Oxytocinergic (OTergic) neurons are largely present within the paraventricular nucleus of the hypothalamus (PVN). Several studies have shown that OTergic drive from PVN to brainstem increases baroreflex sensitivity and improves autonomic control of the circulation. Since preautonomic PVN neurons express different types of estrogen receptors, we hypothesize that ovarian hormone deprivation causes baroreflex impairment, autonomic imbalance and hypertension by negatively impacting OTergic drive and oxytocin levels in pre-autonomic neurons. Here, we assessed oxytocin gene and protein expression (qPCR and immunohistochemistry) within PVN subnuclei in sham-operated and ovariectomized Wistar rats. Conscious hemodynamic recordings were used to assess resting blood pressure and heart rate and the autonomic modulation of heart and vessels was estimated by power spectral analysis. We observed that the ovarian hormone deprivation in ovariectomized rats decreased baroreflex sensitivity, increased sympathetic and reduced vagal outflows to the heart and augmented the resting blood pressure. Of note, ovariectomized rats had reduced PVN oxytocin mRNA and protein expression in all pre-autonomic PVN subnuclei. Furthermore, reduced PVN oxytocin protein levels were positively correlated with decreased baroreflex sensitivity and negatively correlated with increased LF/HF ratio. These findings suggest that reduced oxytocin expression in OTergic neurons of the PVN contributes to the baroreflex dysfunction and autonomic dysregulation observed with ovarian hormone deprivation.

Keywords: baroreflex, oxytocin, arterial pressure, autonomic nervous system, ovariectomy

## INTRODUCTION

Cardiovascular diseases are the main cause of death in women (Mozaffarian et al., 2015). Menopause, characterized by a reduction in the circulating levels of the ovarian hormones progesterone and estrogen, is an important risk factor for cardiovascular diseases. More specifically, ovarian hormone deprivation has been shown to lead to hypertension, abnormal plasma lipids, endothelial dysfunction, elevated oxidative stress, autonomic imbalance and baroreflex impairment, which collectively result in high cardiovascular morbidity and mortality (Jensen et al., 1990; Taddei et al., 1996; Mercuro et al., 2000; Irigoyen et al., 2005; Flues et al., 2010).

It is well known that brainstem nuclei and the paraventricular nucleus of hypothalamus (PVN) are major sites for regulation of cardiovascular autonomic responses. A transient blood pressure rise activates neurons within the nucleus tractus solitarii (NTS), which stimulate parasympathetic neurons in the nucleus ambiguus (NA) and the dorsal motor nucleus of vagus (DMV) (Dampney, 1994). The NTS also projects and excites the GABAergic neurons in the caudal ventrolateral medulla (CVLM), which reduce the activity of sympathetic premotor neurons within the rostral ventrolateral medulla (RVLM) projecting to heart and vessels (Dampney, 1994) This information is continuously conveyed to pre-autonomic PVN neurons, via ascending cathecolaminergic afferents arising from NTS and CVLM, thus regulating their neurosecretory activity (Michelini and Stern, 2009; Sladek et al., 2015). Oxytocinergic (OTergic) pre-autonomic neurons within the dorsal cap, ventromedial and posterior PVN subnuclei are activated by the ascending afferents and project back to brainstem nuclei and spinal cord, thus modulating the autonomic circulatory control (Buijs, 1978; Michelini and Stern, 2009; Geerling et al., 2010; Cruz et al., 2013; Sladek et al., 2015). The physiological relevance of central oxytocin-dependent signaling in control of cardiovascular responses has been demonstrated by several studies. Augmented oxytocinergic drive to brainstem areas sensitizes the baroreceptor reflex control of heart rate facilitating bradycardic responses (Higa et al., 2002; Cavalleri et al., 2011).

Several types of estrogen receptors are expressed in the PVN. The nuclear estrogen receptor β (ER-β), a G proteincoupled estrogen receptor, is specifically expressed in OTergic neurons (Alves et al., 1998; Hrabovszky et al., 1998; Brailoiu et al., 2007); however, a functional link between ovarian hormones and central oxytocin signaling remains to be determined. Therefore, we hypothesized that ovarian hormone deprivation blunts oxytocin expression and signaling in preautonomic areas of the PVN, thus contributing to baroreflex impairment, autonomic imbalance and hypertension. In the present study, we used ovariectomized rats to investigate whether ovarian hormones are required for normal oxytocin mRNA expression and protein content in the posterior, ventromedial and dorsal cap subnuclei of the PVN. Conscious hemodynamic recordings were used to assess resting blood pressure, heart rate and autonomic modulation of heart and vessels.

## METHODS

### Animals

All surgical procedures and experimental protocols (11/2014) were approved by the Ethics Committee on Animal Research of Federal University of Sergipe, in accordance with the guide for the care and use of laboratory animals published by National Institute of Health (National Research Council (US) Committee for the Update of the Guide for the Care and Use of Laboratory Animals, 2011). Thirty-three female Wistar rats (193 ± 5 g) were housed in propylene cages with controlled environmental temperature (22 ± 1 ◦C), 12 h dark/light cycle and water and chow ad libitum. Animals were randomly divided into two groups and submitted to ovariectomy (OVX, n = 16) or sham surgery (SHAM, n = 17).

### Surgical Procedures

At 10 weeks of age, rats were anesthetized (Ketamine: Fort Dodge IA, USA, 80 mg.kg−<sup>1</sup> plus Xylazine: Fort Worth TX, USA, 12 mg.kg−<sup>1</sup> , i.p.) and an abdominal incision was made. Ovaries were then exposed and removed through the oviduct section. SHAM rats were submitted to the same procedures without ovaries' removal. Rats were treated with ketoprofen (Biofen 1%, 2 mg.kg−<sup>1</sup> ; Biofarm, Jaboticabal, Brazil) and penicillin (Pentabiotico Veterinario 24,000 i.u.kg−<sup>1</sup> ; Fontoura Wyeth, Sao Paulo, Brazil) and allowed to recover for 1 week (Irigoyen et al., 2005). Efficiency of ovariectomy was confirmed by analysis of vaginal smears collected for 4 consecutive days. Essentially, only rats that exhibited diestrus phase in all days were included in the OVX group. All animals allocated to the SHAM group were in a regular estrous cycle and were euthanized with the same age.

Eight weeks after OVX or SHAM surgeries, rats were anesthetized (ketamine, 80 mg.kg−<sup>1</sup> , Fort Dodge IA, USA, plus xylazine, 12 mg.kg−<sup>1</sup> , Fort Worth TX, USA, i.p.) and a polyethylene catheter was implanted (PE-10/PE-50, Intramedic, Becton Dickinson Company, Sparks, MD, USA) into the left femoral artery. Twenty-four hours after the procedure, rats are freely moving in their cages and did not exhibit signs of stress. The arterial catheter was then connected to a pressure transducer coupled to the preamplifier (FE221, Bridge Amp, ADInstruments, Bella Vista, NSW, Australia) and the recording system (Powerlab, ADInstruments, Bella Vista, NSW, Australia). Resting pulsatile and mean arterial pressure (AP) were continuously recorded for 30 min and processed using a dedicated software (LabChart 7, ADInstruments, Bella Vista, NSW, Australia). The inflection points of pressure signal were identified to generate beat-to-beat time series of mean arterial pressure (MAP), systolic arterial pressure (SAP), diastolic arterial pressure (DAP) and pulse interval (PI). Heart rate (HR) was calculated as 1/PI. To avoid any bias, all rats included in this study were submitted to the same procedures.

## Assessment of Cardiovascular Autonomic Control

The analyses of PI and SAP variability were performed using the CardioSeries software (v2.4) as previously described (Oliveira et al., 2012). Beat-to-beat series were obtained from pulsatile arterial pressure and converted into discrete points every 100 ms, using cubic spline interpolation (10 Hz). Ten-minutes record of each rat was used for this analysis. Prior to the calculation of the spectral density, data was visually inspected and the nonstationary segments were disregarded. Data was then divided into half-overlapping sequential sets of 512 data points (51.2 s). Segments were windowed with a Hanning window and then the spectrum of each segment was calculated by the FFT algorithm. The PI spectrum, representing the variability of autonomic control of the heart, is composed by bands of very low frequency (VLF; <0.02 Hz), low frequency (LF; 0.2–0.75 Hz) and high frequency (HF; 0.75–3.0 Hz). These values are usually expressed in normalized (nu). LF and HF units, obtained through the division of respective LF and HF power by the total power minus VLF. VLF of PI represents humoral factors that influence heart rate, HF of PI indicates the cardiac parasympathetic modulation, while LF of PI is generally accepted as an index of cardiac sympathetic modulation and LF/HF ratio represents the sympatho-vagal balance to the heart (Malliani et al., 1991; Stauss, 2007; Oliveira et al., 2012). The SAP spectrum reflects arterial pressure variance and is quantified in mmHg<sup>2</sup> . Its VLF component is affected by myogenic vascular function, reninangiotensin system and nitric oxide, LF of SAP represents the vasomotor sympathetic modulation plus endothelial nitric oxide modulation, while HF of SAP is mainly influenced by alterations in cardiac output coupled to changes in the venous return during respiration (Janssen et al., 1995; Heart rate variability: standards of measurement, physiological interpretation and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, 1996; Stauss, 2007).

Spontaneous baroreflex sensitivity (sBRS) was measured in the time domain using the sequence method (Rienzo, 1995). Beat-to-beat arterial pressure series were analyzed with the CardioSeries software (v2.4) to detect sequences of at least 4 beats with increased SAP followed by PI lengthening or decreased SAP with PI shortening that showed correlations greater than 0.85. The slope of the linear regression between SAP and PI was used as a measure of sBRS (mmHg/s).

### Tissue Sampling

After the functional measurements, rats were deeply anesthetized with ketamine and xylazine (300 and 60 mg/kg, respectively, ip) leading to the respiratory arrest. Rats assigned to the PCR experiments were immediately subjected to transcardiac perfusion with saline solution (0.09%, 40 ml/min, 5 min) and decapitated to remove the brain, which was quickly transferred to dry-ice. A slice including the medial and caudal parts of the nucleus (from 1.40 to 2.30 mm caudal to bregma, 800–1000µm) was taken at the hypothalamic level and immediately frozen for bilateral punching of the PVN. Samples were then stored at −80◦C for subsequent analyses. Rats assigned to immunohistochemistry assay were subjected to transcardiac perfusion with Dulbecco's Modified Eagle's Medium (DMEM-Sigma, 40 ml/min, ∼300 ml), followed by infusion of 4% paraformaldehyde in 0.01M PBS (pH 7.4, 40 ml/min, ∼300 ml) and decapitated for brain removal. The brain was post-fixed in 4% paraformaldehyde for 48 h at room temperature, cryoprotected in Tris-PBS (10 mM Tris, 0.9% NaCl, 10 mM phosphate buffer, pH 7.4, containing 0.05% merthiolate) containing 20% sucrose at room temperature for 24–30 h, and then incubated in 0.01M PBS that contained 30% sucrose solution and stored at 4◦C for 3–4 days before further processing.

## Quantitative Real-Time PCR

mRNA expression was assessed via quantitative real time PCR (qPCR). TRizol reagent (0.5 ml) was added to samples and RNA extraction was performed according to the manufacturer's instructions. After extraction, RNA was dissolved in 10µl of DEPC water and stored at −80◦C. Reverse transcriptase reaction was performed only after DNase I was added to samples, and firststrand cDNA synthesis was made with 1µg RNA/reaction, using ImProm-II Reverse transcriptase (Promega, USA), according to the manufacturer's instructions. RNaseOUT was also present during this process and cDNA was stored at −20◦C. qPCR was performed in the Applied Biosystems 7500 Fast Real-Time PCR System (ThermoFisher, California, USA) using Platinum SYBR QPCR Supermix-UDG and specific oligonucleotides for Oxytocin (OT, sense primer, TAGACCTGGATATGCGCAAG; antisense primer, CTCGGAGAAGGCAGACTCAG) and Glyceraldehyde 3-phosphate dehydrogenase (GAPDH, sense primer, GGGCAGCCCAGA ACATCAT; antisense primer, CCGTTCAGCTCTGGGATGAC). OT mRNA expression, normalized to GAPDH, was calculated using the 11Ct method (Pfaffl, 2001) and expressed as fold change in relation to values exhibited by the SHAM-operated rats. All reagents and oligonucleotides were purchased from Invitrogen (San Diego, CA).

### Immunohistochemical Analyses

Sequential hypothalamic coronal sections (30µm, −1.80 to −2.12 caudal to the Bregma) were obtained as previously described (Paxinos and Watson, 2006) using a cryostat (Leica CM 1850; Nussloch, Germany). Sections were collected in tissue culture wells with 0.01 M PBS and then incubated with 0.03% Triton X-100 and 10% normal donkey serum for 30 min. For the immunofluorescence assay, sections were incubated overnight with primary antibody (polyclonal guinea pig anti-oxytocin, 1:200,000 dilution; Bachem, Bubendorf, Switzerland), followed by a 2-h incubation with secondary antibody (donkey antiguinea pig Cy3-labeled, 1:500 dilution; Jackson ImmunoResearch Laboratories, West Grove, PA) diluted in T-PBS containing 0.03% Triton X-100 at room temperature. Slices were placed on slides and mounted with a coverslip and SlowFade Gold anti-fade reagent. Specificity of the antibody was tested by processing sideby-side slices without the incubation with the primary or the secondary antibody.

### Image Acquisition and Analysis

OT immunoreactivity was captured using an epifluorescence microscope (Leica DMLB, Wetzlar, Germany; ×200 magnification) coupled to a digital camera (Axio-Cam HRC; Carl Zeiss, Vision GmbH, Aalen, Germany). Slides were visually inspected to localize the dorsal cap, ventromedial, posterior and magnocellular PVN subnuclei, and image analyses were performed using the ImageJ software (NIH) as previously described (Higa-Taniguchi et al., 2007; Cavalleri et al., 2011; Cruz et al., 2013). The relative OT density and integrated OT density were used as indexes of protein content in specific PVN subnuclei. Relative OT density was calculated as the ratio of the area occupied by the thresholded signal and the total area of interest, and expressed as a percentage (% of total area). Integrated OT density (in arbitrary units, AU) was obtained by the product of OT density and the signal intensity.

### Statistical Analysis

Results were expressed as means ± SEM. Body mass was analyzed by two-way ANOVA followed by the Bonferroni post hoc test. Potential differences in OT mRNA and OT protein content within the PVN subnuclei, as well as hemodynamic parameters and baroreflex sensitivity were compared between SHAM and OVX rats and analyzed by unpaired t-test. Relationships between OT content within the ventromedial PVN and sBRS was assessed by the Pearson correlation coefficient. Differences were considered significant at p < 0.05.

## RESULTS

### Body Mass and Autonomic Cardiovascular Control

Although there were no differences between groups in body mass at beginning of experiments, OVX rats gained significantly more weight than SHAM rats by the end of the study (∼50% vs. ∼26%, respectively; **Table 1**). OVX rats also exhibited higher basal MAP (p = 0.019) accompanied by a trend toward higher HR when compared to the SHAM group (p = 0.062, **Table 1**). PI variance, VLF, LF and HF in absolute units were similar in OVX and SHAM rats (p = 0.344, 0.667, 0.474 and 0.551, respectively); however, these rats exhibited reduced HF of PI (p = 0.015) and elevated LF of PI (p = 0.015) yielding an elevated LF/HF ratio (p = 0.015). SAP variance was higher in OVX vs. SHAM rats (p = 0.011), without significant differences between groups in the VLF, LF or HF of SAP (p = 0.059, 0.210 and 0.730, respectively). In addition, sBRS was largely reduced in OVX group when compared to SHAM (−59%, p = 0.005).

### Effects of Ovarian Hormone Deprivation on PVN OTergic Neurons

Relative OT mRNA expression in the PVN was 45% lower in OVX rats compared to SHAM rats (p = 0.030) (**Figure 1**, upper panel). Accordingly, ovarian hormone deprivation significantly reduced OT protein expression levels, quantified by both relative and integrated OT densities within the posterior (p = <0.001 and p = 0.024, respectively), ventromedial (p = 0.009 for both comparisons) and dorsal cap (p = 0.041 and p = 0.029, respectively) PVN subnuclei (**Figures 1A–C**). In contrast, no differences were observed in relative and integrated OT densities in the magnocellular neurons of OVX rats compared to SHAM controls (p = 0.778 and p = 0.852, respectively) (**Figures 1A–C**).

Interestingly, reduced oxytocin protein levels within the posterior and ventromedial PVN subnuclei were strongly correlated with decreased autonomic control of the heart in OVX rats, as indicated by both decreased baroreflex sensitivity TABLE 1 | Body mass (before and after surgeries), baseline mean arterial pressure (MAP) and heart rate (HR) and cardiovascular autonomic evaluation in rats submitted to ovariectomy (OVX) or SHAM surgery.


*Values are expressed as mean* ± *SEM. Body weight was evaluated in 16–17 rats/group and hemodynamic/autonomic measurements were made in 6–7 rats/group. VLF, LF, and HF represent the very low frequency, low frequency and high frequency bands of pulse interval (PI) and systolic arterial pressure (SAP) variance. sBRS, spontaneous baroreflex sensitivity. Significances (*\**P* < *0.05),* \*\**P* < *0.01;* \*\*\**P* < *0.001 are* \**vs. SHAM; † vs. before.*

and increased LF/HF ratio (**Figure 2**, **Table 2**). Despite not reaching statistical significance, a similar relationship was also observed between OT expression levels at the dorsal cap and sBRS (**Table 2**).

## DISCUSSION

The present findings provide further support for a dysfunctional autonomic cardiovascular control caused by ovarian hormone deprivation. More specifically, our results demonstrate that lack of ovarian hormones impairs baroreceptor reflex control of heart rate, causes autonomic imbalance and increases blood pressure. Although several cellular and molecular mechanisms are likely involved, the present findings further indicate that reduced expression of oxytocin in OTergic neurons of the pre-autonomic PVN subnuclei are involved in the defective autonomic regulation observed. An important role for oxytocin in autonomic regulation has been previously documented (Mercuro et al., 2000; Cavalleri et al., 2011; Cruz et al., 2013). However, to our knowledge, this is the first study demonstrating that ovarian hormone deprivation reduces oxytocin gene and protein expression in pre-autonomic PVN neurons, and that these changes are closely correlated with autonomic dysfunction in ovariectomized rats.

The incidence of cardiovascular diseases is considerably lower in pre-menopausal women than in men. Nevertheless, this difference is mitigated by the reduction of circulating estrogen levels as women age (Becker and Corrao, 1990; Wenger et al., 1993; Tunstall-Pedoe et al., 1994; Golden et al.,

2002). Consistent with our findings, previous studies have shown that estrogen promotes cardioprotection and metabolic homeostasis and that ovarian hormone deprivation, in addition to cardiovascular deficits, causes body weight gain with increased abdominal fat and reduced signaling in central nuclei that control appetite and satiety (Irigoyen et al., 2005; Flues et al., 2010; Lizcano and Guzmán, 2014). Oxytocin is an important peptide involved in the food intake control. Oxytocin or oxytocin agonist centrally administered decreased food consumption while oxytocin antagonist pretretment failed to increase chow intake (Olson et al., 1991; Mullis et al., 2013). In fact, weight gain induced by abnormal central oxytocin expression could per se contribute to baroreflex dysfunction and hypertension development (Skrapari et al., 2007; Re, 2009).

Previous studies indicate that ovarian hormone deprivation is able to reduce oxytocin mRNA expression in the brain, including the PVN (Miller et al., 1989; Patisaul et al., 2003). This study, however, is the first to indicate that lack of female hormones is specific to decrease oxytocin expression in preautonomic PVN subnuclei and that this effect is related to autonomic impairment.

Reduced baroreflex sensitity and increased sympathetic outflow to the heart are predictors of morbidity and mortality in several cardiovascular diseases (Billman et al., 1982; Becker and Corrao, 1990; Mercuro et al., 2000). Here, we observed

TABLE 2 | Regression equations correlating OT immunoreactivity (OTir, measured as relative density and integrated density in different PVN subnuclei) with spontaneous baroreflex sensitivity (BRS) and sympatho-vagal balance to the heart (LF/HF ratio) in rats submitted to SHAM and OVX surgery.


*OTir x BRS and OTir x LF/HF ratio correlations were made with 8–10 rats.* \**(P* < *0.05),* \*\**(P* < *0.01) and* \*\*\**(P* < *0.001) denote significant correlations.*

significant reduction in sBRS coupled with increased sympathetic modulation to the heart and elevated blood pressure in rats deprived of ovarian hormones. These findings are in line with previous studies in ovariectomized rats, in which these changes were associated with augmented oxidative stress in the heart (Irigoyen et al., 2005; Flues et al., 2010). It should be noted that some studies did not find blood pressure changes and autonomic misbalance in ovariectomized rats (Nickenig et al., 1998; Dias et al., 2010). This inconsistency may be explained by the time that rats were exposed to ovarian hormone deprivation. In these studies, the ovariectomy lasted 3–5 weeks. To our knowledge, studies showing high blood pressure levels and dysautonomia were performed at least 8 weeks after ovariectomy induction (Hernández et al., 2000; Irigoyen et al., 2005; Flues et al., 2010 and the present set of data).

Our results together with previous studies suggest that deficient oxytocin expression within preautonomic PVN subnuclei may be an important mechanism of central autonomic deregulation, which contributes to cardiac oxidative stress and may lead to heart dysfunction. Future studies are necessary to test this proposition. In fact, heart disease is a major cause of morbidity and mortality in post-menopausal women (Mozaffarian et al., 2015).

Accumulating evidence indicate a critical role for preautonomic PVN OTergic neurons in the modulation of baroreceptor reflex control of heart rate. The nucleus tractus solitarii/dorsal motor nucleus of vagus (NTS/DMV) complex receives dense PVN OTergic projections, whose activation facilitates vagal outflow to the heart, thus improving reflex bradycardia during transient pressure increases (Buijs, 1978; Higa et al., 2002). It has also been shown that oxytocin released within the NTS/DMV during an acute bout of exercise reduces exercise tachycardia and causes resting bradycardia in trained rats (Braga et al., 2000; Higa-Taniguchi et al., 2009). Collectively, these observations indicate that adequate levels of oxytocin expression in PVN subnuclei are required for effective autonomic regulation of the cardiovascular system. In fact, our current results demonstrate reduced oxytocin expression within the ventromedial and posterior PVN and baroreflex impairment in OVX rats. It was shown that reduced oxytocin content, the neurotransmitter co-released with glutamate in those preautonomic neurons, blunts the activation of OTergic projections to dorsal brainstem areas (Piñol et al., 2014). Peters et al. (2008) also showed that activation of these projections augments glutamate release probability and the frequency of miniature excitatory post-synaptic currents in 2nd order NTS neurons while oxytocin antagonist pretreatment completely blocks this effect. Indeed, in a previous study in conscious rats we observed that oxytocin administration within the NTS/DMV area, mimicking the activation of the long-descending PVN oxytocinergic projections, augments the reflex bradycardia during baroreceptors loading, while its endogenous blockade reduces the bradycardic response (Higa et al., 2002). We also demonstrated that atropine, but not propranolol pretreatment, abrogates the augmentation of reflex bradycardia, indicating that improvement of baroreflex gain is mediated by oxytocininduced increase in the vagal tonus to the heart (Higa et al., 2002; Michelini, 2007). The present set of data also showed strong positive correlations between OT content and baroreflex sensitivity. Although the correlations per se are not proof of causality, our results taken together with previous data on oxytocin and autonomic control strongly suggest that the reduction of PVN oxytocinergic drive in ovariectomized rats may be responsible for both the blunting of baroreflex sensitivity and the increased sympatho-vagal balance to the heart, as shown by the present set of data.

Heart rate variability has been widely accepted as an index of cardiovascular autonomic function. HF has always been associated to parasympathetic modulation; LF represents the sympathetic modulation, but there is evidence that parasympathetic component is also partially aggregated to this band (Appel et al., 1989; Burr, 2007). Some studies verified that LF of PI does not correlate with cardiac norepinephrine spillover and is very low in heart failure individuals (Adamopoulos et al., 1992; Guzzetti et al., 1995; Eisenhofer et al., 1996; Moak et al., 2007). However, the normalization procedures applied to absolute results yield values that are exchangeable across to different evaluation methods. Helpfully, presentation of data in normalized units mitigated several differences in the computed band power (Burr, 2007) and LF (in normalized units) is generally accepted as an index for sympathetic variability. In addition, our results corroborate previous studies that performed cardiac autonomic evaluation by pharmacological blockade (a gold standard method) in a similar protocol of ovarian hormone deprivation (Irigoyen et al., 2005; Flues et al., 2010).

OTergic neurons express β estrogen receptors (ER-β), which may control neuronal oxytocin gene/protein expression within PVN subnuclei (Alves et al., 1998; Hrabovszky et al., 1998). Stern and Zhang (2003) showed that preautonomic neurons within the posterior PVN subnucleus projecting to the rostral ventrolateral medulla exhibited high ER-β density and reduced

### REFERENCES


excitability after ovarian hormone deprivation. It has also been shown that G-protein coupled estrogen receptors (GPERs) are located in several areas of the central nervous system, including the ventromedial and dorsal cap parvocellular neurons (Brailoiu et al., 2007). GPERs have been shown to co-localize with oxytocin in magnocellular PVN and supraoptic neurons (Brailoiu et al., 2007). Whether a direct physical interaction between GPERs and oxytocin in PVN subnuclei indeed occurs, and how such interaction may potentially contribute to the maintenance of oxytocin levels (e.g., via stabilization) remain to be determined. A potential transcriptional regulation of the oxytocin gene by ER-β may also be in place and deserves further investigation.

In summary, our results showed that ovarian hormone deprivation decreases oxytocin gene and protein expression within PVN pre-autonomic neurons involved in circulatory control. The observed deficits in OTergic modulation were accompanied by reduced vagal and increased sympathetic modulation to the heart and augmented SAP variability. Finally, oxytocin content in the PVN was closely correlated with autonomic control of the heart suggesting that depressed hypothalamic OTergic modulation significantly contributes to the cardiovascular deficits observed in ovarian hormone deprivation.

### AUTHOR CONTRIBUTIONS

VUD, RS, and CD: Performed the experiments; VUD and RS: Analyzed the data; VUD, VL, JD, and LM: Wrote, edited and revised the manuscript; VL, VJD, and LM: Approved the final version of the manuscript.

## FUNDING

This work was supported by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES/ Proc BEX 6292/15-1), Fundação de Apoio à Pesquisa e Inovação Tecnológica do Estado de Sergipe (FAPITEC-SE) and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP, proc 2011/51410-9). LM is a Research Fellow of CNPq.


nervous system. Am. J. Cardiol. 85, 787–789, A9. doi: 10.1016/s0002- 9149(99)00865-6


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 De Melo, Saldanha, Dos Santos, Cruz, Lira, Santana-Filho and Michelini.. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Increased Blood Pressure Variability Prior to Chronic Kidney Disease Exacerbates Renal Dysfunction in Rats

Frederico F. C. T. Freitas <sup>1</sup> , Gilberto Araujo<sup>1</sup> , Marcella L. Porto<sup>2</sup> , Flavia P. S. Freitas <sup>2</sup> , Jones B. Graceli <sup>3</sup> , Camille M. Balarini <sup>4</sup> , Elisardo C. Vasquez 2, 5, Silvana S. Meyrelles <sup>2</sup> and Agata L. Gava2, 6 \*

*<sup>1</sup> Biotechnology Graduate Program, Health Sciences Center, Federal University of Espirito Santo, Vitoria, Brazil, <sup>2</sup> Physiological Sciences Graduate Program, Health Sciences Center, Federal University of Espirito Santo, Vitoria, Brazil, <sup>3</sup> Morphology Department, Health Sciences Center, Federal University of Espirito Santo, Vitoria, Brazil, <sup>4</sup> Department of Physiology and Pathology, Health Sciences Center, Federal University of Paraiba, Joao Pessoa, Brazil, <sup>5</sup> Pharmaceutical Sciences Graduate Program, University of Vila Velha, Vila Velha, Brazil, <sup>6</sup> Division of Nephrology, McMaster University, Hamilton, ON, Canada*

### Edited by:

*Ovidiu Constantin Baltatu, Universidade Camilo Castelo Branco, Brazil*

### Reviewed by:

*Ulla Kopp, University of Iowa, USA Cara Hildreth, Macquarie University, Australia*

> \*Correspondence: *Agata L. Gava agatagava@hotmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *12 May 2016* Accepted: *09 September 2016* Published: *23 September 2016*

### Citation:

*Freitas FFCT, Araujo G, Porto ML, Freitas FPS, Graceli JB, Balarini CM, Vasquez EC, Meyrelles SS and Gava AL (2016) Increased Blood Pressure Variability Prior to Chronic Kidney Disease Exacerbates Renal Dysfunction in Rats. Front. Physiol. 7:428. doi: 10.3389/fphys.2016.00428* Increased blood pressure variability (BPV), which can be experimentally induced by sinoaortic denervation (SAD), has emerged as a new marker of the prognosis of cardiovascular and renal outcomes. Considering that increased BPV can lead to organ-damage, the goal of the present study was to evaluate the effects of SAD on renal function in an experimental model of chronic kidney disease (CKD). SAD was performed in male Wistar rats 2 weeks before 5/6 nephrectomy and the animals were evaluated 4 weeks after the induction of CKD. Our data demonstrated that BPV was increased in SAD and CKD animals and that the combination of both conditions (SAD+CKD) exacerbated BPV. The baroreflex sensitivity index was diminished in the SAD and CKD groups; this reduction was more pronounced when SAD and CKD were performed together. 5/6 nephrectomy led to hypertension, which was higher in SAD+CKD animals. Regarding renal function, the combination of SAD and CKD resulted in reduced renal plasma and blood flow, increased renal vascular resistance and augmented uraemia when compared to CKD animals. Glomerular filtration rate and BPV were negatively correlated in SAD, CKD, and SAD+CKD animals. Moreover, SAD+CKD animals presented a higher level of glomerulosclerosis when compared to all other groups. Cardiac and renal hypertrophy, as well as oxidative stress, was also further increased when SAD and CKD were combined. These results show that SAD prior to 5/6 nephrectomy exacerbates renal dysfunction, suggesting that previous augmented BPV should be considered as an important factor to the progression of renal diseases.

Keywords: blood pressure variability, sinoaortic denervation, chronic kidney disease, 5/6 nephrectomy, renal function

## INTRODUCTION

It is well-established that the maintenance of blood pressure (BP) at stable levels is a sine qua non condition for adequate tissue perfusion (Vasquez et al., 2012). Increased blood pressure variability (BPV) is closely associated with the development, progression and severity of cardiac, vascular, and renal organ damage (Sander et al., 2000; Sega et al., 2002; Mancia and Parati, 2003; Tatasciore et al., 2007) as well as with an augmented risk of cardiovascular and renal outcomes (Kikuya et al., 2000; Pringle et al., 2003; Hansen et al., 2010; Stolarz-Skrzypek et al., 2010). Focusing on the kidney, renal function is linearly negatively associated with BPV, and changes in this parameter, regardless the mean BP levels, may predict the development and progression of renal damage (Parati et al., 2012).

Experimentally, increased BPV can be induced by the bilateral disruption of the afferent pathway of the arterial baroreflex system (Kudo et al., 2009), also known as sinoaortic denervation (SAD). This model results in a transient elevation of BP, followed by its normalization within a few days (Osborn and England, 1990) and increased BPV (Norman et al., 1981). In rats, SAD also leads to cardiac hypertrophy with impaired diastolic and systolic function, as well as pulmonary hypertension (Flues et al., 2012). In the kidney, SAD causes significant alterations in renal structures, such as patchy focal sclerotic changes associated with glomerular and tubular atrophy, and interstitial fibrosis in the renal cortex. The interlobular and afferent arterioles adjacent to the sclerotic lesions present arteriolar remodeling characterized by VSMC proliferation and extracellular matrix deposition, leading to the luminal narrowing and occlusion (Aoki et al., 2014).

Chronic kidney disease (CKD) is a serious disorder, and its prevalence is increasing worldwide (James et al., 2010). The progressive nature of CKD and the ensuing end-stage renal disease put a substantial burden on global health-care resources (Meguid El Nahas and Bello, 2005). The classical mechanisms involved in the progression of CKD include activation of the renin-angiotensin system, increased oxidative stress, inflammatory cytokines and deposition of extracellular matrix, usually due to hypertension and/or diabetes (Pirkle and Freedman, 2013; Miranda-Díaz et al., 2016). However, increasing evidence has showed that augmented BPV can also be related to the clinical outcomes in patients with CKD (Mallamaci and Tripepi, 2013). Considering that, we hypothesized that increased BPV prior to the onset of kidney disease could accelerate the disease progression. To test this hypothesis, we evaluated the effects of SAD previously to 5/6 nephrectomyinduced CKD on renal function, glomerulosclerosis, oxidative stress, and cardiovascular parameters. Our data demonstrate that augmented BPV prior to CKD exacerbates kidney dysfunction and should be considered as an important risk factor to the progression of renal diseases.

### MATERIALS AND METHODS

### Animals

Experiments were conducted in male Wistar rats (8–10 weeks old), bred and maintained in the animal care facility at the Federal University of Espirito Santo, Brazil. The animals were housed in individual cages with a controlled temperature (22–23◦C) and humidity (60%) and exposed to a 12:12-h light-dark cycle. All of the experimental procedures were performed in accordance with the National Institutes of Health (NIH) guidelines, and the experimental protocols were previously approved by the Institutional Animal Care and Use Committee (CEUA-UFES Protocol n◦ . 03/2012).

### Experimental Groups

The animals were randomly divided into 4 groups: control (Sham), sinoaortic denervated (SAD), 5/6 nephrectomy (CKD) and SAD + 5/6 nephrectomy (SAD+CKD).

SAD was performed bilaterally in rats under anesthesia with a mixture of ketamine (50 mg/Kg, ip) and xylazine (10 mg/Kg, ip). After a midline neck incision, sternocleidomastoid muscles were reflected laterally, exposing the neurovascular sheath. Then, the aortic fibers traveling along the sympathetic trunk or as isolated fibers were resected as well as the superior laryngeal nerve. The superior cervical ganglia were also removed. To complete SAD, the carotid bifurcation was widely exposed, and the surrounding connective tissue was stripped off. The carotid fiber and the carotid body were sectioned (Krieger, 1964; Miao et al., 2006). The control group underwent a sham operation.

A well-established experimental model of CKD is the 5/6 nephrectomy, which was performed in this study. Two weeks after the SAD or sham procedure, the animals were anesthetized with a mixture of ketamine (50 mg/Kg, ip) and xylazine (10 mg/Kg, ip). Two of the three ramifications of the left renal artery were ligated, causing an infarction of ∼2/3 of the renal mass. The right renal artery and vein were ligated and the right kidney was removed. After these procedures, the peritoneum and the skin were sutured.

### Haemodynamic Measurements, Baroreflex Sensitivity Index (BSI) and BPV Determination

Four weeks after CKD induction, the animals underwent catheterization procedures. After anesthesia (ketamine 50 mg/Kg and xylazine 10 mg/Kg, ip.), a polyethylene catheter filled with heparin solution (50 UI/mL saline) was inserted into the femoral artery to measure mean arterial pressure (MAP), systolic blood pressure (SAP), diastolic blood pressure (DAP) and heart rate (HR), as well as for blood sampling. The vein was also catheterized for drug administration. The rats were allowed to recover during a 24-h period after the catheterization. To record arterial pressure, the arterial catheter was connected to a pressure transducer (Cobe Laboratories, USA) plugged into a pressureprocessor amplifier and data acquisition system (MP100, Biopac Systems, USA). A 45-min recording of MAP, SAP, DAP, and HR was obtained from conscious and freely moving rats. The BPV was quantified using the standard deviation of MAP during the recorded period.

The effectiveness of SAD was confirmed by testing the reflex heart rate responses to changes in arterial pressure during intravenous injections of phenylephrine (0.25–32.0µg/kg) and sodium nitroprusside (0.05–1.6µg/kg), in order to achieve changes in blood pressure ranging from 5 to 50 mmHg. For each animal, BSI was calculated using the mean of changes in HR/mean of changes in MAP elicited by the different doses of phenylephrine and sodium nitroprusside. BSI for each group was calculated by averaging the BSI of the all evaluated animals.

### Renal Function Studies

Renal function was determined using inulin (IN) and sodium para-aminohippurate (PAH) clearance to estimate the glomerular filtration rate (GFR) and renal plasma flow (RPF), respectively (Smith et al., 1945). Four weeks after CKD induction, a separate set of animals were anesthetized with sodium thiopental (50 mg/Kg ip.), the trachea was catheterized with a polyethylene tube (PE-90) to facilitate breathing, and a catheter (PE-240) was introduced into the bladder for urine sampling. The arterial catheter was connected to a pressure transducer (Cobe Laboratories, USA) plugged into a pressureprocessor amplifier and data acquisition system (MP100, Biopac Systems, USA) for continuous monitoring of MAP, SAP, DAP, and HR. The venous catheter was connected to an infusion pump (0.1 mL/min) and a saline solution (0.9%) containing 3% of mannitol was infused over 30 min. After this stabilization period, the animals received an intravenous injection of prime solution containing IN (300 mg/Kg) and PAH (6.66 mg/Kg) and were maintained on a continuous infusion of saline (0.9%) containing IN (15 mg/mL), PAH (4 mg/mL) and mannitol (3%) until the end of the experiment. At 30-min intervals, urine and blood samples were taken, for a total of 4 samples. Haematocrit was measured using a heparinized capillary tube. Plasma and urinary IN and PAH concentrations were measured using a colorimetric assay (Rocco et al., 2008). Blood samples were also used for plasma urea quantification through spectrophotometry.

Renal blood flow (RBF) and renal vascular resistance (RVR) were determined as previously described (Magalhães et al., 2006). Briefly, RBF was calculated by the equation RBF = RPF/(1 haematocrit), and RVR was calculated using the equation RVR = MAP/RBF.

In order to be able to correlate BPV with renal function, some animals were submitted to both procedures. For these animals, BPV was determined as above mentioned and inulin and PAH clearance was performed in the next day. Due to the high mortality rate when the two procedures were performed, especially in the SAD+CKD group, the number of animals in correlation analysis is limited to 4–6.

Urinary water excretion (24 h) was obtained using a metabolic cage to avoid the effects of anesthesia and/or mannitol required to perform renal haemodynamic evaluation. By the end of the treatment, a separated group of animals were placed on metabolic cages during a 24-h accommodation period, followed by urine sampling also during 24 h.

### Quantification of Superoxide Production

Reactive oxygen species generation was performed in blood cells using flow cytometry. Briefly, blood samples were lysed with lysing buffer 1X (Becton Dickinson) for 10 min at 37◦C to remove erythrocytes. The cell suspension was then washed twice in phosphate-buffered saline (PBS) plus 1% Foetal Bovine Serum (FBS) for 10 min and centrifuged at 1200 rpm; and the supernatant was discarded. The cells were collected and resuspended in 1 mL PBS for flow cytometry analysis. For intracellular superoxide anion generation measurements, DHE (160 mM) was added to the cell suspension (10<sup>6</sup> cells), which was then incubated at 37◦C for 30 min in the dark. Samples were treated with 10 mM doxorubicin for 5 min to create oxidative stress without cell toxicity for the positive control, and the negative control cells were incubated with ethanol. After washing and resuspending in PBS, the cells were maintained on ice for immediate detection by flow cytometry (FACSCanto II, Becton Dickinson, San Juan, CA, USA). Data were analyzed using FACSDiva software (Becton Dickinson). For DHE fluorescence quantification, samples were acquired in duplicate, and 10,000 events were used for each measurement. Red fluorescence was detected between 564 and 606 nm using a 585/42 bandpass filter. Data are expressed as the median fluorescence intensity.

### Cardiac and Kidney Hypertrophy and Glomerular Collagen Deposition

At the end of the experiments, the animals were euthanized with an overdose of sodium thiopental and perfused via the left ventricle with Krebs-Hepes buffer (pH 7.4). The left kidney and the heart were removed, cleaned of connective tissue and weighed. To generate the hypertrophy index, the ratio of kidney or heart weight to body weight was calculated. Immediately after being weighed, the kidneys were cut longitudinally, fixed in Bouin solution and then embedded in paraffin. Five micrometerthick sections were obtained and stained with Masson' trichrome for glomerular collagen deposition quantification. The glomeruli were photographed for later analysis. Images were captured with color video camera (VKC150; Hitachi, Tokyo, Japan) connected to a microscope (AX70; Olympus, Center Valley, PA) and analyzed with a specific image program (2100 Leica EWS; Leica, Wetzlar, Germany) by a person blinded to the experimental groups. To determine glomerular sclerosis, at least 30 glomeruli were analyzed in Masson's trichromestained sections using the Image J program. The mean of the glomerular-stained areas (%) was used to determine the collagen deposition for each animal. The glomerulosclerosis index was determined using a semi-quantitative scale based on % stained area of the glomerulus: (1) 0–25; (2) 25–50; (3) 50–75; and (4) >75%.

### Statistical Analysis

Values are expressed as means ± S.E.Ms. Statistical comparisons between the different groups were performed by Student's ttest or two-way analysis of variance (ANOVA) followed by Bonferroni's post-hoc test. The correlation between BPV and GFR variables was examined using Pearson's correlation coefficient. Statistical significance was assessed using a linear regression model. The statistical analyses were performed using Prism software (Prism 5, GraphPad Software, Inc., San Diego, CA, USA). A value of p < 0.05 was regarded as statistically significant.

### RESULTS

### Haemodynamic Measurements, Baroreflex Sensitivity Index (BSI) and BPV Determination

**Table 1** summarizes the results of systolic (SAP), diastolic (DAP), and mean (MAP) arterial pressure and heart rate (HR) in conscious animals 24 h after the catheterization surgery. No differences were observed in any parameters between sham and SAD group. On the other hand, CKD animals presented a greater SAP, DAP, and MAP when compared to the sham and SAD. Interestingly, these changes were exacerbated in the SAD+CKD animals, which also displayed an increased HR.

The results of the baroreflex sensitivity index (BSI) determination are displayed in **Figure 1**. **Figure 1A** shows typical recordings of phenylephrine-induced bradycardic responses (left) and BSI quantification (right) in all studied groups. As expected, the BSI was significantly reduced in the SAD group (−0.45 ± 0.03 bpm/mmHg, p < 0.01) when compared to the sham animals (−1.7 ± 0.06 bpm/mmHg). CKD animals also displayed a slightly reduction in BSI (−1.20 ± 0.03 bpm/mmHg, p < 0.01). The combination of SAD and CKD led to a greater reduction in BSI (−0.22 ± 0.03 bpm/mmHg, p < 0.01) when compared to all other groups. **Figure 1B** shows typical recordings of sodium nitroprusside-induced tachycardic responses (left) and BSI quantification (right) in all studied groups. Like in the bradycardic responses, SAD (0.50 ± 0.02 bpm/mmHg, p < 0.01) and CKD (2.00 ± 0.09 bpm/mmHg, p < 0.01) animals presented a diminished BSI when compared to the sham group (2.93 ± 0.08 bpm/mmHg). Once again, the combination of SAD and CKD worsened this parameter (0.26 ± 0.03 bpm/mmHg, p < 0.01).

We also evaluated BP variability using the standard deviation (SD) of MAP. As demonstrated in **Figure 2**, the SAD animals presented an increased SD of MAP (5.53 ± 0.13 mmHg, p < 0.01) when compared to the sham group (2.40 ± 0.05 mmHg). CKD also augmented the SD of MAP (3.69 ± 0.12 mmHg, p < 0.01); however, when CKD was combined with SAD, the increase in BP variability was more pronounced (8.51 ± 0.12 mmHg, p < 0.01). In **Figure 2**, we can also observe the typical recordings of MAP during resting state in all studied groups. Note that BP variability is significantly


HR (bpm) 344 ± 5 366 ± 6 356 ± 8 387 ± 16<sup>a</sup> *SAP, systolic arterial pressure; DAP, diastolic arterial pressure; MAP, mean arterial pressure; HR, heart rate. All values are expressed as means* ± *SEMs. The number in parentheses represents the number of animals in each group. <sup>a</sup>p* < *0.05 vs. Sham; <sup>b</sup>p* < *0.05 vs. SAD; <sup>c</sup>p* < *0.05 vs. CKD.*

DAP (mmHg) 80 ± 2 83 ± 3 128 ± 10ab 145 ± 10abc SAP (mmHg) 140 ± 2 135 ± 3 184 ± 10ab 224 ± 10abc increased in the SAD+CKD animals when compared to all other groups.

### Renal Function Evaluation

We also evaluated the effects of BPV on renal function parameters, including GFR, RPF, RBF and RVR (**Figure 3**). Inulin clearance results (**Figure 3A**) demonstrated that GFR was not modified by SAD (Sham: 6.5 ± 0.4 vs. SAD: 6.6 ± 0.4 mL/min/Kg). As expected, CKD rats presented a pronounced reduction in GFR (1.8 ± 0.2 mL/min/Kg, p < 0.01), which was not worsened by the combination with SAD (1.9 ± 0.4 mL/min/Kg, p < 0.01). RPF (**Figure 3B**), determined using PAH clearance, was decreased in SAD rats (23.7 ± 1.4 mL/min/Kg, p < 0.05) when compared to sham animals (27.9 ± 2.3 mL/min/Kg). Once again, PAH clearance was reduced in CKD rats (4.5 ± 0.3 mL/min/Kg, p < 0.01); however, SAD+CKD group presented a further reduction in RPF (2.9 ± 0.2 mL/min/Kg, p < 0.01). To calculate the RBF, we also quantified the haematocrit. SAD rats presented normal haematocrit (39 ± 2%) when compared to sham group (43 ± 2%). CKD animals demonstrated a decrease in this parameter (29 ± 1%, p < 0.01), which was not worsened by SAD (32 ± 1%, p < 0.01). The RBF analysis (**Figure 3C**) followed the same pattern as the RPF results. SAD resulted in reduced RBF (Sham: 53.3 ± 4.0 vs. SAD: 41.4 ± 2.5 mL/min/Kg, p < 0.05). CKD group displayed a marked decline in RBF (6.4 ± 0.4 mL/min/Kg, p < 0.01), which was worsened by combination with SAD (3.5 ± 0.7 mL/min/Kg, p < 0.01). As expected, based on the renal blood flow results, RVR (**Figure 3D**) was slightly increased in SAD rats (2.8 ± 0.2 a.u., p < 0.05) compared to sham animals (2.1 ± 0.1 a.u.). CKD group presented an enhanced RVR (22.4 ± 2.7 a.u., p < 0.01), with a further increase in SAD+CKD animals (32.2 ± 1.8 a.u., p < 0.01). Urinary water excretion was augmented in SAD (25.2 ± 0.7 mL/24 h, p < 0.05) group when compared to sham (12.5 ± 1.3 mL/24 h) animals. Experimental induction of CKD led to a further increase on urine excretion (38.3 ± 3.2 mL/24 h, p < 0.01), which was much more prominent in SAD+CKD (57.4 ± 6.0 mL/24 h, p < 0.01) group.

Corroborating the inulin and PAH clearance results, the plasma urea levels are demonstrated in **Figure 4**. Sinoaortic denervation did not change uraemia (5.9 ± 0.4 mmol/L) compared to the sham group (6.4 ± 0.3 mmol/L). As expected, CKD animals presented hyperuraemia (15.5 ± 1.3 mmol/L, p < 0.01), which was aggravated in the SAD+CKD group (22.1 ± 2.5 mmol/L, p < 0.01).

Although SAD did not modify inulin clearance, the correlation analysis between BPV and GFR demonstrated a negative relationship between these parameters in the SAD (r = −0.79, p < 0.05), CKD (r = −0.99, p < 0.05) and SAD+CKD (r = −0.81, p < 0.05) groups (**Figure 5**).

### Quantification of Superoxide Production

Superoxide production was assessed using flow cytometry with DHE (median of fluorescence intensity). As illustrated in **Figure 6**, SAD caused an increase in ROS generation (2757 ± 374, p < 0.01) when compared to the sham group (1045 ± 52). CKD also resulted in augmented superoxide production (1830 ±

66, p < 0.01), which was worsened in the SAD+CKD animals (4468 ± 529, p < 0.01).

### Cardiac and Kidney Hypertrophy and Glomerular Collagen Deposition

**Table 2** presents the results of cardiac and kidney hypertrophy. SAD and CKD groups showed an increased heart weight/body weight ratio when compared to sham group. The cardiac hypertrophy was worsened in SAD+CKD group. The kidney weight/body weight ratio was elevated only in SAD+CKD animals.

**Figure 7** demonstrates the collagen deposition in the glomerulus determined by Masson's Trichrome staining. Typical photomicrographs of all studied groups are displayed in **Figure 7A**, showing that, as expected, CKD resulted in glomerulosclerosis; however, collagen deposition was further increased when SAD and CKD occurred together. The quantification of the percentage glomerular area stained with Masson's Trichrome (**Figure 7B**) demonstrates that SAD did not alter glomerular collagen deposition (sham: 9.3 ± 1.5, SAD: 11.0 ± 0.8%). On the other hand, CKD resulted in glomerulosclerosis (44.8 ± 1.2%, p < 0.01), which was exacerbated in the SAD+CKD group (49.9 ± 1.2%, p < 0.01). The glomerulosclerosis score also confirmed these results. No differences were observed between the sham (1 ± 0) and SAD (1 ± 0) groups. CKD increased the glomerulosclerosis index (2.2 ± 0.07, p < 0.01) and the occurrence of SAD and CKD together worsened this parameter (2.5 ± 0.07, p < 0.01).

### DISCUSSION

In the past few years, several studies have demonstrated that a variety of diseases, such as atherosclerosis, diabetes and CKD, share a common factor: increased BPV (Roman et al., 2001; Di Iorio et al., 2012; Ushigome et al., 2014). In this scenario, BPV quantification emerges as a possible new marker of target-organ damage. Data from clinical studies demonstrated a significant association between BPV and cardiovascular mortality risk and renal death (Di Iorio et al., 2015). BPV is also correlated with a worse prognosis in cardiovascular and renal diseases (Parati et al., 2012; Di Iorio et al., 2013). However, both experimental and clinical studies have focused in evaluating BPV after the onset of the disease, and it is likely the observed dysfunction may have occurred due to other factors besides increased BPV. In the present study we developed an experimental design that was able to induce increased BPV prior to the onset of kidney disease and our results show that SAD prior to CKD induction worsens renal function, with diminished RBF, increased RVR, as well as augmented uraemia and glomerulosclerosis. Taken together, our results showed that increased BPV may accelerate the progression of CKD, indicating that BPV seems to be an important risk factor to the progression of kidney injury and it should be taken in consideration in the clinical practice.

Although SAD was initially considered a neurogenic hypertension model (Krieger, 1964), it can be considered an excellent experimental model to study the effects of augmented BPV, since the removal of the baroreflex afferent pathway, leads to a diminished BSI and increased BPV with no long-term changes in blood pressure (Su and Miao, 2002). The induction of CKD through 5/6 nephrectomy also resulted in a reduction of BSI and augmented BPV. Similar to our data, Griffin et al. (2004) demonstrated an increased BPV in nephrectomised rats. Clinical studies also showed an augmented BPV and diminished BSI in patients with chronic renal failure (Tozawa et al., 1999; Studinger et al., 2006; Di Iorio et al., 2012). Although CKD animals presented higher values of blood pressure but lower BSI than SAD group, we cannot rule out the effects on hypertension per se on BSI. It is well-established in the literature that increased values of blood pressure decrease baroreflex sensitivity in different experimental hypertension models (Vasquez et al., 1994; Peotta et al., 2007; Klippel et al., 2016) and when comparing sham to CKD animals as well as SAD to SAD+CKD group, it appears that increased blood pressure may

groups. Once again, SAD and CKD animals also presented reduced RBF; this reduction was greater in animals that underwent 5/6 nephrectomy. The association of SAD and CKD worsened this parameter. (D) Renal vascular resistance (RVR) in all studied groups. In accordance with the previous results, RVR was augmented in SAD and CKD animals; this increase was more pronounced in the CKD group. When SAD and CKD coexist, the rise in RVR is exacerbated. Sham *n* = 8, SAD *n* = 10, CKD and SAD+CKD *n* = 5. Values are means ± SEMs. <sup>a</sup>*p* < 0.01 vs. Sham group; <sup>b</sup>*p* < 0.01 vs. SAD group; <sup>c</sup>*p* < 0.01 vs. CKD group. Two-way ANOVA. Two-way ANOVA, Student's *t-*test.

have contributed to decrease BSI. Another factor that should be considered is the effects of cardiac hypertrophy on BSI. Studies have demonstrated that increased cardiac mass, without hypertension, leads to an impairment of baroreflex function in different animal models (Meyrelles et al., 1998; Gava et al., 2004). In agreement with this, in experimental models of hypertension, the development of baroreflex dysfunction coincides with the onset of cardiac hypertrophy (Head, 1994). In our study, both SAD and CKD groups present cardiac hypertrophy and reduced BSI, and these alterations were further increased in SAD+CKD animals, indicating that, besides the aforementioned factors, cardiac hypertrophy may have contributed to decreased baroreflex sensitivity.

One of the most important findings of our study concerns the association of SAD and CKD. In almost all analyzed parameters, when SAD and CKD were concomitant, the animals displayed worsened function. Occurring together, SAD and CKD resulted in a two-fold higher reduction in BSI and a 1.5-fold higher increase in BPV. The mechanisms involved in these alterations are not fully understood; however, studies have noted changes in sympathetic activity. According to Irigoyen et al. (1995), chronic SAD rats present several short periods of sympathetic hyperactivity, which could contribute to blood pressure fluctuations. Spectral analysis of the arterial pressure

demonstrated a rise in the low frequency component in both mice (Fazan et al., 2005) and rats (Mostarda et al., 2010) subjected to SAD, indicating an increase in the sympathetic activity. Comparable to these data, Shan et al. (2004) showed that rats with SAD presented higher levels of noradrenaline in areas

FIGURE 5 | Scatterplot linear regression and correlation analysis between glomerular filtration rate (GFR) and blood pressure variability (BPV). Data show a negative relationship of GFR and BPV in SAD, CKD and SAD+CKD group, but not in sham group. Sham and SAD *n* = 6, CKD *n* = 4, and SAD+CKD *n* = 5. The correlation between variables was determined using Pearson's correlation coefficient and a linear regression model.

involved in the vasomotor control, such as medulla oblongata and hypothalamus. The increase in sympathetic activity may also be involved in augmented BPV in the CKD animals because 5/6 nephrectomy also results in an elevation of circulating

### TABLE 2 | Body weight, cardiac, and kidney hypertrophy.


*CW/BW, cardiac weight to body weight ratio; KW/BW, kidney weight to body weight ratio. All values are expressed as means* ± *SEMs. The number in parentheses represents the number of animals in each group. <sup>a</sup>p* < *0.05 vs. Sham; <sup>b</sup>p* < *0.05 vs. SAD; <sup>c</sup>p* < *0.05 vs. CKD.*

catecholamines (Amann et al., 2000; Leineweber et al., 2002). The kidneys present both afferent and efferent sympathetic fibers, and in addition to being the target of sympathetic activity, they may also play a role as the source sympathetic activity (Campese, 2000), including during chronic renal failure. Corroborating this idea, Bigazzi et al. (1994) demonstrated that 5/6 nephrectomised rats present augmented turnover of noradrenaline in the posterior hypothalamic nucleus, which was attenuated after bilateral rhizotomy. Therefore, because sympathetic nervous system hyperactivity is a common feature of both SAD and CKD, the combination of both interventions may lead to even higher sympathetic activation, explaining the

reduction in BSI and elevation in BPV found to be more prominent in the SAD+CKD group.

Regarding the haemodynamic parameters, the SAD animals did not demonstrate any changes in arterial pressure nor heart rate. This result is in accordance with the literature because the SAD animals show hypertension only in the initial phase of SAD. In the chronic phase, arterial pressure returns to normal levels, although BPV remains high (Norman et al., 1981). Nephrectomised animals also presented hypertension, which was also an expected result because studies have demonstrated that the infarction of the renal poles results in an elevation of arterial pressure (Griffin et al., 2004). However, the resting values of systolic, diastolic, and MAP, as well as heart rate, were more elevated in the SAD+CKD group. The hyperactivation of the sympathetic nervous system by the combination of SAD and CKD may play a role in the development of these higher levels of hypertension, although we cannot rule out the role of the renin angiotensin system (RAS). Nephrectomised animals present an elevation of plasma (2 weeks after nephrectomy) and tecidual angiotensin II (Mackie et al., 2001; Vaziri et al., 2007). Nishimura et al. (2007) demonstrated that renin, angiotensin-converting enzyme (ACE) and AT1 receptor mRNA expression is elevated in the hypothalamus and brainstem of nephrectomised animals. It is interesting to note that these regions are involved in sympathetic activation, and the increase of tecidual RAS in these areas may play a role in the sympathetic hyperactivity observed in CKD. SAD also leads to an increase in the RAS system. Angiotensin II levels, AT1 receptors mRNA expression and ACE activity are elevated in the heart, aorta and kidney of SAD animals (Miao et al., 2003; Shan et al., 2004; Feng et al., 2011). Because both SAD and CKD result in augmented RAS, once again, the combination of both conditions may be additive and lead to higher levels of hypertension in the SAD+CKD group. In addition, activation of tecidual RAS may also be involved in the cardiac hypertrophy presented in the SAD and the CKD animals. The increase in cardiac weight/body weight ratio was even higher in the SAD+CKD group, and the mechanisms involved in this alteration may include activation of SNS, augmented tecidual RAS and the higher level of hypertension presented by the animals.

In the renal function studies, we observed that SAD per se did not produce any changes in the GFR. As expected, CKD animals presented a remarkable reduction in the GFR, and the combination of SAD and CKD did not worse this parameter. However, in the SAD+CKD group, the reduction of renal plasma (RPF) and blood (RBF) flow and the increase in RVR were statistically greater than in all other groups.

The maintenance of the GFR in the SAD+CKD group compared to CKD animals, even with a smaller RBF, may involve the RAS system. It is well-established that angiotensin II is able to cause a preferential constriction of the efferent arteriole, increasing glomerular hydrostatic pressure and maintaining the GFR (Bidani et al., 2013). As both SAD and CKD increase intrarenal angiotensin II (Shan et al., 2004; Vaziri et al., 2007), we can speculate that a greater activation of RAS in SAD+CKD may prevent a greater reduction of GFR in this group. Although SAD+CKD animals did not present a further reduction in GFR, it seems that the combination of both procedures resulted in a worsened renal function, because other analyzed parameters such as hyperuraemia and glomerulosclerosis were exacerbated in this group. Additionally, the correlation analysis showed that there is a negative relationship between GFR and BPV.

Another pathway that might be involved in the observed changes in renal function is the increased sympathetic nerve activity. Even thought we did not quantify renal sympathetic nerve activity in the present investigation, studies have demonstrated that both SAD and CKD result in increased sympathetic drive (Irigoyen et al., 1995; Amann et al., 2000; Leineweber et al., 2002; Fazan et al., 2005). It is well-stablished that SNS activity plays an important role in the genesis of hypertension; however, the deleterious effects of increased sympathetic drive on the kidneys are not only caused by higher blood pressure levels (Joles and Koomans, 2004). Prolonged SNS hyperactivity can damage intrarenal blood vessels by inducing proliferation of smooth muscle cells and fibroblasts in the vessel wall (Zhang and Faber, 2001; Erami et al., 2002). In addition, activation of renal sympathetic fibers result in vasoconstriction and reduce RBF (Johns et al., 2011), leading to renal injury. In our study, it is possible that the combination of SAD and CKD resulted in an even higher sympathetic drive than when these situations occurred alone, contributing to a worsened renal function on SAD+CKD group.

Although we speculate that increased BPV in SAD+CKD may be responsible for the observed changes in renal function, we cannot rule out the role of hypertension, since this group presented higher levels of systolic, diastolic and MAP. However, the effects of elevated BPV in renal organ damage may be more important than the classic risk factors of a high blood pressure. Corroborating this idea, Miao et al. (2006) identified that increased short-term BPV is a more critical determinant for renal damage than mean BP levels. An important clinical study developed by Parati et al. (1987) demonstrated that, for nearly any level of blood pressure, the patients who presented higher BPV also had increased prevalence and severity of organ damage, in both short and long term evaluation of BPV. Crosssectional studies in non-treated hypertensive patients have found an increased short term BPV to be positively correlated with impaired renal function (Parati et al., 2012). Taken together, these data indicate that, regardless the blood pressure levels, there is a relationship between BPV and the severity of organ damage, including in hypertensive states.

The alterations in renal haemodynamic observed in our study may occur also due to changes in the nitric oxide (NO) system, which plays an important role in regulating RBF (Mattson and Meister, 2005; Toda and Okamura, 2011), especially counterbalancing the effects of increased sympathetic activity. When SNS activity is augmented, NO formation also increases to prevent kidney ischaemia, mainly in medullar levels (Zou and Cowley, 2000). However, a higher NO generation does not necessarily leads to more effects of this molecule, especially if oxidative stress is increased, as observed in the SAD+CKD group. Similar to these data, studies have demonstrated that both experimental SAD and CKD induce a smaller NO bioavailability (Nakayama et al., 2009; Wu et al., 2013). Our data shows that SAD+CKD animals present higher oxidative stress in the blood, indicating that renal oxidative could also be increased. Therefore, the greater increase in RVR and, consequently, the reduction in RBF, may be attributed to the augmented oxidative stress observed in the SAD+CKD group, reducing NO bioavailability. Thus, the prominent effects of the association of SAD and CKD on RBF and RVR may be attributed to activation of renal vasoconstrictive systems (SNS and SRA) and reduced NO bioavailability.

The reduction of NO bioavailability may also have played a role in the enhanced levels of glomerulosclerosis presented by the SAD+CKD rats. Corroborating this hypothesis, previous studies have demonstrated that eNOS <sup>−</sup>/<sup>−</sup> mice exhibit exacerbated renal interstitial injury and global glomerulosclerosis (Nakayama et al., 2009) and that iNOS <sup>−</sup>/<sup>−</sup> mice show higher levels of tubular apoptosis (Miyajima et al., 2001) and interstitial fibrosis (Hochberg et al., 2000). In the kidney, NO reduced mesangial proliferation and extracellular matrix synthesis (Zhou et al., 2004) and the mechanisms involved in these effects include inhibition of TGF-β<sup>1</sup> and its downstream effector molecule fibronectin (Zhou et al., 2008), inhibition of TNFα (Whiting et al., 2013) and modulation of cytokine-induced metalloproteinases and inhibitors of metalloproteinases (Yang et al., 2008).

Taken together, these data indicate that increased BPV prior to renal dysfunction induced by SAD, as well as hypertension, exacerbates renal injury, and glomerulosclerosis in an experimental model of CKD. These results reinforce the role of increased BPV and hypertension as important markers to the progression of renal diseases.

### AUTHOR CONTRIBUTIONS

FFF performed experiments and drafted manuscript, GA performed experiments, MP performed experiments, FPF performed experiments, JG performed experiments, CB analyzed data, interpreted results of experiments, edited and revised manuscript, EC analyzed data, interpreted results of experiments, edited and revised manuscript, SM analyzed data, interpreted results of experiments, edited and revised manuscript, AG conception and design of research, analyzed data, interpreted results of experiments, prepared figures, drafted manuscript, edited and revised manuscript, approved final version of manuscript.

### ACKNOWLEDGMENTS

This research was supported by the National Council for the Development of Science and Technology (CNPq/473177/2013- 7) and State Agency for the Development of Science and Technology (FAPES/54674166/2011).

### REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Freitas, Araujo, Porto, Freitas, Graceli, Balarini, Vasquez, Meyrelles and Gava. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Dipeptidyl Peptidase IV Inhibition Exerts Renoprotective Effects in Rats with Established Heart Failure

Daniel F. Arruda-Junior <sup>1</sup> , Flavia L. Martins <sup>1</sup> , Rafael Dariolli <sup>1</sup> , Leonardo Jensen<sup>1</sup> , Ednei L. Antonio<sup>2</sup> , Leonardo dos Santos <sup>3</sup> , Paulo J. F. Tucci <sup>2</sup> and Adriana C. C. Girardi <sup>1</sup> \*

*<sup>1</sup> Heart Institute (InCor), University of São Paulo Medical School, São Paulo, Brazil, <sup>2</sup> Cardiology Division, Department of Medicine, Federal University of São Paulo, São Paulo, Brazil, <sup>3</sup> Department of Physiological Sciences, Federal University of Espírito Santo, Vitória, Brazil*

Circulating dipeptidyl peptidase IV (DPPIV) activity is associated with worse cardiovascular outcomes in humans and experimental heart failure (HF) models, suggesting that DPPIV may play a role in the pathophysiology of this syndrome. Renal dysfunction is one of the key features of HF, but it remains to be determined whether DPPIV inhibitors are capable of improving cardiorenal function after the onset of HF. Therefore, the present study aimed to test the hypothesis that DPPIV inhibition by vildagliptin improves renal water and salt handling and exerts anti-proteinuric effects in rats with established HF. To this end, male Wistar rats were subjected to left ventricle (LV) radiofrequency ablation or sham operation. Six weeks after surgery, radiofrequency-ablated rats who developed HF were randomly divided into two groups and treated for 4 weeks with vildagliptin (120 mg/kg/day) or vehicle by oral gavage. Echocardiography was performed before (pretreatment) and at the end of treatment (post-treatment) to evaluate cardiac function. The fractional area change (FAC) increased (34 ± 5 vs. 45 ± 3%, *p* < 0.05), and the isovolumic relaxation time decreased (33 ± 2 vs. 27 ± 1 ms; *p* < 0.05) in HF rats treated with vildagliptin (post-treatment vs. pretreatment). On the other hand, cardiac dysfunction deteriorated further in vehicle-treated HF rats. Renal function was impaired in vehicle-treated HF rats as evidenced by fluid retention, low glomerular filtration rate (GFR) and high levels of urinary protein excretion. Vildagliptin treatment restored urinary flow, GFR, urinary sodium and urinary protein excretion to sham levels. Restoration of renal function in HF rats by DPPIV inhibition was associated with increased active glucagon-like peptide-1 (GLP-1) serum concentration, reduced DPPIV activity and increased activity of protein kinase A in the renal cortex. Furthermore, the anti-proteinuric effect of vildagliptin treatment in rats with established HF was associated with upregulation of the apical proximal tubule endocytic receptor megalin and of the podocyte main slit diaphragm proteins nephrin and podocin. Collectively, these findings demonstrate that DPPIV inhibition exerts renoprotective effects and ameliorates cardiorenal function in rats with established HF. Long-term studies with DPPIV inhibitors are needed to ascertain whether these effects ultimately translate into improved clinical outcomes.

Edited by:

*Valdir Andrade Braga, Federal University of Paraiba, Brazil*

### Reviewed by:

*Ravi Nistala, University of Missouri-Columbia, USA Berthold Hocher, University of Potsdam, Germany*

> \*Correspondence: *Adriana C. C. Girardi adriana.girardi@incor.usp.br*

> > Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *15 May 2016* Accepted: *27 June 2016* Published: *12 July 2016*

### Citation:

*Arruda-Junior DF, Martins FL, Dariolli R, Jensen L, Antonio EL, dos Santos L, Tucci PJF and Girardi ACC (2016) Dipeptidyl Peptidase IV Inhibition Exerts Renoprotective Effects in Rats with Established Heart Failure. Front. Physiol. 7:293. doi: 10.3389/fphys.2016.00293*

Keywords: vildagliptin, cardiorenal dysfunction, fluid retention, glucagon-like peptide-1, proteinuria, megalin

## INTRODUCTION

Dipeptidyl peptidase IV (DPPIV) is a widely expressed ectopeptidase that exists anchored as a transmembrane protein or in a soluble form in plasma and other body fluids (Lambeir et al., 2003). The kidney is the main source of DPPIV, where the enzyme is highly concentrated in the apical microvilli of the proximal tubule cells (Kenny et al., 1976; Girardi et al., 2001). DPPIV catalyzes the release of N-terminal dipeptides from polypeptides with proline or alanine in the second position. Among the many bioactive peptides cleaved by DPPIV, there is the gut hormone glucagon-like peptide-1 (GLP-1), which is the major incretin responsible for post-prandial insulin secretion. Indeed, DPPIV inhibitors, commonly called gliptins, increase the bioavailability of GLP-1 and improve systemic glucose homeostasis, thereby constituting the second line oral therapy in type 2 diabetes.

In addition to their effects on glycemic control, DPPIV inhibitors have been shown to produce cardioprotective and renoprotective effects. These beneficial actions may also be attributed, at least partially, to increased GLP-1 bioavailability. Upon binding of GLP-1 to its receptor (GLP-1R) in the heart, it induces activation of the cytoprotective signaling pathways cAMP/PKA and PI3K/Akt (Ussher and Drucker, 2012), reduces apoptosis (Ravassa et al., 2011), increases heart glucose uptake (Bao et al., 2011), reduces infarct size (Timmers et al., 2009) and improves coronary blood flow by means of its vasodilating actions (Ban et al., 2008). Moreover, GLP-1 induces diuresis and natriuresis by increasing renal blood flow and glomerular filtration rate (GFR) and by reducing NHE3-mediated sodium reabsorption in the renal proximal tubule sodium via cAMP/PKA activation (Crajoinas et al., 2011; Rieg et al., 2012; Farah et al., 2016).

Recent translational studies have suggested that increased DPPIV activity may play an important role in the pathophysiology of heart failure (HF) (Shigeta et al., 2012; dos Santos et al., 2013; de Almeida Salles et al., 2016). We have previously showed that increased circulating DPPIV activity is associated with worse cardiovascular outcomes in human and experimental HF and that long-term DPPIV inhibition exerts cardioprotective actions that prevent the development/progression of HF in rats (dos Santos et al., 2013; de Almeida Salles et al., 2016).

Worsening renal function in the setting of HF is widely accepted as an independent risk factor for a poor prognosis. HF is often associated with sodium and water retention, a reduction in renal blood flow and GFR (Laskar and Dries, 2003), renal tubular damage and proteinuria (Udani and Koyner, 2010; Boerrigter et al., 2013). Despite its natriuretic actions, DPPIV inhibitors have also been shown to confer renoprotection by reducing urinary protein excretion and ameliorating renal damage in a variety of clinical and experimental studies in diabetic, obese and renal failure patients and rodents (Hattori, 2011; Kanasaki et al., 2014; Scirica et al., 2014; Sharkovska et al., 2014; Nakamura et al., 2016; Tsuprykov et al., 2016). However, it remains to be established whether DPPIV inhibitors are capable of reversing cardiorenal dysfunction after the onset of HF.

Given the inextricable importance of heart-kidney interactions in HF, the main purpose of the present study was to test the hypothesis that DPPIV inhibition by vildagliptin improves renal water and salt handling and exerts antiproteinuric effects in rats with established HF. The molecular mechanisms underlying these renoprotective effects were also investigated.

### MATERIALS AND METHODS

### Reagents and Antibodies

The dipeptidyl peptidase IV (DPPIV) inhibitor vildagliptin was a gift from Novartis Pharmaceuticals (Basel, Switzerland). The monoclonal antibody (mAb) against the Na+/H<sup>+</sup> exchanger isoform 3 (NHE3), clone 3H3 (Kocinsky et al., 2005), and a polyclonal antibody against megalin were kindly provided by Dr. Peter Aronson and Dr. Daniel Biemesderfer, respectively (Yale University, New Haven, CT). A mAb antibody against DPPIV (clone 5H8), a mouse mAb, clone 14D5, that recognizes NHE3 only when it is phosphorylated at serine 552 (Kocinsky et al., 2005) and polyclonal antibodies against the glucagon-like peptide-1 receptor (GLP-1R), cubilin, nephrin, and podocin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The polyclonal antibody against phospho-(Ser/Thr) protein kinase A (PKA) substrates (Gronborg et al., 2002) was obtained from Cell Signaling (Beverly, MA). The mAb against actin (JLA20) was purchased from Merck Millipore (Darmstadt, Germany). A polyclonal antibody against CD31 was purchased from Abcam (Cambridge, MA). Horseradish peroxidase-conjugated goat anti-mouse, goat anti-rabbit and rabbit anti-goat secondary antibodies were purchased from Jackson ImmunoResearch (West Grove, PA). Chemicals were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise specified.

### Animal Model

Experiments were carried out using male Wistar rats (2–3 months old, 200–250 g) obtained from the University of São Paulo Medical School, São Paulo, SP, Brazil. Animals were maintained in the Heart Institute animal facility in compliance with the Ethical Principles of the Brazilian College of Animal Experimentation and were approved by the Institutional Animal Care and Use Committee.

Experimental HF was induced by left ventricular (LV) myocardial injury after radiofrequency catheter ablation as previously described (Antonio et al., 2009; Inoue et al., 2012; dos Santos et al., 2013; de Almeida Salles et al., 2016). Briefly, rats were anesthetized with isoflurane, intubated and mechanically ventilated with oxygen enriched air and kept warm during surgical procedures (37◦C). Subsequently, a leftside thoracotomy was performed at the fourth intercostal space. LV injury was created by delivering high-frequency currents (1000 KHz, 12 w, during 10 s) generated by a conventional radiofrequency generator (model TEB RF10; Tecnologia Eletrônica Brasileira, São Paulo, Brazil). The sham group underwent similar surgical procedures, but was not subjected to radiofrequency myocardial ablation. Six weeks after surgery, radiofrequency-ablated rats that developed HF

were randomly divided into two groups and treated during 4 weeks with vildagliptin (120 mg/kg/day bid) or vehicle by oral gavage. Vehicle-treated sham rats were used as controls. Echocardiographic evaluation of left ventricle systolic and diastolic function as well as serum levels of BNP were used to characterize HF. HF was considered when the fractional area change (FAC) was lower than 40% and circulating levels of brain natriuretic peptide (BNP) were higher than 1.0 ng/mL. For determination of BNP serum concentration, blood samples were withdrawn from the rat retro-orbital sinus under isoflurane anesthesia and immediately transferred into chilled tubes containing clot activator and gel for serum separation (BD vacutainer <sup>R</sup> SST <sup>R</sup> II Advance <sup>R</sup> ). Serum levels of BNP were measured by enzyme-linked immunoassay (ELISA) (BNP 32 Rat ELISA kit, Abcam) according to the manufacturer's instructions. Mortality rate between HF rats treated with vehicle or with vildagliptin was lower than 5% and did not differ between these two groups of rats.

### Echocardiography

Doppler echocardiography was performed 6 weeks after LV-radiofrequency ablation or sham surgery (pretreatment) and after treatment with vehicle or vildagliptin (post-treatment). Animals were anesthetized with ketamine and xylazine (50 mg/kg and 10 mg/kg, respectively, i.p.) and placed in the left lateral decubitus position (45◦ angle) to obtain cardiac images. Images were captured and analyzed using Sonos 5500 ultrasound equipment (Philips Medical System, Bothell, WA) with a 12–14 MHz transducer (2 cm depth with fundamental and harmonic imaging). Echocardiographic images were acquired placing the cursor of pulsed wave Doppler in the LV outflow tract to display the end of aortic ejection and the onset of mitral inflow. End-diastolic (DA) and end-systolic (SA) transverse areas of LV were measured in parasternal-papillary short-axis images. FAC was calculated based on the areas using the following equation: FAC (in %) = (DA − SA)/DA × 100. Diastolic function was obtained at apical plane images. Isovolumetric relaxation time (IVRT) was measured in tissue Doppler images in the lateral wall of the ventricle. Transthoracic echocardiography was performed by two investigators who were blind to the experimental groups.

### Renal Function

Renal function was evaluated as described previously (Inoue et al., 2013). In brief, rats were housed individually in metabolic cages (Tecniplast, Buguggiate, VA, Italy) for four consecutive days. The first day was used to adapt the rats to the cages, and the following 3 days were used to assess renal function. Food consumption and water intake were monitored daily. Urine samples collected during each 24-h period were used to determine urine output, creatinine, sodium and protein excretion. At the end of the experiment, terminal arterial blood samples were collected from the abdominal aorta and transferred into vacutainer tubes (BD vacutainer <sup>R</sup> SST <sup>R</sup> II Advance <sup>R</sup> , Franklin Lakes, NJ) to obtain serum. Serum was separated by centrifugation at 4000 rpm for 15 min for the measurement of biochemical parameters. Urine output was measured gravimetrically. Creatinine clearance was used to estimate GFR. Creatinine concentration was measured by the kinetic alkaline picrate (Jaffe) reaction (LabTest, Lagoa Santa, MG, Brazil). Serum and urine sodium concentrations were measured by electrolyte analyzer (ABL800 FLEX, Radiometer Copenhagen). Total urinary protein excretion was measured using a commercially available kit based on the pyrogallol redmolybdate method (Sensiprot, Labtest). All experiments were performed following the manufacturer's instructions.

## Determination of Active GLP-1 Serum Levels

For the quantitative determination of active GLP-1, blood was collected and transferred to vacutainer collecting tubes containing the DPPIV inhibitor P32/98 (10 µM) (Abcam). Serum levels of active GLP-1 (7–36) were measured by ELISA (Glucagon-Like Peptide-1 (Active) ELISA, Merck Millipore) according to the manufacturer's instructions.

### Measurement of Blood Glucose

Sham and HF rats treated with vehicle or vildagliptin were fasted for 8 h. The blood glucose level was determined from their tail vein using the ACCU-CHECK <sup>R</sup> Performa meter (Roche Diagnostics GmbH, Mannheim, Germany).

### Biometric and Morphometric Analysis

Anesthetized rats (ketamine and xylazine 50 mg/kg and 10 mg/kg, respectively, i.p.) were killed by decapitation, and their hearts, kidneys and lungs were immediately removed and weighed. The heart and kidney weight to body weight ratio were used as an index of organ hypertrophy. The relative water content of lung tissue was calculated using the following equation: Lung water content (in %) = (lung wet weight − lung dry weight)/lung wet weight × 100.

The apex of the heart of each rat was separated and prepared for histological and immunohistochemical assays. The remainder of the heart was frozen at −80◦C for molecular analysis. Five-micro meter sections of paraffin-embedded tissue were mounted onto slides and stained with hematoxylin and eosin (for determination of nuclear cell volume) or picrosirius red (for the determination of interstitial collagen density). A computerized image acquisition system (Leica Imaging Systems, Bannockburn, IL) was used for the analyses. As an estimate of myocyte hypertrophy, the average nuclear volume was determined in 70−90 cardiomyocytes cut longitudinally, acquired in 5 randomized 400× magnification fields per each animal, and calculated according to the following equation: nuclear volume = π × D × d 2 /6 (d = shorter nuclear diameter; D = longer diameter) (Gerdes et al., 1994; dos Santos et al., 2013). Interstitial fibrosis in the remodeled LV was evaluated as the percent area occupied by collagen fibers, excluding stained ablation scars and perivascular fibers. After digitalization, the red-stained areas were quantified as the average percentage of the total area from each of five randomized 200 × magnification fields per animal. The observer was blinded to the experimental groups. Sections from the midventricular level of each heart (4 µm) stained with picrosirius red were scanned, and the circumference of the fibrotic infarct area was measured with Image J software (NIH). The size of the injured myocardial area was determined by calculating the percent of the length of the area occupied by the scar at the midventricular section. In this regard, we previously demonstrated that this single measure is a satisfactory prediction of average infarct extension (dos Santos et al., 2008).

Immunohistochemical staining was performed to measure capillary density and to define the localization of DPPIV in the heart. Endogenous peroxidase activity was blocked by 3 min incubation in 3% H2O<sup>2</sup> (seven times at room temperature) and then rinsed with PBS (137 mM NaCl, 2.5 mM KCl, 10 mM Na2HPO4, and KH2PO<sup>4</sup> 176 mM, pH 7.4). Non-specific reactions were blocked in 2% goat serum for 20 min and then incubated with the primary antibodies. The primary antibodies used were the mAb anti-DPPIV antibody or the rabbit polyclonal anti-CD31 antibody, and both of them were diluted 1:50 in the blocking buffer containing 5% BSA. Negative controls were not incubated with primary antibodies. After 18 h incubation at 4◦C, tissues were washed 3 times for 5 min with PBS and incubated with secondary antibody. After washing in PBS, tissue sections were incubated with an HRP solution Universal LSAB 2 kit containing biotin-streptavidin complex for signal amplification of the primary antibody. Immunoreactions were detected with 3,3′ -diaminobenzidine tetrahydrochloride (DAB) for 7 min. Immunostaining was visualized under a microscope and positive staining (brown color) analyzed under 400 × magnification. For capillary density evaluation, the number of capillaries CD31+ was counted from 10 randomized fields per animal at 400 × magnification. Image analysis software (Leica Imaging Systems, Bannockburn, IL, USA) was used to measure the capillary density, calculated as the number of capillaries per tissue area in the remote LV wall. The measured total tissue area was corrected for the remaining interstitial space.

## Determination of DPPIV Activity and Abundance

DPPIV activity was assayed in rat serum, kidney and heart homogenates using a colorimetric method that measures the release of p-nitroaniline resulting from the hydrolysis of glycylproline p-nitroanilide tosylate (Pacheco et al., 2011). Renal and heart DPPIV activity was normalized to total protein levels, and DPPIV abundance in the rat kidney and heart homogenates were analyzed by immunoblotting.

## Protein Extraction from Heart and Renal Cortex

Harvested hearts from rats were homogenized in a Polymix PX-SR 50 E homogenizer (Kinematica, AG, Switzerland) in icecold phosphate buffered saline (PBS) (10 mmol/L phosphate, 140 mmol/L NaCl, pH 7.4), including phosphatase inhibitors (15 mM NaF and 50 mM sodium pyrophosphate) and Halt Protease Inhibitor Cocktail (Thermo Fisher Scientific, Rockford, IL). Renal cortical homogenates were prepared as previously described (Crajoinas et al., 2014).

## Determination of Protein Kinase A (PKA) Activity in Renal Cortical Homogenates

Equal amounts (25 µg) of renal cortical homogenates were resolved by SDS-PAGE and analyzed by immunoblotting using an antibody specific for phosphorylated PKA substrates (Gronborg et al., 2002; Crajoinas et al., 2014).

## SDS-Page and Immunoblotting

Equal protein amounts of heart, renal cortical homogenate or a volume of urine containing 25 µg of creatinine were solubilized in SDS sample buffer (2% SDS, 10% glycerol, 0.1% bromophenol blue, 50 mmol/L Tris, pH 6.8), and subjected to 7.5 or 10% SDS-PAGE polyacrylamide gel. The separated proteins were transferred from the gel to a polyvinylidene difluoride membrane (PVDF) (Immobilon-P, Merck Millipore, Darmstadt, Germany) at 350 mA for 8–10 h at 4◦C with a TE 62 Transfer Cooled Unit (GE HealthCare, Piscataway, NJ, USA), and stained with Ponceau S. PVDF membranes containing transferred proteins were subsequently blocked with 5% non-fat dry milk or 5% bovine serum albumin and 0.1% Tween 20 in PBS at a pH of 7.4 for 1 h to block non-specific binding of the antibody, followed by overnight incubation in the primary antibody. The membranes were then washed five times in blocking solution and incubated for 1 h at room temperature with an appropriated horseradish-peroxidase-conjugated immunoglobulin secondary antibody (1:2000). After washing five times in blocking solution and twice in PBS (pH 7.4), the protein bands were detected using enhanced chemiluminescence system (GE Healthcare) according to the manufacturer's protocols. The visualized bands were digitized using an ImageScanner (GE HealthCare) and quantified using the Scion Image Software package (Scion Corporation, Frederick, MD). Gels containing samples of urine were silver stained using ProteoSilver Plus Kit.

### Quantitative Real Time RT-PCR

Total RNA was isolated from hearts using Trizol (Thermo Fisher Scientific, Carlsbad, CA) according to the manufacturer's instructions, quantified (ND-1000 spectrophotometer— NanoDrop Technologies, Inc.), and treated with DNase-I. First-strand cDNA synthesis was performed using Super-Script III Reverse Transcriptase (Invitrogen) following the manufacturer's guidelines. The oligonucleotide primers CCAACTCCAGAGGACAACCT (forward) and TCTTCGTCCGTGTACCACAT (reverse) were used to detect DPPIV, and GATTCTGCTCCTGCTTTTCC (forward) and TCTTTTGTAGGGCCTTGGTC (reverse) were used to detect BNP. PCR products were visualized on 0.8% agarose gels with ethidium bromide. Reactions were carried using SYBR Green PCR Master Mix-PE (Thermo Fisher Scientific) on an ABI Prism <sup>R</sup> 7500 Fast Sequence Detection System (Applied Biosystem, Foster City, CA). The comparative threshold cycle method was used for data analyses. All samples were assayed in triplicate. Transcripts for three reference genes were determined: beta-actin (forward: CGTTGACATCCGTAAAGACC; reverse GCCACCAATCCACACAGA), GAPDH (forward: ATGGTGAA GGTCGGTGTG; reverse: GAACTTGCCGTGGGTAGAG) and cyclophilin A (forward: AATGCTGGACCAAACACAAA-30; reverse: CCTTCTTTCACCTTCCCAAA). The BestKeeper software (Pfaffl et al., 2004) was used to identify the best suit reference gene (Cyclophilin A) for data normalization under our experimental conditions. Relative expression was analyzed by the 2−11CT method.

### Statistical Analysis

All values are expressed as the means ± standard error of the mean (SEM). Comparisons between two groups were performed using unpaired t-tests. If more than two groups were compared, the statistical significance was determined using one-way analysis of variance (ANOVA) followed by Tukey's post-hoc test. The results were considered statistically significant when p < 0.05.

### RESULTS

### DPPIV Inhibition Improves Cardiac Function in Rats with Established HF

As seen in **Table 1**, treatment with the DPPIV inhibitor vildagliptin reduced serum DPPIV activity by >70% in HF rats compared to HF rats treated with the vehicle.

At the end of the 4-week treatment period (post-treatment), cardiac dysfunction was aggravated in HF rats who received the vehicle compared to the pretreatment period (**Figure 1**). Worsening of cardiac dysfunction in these animals were evidenced by a decline in FAC (37 ± 2 vs. 30 ± 2%, p < 0.05) (**Figure 1A**), which reflects a progressive reduction in the overall LV contractility, as well as an increase in IVRT (33.1 ± 0.7 vs. 36.5 ± 2.4 ms, p < 0.05) (**Figure 1B**), which reflects an aggravated diastolic dysfunction. Additionally, vehicle-treated HF rats displayed a remarkable increase in BNP serum levels from pretreatment to post-treatment (0.94 ± 0.01 vs. 2.65 ± 0.46 ng/mL p < 0.001). Conversely, HF rats treated with the DPPIV inhibitor vildagliptin exhibited an increase in FAC (34 ± 5 vs. 45 ± 3%, p < 0.05) (**Figure 1A**) and a reduction in IVRT (33 ± 2 vs. 27 ± 1 ms, p < 0.05) (**Figure 1B**) and in serum BNP 32 levels (0.93 ± 0.07 vs. 0.55 ± 0.02 ng/mL, p < 0.001) (**Figure 1C**) compared with the pretreatment period. As expected, cardiac function and serum BNP 32 levels were similar between pretreatment and post-treatment periods in sham rats.

### DPPIV Inhibition Attenuates Cardiac Remodeling and Increases Capillary Density in Rats with Established HF

The biometric characteristics of the rats are shown in **Table 1**. Average body weight gain was similar among the three experimental groups. Vildagliptin administration did not influence the size of the injured myocardial area. Vehicletreated HF rats showed higher heart weight-to-body weight ratio, indexed lung mass and the percent of water content when compared with sham-operated controls. In vildagliptin-treated rats, the heart weight-to-body weight ratio was lower than in vehicle treated HF rats but remained higher than in sham rats. The relative water content of lung tissue in vildagliptin treated HF rats was reduced to sham levels.

TABLE 1 | Biometric parameters and serum DPPIV activity in sham and HF rats treated with vehicle (HF) or vildagliptin for 4 weeks.


*Values are means* ± *SEM. BW, body weight; OD, optical density.* \*\*\**p* < *0.001 vs. Sham.* ##*p* < *0.01 and* ###*p* < *0.001 vs. HF.*

(C) Representative immunohistochemical staining of CD31 in sections of the LV viable wall showing individual capillaries (small dark circles) (400×original magnification). *n* = 7–9 rats/group. Values are means ± SEM. \*\**p* < 0.01 and \*\*\**p* < 0.001 vs. Sham. #*p* < 0.05 and ###*p* < 0.001 vs. HF.

The potential anti-hypertrophic effects of DPPIV inhibition in rats with established HF were further evaluated by histological analysis of hematoxylin and eosin-stained cardiac sections (**Figure 2A**). The results of these analyses demonstrated that the mean cardiomyocyte nuclear volume in vehicle-treated HF rats was larger than that of sham- and vildagliptin-treated rats, which significantly attenuated this increase (**Figure 2A**). In addition, vehicle-treated HF rats had a higher percentage of interstitial collagen in the reminiscent myocardium than sham rats, which was significantly attenuated by DPPIV inhibition (**Figure 2B**). Representative photomicrographs of CD31 stained sections in the LV viable wall showing individual capillaries are presented in the **Figure 2C**. HF rats treated with vildagliptin showed a slight but significant increase in the amount of CD31+ capillaries compared to the vehicle-treated group in the remote area of injured hearts indicating a reduced capillary rarefaction. Moreover, the non-treated HF rats had a significant decrease in the number of capillaries per mm<sup>2</sup> compared to sham group (**Figure 2C**).

### DPPIV Inhibition Suppresses Cardiac DPPIV Activity and Expression in Rats with Established HF

Vehicle-treated HF rats exhibited higher levels of cardiac DPPIV activity compared to sham and vildagliptin, which markedly decreased DPPIV activity (**Figure 3A**). Increased heart DPPIV expression in HF rats was observed both in cardiac endothelial cells as well as in the pericardium membrane (**Figure 3B**). Notwithstanding, vildagliptin reduced DPPIV expression at both sites. Higher DPPIV activity and protein expression was accompanied by higher levels of DPPIVmRNA expression in the heart of HF rats compared to sham rats, which was also attenuated by vildagliptin treatment (**Figure 3C**).

### DPPIV Inhibition Diminishes DPPIV Activity but Not Expression in the Kidney of Rats with Established HF

The activity and protein expression of DPPIV in the renal cortex of sham and HF rats treated with vildagliptin or with the vehicle are illustrated at the **Figure 4**. Renal cortical DPPIV activity was lower in HF rats treated with vildagliptin compared to both sham and vehicle-treated HF rats (**Figure 4A**). As shown in **Figure 4B**, the expression of DPPIV in the renal cortex was similar among the three groups of rats.

### DPPIV Inhibition Increases Circulating GLP-1 Active and Restores PKA Signaling in the Kidneys of Rats with Established HF

The levels of active GLP-1 in HF rats treated with vildagliptin (20.9 ± 2.9 pM) were approximately three times higher than those of vehicle-treated HF rats (6.9 ± 0.4 pM p < 0.001) and sham rats (8.2 ± 0.5 pM P < 0.001) (**Figure 5A**). However,

despite the differences in GLP-1 levels between vehicle-treated-HF and HF rats treated with vildaglitpin, the fasting blood glucose levels were similar among the groups (**Figure 5B**).

The effect of DPPIV inhibition on renal cortical PKA activity, a downstream effector for the GLP-1R (Farah et al., 2016), was estimated by immunoblotting. As shown in **Figure 5C**, vehicletreated HF rats exhibited lower renal cortical PKA activity than sham rats. Vildagliptin treatment increased the levels of phosphorylated PKA substrates in the renal cortex of HF rats compared to vehicle-treated HF rats.

## DPPIV Inhibition Improves Renal Sodium and Water Handling in Rats with Established HF

The effects of DPPIV inhibition on sodium and water handling are shown in **Figure 7**. HF rats exhibited lower urinary flow (**Figure 6A**) and sodium excretion (**Figure 6B**) compared to sham rats, whereas treatment with vildagliptin significantly restored urine output and sodium excretion in rats with established HF (**Figures 6A,B**). The mean values of water and sodium intake were not significantly different among the three experimental groups of rats (**Figures 6C,D**). Accordingly, HF rats treated with the vehicle exhibited higher positive water (**Figure 6E**) and sodium balance (**Figure 6F**) compared to sham rats. Daily water and sodium balance in HF rats treated with vildagliptin were similar to sham rats and lower than non-treated HF rats.

## DPPIV Inhibition Increases NHE3 Phosphorylation at Serine 552 in the Renal Cortex of HF Rats

Lower levels of phosphorylated NHE3 at the PKA consensus site serine 552 in the proximal tubule has been associated with higher NHE3 transport activity (Crajoinas et al., 2010, 2014; Girardi and Di Sole, 2012; Pontes et al., 2015; Farah et al., 2016). As show in **Figures 7A,B**, the levels of NHE3 phosphorylation at serine 552 (PS552-NHE3) were much lower in HF than in sham-operated rats (39 ± 4 vs. 100 ± 3%, P < 0.01) (**Figures 7A,B**). In rats treated with vildagliptin, PS552-NHE3 was higher than in vehicle treated HF rats and similar to sham rats. Total NHE3 protein expression was slightly higher in the renal cortex of vehicle-treated HF rats than in sham-operated rats and vildagliptin-treated HF rats (**Figures 7A,C**).

### DPPIV Inhibition Improves GFR and Exerts Anti-Proteinuric Effects in Rats with Established HF

As depicted in **Figure 8**, vehicle-treated HF rats exhibited a much lower GFR (**Figure 8A**) and a much higher excretion of protein in the urine than sham rats (**Figures 8B,C**). Vildagliptin treatment restored both GFR and proteinuria to sham levels. These changes on GFR were not accompanied by changes on the kidney/body weight ratios (**Table 1**). The profile of urinary proteins excreted by HF and sham rats was evaluated by SDS-PAGE, and the amount of intact albumin was semiquantitatively determined by densitometry. As seen in **Figures 8C,D**, the urinary excretion of intact albumin was remarkably higher in vehicle-treated HF rats compared to both sham and HF rats treated with vildagliptin. Moreover, the levels of albumin excretion were not significantly different between vildagliptin-treated HF rats and sham rats. Analysis of the pattern of proteinuria excreted by vehicle-treated HF rats (**Figure 8C**) suggests a mixed origin of proteinuria because both glomerular and tubular protein fractions, i.e., high and low molecular weight proteins, were present in the urine.

### DPPIV Inhibition Increases Renal Cortical Expression of Megalin, Nephrin and Podocin Expression in Rats with Established HF

Given the anti-proteinuric effect of vildagliptin treatment, we next tested the hypothesis that DPPIV inhibition could regulate the expression of components of the apical endocytic machinery in the renal proximal tubule (megalin and cubilin) (Willnow et al., 1996; Leheste et al., 1999; Birn and Christensen, 2006) and/or components of the glomerular filtration barrier, including nephrin and podocin (Kestila et al., 1998; Tryggvason, 1999;

Boute et al., 2000; Luimula et al., 2000; Agrawal et al., 2013). As shown in **Figure 9A**, the protein expression of the endocytic receptor megalin was significantly reduced in the cortex of HF rats compared to sham rats (69 ± 6 vs. 100 ± 3%, p < 0.05). Interestingly, vildagliptin treatment upregulated megalin to levels higher than sham rats (142 ± 11 vs. 100 ± 3%, p < 0.01). Conversely, the abundance of cubilin in the renal cortex was similar among the three groups of rats (**Figure 9B**). Although the podocyte main slit diaphragm proteins (Kestila et al., 1998; Tryggvason, 1999; Boute et al., 2000; Luimula et al., 2000; Agrawal et al., 2013), nephrin and podocin, were similar between sham and vehicle-treated HF rats, they were upregulated by DPPIV inhibition in the renal cortex of rats with established HF (**Figures 9C,D**).

### DISCUSSION

The heart and the kidney are very closely related. Cardiac impairment can result in renal dysfunction, and worsening of renal function is a strong predictor of long-term adverse outcomes in patients with HF. In the present study, we demonstrated that chronic treatment with the DPPIV inhibitor vildagliptin exerts renoprotective and cardioprotective effects in rats with established HF, reversing cardiac remodeling and improving both LV systolic and diastolic function. Longterm renoprotection conferred by DPPIV inhibition involves improved renal handling of sodium and water which may have ultimately led to relief of volume expansion and pulmonary congestion in HF. Additionally, DPPIV inhibition significantly reduces proteinuria in HF rats. The anti-proteinuric effects of DPPIV inhibition were associated with upregulation of the apical proximal tubule endocytic receptor megalin as well as of the podocyte main slit diaphragm proteins nephrin and podocin.

We have recently reported (dos Santos et al., 2013) that humans and rats with HF exhibit higher plasma levels of DPPIV activity and abundance compared to their healthy counterparts. Additionally, previous studies from our laboratory (dos Santos et al., 2013) and others (Sauvé et al., 2010; Gomez et al., 2012; Shigeta et al., 2012; Aoyama et al., 2016) have demonstrated that genetic deletion or pharmacologic inhibition of DPPIV

prevents the onset of HF after myocardial infarct/injury in rodents and large animal models. However, to the best of our knowledge, this is the first report that reveals that DPPIV inhibition can exert not only preventive but also therapeutic effects in HF by restoring myocardial structure and function. Our results that vildagliptin ameliorated cardiorenal function in rats with established HF contrast with those from Yin et al. (2011). These authors did not observe any beneficial effect of vildagliptin at a daily dose of 15 mg/kg/day once daily on either preventing or reversing cardiac remodeling and dysfunction in post-myocardial infarcted rats. One plausible explanation of this discrepancy between our findings and theirs is the higher dose (120 vs. 15 mg/kg/day) and higher frequency of administration of vildagliptin we employed (twice daily vs. once daily dosing). Indeed, we have noticed that although chronic treatment with 20 mg/kg/day vildagliptin inhibits plasma DPPIV activity, it fails to inhibit the activity of the peptidase in the heart and in the kidneys of HF rats. Moreover, at this low dose, vildagliptin is unable to ameliorate cardiac and renal function and to reduce pulmonary congestion in rats with established HF (unpublished observations). In concert, these observations suggest that local inhibition of DPPIV activity in the heart and possibly in the kidney may have a pivotal role in mediating the therapeutic effects of DPPIV inhibitors in rats with HF.

An intriguing finding of the present work is that administration of vildagliptin, a competitive inhibitor of DPPIV catalytic activity, not only inhibits the activity of DPPIV but also reduces the protein and mRNA-expression of the peptidase in the heart of HF rats. Similarly, Kanasaki et al. (2014) have recently reported that administration of the DPPIV inhibitor linagliptin reduces DPPIV activity and expression in the kidney and endothelial cells of streptozotocin-induced diabetic mice. Reduced DPPIV expression in these diabetic mice was associated with an upregulation of components of the microRNA (miRNA)

29 family. Interestingly, the miRNA 29 family has been shown to be downregulated after myocardial infarction (van Rooij et al., 2008; Melo et al., 2014), and this downregulation contributes to cardiac fibrosis (van Rooij et al., 2008). However, it remains to be established whether this post-transcriptional mechanism is also involved in the upregulation of DPPIV activity/expression in the heart of experimental models of HF.

The cardioprotective effects of DPPIV inhibition are often attributed to increased bioavailability of GLP-1, BNP, and SDF-1α that ameliorate cardiac performance and contractility, reduce hypertrophy, fibrosis and apoptosis, and improve stem cell mobilization and angiogenesis to the myocardium (Zaruba et al., 2009; Shigeta et al., 2012; dos Santos et al., 2013; Hocher et al., 2013). The present data suggest a role for the renoprotective actions of vildagliptin in the outcomes of rats with established HF. Indeed, it is undeniable that neurohumoral activation in response to the low output in decompensated HF and the consequent water and salt retention by the kidneys are important factors that increase circulatory filling pressure and restore cardiac ejection (Cadnapaphornchai et al., 2001; Chatterjee, 2005; Brum et al., 2006). However, chronic and excessive volume expansion with increased preload can lead to cardiac remodeling and dilation, which has deleterious effects on cardiac function. In fact, changes on renal function are not only a marker of HF but may also be a pathogenic factor in causing the progression of cardiac deterioration (Boerrigter et al., 2013). One may therefore speculate that improvement of cardiac remodeling and function in vildagliptin-treated HF rats may result, at least in part, from restored renal function.

Derangements in several hormonal systems contribute to sodium and water retention in HF. The results from our previous (dos Santos et al., 2013) and present work suggest that restoration

of the GLP-1/GLP-1R signaling activation in the renal cortex may be implicated in the vildagliptin-mediated amelioration of renal function in HF rats. The acute diuretic and natriuretic actions of GLP-1 have been consistently demonstrated by a variety of studies in rodents (Moreno et al., 2002; Crajoinas et al., 2011; Rieg et al., 2012; Thomson et al., 2013; Farah et al., 2016) and humans (Gutzwiller et al., 2004, 2006; Skov et al., 2013). The effects of GLP-1 on sodium and water handling in both rodents and humans is mediated, at least in part, by inhibition of NHE3-mediated renal proximal tubule sodium reabsorption via activation of the cAMP/PKA signaling pathway (Crajoinas et al., 2011; Rieg et al., 2012; Thomson et al., 2013; Farah et al., 2016). Accordingly, in the present work, we found that vildagliptin administration in rats with established HF elevated the concentration of circulating active GLP-1, and most likely, the concentration of this incretin hormone in the renal tubular fluid restores PKA activation to the levels of sham rats. These events were associated with higher levels of renal cortical NHE3 phosphorylation at the PKA consensus site serine 552, a surrogate for reduced NHE3 transport activity in the proximal tubule (Crajoinas et al., 2010; Girardi and Di Sole, 2012). It is noteworthy that GLP-1 also exerts diuretic and natriuretic effects in rodents through increments in both renal blood flow and GFR (Jensen et al., 2015). Indeed, recent studies have shown that GLP-1R is expressed in the afferent arterioles and that stimulation of GLP-1R by specific agonists increases renal blood flow (Thomson et al., 2013; Jensen et al., 2015). Therefore, our finding that HF rats treated with vildagliptin display higher GFR than vehicle-treated HF rats may also be due, at least in part, to the activation of GLP-1/GLP-1R in the renal vasculature.

DPPIV modulates multiple substrates other than GLP-1 that exert natriuretic and/or renoprotective effects, such as BNP, SDF-1α, substance P, among others (Makino et al., 2015). Thus, it is important to emphasize that the beneficial effects of DPPIV inhibition on sodium and water balance in HF shown herein might, in part, occur via GLP-1 independent mechanisms. In this context, Rieg and colleagues found that administration of the DPPIV inhibitor alogliptin was capable of inducing diuresis and natriuresis in GLP-1R knockout mice (Rieg et al., 2012). Moreover, we have previously found that inhibition of the catalytic activity of DPPIV inhibits NHE3 activity in the opossum kidney clone P (OKP) proximal tubule cell line, and it is well known that the proximal tubule does not synthesize GLP-1 (Girardi et al., 2008).

In addition to the effects on sodium and water homeostasis, pharmacological administration of DPPIV inhibitors confers renoprotection by reducing proteinuria and ameliorating renal damage in experimental models of diabetic nephropathy (Liu et al., 2012; Kanasaki et al., 2014; Eun Lee et al., 2016). The

### REFERENCES

Agrawal, V., Prasad, N., Jain, M., and Pandey, R. (2013). Reduced podocin expression in minimal change disease and focal segmental glomerulosclerosis is related to the level of proteinuria. Clin. Exp. Nephrol. 17, 811–818. doi: 10.1007/s10157-013- 0775-y

anti-proteinuric effects of DPPIV inhibition have also been observed in type 2 diabetes patients (Hattori, 2011; Groop et al., 2013; Kawasaki et al., 2015; Nakamura et al., 2016). In the present study, we found that HF rats display tubular and glomerular proteinuria and that tubular proteinuria in these animals was associated with reduced expression of the endocytic receptor megalin. On the other hand, the origin of glomerular proteinuria in these rats remains obscure. It is worth mentioning that glomerular proteinuria has not often been associated with lower expression of nephrin but instead with changes in the levels of tyrosine phosphorylation of its cytoplasmic tail (Carney, 2016; New et al., 2016). Surprisingly, treatment with vildagliptin upregulated megalin as well as the podocyte slit diaphragm proteins nephrin and podocin in the kidneys of HF rats to levels higher than sham. The molecular mechanisms by which vildagliptin increases the abundance of megalin, nephrin and podocin in the renal cortex of HF rats is yet to be established.

In summary, our findings demonstrated that the DPPIV inhibitor vildagliptin exerts renoprotective effects and ameliorates cardiorenal function in rats with established HF. Moreover, our data suggest that DPPIV may constitute one of the pathophysiological connections between the failing heart and kidneys. Given the lack of pharmacological agents that directly improve renal function in patients with HF, long-term studies with DPPIV inhibitors are warranted to ascertain whether the renoprotective effects of DPPIV inhibition ultimately translate into improved clinical outcomes.

### AUTHOR CONTRIBUTIONS

DA, performed experiments, analyzed data, interpreted the results of the experiments, prepared figures, and drafted the manuscript. FM, performed experiments, analyzed data, and interpreted the results of the experiments. RD, performed experiments, analyzed data, interpreted the results of the experiments, and edited and revised the manuscript. LS, performed experiments and analyzed data. EA, performed experiments. LD, interpreted the results of the experiments and drafted, edited and revised manuscript. PT, developed the experimental model of HF and edited and revised manuscript. AG, conceived of and designed the research, prepared figures, and drafted, edited and revised manuscript. All, approved final version of the manuscript.

### ACKNOWLEDGMENTS

This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Grant 2013/10619-8 and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq).


glucagon-like peptide-1 in heart failure and mitigates myocardial remodeling and apoptosis via the exchange protein directly activated by cyclic AMP 1/ras-related protein 1 axis. Circ. Heart Fail. 9:e002081. doi: 10.1161/CIRCHEARTFAILURE.115.002081


peptidase-4 improves cardiovascular outcomes after myocardial infarction in mice. Diabetes 59, 1063–1073. doi: 10.2337/db09-0955


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Arruda-Junior, Martins, Dariolli, Jensen, Antonio, dos Santos, Tucci and Girardi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# New Insights on the Use of Dietary Polyphenols or Probiotics for the Management of Arterial Hypertension

José L. de Brito Alves 1, 2 \*, Vanessa P. de Sousa<sup>1</sup> , Marinaldo P. Cavalcanti Neto<sup>3</sup> , Marciane Magnani <sup>4</sup> , Valdir de Andrade Braga<sup>5</sup> , João H. da Costa-Silva<sup>6</sup> , Carol G. Leandro<sup>6</sup> , Hubert Vidal <sup>2</sup> and Luciano Pirola<sup>2</sup>

*<sup>1</sup> Department of Nutrition, Health Sciences Center, Federal University of Paraíba, João Pessoa, Brazil, <sup>2</sup> CarMeN (Cardio, Metabolism, Diabetes and Nutrition) Laboratory, Institut National de la Santé et de la Recherche Médicale U1060, INRA 1397, Université Claude Bernard Lyon 1, Oullins, France, <sup>3</sup> Departments of Clinical Analyses, Toxicology and Food Sciences, University of São Paulo, Ribeirão Preto, Brazil, <sup>4</sup> Department of Food Engineering, Technology Center, Federal University of Paraíba, João Pessoa, Brazil, <sup>5</sup> Biotechnology Center, Federal University of Paraíba, João Pessoa, Brazil, <sup>6</sup> Department of Physical Education and Sport Sciences, Federal University of Pernambuco, Vitoria de Santo Antão, Brazil*

### Edited by:

*Jean-Pierre Montani, University of Fribourg, Switzerland*

### Reviewed by:

*Sergio Davinelli, University of Molise, Italy Arrigo Francesco Cicero, University of Bologna, Italy*

\*Correspondence: *José L. de Brito Alves jose\_luiz\_61@hotmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *22 May 2016* Accepted: *20 September 2016* Published: *06 October 2016*

### Citation:

*de Brito Alves JL, de Sousa VP, Cavalcanti Neto MP, Magnani M, Braga VA, Costa-Silva JH, Leandro CG, Vidal H and Pirola L (2016) New Insights on the Use of Dietary Polyphenols or Probiotics for the Management of Arterial Hypertension. Front. Physiol. 7:448. doi: 10.3389/fphys.2016.00448* Arterial hypertension (AH) is one of the most prevalent risk factors for cardiovascular diseases (CD) and is the main cause of deaths worldwide. Current research establish that dietary polyphenols may help to lower blood pressure (BP), thus contributing to the reduction of cardiovascular complications. In addition, the health benefits of probiotics on BP have also attracted increased attention, as probiotics administration modulates the microbiota, which, by interacting with ingested polyphenols, controls their bioavalability. The aim of the present mini-review is to summarize and clarify the effects of dietary polyphenols and probiotics administration on BP using combined evidence from clinical and experimental studies, as well as to discuss the current debate in the literature about the usefulness of this nutritional approach to manage BP. Clinical trials and experimental studies have demonstrated that consuming dietary polyphenols or probiotics in adequate amounts may improve BP, ranging from modest to greater effects. However, the mechanisms linking probiotic intake and reduced BP levels need to be further elucidated as a definitive consensus on the link between intake of polyphenols or probiotics and improvement of AH has not been reached yet.

### Keywords: probiotics, hypertension, blood pressure, dietary polyphenols

## INTRODUCTION

Arterial hypertension (AH) affects more than 1 billion people and is the major risk factor for CD (Hedner et al., 2012). The cause of AH has been difficult to identify due to its multi-factorial nature, which involves genetic and environmental factors. A balanced and healthy diet plays a key role in the maintenance of the cardiovascular health status, which is a major determinant of lifespan. In the last decades, in both western and developing countries, the consumption of highly caloric—rich in fat and carbohydrates—and sodium-rich diets have become predominant, increasing the incidence of AH (Bjerregaard, 2010; Popkin, 2011).

The regulation of BP is one of the most complex physiological functions and depends on the integrated actions of cardiovascular, renal, neural and endocrine systems (Corry and Tuck, 1999; de Brito Alves et al., 2015). In addition, augmentation of proinflammatory markers, reactive oxygen species and dysfunction in energy metabolism are related to hypertensive conditions (Carthy, 2014).

Despite the advances in the understanding of the pathophysiology and pharmacotherapy of AH, interventional strategies helping to reduce BP levels remain one of the great problems to be developed. Studies have investigated the beneficial effect of dietary changes on BP levels and also identified a BP-lowering activity on different dietary compounds.

The intestine and its microbiota constitute an important site of interaction with the dietary compounds (Bäckhed et al., 2004), and in particular polyphenols, whose bioavailability is dependent on prior metabolization by the microbiota (Ozdal et al., 2016; Stevens and Maier, 2016). The interaction between gut microbiota and diet can affect the gut-immune homeostasis, cell proliferation, metabolism, and intestinal permeability (Ding et al., 2010). Growing evidence supports the notion that the gut microbiota plays an important role in the development of CD and AH (Khalesi et al., 2014; Jose and Raj, 2015), and dietary compounds might modulate the gut microbiota favoring or lowering the BP (Buettner et al., 2007; Anhê et al., 2015).

Accordingly, findings have suggested the notion that dietary polyphenols and probiotic supplementation could help alleviating AH, by altering the gut microbiota and favoring an antioxidant activity (Anhê et al., 2015; Gómez-Guzmán et al., 2015; Grosso et al., 2016), although null and contrary findings have also been reported (Ras et al., 2013). In the present mini-review, we summarize the current understanding emerging from experimental, clinical and epidemiological studies, on how dietary polyphenols and probiotics intake may help to lower BP in hypertensive conditions.

### DIETARY POLYPHENOLS AND THE CONTROL OF BP

Polyphenols are a large and heterogeneous family of bioactive molecules found in numerous food sources. Generally defined as dietary antioxidants, polyphenols have been established as bioactive compounds that benefit human health via modulation of metabolism (Choi et al., 2014). Dietary polyphenols are mainly classified in catechins (proanthocyanidins), flavonols, flavanones, ellagitannins, and isoflavones. Studies have investigated the effects of dietary polyphenols, or their metabolites supplementation either via administration as polyphenol-enriched diets, polyphenol extracts from foods or as administration of specific polyphenolic compounds (e.g., quercetin, rutin, resveratrol, hesperidin, cinnamon) (Mendes-Junior et al., 2013; Amiot et al., 2016).

Dietary polyphenols have been shown to exert beneficial effects on markers for cardiovascular risk factors, including reduction of BP, improvement of endothelial function and lowering of plasma lipids. Mechanistically, it has been suggested that dietary polyphenols can alleviate hypertension through antiinflammatory and anti-oxidant effects, and increased oxide nitric (NO) production (Davinelli and Scapagnini, 2016). The antiinflammatory effect is associated with a reduced expression of the redox-sensitive nuclear factor-kB (NF-κB), while that the anti-oxidant effect of polyphenols is related to improved enzymatic activities of superoxide dismutase, catalase and glutathione peroxidase. In addition, polyphenols participate in the activation of the redox-sensitive phosphoinositide 3 (PI3)-kinase/Akt pathway, leading to increased formation of NO (Davinelli and Scapagnini, 2016). Taken together, all these pathways help to reduce blood pressure in hypertensive conditions.

Other studies have, however, reached negative or null effects. These effects have been investigated in several cohorts and randomized clinical trials (**Table 1**).

In a small-scale randomized nutritional trial, the administration of a polyphenol-rich diet (approximately 3000 mg polyphenol/day) reduced postprandial triglyceride-rich lipoprotein plasma concentrations and oxidative stress in study participants with a high risk of CD (Annuzzi et al., 2014).

The beneficial effects of polyphenol supplementation have also been demonstrated in the larger PREDIMED (Prevention with Mediterranean Diet) cohort. In this study, the consumption of a Mediterranean diet—supplemented with extra-virgin olive oil or nuts—resulted in reduced incidence of cardiovascular events (myocardial infarction, stroke, or death from cardiovascular causes) (Estruch et al., 2013). Within the PREDIMED cohort, a sub-study on 1139 high-risk participants was performed in which two different polyphenol-rich diets (based on supplementation with extra-virgin olive oil or nuts) were randomly assigned. The increase in polyphenol intake—which was unequivocally identified as increased total urinary polyphenol excretion was associated with decreased inflammatory biomarkers and a decrease of systolic and diastolic BP (Medina-Remón et al., 2016).

In the Polish population of the non-interventional HAPIEE cohort (Health, Alcohol and Psychosocial factors In Eastern Europe), elevated dietary intake of polyphenols was associated with lower body mass index (BMI), waist circumference (WC), BP and triglycerides, further suggesting that high polyphenols intake is inversely associated to metabolic syndrome and its clinical manifestations (Grosso et al., 2016). To ascertain whether olive oil polyphenols alleviate AH independently from the lipid component of olive oil, which is rich in monounsaturated fatty acids (MUFA), a doubleblind, crossover dietary-intervention study was performed in which mildly hypertensive women received polyphenol-rich olive oil (approximately 30 mg/day) in a first dietary period, and polyphenol-free olive oil in a second dietary period. Interestingly, only polyphenol-rich olive oil decreased BP and improved endothelial function, underscoring the specific role of polyphenols within the olive oil (Moreno-Luna et al., 2012). In addition, in a randomized, double blind, controlled, crossover trial, the hypotensive and lipid-lowering capacity have also been demonstrated for olive leaf extract (Lockyer et al., 2015, 2016).

Although the studies presented above indicate that administration of dietary polyphenols might contribute to the control of BP, a definitive consensus has not been reached yet, as independent investigations did not support the hypothesis that dietary intake of polyphenols is beneficial to cardiovascular health (Ras et al., 2013).


### TABLE 1 | Summary of the major studies investigating the relationship between dietary polyphenols intake and blood pressure/cardiovascular health.

### RESVERATROL

Resveratrol (3,5,4′ -trihydroxy-trans-stilbene) is a stilbenoid. Studies demonstrated that resveratrol possess an intrinsic cancer chemopreventive activity (Jang et al., 1997), and further interest on this molecule arose with the identification of its capability to increase lifespan when administered to experimental organisms including Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophila melanogaster (Ingram et al., 2006), and mice (Baur et al., 2006). Furthermore, it was soon documented in animal models that resveratrol (i) protects from the development of obesity-dependent metabolic disorders (Fröjdö et al., 2008) and (ii) can be the molecule responsible for the cardio-protective effects of red wine (Wu et al., 2011).

Preclinical studies have investigated the effects of resveratrol administration on cardiovascular health. In spontaneously hypertensive rats (SHR) and in mice rendered hypertensive by angiotensin-II injection, resveratrol administration reduced oxidative stress in the endothelium, improved vascular function and attenuated AH (Dolinsky et al., 2013). Resveratrol supplementation to SHR dams during the perinatal period (from gestation to weaning of the offspring) alleviated the development of AH in the offspring at the adult age (Care et al., 2016). These finding have prompted the development of several small-scale clinical trials evaluating the effects of resveratrol supplementation on cardiovascular systems.

In some trials, resveratrol supplementation (range 200–300 mg/day) improved insulin resistance, glycemic profile and lipid metabolism (Wong et al., 2011; Chen et al., 2015). However, in other studies, 12-week supplementation with 500 mg/day resveratrol (Faghihzadeh et al., 2015) or 150 mg/day for 4 weeks (van der Made et al., 2015) did not change metabolic risk markers related to cardiovascular dysfunction.

A recent meta-analysis of 10 randomized clinical trials failed to show any benefit of resveratrol supplementation on cardiovascular risk factors. In particular resveratrol had no effect on systolic or diastolic BP (Sahebkar et al., 2015). However, an independent meta-analysis performed on more stringent adjudication criteria for study quality, including only six studies comprising a total of 247 subjects, suggested that resveratrol consumption decreased systolic BP (at the higher administered doses) while having no significant effects on diastolic BP (Liu et al., 2015). Given the absence of a clear consensus on the effects on resveratrol on the control of BP in humans, as opposed to the clear conclusions in rodent models, the need for larger and well-designed clinical trials, was solicited by the authors of both meta-analysis to definitely prove, or reject, a causal link between resveratrol administration and the control of BP (Novelle et al., 2015).

### QUERCETIN

Quercetin [2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4Hchromen-4-one] is a polyphenolic compound belonging to the class of flavonoids, which is naturally found in apples, berries, and red wine (Sigel et al., 1977). Epidemiological data and dietary analysis indicated that quercetin could contribute to decrease the risk of coronary heart disease in elderly patients (Hertog et al., 1993). To test the hypothesis that quercetin reduces BP in hypertensive patients, a randomized, double-blind, placebocontrolled, crossover study was performed to test the efficacy of quercetin supplementation (730 mg quercetin/day for 28 days) in pre-hypertensive and stage-1 hypertensive volunteers. Quercetin supplementation reduced systolic, diastolic and mean arterial pressure in stage-1 hypertensive volunteers, while having no effect on pre-hypertensive participants (Edwards et al., 2007). More recently, data obtained from hypertensive patients, also indicated that quercetin supplementation (162 mg/day for 6 weeks) exert cardio-protective effect, with a decrease of 24 h systolic BP of 3.6 mmHg (Brüll et al., 2015). On the other hand, a recent similar study in overweight-to-obese adults with AH did not detect changes in post-prandial BP, nor endothelial function, upon acute administration of quercetin (54 mg in a single dose) (Brüll et al., 2016). The effects of quercetin administration are likely dependent on the doses and administration timespan. A recent meta-analysis of randomized controlled trials investigating the effects of quercetin on BP support the idea that quercetin doses greater than 500 mg/day have a significant effect on the reduction of BP (Serban et al., 2016).

### HESPERIDIN

Hesperidin (30, 5, 9-dihydroxy-40-methoxy-7-orutinosyl flavanone) is an abundant flavonoid found in citrus fruit, particularly in the peel of oranges and lemon (Sharma et al., 2015). Reports suggest that hesperidin exerts a cardio-protective action via its antioxidant and antihypertensive properties (Wilmsen et al., 2005), as also demonstrated in an ischemic heart disease model in diabetic rats (Agrawal et al., 2014) and in SHR (Yamamoto et al., 2008; Ikemura et al., 2012). Over the last years, a large body of studies in cell culture and animal models has elucidated the molecular targets and mechanisms of action of hesperidin. Besides its cardio-protective actions, hesperidin has also shown anticancer, anti-inflammatory, and neuroprotective properties (Roohbakhsh et al., 2015). However, at present, clinical studies regarding the therapeutic effects of hesperidin have not appeared, and pre-clinical testing in humans is warranted to confirm the beneficial effects observed in animal models.

Recently it has been suggested that dietary polyphenols consumption could help maintain intestinal homeostasis and metabolic health. In light of recent discoveries, studies have shown that dietary polyphenols can exert part of their beneficial action through modulation of the microbiota (Anhê et al., 2016). For example, it was demonstrated that a polyphenol-rich cranberry extract prevented obesity and the metabolic syndrome in diet-induced obesity through prebiotic effect on the gut microbiota (Anhê et al., 2015). The investigation on whether dietary polypnehols attenuate BP through beneficial actions on gut microbiota is a newly developing field and the underlying mechanisms remain to be elucidated.

### EFFECTS OF PROBIOTICS ON BP

The term probiotic means "for life" and it was first used to describe compounds produced by protozoa to stimulate the growth of other organisms (Lilly and Stillwell, 1965). Currently, the term probiotics refers to nonpathogenic microorganisms (bacteria or yeast) that, when ingested, are capable to reach the gut in sufficient amounts to confer health benefits (Parvez et al., 2006). Historically, probiotics derived from dairy products were the first to be isolated and studied (Tapsell, 2015). However, in the last decades, potentially probiotic strains from vegetable sources have also been isolated and their health benefits investigated (Rivera-Espinoza and Gallardo-Navarro, 2010; Vitali et al., 2012). Probiotics can be ingested either as supplements or incorporated in food or beverages in the form of dairy or non-dairy probiotic products (Jankovic et al., 2010; Vijaya Kumar et al., 2015). Despite the differences among the quantities of probiotic intake recommended by American or European Agencies to confer generic health claims, probiotic intake of around 106–10<sup>8</sup> CFU/g−<sup>1</sup> (or mL−<sup>1</sup> ) or 108–10<sup>10</sup> CFU/day (CFU, colony forming units) have shown to be efficacious (Champagne et al., 2011).

Saccharomyces boulardii is the main nonpathogenic yeast being used as probiotic. In addition, numerous bacteria belonging to the Enterococcus, Pediococcus, Bacillus, Streptococcus, Lactococcus and Propionibacterium genera are recognized as potential candidates for probiotic status. It is important to highlight that the Lactobacillus (L.) and Bifidobacterium (B.) genera constitute the majority of probiotics found on the market (Wohlgemuth et al., 2010; Champagne et al., 2011). Particularly, the lactic acid bacteria L. acidophilus, L. casei, L. paracasei, L. fermentum, L. reuteri, L. plantarum, L. rhamnosus and L. salivarius, as well as B. bifidum, B. breve, B. infantis, B. lactis, B. longum and B. thermophilum are among the main probiotics species marketed worldwide (Vijaya Kumar et al., 2015).

The benefits of the ingestion of probiotics by humans seem to be related to the improvement of the gut microbiota status, increase of enterocyte's resistance to pathogens, decrease or almost total elimination of pathogenic microorganisms within the intestinal tract, alleviation of nutritional intolerances (e.g., increased tolerance to lactose), enhancement of macroand micro-nutrients bioavailability, and the reduction of the prevalence of allergies in susceptible individuals (Sharma and Devi, 2014).

Beneficial effects on intestinal permeability to macromolecules, including lipopolysaccharides and on gut inflammation are also considered as major mechanisms conferring health benefit to probiotics (Cani, 2014). A recent study in a germ-free mouse model has also shown that administration of probiotics belonging to the Lactobacillus genus may alleviate the negative effects of chronic under-nutrition on postnatal growth (Schwarzer et al., 2016).

Interestingly, experimental and clinical reports have demonstrated that improvement of the gut microbiota though probiotic supplementation might positively help in reducing BP in the hypertensive conditions (Ettinger et al., 2014; Jose and Raj, 2015; Mell et al., 2015).

Despite these findings, other reports has shown that the probiotic supplementation did not induce any significant alterations in BP, heart rate (HR) or cardiovascular risk markers, such as total cholesterol, low-density lipoprotein, proinflammatory markers (Barreto et al., 2014; Mahboobi et al., 2014; Ivey et al., 2015).

### EXPERIMENTAL FINDINGS

By using a hypertensive rat model treated with nitro-Larginine methyl ester (L-NAME), it was demonstrated that the supplementation of fermented blueberries (very rich in polyphenols) containing L. plantarum (2 g/day for 4 weeks, containing 10<sup>9</sup> CFU) reduced systolic (by approximately 45%) and diastolic (by approximately 45%) BP in hypertensive animals (Ahrén et al., 2015). Mechanistically, it has been suggested that fermentation of blueberries by L. plantarum could reduce BP trough a mechanism involving a nitric oxide (NO)-dependent pathway (Ahrén et al., 2015). However, another study showed that adding probiotics to a blueberry-enriched diet did not enhance, and actually might have impaired the anti-hypertensive effect of blueberry consumption (Blanton et al., 2015).

In SHR, long-term administration of L. fermentum or L. coryniformis plus L. gasseri (3.3 × 10<sup>10</sup> CFU/day, for 5 weeks) similarly induced a progressive reduction in the systolic arterial pressure without significant modifications of the HR (Gómez-Guzmán et al., 2015). This finding was linked to improved endothelial function, reduced vascular oxidative stress and decreased vascular inflammation in the aorta of SH rats (Gómez-Guzmán et al., 2015). Interestingly, recombinant L. plantarum expressing angiotensin converting enzyme inhibitory peptide was effective in the diminution of BP in SHR. This finding was linked to increased levels of NO, as well as decreased levels of endothelin and angiotensin II in plasma, heart, and kidney in SHR (Yang et al., 2015).

The treatment with the probiotic formulation termed VSL#3 (Streptococcus thermophilus, B. longum, B. breve, B. infantis, L. acidophilus, L. plantarum, L. casei, L. bulgaricus) prevented endothelial dysfunction and improved vascular oxidative stress most likely by reducing bacterial translocation and the local angiotensin system in the mesenteric artery of rats with portal hypertension (Rashid et al., 2014). Another mechanism involved in the antihypertensive effect of probiotics is the production of bioactive peptides with angiotensin converting enzyme (ACE) inhibitory properties during the fermentation process (Thushara et al., 2016). ACE inhibition, in turn, lowers the synthesis of angiotensin II, which result in attenuation of vasoconstriction and blood pressure.

### CLINICAL FINDINGS

Initial clinical testing of the hypothesis that probiotics could reduce BP has been performed in small-scale, double blind, placebo-controlled studies. For example, supplementation of the diet with L. plantarum for 6 week in a population of smokers of both genders resulted in reduced systolic BP, improvement of metabolic alterations and attenuated generation of reactive oxygen species (Naruszewicz et al., 2002). Nevertheless, in an investigation in postmenopausal women with metabolic syndrome, a 14 days supplementation with fermented or non-fermented milk supplemented with L. plantarum, did not result in an improvement of systolic or diastolic BP. However, a significant reduction in total cholesterol, low-density protein cholesterol, glucose, homocysteine and inflammatory biomarkers was observed (Barreto et al., 2014), although this was not associated to reduced arterial BP, perhaps because of the short duration of the study. Similarly, administration to obese hypertensive patients of an hypocaloric diet (1500 kcal/day), supplemented with cheese containing the probiotic L. plantarum showed a remarkable reduction of body mass index associated with decrease of BP when compared to a control group receiving the same diet without probiotic supplementation (Sharafedtinov et al., 2013).

In a randomized double-blind clinical trial on type 2 diabetic volunteers, the supplementation with probiotic soymilk (containing L. planetarium A7) did not change the anthropometric parameters (represented by body mass index and waist to hip ratio), however, it reduced both systolic and diastolic BP (Hariri et al., 2015).

On the other hand, a 6-weeks randomized, controlled, parallel, double blind, factorial study performed in overweight men and women, demonstrated that consumption of L. acidophilus and B. animalis (at a dose of 3 × 10<sup>9</sup> CFU/day) did not significantly alter BP, HR, total cholesterol, lox density lipoprotein, high density lipoprotein, or triglycerides (Ivey et al., 2015). Similarly, another double-blind randomized controlled study, found that L. acidophilus and B. bifidum supplementation did not reduce systolic or diastolic BP in healthy adults with hypercholesterolemia (Rerksuppaphol and Rerksuppaphol, 2015). These findings suggest that the choice of an appropriate strain, as also the optimal dosage is crucial to achieve ideal beneficial effects from probiotics.

### CONCLUSION

The experimental and clinical findings summarized in this review suggest that dietary polyphenols or probiotic consumption may reduce BP and improve cardiovascular risk markers. Future studies investigating the effects of different polyphenolic compounds and probiotics, optimal dosage, intervention times, and studies on the underlying molecular mechanisms leading to improved control of BP are recommended to clarify the beneficial effects of dietary polyphenols and probiotics on AH. In addition, studies are needed to investigate the combined supplementation with dietary polyphenols and probiotics on the BP levels.

### AUTHOR CONTRIBUTIONS

JB and LP drafted the work and revised critically for important intellectual content, wrote the paper, and performed the final review of the manuscript. Vd, MC, MM, VB, JC, CL, and HV contributed to the conception of the work and performed the final review of the manuscript.

## REFERENCES


thrombosis in stroke-prone spontaneously hypertensive rats. Phytother. Res. 26, 1272–1277. doi: 10.1002/ptr.3724


microbiota and effects on bioaccessibility. Nutrients 8:78. doi: 10.3390/nu 8020078


and slightly obese subjects: a randomized, placebo-controlled crossover trial. PLoS ONE 10:e0118393. doi: 10.1371/journal.pone.0118393


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 de Brito Alves, de Sousa, Cavalcanti Neto, Magnani, Braga, Costa-Silva, Leandro, Vidal and Pirola. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Effects of Kefir on the Cardiac Autonomic Tones and Baroreflex Sensitivity in Spontaneously Hypertensive Rats

Brunella F. Klippel 1 †, Licia B. Duemke2 †, Marcos A. Leal <sup>1</sup> , Andreia G. F. Friques <sup>2</sup> , Eduardo M. Dantas <sup>3</sup> , Rodolfo F. Dalvi <sup>4</sup> , Agata L. Gava<sup>1</sup> , Thiago M. C. Pereira2, 4 , Tadeu U. Andrade<sup>2</sup> , Silvana S. Meyrelles <sup>1</sup> , Bianca P. Campagnaro<sup>2</sup> \* and Elisardo C. Vasquez 1, 2

### Edited by:

*Valdir Andrade Braga, Federal University of Paraiba, Brazil*

### Reviewed by:

*Bruce N. Van Vliet, Memorial University of Newfoundland, Canada Joao Henrique da Costa Silva, Federal University of Pernambuco, Brazil Valter Joviniano Santana Filho, Federal University of Sergipe, Brazil*

### \*Correspondence:

*Bianca P. Campagnaro biancacampagnaro@yahoo.com.br*

*† These authors have contributed equally to this work.*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *07 April 2016* Accepted: *22 May 2016* Published: *07 June 2016*

### Citation:

*Klippel BF, Duemke LB, Leal MA, Friques AGF, Dantas EM, Dalvi RF, Gava AL, Pereira TMC, Andrade TU, Meyrelles SS, Campagnaro BP and Vasquez EC (2016) Effects of Kefir on the Cardiac Autonomic Tones and Baroreflex Sensitivity in Spontaneously Hypertensive Rats. Front. Physiol. 7:211. doi: 10.3389/fphys.2016.00211* *<sup>1</sup> Laboratory of Translational Physiology, Federal University of Espirito Santo, Vitoria, Brazil, <sup>2</sup> Pharmaceutical Sciences Graduate Program, Vila Velha University, Vila Velha, Brazil, <sup>3</sup> Department of Physiology, Federal University of Vale Sao Francisco, Petrolina, Brazil, <sup>4</sup> Department of Biochemistry, Institute of Education, Science and Technology, Vila Velha, Brazil*

Aims: It has been previously shown that the probiotic kefir (a symbiotic matrix containing acid bacteria and yeasts) attenuated the hypertension and the endothelial dysfunction in spontaneously hypertensive rats (SHR). In the present study, the effect of chronic administration of kefir on the cardiac autonomic control of heart rate (HR) and baroreflex sensitivity (BRS) in SHR was evaluated.

Methods: SHR were treated with kefir (0.3 mL/100 g body weight) for 60 days and compared with non-treated SHR and with normotensive Wistar-Kyoto rats. Cardiac autonomic vagal (VT) and sympathetic (ST) tones were estimated through the blockade of the cardiac muscarinic receptors (methylatropine) and the blockade of β1−adrenoceptor (atenolol). The BRS was evaluated by the tachycardia and bradycardia responses to vasoactive drug-induced decreases and increases in arterial blood pressure (BP), respectively. Additionally, spontaneous BRS was estimated by autoregressive spectral analysis.

Results: Kefir-treated SHR exhibited significant attenuation of basal BP, HR, and cardiac hypertrophy compared to non-treated SHR (12, 13, and 21%, respectively). Cardiac VT and ST were significantly altered in the SHR (∼40 and ∼90 bpm) compared with Wistar rats (∼120 and ∼30 bpm) and were partially recovered in SHR-kefir (∼90 and ∼25 bpm). SHR exhibited an impaired bradycardic BRS (∼50%) compared with Wistar rats, which was reduced to ∼40% in the kefir-treated SHR and abolished by methylatropine in all groups. SHR also exhibited a significant impairment of the tachycardic BRS (∼23%) compared with Wistar rats and this difference was reduced to 8% in the SHR-kefir. Under the action of atenolol the residual reflex tachycardia was smaller in SHR than in Wistar rats and kefir attenuated this abnormality. Spectral analysis revealed increased low frequency components of BP (∼3.5-fold) and pulse interval (∼2-fold) compared with Wistar rats and these differences were reduced by kefir-treatment to ∼1.6- and ∼1.5-fold, respectively. Spectral analysis also showed an impairment of spontaneous BRS in SHR, but kefir-treatment caused only a tendency to reverse this result.

Conclusions: The novelty of this study is that daily chronic consumption of a low dose of kefir reduced the impairment of the cardiac autonomic control of HR and of the impaired BRS in SHR.

Keywords: probiotic, hypertension, sympathetic tone, vagal tone, tachycardia, bradycardia

### INTRODUCTION

In recent years, the fermented milks containing lactic acid bacteria, the so-called probiotics, including kefir, emerged as an alternative therapy due to the growing interest for wellbeing and healthy lifestyle. Experimental and clinical studies have demonstrated the beneficial effects of functional foods in cardiovascular diseases (Jakala et al., 2009; Turpeinen et al., 2011; Monteiro et al., 2012; Astrup, 2014; Friques et al., 2015).

The spontaneously hypertensive rat (SHR) have been used as an important tool to understanding cardiovascular dysfunctions, such as high blood pressure (BP), endothelial injury and abnormal neural control of the cardiovascular system (Vasquez et al., 1997; Abreu et al., 1998; Blanco et al., 2015; Friques et al., 2015). In recent years, this hypertensive animal emerged as a promising model to identify alternative or non-pharmacological agents for prevention/treatment of cardiovascular disease because it develops essential hypertension (Ceroni et al., 2009; Monteiro et al., 2012; Barbosa Neto et al., 2013; Friques et al., 2015). In this regard, our group has recently demonstrated the beneficial effects of kefir on BP and endothelial dysfunction in chronically treated SHR rats (Friques et al., 2015). The mechanisms of those actions included a partial correction of the reactive oxygen species/nitric oxide imbalance and the partial restoring of the endothelial architecture due to endothelial progenitor cells attraction.

Another relevant characteristic of the SHR is an abnormal autonomic nervous control of cardiac activity (Ceroni et al., 2009; Monteiro et al., 2012; Barbosa Neto et al., 2013), which is closely correlated with end-organ damage (Su and Miao, 2005; Abboud, 2010; Abboud et al., 2012). The influence of the autonomic nervous system on the heart rate (HR) and blood pressure (BP) in SHR has clearly been demonstrated (Schenberg et al., 1995; Vasquez et al., 1997; Abreu et al., 1998; Dias da Silva et al., 2002; Silva et al., 2009; Abboud, 2010) and appears to be frequencydependent (Dias da Silva et al., 2002; Ceroni et al., 2009; Silva et al., 2009). Despite some discrepancies, multiple studies have demonstrated a decreased cardiac parasympathetic (vagal) tone and an increased cardiac and vessel sympathetic tone in SHR when compared with normotensive Wistar rats (Ceroni et al., 2009; Silva et al., 2009; Hayward et al., 2012; Barbosa Neto et al., 2013). However, a possible beneficial action of the probiotic kefir on the cardiac dysautonomia in hypertensive rats has not yet been evaluated.

The baroreflex sensitivity, which is a marker of the capability of reflexes to increase vagal activity and to decrease sympathetic activity in response to a sudden increase in BP, is diminished in experimental models of essential (Barbosa Neto et al., 2013) and secondary (Moyses et al., 1994; Campagnaro et al., 2012) hypertension, as well as in cardiovascular related diseases, such as atherosclerosis (Vasquez et al., 2012). This reflex has classically been evaluated through invasive approaches using vasoconstrictor and vasodilator agents (Moyses et al., 1994; Campagnaro et al., 2012) and later through procedures using spectral analysis (Dias da Silva et al., 2002; Chapleau and Sabharwal, 2011).

The concept that the BP is affected by baroreflex-mediated changes in autonomic nerve activity in the heart and systemic vasculature, highlights the importance of studying the effects of alternative functional foods therapies on the autonomic and reflex control of the cardiovascular system. Therefore, the present study was designed to test the hypothesis that chronic administration of kefir can ameliorate the abnormal autonomic control of HR and the impaired baroreflex sensitivity in SHR rats.

### MATERIAL AND METHODS

### Animals

The present study was performed in male 4-month-old SHR and in age-matched Wistar-Kyoto rats obtained from the Vila Velha University animal care and that were maintained in the animal care facility of the Federal University of Espirito Santo, Brazil. The rats were housed in individual acclimatized plastic cages with a controlled temperature (22–23◦C), light-dark cycle (12:12-h), and were fed with a standard rat chow and water ad libitum. The study protocols were previously approved by the Institutional Committee on Animal Care (CEUA-UFES, Protocol #040/2014). All experimental procedures were performed in accordance with the guidelines for the care and use of laboratory animals as recommended by the National Institutes of Health (NIH).

### Kefir Preparation, Identification, and Administration

The identification, preparation and administration of kefir were performed as previously described (Friques et al., 2015). Briefly, kefir was obtained from the fermentation of the grains in whole milk. The kefir beverage was prepared by adding kefir grain to pasteurized whole milk in a ratio of 4% (w/v).

The treatment of the animals was started at the age of 4-month-old and lasted 60 days. One group of animals was treated with kefir (0.3 mL/100 g body weight, by gavage, SHRkefir) for 60 days. Another group of SHR was administered whole milk (0.3 mL, pH adjusted to 4.5, SHR) for 60 days for use as the hypertensive controls. The rationale for using 0.3 mL/100 g body weight was based on the dose translation from human to animal studies by a simple method using the body surface area normalization and that dose is compatible with that used in human beings. The Wistar rats were administered whole milk for 60 days and were used as normotensive control groups. The reason for treating the animals with kefir for 60 days was based on a previous study from our group (Friques et al., 2015), demonstrating that the treatment for less than 60 days had no effect on cardiovascular parameters in SHR.

## Instrumentation for Hemodynamic Measurements

After 60 days of kefir administration, the animals were intraperitoneally anesthetized with a mixture of ketamine and xylazine (91 + 9.1 mg/kg) and a polyethylene catheter (PE 50) was positioned into the femoral vein for injection of drugs and another into the inferior aorta for measurement of pulsatile pressure, mean BP and HR 48 h later using a dataacquisition system (Biopac Systems, Santa Barbara, CA, USA) in unrestrained animals.

At the end of the evaluation of hemodynamic parameters, the animals were euthanized by an over-dose of thiopental (100 mg/kg) and the hearts were excised. The right and the left ventricles, including the interventricular septum, were dissected from the remaining cardiac tissues. The tibia bone was also dissected and its length was measured. To evaluate the extent of cardiac hypertrophy for each animal the left ventricle weight was normalized by the animal's body weight and tibia length.

### Evaluation of the Cardiac Autonomic Tones

Cardiac autonomic tones were estimated through selective pharmacological blockers of the muscarinic receptors (methylatropine) and the β1−adrenergic receptors (atenolol) in conscious animals, based on previous studies from our laboratory and from others (Chapleau and Sabharwal, 2011; Campagnaro et al., 2012). As shown in the scheme in **Figure 3** (top), the cardiac parasympathetic tone was estimated by the change in basal HR 15 min after a single injection of methylatropine (1 mg/kg, i.v.), which reaches a plateau effect at approximately this time and lasts for ∼30 min. Immediately afterwards, they were injected with atenolol (1 mg/kg, i.v.), which also reached a maximum effect 15 min later, and this value was considered the intrinsic (pacemaker) HR. the next day, the sequence of the injections was the opposite, and the cardiac sympathetic tone was estimated by the change in the basal HR 15 min after atenolol, and the intrinsic HR was estimated 15 min under the double blockade.

### Pharmacological Baroreflex Sensitivity

The baroreflex control of arterial pressure was evaluated in conscious animals by measuring the tachycardia and bradycardia in response to an equivalent increase in arterial BP in each of the three groups of conscious animals. Based on previous studies from our laboratory demonstrating that the major sensitivity of the baroreflex is at changes in arterial BP close to the resting values and that changes in BP higher than 40 mmHg could elicit complex humoral mechanisms, we decided to challenge the baroreflex with sudden increase and decrease in BP of ∼25 mmHg (Schenberg et al., 1995; Gava et al., 2004). These changes in BP were elicited by a single bolus injection of phenylephrine (1µg/kg) and sodium nitroprusside (1µg/kg), respectively.

The relative contribution of the cardiac parasympathetic and sympathetic nerves was assessed by increase or decrease in HR in response to a sudden decrease or increase in BP, under the blockade of cardiac muscarinic receptors with methylatropine (1 mg/kg, i.v.) on day one and the blockade of the cardiac β1-adrenoceptors with atenolol (1 mg/kg, i.v.) on the following day.

### Spectral Analysis

An aim of this protocol was to characterize patterns of autonomic control in SHR and normotensive Wistar rats by power spectral analysis of pulse interval (PI) variability and BP variability as previously described (Dantas et al., 2012), Power spectra analysis was performed using a Matlab-customized software validated against the software developed by A. Porta in Italy (Linear Analysis version 8.3, University of Milan, Italy). Pulse intervals from systolic arterial pressure were obtained from 30 min of continuous BP records in conscious animals. A parametric method based on autoregressive model of spectral estimation was performed for systolic arterial BP and PI analysis. The oscillatory components in the rat were quantified as low frequency (LF: 0.2–0.8 Hz) mainly corresponding to sympathetic activity and high frequency (HF: 0.8–2.8 Hz) corresponding to vagal activity (influenced by respiration; Silva et al., 2009; Dantas et al., 2013; Quagliotto et al., 2015).

The estimation of spontaneous baroreflex sensitivity was obtained by measuring oscillations in BP and PI in the LF range using spectral analysis. The baroreceptor sensitivity value was provided by α-LF index, which was calculated as the square root of the ratio between the absolute power of PI-LF/BP-LF, expressed in ms/mmHg. The oscillations in BP and PI needed to be coherent (coherence<sup>2</sup> ), with a coherence higher than 0.5, and a negative phase difference was required between the two variables, as recently reviewed (Chapleau and Sabharwal, 2011).

### Statistical Analysis

The values are expressed as means ±SEM. First, a D'Agostino-Pearson omnibus normality test was performed to verify if that the values came from a Gaussian distribution. All data Statistical comparisons between the different groups were performed by a randomized one-way analysis of variance (ANOVA), followed by Bonferroni's post hoc test. A value of p < 0.05 was considered statistically significant. Statistical analysis was performed using GraphPad Prism software version 6.07 (GraphPad, Inc., San Diego, CA, USA).

## RESULTS

### Microbiological Analysis of Kefir

The microbiological analysis of random samples of grains used in the present study, showed a dominant microflora of kefir, which included various bacteria that are known to have beneficial effect (Acetobacter aceti, Acetobacter sp., Lactobacillus delbrueckii delbrueckii, Lactobacillus fermentum, Lactobacillus fructivorans, Enterococcus faecium, Leuconostoc spp.), as well as Lactobacillus kefiranofaciens, and yeasts (Candida famata, Candida krusei, and Candida kefir). The global counting of microorganisms in the samples of milk fermented with kefir grains was 7.6 x 10<sup>7</sup> units forming colonies (UFC)/mL. Fermenter microorganisms averaged 6.6 × 10<sup>7</sup> UFC/mL. Lactobacillus sp and Lactococcus sp represent the largest and most commonly identified lactic acid bacteria isolates. The average number of acetic and fermenter bacterias was 2.0 × 10<sup>6</sup> and 6.6 × 10<sup>7</sup> UFC/mL, respectively and the average number of yeasts was ∼4.1 × 10<sup>5</sup> UFC/mL.

### Basal BP and HR in Conscious Animals

As expected, the conscious SHR exhibited significant high levels of systolic, diastolic and mean BP (204 ± 8, 144 ± 4, and 167 ± 4 mmHg, respectively, p < 0.05) and resting HR (355 ± 9 bpm, p < 0.05) compared with the Wistar rats (140 ± 8, 88 ± 2, and 104 ± 2 mmHg, respectively, and 328 ± 12 bpm; **Figure 1**). Administration of kefir for 60 days caused a significant attenuation of systolic, diastolic and mean BP which was reduced to 169 ± 8, 116 ± 6, and 134 ± 6 mmHg, respectively, and on HR, which was reduced to 322 ± 11 bpm.

### Body Weight and Cardiac Hypertrophy

Administration of kefir did not cause effect on the consumption of water and food. At the moment of the evaluation of the outcomes (i.e., 60 days after kefir administration), body weight was significantly diminished (∼32%, p < 0.05) in non-treated and kefir-treated SHR than in the age-matched Wistar rats (328 ± 3 vs. 480 ± 8 g, p < 0.05). Left and right ventricular cardiac mass were significantly (p < 0.05) greater in non-treated SHR (893 ± 30 mg or 13% and 386 ± 24 mg or 40%, respectively) than in normotensive Wistar rats (787 ± 29 and 274 ± 20 mg, respectively). Kefir administration for 60 days caused a significant and complete normalization in the left ventricle mass (785 ± 25 mg or –13%) and a significant attenuation of the right ventricle mass (352 ± 37 mg or −10%). The normalization of the left ventricle weight by the body weight and tibia length revealed left ventricular hypertrophy in the non-treated SHR group (2.7 ± 0.3 mg/g and 22 ± 1.4 mg/mm, respectively, p < 0.05) compared with the values observed in the Wistar group (1.6 ± 0.4 mg/g and 17 ± 1.4 mg/mm, respectively; **Figure 2**). The treatment of SHR with kefir for 60 days had no significant effect on the left ventricle hypertrophy when the ventricle chamber weight was normalized by the body weight (2.4 ± 0.2 mg/g and 19 ± 1.2 mg/mm, respectively, p < 0.05). However, kefir administration reduced the left ventricle hypertrophy, when the ventricle chamber weight was normalized by the tibia length, which was 41 ± 0.3 mm in both SHR and SHR-kefir vs. 46 ± 0.3 mm in Wistar rats, p < 0.05).

### Cardiac Autonomic Tone

The protocol used to assess the cardiac parasympathetic (vagal) and sympathetic tones is schematically shown in **Figure 3B**. **Figure 3A** shows representative recordings of arterial BP and HR and **Table 1** and **Figure 3C** show the average data. In the normotensive Wistar group, the blockade of muscarinic receptors with methylatropine resulted in a significant increase in basal HR, from 336 ± 15 to 455 ± 11 bpm, indicating a vagal tone of +120 bpm (**Figure 3C**). In the non-treated SHR group, the change in basal HR was from 365 ± 12 to 406 ± 17 bpm (approx. +40 bpm), i.e., a significant decreased vagal tone compared with Wistar rats (**Table 1** and **Figure 3C**). In the SHR group treated with kefir for 60 days, methylatropine caused a significant change from 319 ± 15 to 410 ± 12 bpm (**Table 1** and **Figure 3C**), indicating a partial recovery of the vagal tone (approximately +90 bmp, p < 0.05; **Figure 3C**).

The effects of the blockade of β1-adrenoceptors with atenolol, which was used to evaluate the cardiac sympathetic tone, is illustrated in **Figure 3A** and the average data are shown in **Table 1** and **Figure 3C**. The basal HR was reduced to 306 ± 14 bpm (approximately −30 bpm) in the Wistar group. In the non-treated SHR the magnitude of reduction in basal HR was significantly higher (277 ± 16 bpm, approximately −90 bpm, p < 0.05), but the treatment of SHR with kefir for 60 days abolished this difference between strains; the basal HR was reduced to 294 ± 15 bpm (approximately −25 bpm).

The intrinsic HR, which was considered as the resting HR under the double blockade with methylatropine and atenolol (**Table 1**), was similar in the three groups of animals (**Figure 3C**).

## Classical Pharmacological Analysis of the Baroreflex Function

The effect of kefir adminsitration on the baroreflex sensitivity as well as the relative contribution of the cardiac vagal and sympathetic components in conscious SHR are shown in **Figures 4**, **5**. As expected and illustrated in **Figure 4**, when the baroreflex was challenged by moderate phenylephrine-induced increases in arterial BP (approx. 25 mmHg), the SHR group showed a significantly decrese in the reflex bradycadia (∼50%, p < 0.05), when compared with the Wistar group and, consequently, a significant reduction in the baroreflex gain (∼50%, p < 0.05) was observed (**Figure 4B**). When SHRs were treated with kefir for 60 days, phenylephrine-induced matchedincreases in BP resulted in mild, but significant, improvement of the reflex bradycardia (∼40%, p < 0.05) and consequently also the baroreflex gain (∼35%, p < 0.05), respectively. The blockade of cardiac muscarinic receptors with methylatropine basically abolished the reflex bradycardia and the baroreflex gain in a similar way in the three groups. The disappearence of intergroup

TABLE 1 | Effects of methylatropine and atenolol on resting heart rate in conscious Wistar, non-treated SHR and SHR treated for 60 days with kefir.


*Values are means* ± *SEM. \*p* < *0.05 vs. Wistar group;* #*p* < *0.05 vs. SHR group.*

differences in the reflex bradycardia after methylatropine indicate that they were due to the vagal component of the baroreflex.

**Figures 5A,B** show representative recordings and average data of the baroreflex when it was challenged by a moderate dose of sodium nitroprusside. This vasodilator induced falls in arterial BP of approximatly 25 mmHg in all groups of animals. The SHR group showed a significantly lower reflex tachycadia (23%, p < 0.05), when compared with the Wistar group and, consequently, a significant reduction in the baroreflex gain (23%, p < 0.05) was observed (**Figure 5B**). When SHRs were treated with kefir for 60 days, these differences were reduced to ∼8% and ∼12%, despite the same size of nitroprusside-induced reductions of BP. The blockade of cardiac β1-adrenoceptors with atenolol reduced (but not abolished) the reflex tachycardia in Wistar, SHR and SHR-kefir. The values reached were significantly lower in SHR and significantly recovered in the SHR treated with kefir for 60 days (14 ± 2.4, 7 ± 2.8, and 12 ± 2.4 bpm, respectively). A similar effect was observed in the baroreflex gain (0.53 ± 0.11, 0.26 ± 0.10, and 0.46 ± 0.12 bpm/mmHg, respectively; **Figure 5**). The lower reflex tachycardia and baroreflex gain in SHR under the atenolol effect indicate that part of the reflex tachycardia may be due to a withdrawall of the vagal activity, which was reduced in the SHR.

FIGURE 2 | Effects of chronic administration of kefir on the cardiac hypertrophy in SHR based on the left ventricular weight normalized by the body weight (A: left graph) and tibia length (B: right graph). Values are means ± SEM (*n* = 8 per group). \**p* < 0.05 compared to Wistar group, #*p* < 0.05 compared to non-treated SHR (one-way ANOVA).

FIGURE 3 | Effects of chronic administration of kefir on the cardiac autonomic tones in SHR. (A): typical recordings of pulsatile blood pressure before and respective chronotropic changes after the blockade of muscarinic receptors with methylatropine (1st day) and after the blockade of β-adrenoceptors (2nd day). (B): scheme showing the estimation of the vagal tone through methylatropine and the sympathetic tone through atenolol and the pacemaker or intrinsic heart rate (IHR), which was determined by the double blockade of vagal and sympathetic activity. (C): mean values ± SEM (*n* = 11 to 12 per group) used for determination of each parameter. \**p* < 0.05 vs. Wistar group; #*p* < 0.05 vs. non-treated SHR. (B): upper panel was redrawn from Chapleau and Sabharwal (2011).

### Spectral Analysis

Spectral analysis was applied to study the effect of kefir on the PI and BP variability in SHR-kefir compared to non-treated SHR and normotensive Wistar rats. The data showed a significantly higher PI variance in non-treated SHR (730 ± 150 ms<sup>2</sup> , p < 0.05) when compared with SHR-kefir (478 ± 80 ms<sup>2</sup> ) and with normotensive Wistar rats (247 ± 46 ms<sup>2</sup> ). The power spectral analysis of each component of the PI variabilities showed a remarkable difference in that the PI-LF component, which is widely accepted as an index of sympathetic modulation. SHR showed a significant increase (2-folf, p < 0.01) in PI-LF when compared with the Wistar group (39 ± 9 ms<sup>2</sup> ), suggesting a high sympathetic activity in the hypertensive group. The treatment of SHR with kefir for 60 days caused a significant attenuation of the PI-LF component (∼1.4-fold, p < 0.05; **Figure 6**, left bar graph). However, when comparison was made using the normalized values of PI-LF, no significant differences were observed among the three groups of animals, suggesting that this result is consistent with there being no difference between SHR and WKY in sympathetic control of the heart. The comparison of values of

PI-HF, which is widely associated with vagal modulation, did not show significant differences comparing the row data (**Figure 6**, right bar graph).

The spectral analysis of the BP variability showed a significantly higher variance in the non-treated SHR (73 ± 9 mmHg<sup>2</sup> , p < 0.05) when compared with Wistar rats (22 ± 3 mmHg<sup>2</sup> ) and it was reduced, but not normalized, in the SHR-kefir group (65 ± 8 mmHg<sup>2</sup> ). The BP-LF value, which is linked to vasosympathetic tone, was greater in the non-treated SHR group (13 ± 2 mmHg<sup>2</sup> , p < 0.05) when compared with the normotensive Wistar rats (3.7 ± 0.4 mmHg<sup>2</sup> ; **Figure 6**, right bar graph), providing an evidence of a relationship between the high BP, high vascular sympathetic activity and power spectra BP-LF in the SHR. Kefir treatment of SHR for 60 days significantly reduced that value to 8 ± 2 mmHg<sup>2</sup> (**Figure 6**, right bar graph).

The analysis of the spontaneous baroreflex sensitivity through the LF α-index in the conditions of a coherence higher than 0.5 and a negative phase, showed that the non-treated SHR exhibited significantly lower sensitivity values than the Wistar group (1.36 ± 0.09 vs. 1.75 ± 0.11 ms/mmHg, p < 0.05). Kefir treatment of SHR for 60 days resulted in a tendency (1.48 ± 0.09 ms/mmHg, p > 0.05) to attenuation of that dysfunction (**Figure 6**).

### DISCUSSION

The present study investigated the effects of multi-strain kefir on cardiovascular autonomic control and baroreflex sensitivity in SHR rats, since most of the available studies have focused on isolated bacteria providing limited information on the functional role of the widely consumed beverage obtained by fermentation of milk with kefir grains.

As previously demonstrated by us through microbiological analysis and scanning electron microscopy, the Brazilian kefir grains used in the present study are a complex and a sharing symbiotic mixture of beneficial bacteria and yeasts, including the Lactobacillus kefiranofaciens, Lactobacillus kefir, and Candida kefir (Friques et al., 2015). It has been demonstrated that the chronic consumption of kefir has protective effect against the development of high BP, endothelial dysfunction and endothelial surface damage in the SHR model of essential hypertension (Friques et al., 2015).

Despite a large body of evidence supporting the antihypertensive effects of kefir, it should be take into account that these effects may vary depending on the bacterial strain (Million et al., 2012). Most studies have been conducted using different isolated bacterial strains. For instance, it has been demonstrated that the antihypertensive effects of probiotics Lactobacillus strains in SHR could be through a decrease in pro-oxidative status (Gómez-Guzmán et al., 2015). One of the pro-oxidative contributors could be the activation of the renin-angiotensin system, but in the present study, we did not evaluate the effects of kefir on the angiotensin II in the SHR model. However, the widely consumed kefir beverage presents different microorganisms genera and, in this regard, the probiotic treatment protocol applied in the present work are more like the way humans use kefir.

There is emerging evidence of beneficial effects of probiotics in cardiovascular system. However, it is unclear if the promising positive effects evidenced could be due the regulation of cardiovascular autonomic control. We hypothesized that chronic kefir consumption for 60 days may be beneficial in management of autonomic control, based on our previous study showing that administration for shorter periods did not have significant effects on HR, BP and endothelial function (Friques et al., 2015). Similarly, the present study showed that the chronic administration of kefir for 60 days attenuated high BP and consequently the cardiac hypertrophy. Although the reduction in cardiac hypertrophy could be due to a reduction in afterload and/or in cardiac sympathetic tone, additional studies are necessary to elucidate the mechanism by which kefir administration attenuates the cardiac hypertrophy in this model of hypertension.

Considering that, the imbalance of autonomic nervous system markedly influences the HR and BP and that it has been associated with targeted-organ damage and increased risk of morbi-mortality (Abboud, 2010; Abboud et al., 2012), we assessed the effects of kefir on the cardiac parasympathetic (vagal) tone and cardiac and vascular sympathetic tones in SHR by using two different experimental approaches. First, the parasympathetic cardiovagal and the cardio-vaso-sympathetic tones were evaluated through the classic invasive method illustrated in **Figure 3**. The increase in HR after administering methylatropine, a muscarinic cholinergic receptor blocker, reflects the cardiovagal tone present under baseline resting conditions, and the decrease in HR after atenolol administration, a cardiac β1-blocker, reflects cardiac sympathetic tone; a double blockade enables the determination of the intrinsic HR. The imbalance between the cardiac vagal activity (decreased) and sympathetic activity (augmented) in SHR is in agreement with other studies (Barbosa Neto et al., 2013). The novelty of the present study is the finding that chronic administration of kefir caused a significant attenuation of the cardiac autonomic imbalance of a similar magnitude to that obtained by physical exercise (Barbosa Neto et al., 2013), phototherapy (Monteiro et al., 2012), and pharmacological medication (Dias da Silva et al., 2002).

It has been shown that a disruption in the balance between vagal and sympathetic tones can lead to an impairment in baroreflex sensitivity, as has been demonstrated in different models of arterial hypertension (Moyses et al., 1994; Campagnaro et al., 2012), including the SHR (Dias da Silva et al., 2002; Monteiro et al., 2012). Therefore, the effect of kefir on the impaired baroreflex that characterizes SHR was tested in a separated set of experiments using the classical invasive pharmacological maneuvers by which reflex bradycardia during phenylephrine-induced increases in BP and reflex tachycardia during sodium nitroprusside-induced decreases an in BP are analyzed (Moyses et al., 1994; Campagnaro et al., 2012).

The present study is the first to demonstrate that Kefir treatment for 60 days ameliorated the impaired baroreflex sensitivity observed in SHR. When the procedure was repeated under the blockade action of methylatropine the reflex bradycardia and baroreflex gain was abolished in the three groups of animals, indicating that the impaired baroreflex during

increases in BP was solely due to the vagal component. On the other hand, when the BP was lowered under the action of the blocker atenolol, the reflex tachycardia was markedly reduced, but this reduction was to a much lesser extent in the Wistar group. Therefore, our data demonstrate that reflex tachycardia is a consequence of both the activation of the cardiac sympathetic component and the simultaneous withdrawn of the parasympathetic activity, which is in agreement with others (Moyses et al., 1994; Campagnaro et al., 2012), and favors the notion that physiological interventions always elicits reciprocal changes in sympathetic and parasympathetic nerve activities. Thus, the tachycardia in response to falls in BP observed could, at least in part, be due to a concurrent vagal activity withdrawal in the Wistar group and possibly in the SHR because the vagal tone was impaired and consequently less effective. In agreement with the review of Head (1994), we observed that the diminished sensitivity of the baroreflex in SHR (characterized by hypertension and cardiac hypertrophy) is mainly due to reduced capacity of the cardiac vagal component rather than a change of the sympathetic (our data: −35 and −23%, respectively). Our data show that kefir treatment of SHR was able to attenuate the impaired baroreflex sensitivity by restoring part of the cardiac autonomic activity. A possible mechanism for the beneficial effect of kefir on the vagal-mediated baroreflex sensitivity in this model of hypertension could be through the reduction of the cardiac hypertrophy as observed by using different probiotics (Lin et al., 2012; Gan et al., 2014; Gómez-Guzmán et al., 2015). Therefore, in addition to other non-pharmacological therapies, such as physical exercise (Ceroni et al., 2009), the chronic administration of the nutritional probiotic kefir appears to have beneficial actions targeting the restoration of the balance between the vagal and sympathetic activities and improving the baroreflex function in hypertensive subjects.

The influence of the autonomic nervous system on the BP and PI is frequency-dependent and the power spectral analysis of beat-to-beat BP recordings has been increasingly used as a valuable non-invasive tool to evaluate the autonomic nervous system modulation on HR and BP in both experimental animals (Murphy et al., 1991; Ceroni et al., 2009; Silva et al., 2009; Hayward et al., 2012; Barbosa Neto et al., 2013) and humans (Silva et al., 2009). We used, in a separated set of experiments, spectral analysis to additionally assess the relative contribution of the sympathetic and vagal tones and their contribution in the reflex activities on conscious SHR chronically treated with kefir when compared with non-treated SHR and normotensive Wistar rats.

In the present study, we observed that kefir significantly reduced the increased cardiac sympathetic tone, which was evaluated through pharmacological blockade and that it reduced the HR variability (or the LF spectra power of PI) in the SHR. It has been demonstrated that LF spectral power of HR is modulated by the autonomic nervous system (Parati et al., 1995; Stauss, 2007; Silva et al., 2009; Chapleau and Sabharwal, 2011). Therefore, a plausible mechanism by which this probiotic partly restores the normal cardiac sympathetic tone and the PI variability may include a direct or indirect inhibition of the sympathetic nerve transmission at its origin in the central nervous system or at the innervation site in the heart. This speculation is corroborated by the finding that the ganglionic blockade (Diedrich et al., 2002) and cardiac muscarinic blockade reduce HR variability (Médigue et al., 2001). Specifically, an important mechanism of action of kefir at the autonomic nervous system could include a decrease in the production of reactive oxygen species and increase of nitric oxide availability, linking the present results to the hypotensive effects of kefir in this genetic model of hypertension, recently reported by Friques et al. (2015).

The PI-LF and BP-LF components, which have been considered markers of cardiac and vascular sympathetic neural activity (Parati et al., 1995; Task Force, 1996), were significantly augmented in SHR when compared with Wistar rats. This finding is consistent with other results (Ceroni et al., 2009; Silva et al., 2009; Hayward et al., 2012; Barbosa Neto et al., 2013). However, the use of LF as an estimative of pure sympathetic activity to the heart has been challenged (Reyes del Paso et al., 2013) and considered a reflection of sympathetic and parasympathetic nerve activities. Increased BP variability is closely correlated to end-organ damage in normotensive and hypertensive animals, and the pharmacologic reduction of BP variability attenuates hypertension outcomes to target-organs (Su and Miao, 2001, 2005). Additionally, our data showed that kefir administered for 60 days to SHR partially corrected the abnormal LF component, suggesting that it could become a nonpharmacological alternative for prevention/treatment of organdamages associated with increased BP and PI variabilities.

The PI-HF component, which has often been considered a marker of cardiac vagal activity (Dantas et al., 2010), was similar in the Wistar and SHR, consistent with other findings (Ceroni et al., 2009; Barbosa Neto et al., 2013). However, this component does not seem to be solely attributed to cardiac vagal activity, it is also influenced by respiration (Chapleau and Sabharwal, 2011; Reyes del Paso et al., 2013). Additionally, no significant differences were observed among the three groups when we used the LF-to-HF ratio to estimate the sympathovagal balance (Kuwahara et al., 1994; Parati et al., 1995; Task Force, 1996). Kefir treatment did not altered the above parameters. However, the use of the LF/HF ratio as a measure of sympathovagal balance can be misleading (Chapleau and Sabharwal, 2011). The very low frequency (VLF) component was not analyzed in the present study because its physiological correlates are still not clear.

The ratio between the absolute powers of PI-LF and BP-LF, the LF α-index, which has shown to express the spontaneous baroreflex sensitivity (Silva et al., 2009; Dantas et al., 2010, 2012), was diminished in the SHR, which was consistent with other studies (Silva et al., 2009) and the treatment with kefir restored the normal values. As described by Fazan et al. (2005), baroreflex modulation of HR contributes to LF, but not HF variability and is mediated by both sympathetic and parasympathetic drives. Therefore, increased BP variability, which is closely correlated with end-organ damage in hypertensive conditions (Su and Miao, 2001, 2005), can be ameliorated with pharmacological treatment (Su and Miao, 2001; Dias da Silva et al., 2002) and with nonpharmacological interventions, such as with physical exercise (Ceroni et al., 2009) and treatment with flavonoids (Monteiro et al., 2012). Additionally, we demonstrated for the first time that increased BP variability could also account (alternatively or additionally) for beneficial effects of the probiotic kefir, reducing BP variability attenuates end-organ damage. However, we may take into account that, different from the classical pharmacologic approach, the use of variabilities give us only an indirect estimation of the cardiac sympathetic and vagal tone.

The effects of hypertension in target organs could be associated with impaired autonomic function due to increased cerebrovascular oxidative stress and inflammation (Toth et al., 2013). Although our data do not allow us to state precise mechanisms, we speculate that the imbalance between pro- and anti-inflammatory cytokines and between pro- and anti-oxidant molecules could also occur in central areas of neural control of the cardiovascular system, as based on an extrapolation of data from others showing that the hypothalamic paraventricular nucleus (PVN) contributes to sympathoexcitation, hypertension and cardiac hypertrophy in the SHR (Jia et al., 2014). Because kefir is a probiotic beverage with anti-inflammatory (Carasi et al., 2015) and anti-oxidant (Friques et al., 2015) properties, we suggest that the findings of the present study may possible be due to a decreased production of cytokines and reactive oxygen species in the hypothalamic PVN which attenuates hypertension and end-organ damage by up-regulating anti-inflammatory and anti-oxidant molecules, restoring the normal balance (Tan et al., 2015).

In conclusion, the contribution of the cardiac sympathetic and parasympathetic tones and their relative contribution to the baroreflex activities were assessed through classical pharmacologic approaches and spectral analysis. The novelty of this study is that daily administration of kefir for 60 days partly corrects the alterations in cardiac function (including autonomic tone, baroreflex sensitivity) and PI and BP variability in SHR. Therefore, increased BP variability, which is closely correlated with end-organ damage in hypertensive conditions, now could be ameliorate with the beneficial effects of the probiotic kefir, in addition to pharmacological and non-pharmacological therapies.

## LIMITATIONS AND FUTURE DIRECTIONS

A limitation of the present study is that measurement of changes in HR and BP in the protocols we have designed are not sufficient to identify the mechanism of action by which kefir attenuates the abnormalities in the autonomic control of the heart and vessels. Therefore, the present data provide new insights in the prevention/treatment of abnormal vagal and sympathetic control of heart and vessels and opens new avenues for searching the mechanisms involved in the beneficial effects of kefir of the heart and vessels.

## AUTHOR CONTRIBUTIONS

Preparation and administration of kefir: LD and AF; Microbiological analysis of kefir: TA; Hemodynamic measurements: BK and ML; Acquisition and analysis of data from cardiac tones and baroreceptor reflex sensitivity: LD, BK, and AG; Spectral analysis: ED and RD; Critical revision of the manuscript: SM and TP; Conception, study's design, supervision and co-supervision of the study: EV and BC. All authors read and approved the final version of the manuscript.

### ACKNOWLEDGMENTS

EV is supported by CNPq (Ref. 476525/2012-8 and 303001/2015- 1) and Fapes (Grant Universal 2014 Proc 67597482). SM is supported by CNPq (307584/2015-1). BC is supported by CNPq (445736/2014-3). TP is supported by CNPq (445080/2014-0). TA is supported by FAPES/CNPq/MS-Decit/SESA (10-2013-PPSUS-Ref 65835131).

## REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Klippel, Duemke, Leal, Friques, Dantas, Dalvi, Gava, Pereira, Andrade, Meyrelles, Campagnaro and Vasquez. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# C-Type Natriuretic Peptide Induces Anti-contractile Effect Dependent on Nitric Oxide, Oxidative Stress, and NPR-B Activation in Sepsis

Laena Pernomian<sup>1</sup> , Alejandro F. Prado<sup>1</sup> , Bruno R. Silva<sup>2</sup> , Aline Azevedo<sup>3</sup> , Lucas C. Pinheiro<sup>1</sup> , José E. Tanus-Santos <sup>1</sup> and Lusiane M. Bendhack <sup>2</sup> \*

*<sup>1</sup> Department of Pharmacology, School of Medicine of Ribeirão Preto (FMRP), University of São Paulo, Ribeirão Preto, Brazil, <sup>2</sup> Department of Physics and Chemistry, Faculty of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Ribeirão Preto, Brazil, <sup>3</sup> Department of Biomechanics, School of Medicine of Ribeirão Preto (FMRP), Medicine and Rehabilitation of the Locomotor System, University of São Paulo, Ribeirão Preto, Brazil*

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Maria Socorro França-Silva, Federal University of Paraíba, Brazil Marcio R. V. Santos, Federal University of Sergipe, Brazil*

> \*Correspondence: *Lusiane M. Bendhack bendhack@usp.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *16 March 2016* Accepted: *30 May 2016* Published: *23 June 2016*

### Citation:

*Pernomian L, Prado AF, Silva BR, Azevedo A, Pinheiro LC, Tanus-Santos JE and Bendhack LM (2016) C-Type Natriuretic Peptide Induces Anti-contractile Effect Dependent on Nitric Oxide, Oxidative Stress, and NPR-B Activation in Sepsis. Front. Physiol. 7:226. doi: 10.3389/fphys.2016.00226* Aims: To evaluate the role of nitric oxide, reactive oxygen species (ROS), and natriuretic peptide receptor-B activation in C-type natriuretic peptide-anti-contractile effect on Phenylephrine-induced contraction in aorta isolated from septic rats.

Methods and Results: Cecal ligation and puncture (CLP) surgery was used to induce sepsis in male rats. Vascular reactivity was conducted in rat aorta and resistance mesenteric artery (RMA). Measurement of survival rate, mean arterial pressure (MAP), plasma nitric oxide, specific protein expression, and localization were evaluated. Septic rats had a survival rate about 37% at 4 h after the surgery, and these rats presented hypotension compared to control-operated (Sham) rats. Phenylephrine-induced contraction was decreased in sepsis. C-type natriuretic peptide (CNP) induced anti-contractile effect in aortas. Plasma nitric oxide was increased in sepsis. Nitric oxide-synthase but not natriuretic peptide receptor-B expression was increased in septic rat aortas. C-type natriuretic peptide-anti-contractile effect was dependent on nitric oxide-synthase, ROS, and natriuretic peptide receptor-B activation. Natriuretic peptide receptor-C, protein kinase-Cα mRNA, and basal nicotinamide adenine dinucleotide phosphate (NADPH)-dependent ROS production were lower in septic rats. Phenylephrine and CNP enhanced ROS production. However, stimulated ROS production was low in sepsis.

Conclusion: CNP induced anti-contractile effect on Phenylephrine contraction in aortas from Sham and septic rats that was dependent on nitric oxide-synthase, ROS, and natriuretic peptide receptor-B activation.

Keywords: C-type natriuretic peptide, sepsis, natriuretic peptide receptor B, reactive oxygen species, nitric oxide, phenylephrine-induced contraction

## INTRODUCTION

One of the most important clinical characteristics of sepsis and septic shock is the vascular hyporesponsivity to contractile agonists (Donaldson and Myers, 1996; Vromen et al., 1996; Strunk et al., 2001). It represents an important condition for patient survival. Different therapeutic strategies aimed to improve vital organ function (Leone and Martin, 2008). The identification of new intracellular signaling related to sepsis progression might contribute to the development of therapeutic strategies to reduce sepsis-associated mortality.

Sepsis cardiovascular dysfunction involves excessive nitric oxide (NO) production (Strunk et al., 2001) by NO-synthases (NOS) (Thomas et al., 2008). The endothelial NOS (NOS3) plays main role on organ blood flow distribution and is related to microvascular permeability regulation and cell interaction. Sand et al. (2015) showed that mesenteric blood flow decreases in a time-dependent manner and in parallel with the development of metabolic acidosis and organ dysfunction after induction of sepsis. Inducible NOS (NOS2) is constitutively expressed on few quantities and its protein expression is enhanced after inflammatory stimuli in different cells (Strunk et al., 2001). Sustained NO production following NOS2 induction is associated with sepsis hyporesponsivity (Donaldson and Myers, 1996; Vromen et al., 1996; Strunk et al., 2001). Furthermore, the role of neuronal NOS (or NOS1) seems to contribute to the hyporesponsivity of contractile adrenergic agonists in sepsis (Nardi et al., 2014). The inhibition of soluble guanylyl cyclase (sGC), enzyme which is the main target to NO, is able to recover mean arterial pressure (MAP) and enhance cardiac contractility in septic patients (Fernandes et al., 2009).

The oxidative stress is associated with impaired vasoconstriction in sepsis (Szabó et al., 1995; Salvemini and Cuzzocrea, 2002; Wu et al., 2004). According to Wu et al. (2004), antioxidant treatment, before cecal ligation and puncture (CLP) surgery to induce sepsis, increases mice survival, and decreases hypotension, plasma NO metabolites, oxidative stress, NOS2 mRNA, and angiotensin II (AngII) hyporesponsivity. Regarding to endothelial cells, Li et al. (2016) showed that lipopolysaccharide (LPS) treatment of human umbilical vein endothelial cells (HUVEC) increases oxidative stress, malondialdehyde levels, superoxide dismutase 2 (SOD2) protein expression and phosphorylation of c-Jun N-terminal kinases (JNK), and decreases SOD1 expression. Moreover, Zhou et al. (2012) observed that intravenous bolus injection of ascorbate, 30 min prior to CLP surgery or 3 h after CLP surgery, attenuated NOS activity and prevented vascular permeability increases.

The C-type natriuretic peptide (CNP) is a member of the natriuretic peptide family with high conserved amino acid sequence (Inoue et al., 2003). It can be synthesized, stored, and released from endothelial cells. CNP is present in LPS-, cytokines-, and growth factors-stimulated endothelial cells (Suga et al., 1993). However, CNP remains low on the plasma concentration in healthy subjects (Stingo et al., 1992). CNP is an endothelium-derived vasodilator agent (Suga et al., 1992), it induces vasodilation in an endothelium-dependent manner (Amin et al., 1996; Andrade et al., 2014) and it can hyperpolarize plasma membrane of vascular smooth muscle (VSM) cell through natriuretic peptide receptor (NPR)-C activation (Chauhan et al., 2003; Garcha and Hughes, 2006). Natriuretic peptides, i.e., atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), and CNP are able to induce intravascular shedding of glycocalyx and increase vascular permeability (Jacob et al., 2013). Furthermore, the authors suggest that this condition might be present in the pathophysiology of sepsis.

ANP, BNP (Piechota et al., 2009), and CNP (Hama et al., 1994) are elevated on plasma samples from septic subjects. Until now, three different NPR were characterized: NPR-A, NPR-B, and NPR-C (Kone, 2001). ANP and BNP present affinity for NPR-A, whereas CNP is the endogenous agonist for NPR-B. Both NPR-A and NPR-B are associated with particulate guanylyl cylclase (pGC), leading to the intracellular cyclic guanosinemonophosphate (cGMP) formation (Rautureau et al., 2010). ANP, BNP, and CNP have similar affinity for NPR-C, which is associated with G<sup>i</sup> protein, adenylyl cyclase inhibition, and phospholipase C-β (PLC-β) activation (Matsukawa et al., 1999), as well as activation of NOS induced by calcium increase rather than NOS phosphorylation (Murthy et al., 1998; Costa et al., 2007; Chen et al., 2008; Andrade et al., 2014).

In VSM cells stimulated with LPS or tumor necrosis factorα (TNF-α), natriuretic peptides increase nitrite formation while the NOS inhibitor reduces CNP responses, suggesting the involvement of NO on CNP effects (Marumo et al., 1995). NPR-A/B antagonist infusion attenuated the hypotension in a sepsis model (Hinder et al., 1997). According to Panayiotou et al. (2010), knockout mice to NPR-A exposed to LPS have low systemic hypotension, and lower levels of NO metabolites, cGMP levels and NOS2 expression compared to wild-type septic animals. Therefore, the contribution of natriuretic peptide system to the vascular dysfunction seems to be evident on sepsis. However, little is known about the role of CNP in this dysfunction and if it could act as a modulator of α1-adrenergic vascular contraction on animals septic vessels. The hypothesis of the present work was that CNP induces anti-contractile effect on phenylephrine (PE) contraction in septic vessels, which effect

**Abbreviations:** Ana, Anantin; AngII, Angiotensin II; ANP, Atrial natriuretic peptide; AUC, Area under the curve; BNP, Brain natriuretic peptide; cGMP, Cyclic guanosine-monophosphate; CLP, Cecal ligation and puncture; CNP, C-type natriuretic peptide; CO, Control; DAG, Diacylglycerol; DPI, Diphenylene iodonium chloride; DTPA, Diethylene triaminepentaacetic acid; FI, Fluorescence intensity; FITC, Fluorescein isothiocyanate; g.g−<sup>1</sup> , Gram of contraction per gram of dry tissue weight; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GK, cGMP-dependent kinase; H2O2, Hydrogen peroxide; HRP, Horseradish peroxidase; HUVEC, Human umbilical vein endothelial cells; IL, Interleukin; IP3, Inositol-trisphosphate; JNK, c-Jun N-terminal kinases; L-NAME, Nω-nitro-Larginine methyl ester hydrochloride; LPS, Lipopolysaccharide; MAP, Mean arterial pressure; ME, Maximum effect; mmHg, Milimeters of mercury; mN.mm−<sup>1</sup> , Milinewton per milimeter; NADPH, Nicotinamide adenine dinucleotide phosphate; NaOH, Sodium hydroxide; NEM, N-ethylmaleimide; NO, Nitric oxide; NOS, Nitric oxide synthase; NOx, nitrite + nitrate; Nox1, NADPH oxidase 1; NPR, Natriuretic peptide receptor; O<sup>−</sup> 2 , Superoxide anion; ONOO−, Peroxynitrite; PE, Phenylephrine; PEG, Polyethylene glycol; PKC, Protein kinase C; PLC, Phospholipase C; qPCR, Quantitative polymerase chain reaction; RLU, Relative luminescence unit; RMA, Resistance mesenteric arteries; ROS, Reactive oxygen species; Sham, Sham-operated rats; SOD, Superoxide dismutase; pGC, Particulate guanylyl cyclase; sGC, Soluble guanylyl cyclase; TNF, Tumor necrosis factor; U, Units; VSM, Vascular smooth muscle.

would be dependent of NO and reactive oxygen species (ROS). This study aimed to evaluate the contribution of intracellular NO and ROS, and the role of activation of NPR-B in the CNP-anticontractile effect on PE-contraction in aortas isolated from rats submitted to CLP surgery to induce sepsis.

### METHODS

### Animals

All the procedures were performed in accordance with the standards and policies of the Ethics Committee on Animal Care and Use of the University of São Paulo (#144/2011). Male rats (200 g) were anesthetized with Tribromoethanol (0.25 g.Kg−<sup>1</sup> , i.p.) before cathetherization surgery on femoral artery. On the next day, the MAP was measured for basal values (before surgery), using pressure transducer. Then, the rats were anesthetized with Tribromoethanol (0.25 g.Kg−<sup>1</sup> , i.p.) and randomly submitted to CLP surgery to induce sepsis, with 12 punctures using a 16 gauge-needle on anti-mesenteric border of the cecum (Wichterman et al., 1980; Araújo et al., 2011), or the control group was submitted to a medial laparotomy only (Sham). Survival and MAP were evaluated for 24 h after surgeries. Animals were maintained on standard rat chow and water.

### Functional Studies by Vascular Reactivity

Sham and CLP rats were killed by decapitation under anesthesia (inhaled Isoflurane) and thoracic aorta and resistance mesenteric artery (RMA) were used to measure the isometric tension.

### Vascular Reactivity Studies on Aorta

Aorta was cut into rings (4 mm) and placed between two stainless-steel stirrups and connected to an isometric force transducer to measure tension. Cumulative concentration-effect curves were constructed to PE (0.1 nmol.L−1−10 µmol.L−<sup>1</sup> ) in intact endothelium rat aortas. These curves were obtained in the absence (CO) or after 30 min incubation with CNP (10 nmol.L−<sup>1</sup> ), Nω-nitro-L-arginine methyl ester hydrochloride (L-NAME, 100 µmol.L−<sup>1</sup> ), Nω-propyl-L-arginine (50 nmol.L−<sup>1</sup> ), 1400 W (10 nmol.L−<sup>1</sup> ), Tiron (100 µmol.L−<sup>1</sup> ), polyethylene glycol-Catalase (PEG-Catalase, 250 U.mL−<sup>1</sup> ), Anantin (Ana, 0.1 µmol.L−<sup>1</sup> or 1 µmol.L−<sup>1</sup> ) alone or in the combination of each inhibitor/antagonist and CNP. The PE potency (pD2) and contractile maximum effect (ME) were evaluated.

### Vascular Reactivity on RMA

RMA were isolated and cut into rings (2 mm) using a dissection magnifier in cold Krebs solution for small arteries. Only the second branch of RMA with 200–350 µm of internal circumference were used. A tungsten wire (40 µm) was passed into the vessel and tied in myograph to resistance vessels and another was connected to isometric force transducer. Cumulative concentration-effect curves were constructed to PE (0.1 nmol.L−1−100 µmol.L−<sup>1</sup> ) in intact endothelium RMA in the absence or in the presence of CNP (10 nmol.L−<sup>1</sup> , 30 min). The area under the curve (AUC) of PE contraction was evaluated.

### Measurement of Plasma NO Metabolites

Cardiac blood samples were collected from Sham or CLP rats in heparin-containing tubes. Plasma aliquots of Sham or CLP rats were analyzed for their nitrite content using an ozone-based reductive chemiluminescence assay (Pinheiro et al., 2012). The plasma NOx concentration was determined by using the Griess reaction (Pinheiro et al., 2012).

### Western Blotting Analysis

Protein expression of NOS2, NOS3, NPR-B, or Nox1 were analyzed in Sham or CLP rat aorta homogenate. Membranes were incubated with mouse primary antibody anti-NOS2 (1:2500), or anti-NOS3 (1:2500), and rabbit primary antibody anti-NPR-B (1:5000), or anti-Nox1 (1:2000), overnight at 4◦C. Then, membranes were incubated with a HRP-conjugated goat antirabbit (1:5000) or goat anti-mouse secondary antibody (1:5000) for 60 min at room temperature. Protein bands were visualized by means of chemiluminescence. Protein expression levels were normalized by mouse anti-β-actin (1:2000).

### Confocal Microscopy Analysis

Aorta was isolated from Sham or CLP rats and frozen on cryoprotection liquid. Sheets were prepared with 10 µm of thickness and immunofluorescence was performed to evaluate endogenous CNP (1:100) or NPR-C (1:100) expression. The secondary antibody mouse anti-goat Alexafluor 647 (1:1000) or sheep anti-rabbit Alexafluor 647 (1:1000) was incubated for 1 h, at room temperature. Fluoroshield <sup>R</sup> with DAPI was applied overnight at 4◦C. Images were acquired using confocal microscopy.

### Lucigenin Chemiluminescence Analysis

Vascular nicotinamide adenine dinucleotide phosphate (NADPH)-dependent ROS production was assessed in intactendothelium aorta from Sham or CLP rats, previously stimulated with PE (0.1 µmol.L−<sup>1</sup> ) in the absence or in the presence of CNP (10 nmol.L−<sup>1</sup> , 30 min) or not stimulated (basal). After that, aortas were collected and frozen on liquid nitrogen. The luminescence was measured using Single Tube Luminometer Berthold FB12, at 37◦C. Data were presented as relative luminescence units (RLU).mg−<sup>1</sup> .min−<sup>1</sup> .

### Quantitative Polymerase Chain Reaction (qPCR) Analysis

Total ribonucleic acid (RNA) was extracted from aorta homogenates isolated from Sham or CLP rats with TRIzol <sup>R</sup> reagent according to manufacturer's instruction. The cDNA was synthesized with 1 µg of RNA using anchored random hexadeoxynucleotide in the reaction conditions for transcriptor First-strand cDNA Synthesis Kit. qPCR was performed using SYBR <sup>R</sup> FAST qPCR kit Master Mix Universal with 50 nmol.L−<sup>1</sup> of primers for PKCα (protein kinase-Cα) or GAPDH (glyceraldehyde-3-phosphate dehydrogenase). The samples were incubated for 2 min at 50◦C and 10 min at 95◦C followed by 40 cycles at 95◦C for 15 s and 60◦C for 1 min.

## Drugs and Solutions

Drugs and solutions are presented on Supplemental Material.

### Statistical Analysis

Results are presented as mean ± S.E.M. Each experimental n represents samples isolated from different animals. Comparisons between groups were conducted by the Student's t-test, Oneway analysis of variance followed by Newman-Keuls, or Twoway analysis of variance using Bonferroni correction for multiple comparisons, as appropriate. The level of statistical significance was defined as P < 0.05.

## RESULTS

### Survival and Mean Arterial Pressure Analysis

Sham rats survival rate was 100% at 24 h after surgery (n = 20). CLP rats had lower survival rate than Sham, and 4 h after the CLP surgery it was 37% of survival, reaching no longer than 2% of survival at 24 h (n = 30) (**Figure 1**). Basal MAP was similar in both groups (Sham: 96.7 ± 0.6 mmHg, n = 6; CLP: 96.1 ± 0.4 mmHg, n = 6). Four hours after the surgeries, MAP decreased in both groups which values were lower on CLP (51.4 ±0.9 mmHg, n = 6) than in Sham rats (86.2 ± 0.3 mmHg, n = 6) (**Figure 1**).

## PE Induced Contraction and CNP Induced Anti-contractile Effect in Aorta but Not in RMA

The ME and potency (pD2) to PE were lower on CLP (ME: 753 ± 47 g.g−<sup>1</sup> , pD2: 6.74 ± 0.14, n = 11) than in Sham aortas (ME: 1232 ± 30 g.g−<sup>1</sup> , pD2: 7.32 ± 0.10, n = 11) (**Figure 1**). Similarly, in RMA, PE-induced contraction was lower in CLP (4.73 ± 0.52 mN.mm−<sup>1</sup> ; n = 7) than in Sham (9.68 ± 1.30 mN.mm−<sup>1</sup> ; n = 6). The AUC of PE-induced contraction on RMA was also lower in CLP (9.55 ± 0.86) than in Sham (24.27 ± 2.75). CNP induced anti-contractile effect in aortas but not in RMA (**Figure 1**). CNP reduced the PE-induced contraction in Sham aortas (ME: 877 ± 25 g.g−<sup>1</sup> ; pD2: 6.68 ± 0.10; n = 8). However, only ME was decreased by CNP in CLP aortas (ME: 365 ± 74 g.g−<sup>1</sup> ). CNP anti-contractile effect was higher in CLP than Sham aortas.

### NPR-C was Decreased in CLP Compared to Sham Aortas

As shown in the photomicrographs (**Figure 2**) and in the measurement of fluorescence intensity (FI) emitted by anti-NPR-C antibody, the NPR-C expression was lower in CLP (4.35 ± 0.46 U, n = 6) compared to Sham aortas (7.81 ± 0.80 U, n = 6).

## Plasma NO and Protein Expression of NOS3 and NOS2 were Increased, and NPR-B was Not Changed in CLP Aortas

NOS3 protein expression in CLP aorta was greater (0.20 ± 0.02 U, n = 4) than in Sham aorta (0.10 ± 0.01, n = 4) (**Figure 3**). Similarly, NOS2 protein expression in CLP aorta was higher (0.06 ± 0.002 U, n = 4) than in Sham aorta (0.03 ± 0.001 U, n = 4). However, NPR-B protein expression was not different between Sham (0.79 ± 0.07 U, n = 4) and CLP (0.92 ± 0.06 U, n = 6). Plasma NO metabolites such as nitrite and nitrate were higher on CLP (nitrite: 4.16 ± 0.79 µmol.L−<sup>1</sup> , n = 4; nitrate: 69.12 ± 10.58 µmol.L−<sup>1</sup> , n = 7) than in Sham (nitrite: 0.69 ± 0.09 µmol.L−<sup>1</sup> , n = 4; nitrate: 21.97 ± 2.38 µmol.L−<sup>1</sup> , n = 7) rats.

## NPR-C was Lower and NPR-B was Not Changed in CLP in Relation to Sham Aortas

NPR-B protein expression was not different between Sham (0.79 ± 0.07 U, n = 4) and CLP (0.92 ± 0.06 U, n = 6), whereas NPR-C expression was lower in CLP (4.35 ± 0.46 U, n = 6) aortas as compared to Sham (7.81 ± 0.80 U, n = 6), as shown in **Figure 3**.

### CNP-Induced Anti-contractile Effect was Dependent on NOS Activity in Rat Aortas

PE-induced contraction was increased by the non-selective inhibition of NOS with L-NAME in Sham (2097 ± 97 g.g−<sup>1</sup> , n = 11) and CLP (1365 ± 116 g.g−<sup>1</sup> , n = 12) aortas. But it was still lower in CLP aortas (**Figure 4**). However, pD<sup>2</sup> value increased in CLP aortas (7.30 ± 0.09). CNP-anti-contractile effect on PEinduced contraction was reversed by L-NAME in both groups (Sham: 1876 ± 83 g.g−<sup>1</sup> , n = 11; CLP: 1936 ± 82 g.g−<sup>1</sup> , n = 12). L-NAME and CNP increased pD<sup>2</sup> values in Sham (7.58 ± 0.16) and in CLP (7.10 ± 0.28) aortas.

## NOS1 and NOS2 Also Contributes to CNP-Induced Anti-contractile Effect in Rat Aortas

The selective NOS1 inhibitor Nω-propyl did not modify the potency to PE in Sham (7.20 ± 0.10, n = 8) or CLP (7.05 ± 0.20, n = 7) aortas (**Figure 5**). This inhibitor increased ME to PE only in CLP aortas (1109 ± 79 g.g−<sup>1</sup> ), whereas this contraction remained unchanged in Sham (1349 ± 53 g.g−<sup>1</sup> ). CNP-anticontractile effect on PE-induced contraction was reversed by Nω-propyl in Sham (1342 ± 67 g.g−<sup>1</sup> , n = 9) or CLP (1060 ± 126 g.g−<sup>1</sup> , n = 5) aortas (**Figure 5**). PE-induced maximum contraction in CLP (1060 ± 126 g.g−<sup>1</sup> , n = 5) was lower than in Sham (1342 ± 67 g.g−<sup>1</sup> ) aortas in the presence of CNP and Nωpropyl. In addition, CNP-anti-contractile effect on PE potency was reversed in Sham (7.08 ± 0.09) and it was not modified in CLP (6.71 ± 0.21) aortas.

The selective NOS2 inhibitor 1400 W, increased PE pD<sup>2</sup> value in CLP (8.04 ± 0.34) and the ME to PE in Sham (1427 ± 65 g.g−<sup>1</sup> , n = 7) and CLP (954 ± 52 g.g−<sup>1</sup> , n = 7), which effect was lower in CLP than in Sham aortas (**Figure 5**). The double exposure CNP and 1400 W increased PE potency in Sham (7.19 ± 0.15, n = 8) and CLP aortas (7.63 ± 0.34, n = 6). The ME to PE was greater in CLP (789 ± 86 g.g−<sup>1</sup> ) after the double exposure to CNP and 1400 W, but not in Sham aortas (941 ± 57 g.g−<sup>1</sup> ).

## CNP-Induced Anti-contractile Effect was Dependent on ROS in Rat Aortas

The superoxide (O<sup>−</sup> 2 ) scavenger Tiron reduced the potency to PE in Sham (6.70 ± 0.15) but not in CLP (6.69 ± 0.20) aortas (**Figure 6**). However, ME to PE was enhanced on CLP (1034 ± 70 g.g−<sup>1</sup> ) aortas. Moreover, CNP-anti-contractile effect on PE

FIGURE 2 | NPR-C staining in vascular layers in Sham or CLP rat aortas. Photomicrograph of immunofluorescence of NPR-C staining in Sham (*n* = 6) or CLP (*n* = 6) rat aortas. DIC represents dichroic contrast phase, DAPI represents nuclei staining, α-actin from α-actin smooth muscle-FITC staining. Bar represents 50 µm. *L*, *M,* and *Adv* represent lumen, media layer, and adventitia, respectively. Each experimental *n* represents samples isolated from different animals.

rat aorta. Maximum effect values (D) (g.g−1) of PE curves of Sham or CLP rat aorta. Data are presented as mean ± S.E.M. \*different from Sham CO; #different from CLP CO; <sup>a</sup>different from Sham CNP; <sup>c</sup>different from CLP CNP. Each experimental *n* represents samples isolated from different animals. Two-way Anova, Bonferroni correction to evaluate interaction factor; One-way Anova, *pos-hoc* Newman-Keuls to three or more comparisons (*P* < 0.05).

contraction was reversed by Tiron on pD<sup>2</sup> values (Sham: 7.18 ± 0.10, n = 6; CLP: 7.01 ± 0.13, n = 13) and ME (Sham: 1822 ± 101 g.g−<sup>1</sup> , n = 6; CLP: 1328 ± 104 g.g−<sup>1</sup> , n = 13). Although ME to PE in CLP aortas in the presence of CNP and Tiron induced a greater value compared to CNP alone, this value remained lower than that one on Sham aortas.

The intracellular breakdown of hydrogen peroxide (H2O2) with PEG-Catalase did not change the contractile response to PE in both groups (**Figure 6**). However, the presence of PEG-Catalase and CNP reversed CNP-anti-contractile effect on PE potency (Sham: 7.07 ± 0.10, n = 5; CLP: 7.20 ± 0.25, n = 11) and ME (Sham: 1199 ± 60 g.g−<sup>1</sup> , n = 5; CLP: 970 ± 80 g.g−<sup>1</sup> , n = 11).

### CNP-Induced Anti-Contractile Effect was Dependent on NPR-B Activation and PE Contraction was Modulated by NPR-B in Rat Aortas

The non-selective NPR-A/B antagonist Anantin (0.1 µmol.L−<sup>1</sup> ) did not modify the PE pD<sup>2</sup> values (Sham: 7.23 ± 0.23, n = 8; CLP: 7.05 ± 0.19, n = 13). However, it increased ME to PE in CLP (1093 ± 107 g.g−<sup>1</sup> ) but not in Sham (1198 ± 114 g.g−<sup>1</sup> ) aorta (**Figure 7**). CNP-anti-contractile effect was reversed in Sham and CLP (pD2: Sham 7.50 ± 0.10; CLP: 7.35 ± 0.30; ME: Sham 1345 ± 115 g.g−<sup>1</sup> , n = 8; CLP: 891 ± 86 g.g−<sup>1</sup> , n = 13) aortas. When Anantin was used at the concentration of 1 µmol.L−<sup>1</sup> , PEinduced contraction was potentiated in both groups (Sham: 8.49 ± 0.30, n = 5; CLP: 8.06 ± 0.80, n = 11), but it did not change ME to PE. In addition, the PE pD<sup>2</sup> values in the presence of 1 µmol.L−1Anantin were greater than the values observed with 0.1 µmol.L−1Anantin in Sham.

### Endogenous Expression of CNP was Lower in Endothelium and Higher in VSM Layer in CLP Than in Sham Rat Aortas

The endothelial expression of endogenous CNP, as shown in the photomicrographs (**Figure 8**) and the measurement of FI of anti-CNP antibody (**Figure 9**), was lower in CLP (6.50 ± 0.66 U, n = 6) than in Sham aortas (9.78 ± 0.50 U, n = 4). However, CNP expression on VSM was higher in CLP (9.72 ± 0.60 U, n = 6) compared to Sham aortas (7.32 ± 0.11 U, n = 4). It was not different in the adventitia isolated from Sham (13.35 ± 1.00 U, n = 4) and CLP aortas (15.02 ± 1.20 U, n = 6). It can be noted that endogenous CNP is higher in the adventitia than in the endothelium which is higher than in the VSM in Sham. CNP is higher in adventitia than in the VSM that is higher than in the endothelium in CLP aortas.

### PE and CNP Enhanced NADPH-Dependent ROS Production

Nox1 expression was higher in CLP (1.61 ± 0.17 U, n = 4) than in Sham aortas (0.73 ± 0.08 U, n = 4). Basal NADPHdependent ROS production was lower in CLP (36.95 ± 3.27 RLU.mg−<sup>1</sup> .min−<sup>1</sup> , n = 3) than in Sham aortas (204.07 ± 20.61 RLU.mg−<sup>1</sup> .min−<sup>1</sup> , n = 5). PE increased NADPH-dependent ROS production in Sham (303.69 ± 30.78 RLU.mg−<sup>1</sup> .min−<sup>1</sup> , n = 8) and CLP aortas (145.66 ± 14.05 RLU.mg−<sup>1</sup> .min−<sup>1</sup> , n = 5). CNP and PE further increased NADPH-dependent ROS production

\*different from Sham CO; #different from CLP CO; <sup>a</sup>different from Sham CNP; <sup>c</sup>different from CLP CNP; <sup>e</sup>different from Sham Ana. Each experimental *n* represents samples isolated from different animals. Two-way Anova, Bonferroni correction to evaluate interaction factor; One-way Anova, *pos-hoc* Newman-Keuls to three or more comparisons (*P* < 0.05).

in Sham (690.34 ± 60.05 RLU.mg−<sup>1</sup> .min−<sup>1</sup> , n = 6) and CLP rat aortas (382.78 ± 32.45 RLU.mg−<sup>1</sup> .min−<sup>1</sup> , n = 5). However, all the values obtained in CLP were lower than in Sham. The non-selective flavoproteins inhibitor DPI, which was used as a negative control, inhibited the basal NADPH-dependent ROS production in Sham rat aorta (−32.44 ± 2.91 RLU.mg−<sup>1</sup> .min−<sup>1</sup> , n = 3) (**Figure 9**).

## PKCα mRNA Expression was Lower in CLP Than in Sham Rat Aortas

The relative PKCα mRNA expression (**Figure 9**) was lower in CLP (0.43 ± 0.13, n = 4) than in Sham (1.55 ± 0.31, n = 4) aortas.

## DISCUSSION

The main findings of this work were: (i) CNP induced anticontractile effect on PE contraction in rat aortas but not in rat RMA, (ii) this CNP anti-contractile effect was dependent on NOS activity, ROS production and NPR-B activation in Sham and CLP aortas, (iii) endogenous CNP was reduced on endothelium and increased on the VSM cells in CLP aortas, (iv) NPR-C expression was low on VSM layer of CLP aortas, (v) basal NPR-B activity was able to modulate PE-induced vasoconstriction, and (vi) basal NADPH-dependent ROS production was low in CLP aortas, and PE and CNP increased ROS production. It is the first report that CNP can modulate PE contraction, which effect was higher in CLP than Sham aortas.

CLP experimental sepsis model represents clinical conditions observed on patients with intestinal puncture, developing severe sepsis within few hours or some days (Lundblad et al., 1995). Moreover, it was observed that CLP group had a pronounced systemic hypotension and low survival rate, corroborating with findings from Tyml et al. (1998), Liaw et al. (2005), and Tsao et al. (2010). However, we have shown that Sham rats presented lower hypotensive effect than CLP group. According to Jong et al. (2002) general anesthetics can change cardiovascular parameters. Therefore, in our study, the decreased MAP observed on Sham rats could be due to tribromoethanol effect. Besides systemic hypotension showed on CLP rats, it was noted that PE-induced contraction on aorta and RMA was lower in CLP than in Sham rats.

In accordance to Vromen et al. (1996) and Wang et al. (2004), in CLP rat aortas there is low vasoconstriction induced by norepinephrine. Silva-Santos et al. (2002) noted LPS-exposure of aortas during 12 h, at isolated organ bath, reduced contraction to PE. Usually, such reduced contraction to different agonists have been associated with larger NO production by NOS1 and/or NOS2 (Vromen et al., 1996; Silva-Santos et al., 2002; Wang et al., 2004; Enkhbaatar et al., 2009; Tsao et al., 2010; Nardi et al., 2014), oxidative stress (Tsao et al., 2010), and greater inflammatory mediators production (Panayiotou et al., 2010). Enkhbaatar et al. (2009) showed that specific NOS1 inhibitor, called as ZK234238, attenuated hypotension and the fall in systemic vascular resistance, inhibited the increase in plasma nitrate/nitrite, and attenuated inflammatory markers such as myeloperoxidase activity, IL (interleukin)-6 mRNA, and reactive

nitrogen species. Furthermore, Nardi et al. (2014) demonstrated that the treatment with 7-nitroindazole, another NOS1 inhibitor, reverted the hyporesponsivity to PE and norepinephrine, and reduced the production of cGMP stimulated by PE in CLP rat aortas. Our results are in agreement with the others and it was noted raise on plasma NO levels, as nitrite and nitrate, and raise on protein expression of NOS2 and NOS3 in CLP rat aortas. Moreover, the non-selective NOS inhibitor L-NAME, or the selective NOS1 inhibitor Nω-propyl, or the selective NOS2 inhibitor 1400 W, increased PE contraction in both groups; but PE-induced maximum contraction was still lower in CLP aortas, suggesting additional mechanisms which are responsible for PE hyporresponsivity.

CNP-induced anti-contractile effect was greater in CLP than in Sham aortas. However, we did not observe CNP-induced modulation of PE contraction in RMA from both groups, using a single concentration of CNP in these studies. Villar et al. (2007) showed CNP-induced relaxation in RMA isolated from rats that was dependent on NPR-C activation. Besides, this group demonstrated all the three NPRs in rat aorta and RMA. We cannot exclude the possibility that higher CNP concentration could induce the anti-contractile effect on these resistance vessels, since the contraction elicited by PE needed higher concentration in RMA than in aorta. Since this CNP anti-contractile effect was observed on aortas at low concentration (10 nmol.L−<sup>1</sup> ), we decided to continue to investigate this vessel. CNP anticontractile effect was completely reversed by NOS inhibitors showing CNP could activate NOS and NO production could be responsible for CNP anti-contractile effect in aortas. In the same way, Murthy et al. (1998) had demonstrated that a selective agonist of NPR-C, called cANP(4-23), can increase NOS activity via Gi protein and calcium influx. Costa et al. (2007) reported that CNP increased NOS activity in aorta homogenate and our research group (Andrade et al., 2014) has described endothelium-dependent relaxation induced by CNP in rat aortas that seems to lead to NPR-C activation, but with no alteration in phosphorylation status of Ser<sup>1177</sup> of NOS3. As described by Anand-Srivastava et al. (1996), the activation of cytosolic domain of NPR-C leads to adenylyl cyclase inhibition via Pertussis toxin sensitive-G protein. Furthermore, Anand-Srivastava (2005) reviewed NPR-C intracellular signaling with inositol-trisphosphate (IP3) and diacylglycerol (DAG) formation, calcium mobilization, potassium channel, NOS, and PKC activation.

Although Nox1 protein expression was higher in CLP aortas, the basal activity of NADPH-dependent ROS production was low in CLP aortas, suggesting another source of ROS in CLP aortas. Moreover, Brandes et al. (1999) determined after LPS stimulus, that there are increased NADPH oxidase and xanthine oxidase expression in rat aortas, leading to O<sup>−</sup> 2 , H2O<sup>2</sup> production and after reaction with NO to form peroxynitrite (ONOO−). Since NADPH oxidase and xanthine oxidase produce O<sup>−</sup> 2 and <sup>H</sup>2O2, these results suggest the contribution of O<sup>−</sup> 2 to the contractile effect to PE in Sham and CLP rat aortas. However, when intracellular H2O<sup>2</sup> was removed, there were no changes on PE-induced contraction in both groups. It is possible that ROS production in VSM cells could compensate the removal of endothelial ROS on PE-induced contraction.

NADPH-dependent ROS production was increased after PE alone or CNP and PE stimuli, suggesting that CNP can positively modulate ROS production in rat aortas. In relation to the CNP anti-contractile effect in rat aortas, it seems dependent on H2O2, possibly through O<sup>−</sup> 2 dismutation by SOD. As presented by Anand-Srivastava (2005), CNP could stimulate PKC which in turn activates NADPH oxidase (Brandes et al., 2014). Thus, NADPH-dependent ROS production increased by CNP and PE could be due to NPR-C activation dependent on PKC. Jao et al. (2001) reported low protein expression of PKCα in rat model of sepsis. The reduced effect of PE or CNP and PE on NADPH-dependent ROS production could be due to reduced PKC expression and/or PKC deficient activity in CLP aortas. In fact, PKCα mRNA expression was lower in CLP aortas compared to Sham. Furthermore, protein expression of NPR-C on VSM cells from CLP aortas was lower compared to Sham. It might suggest a reduction on protein expression or activity of the NPR-C-PKCα-NADPH oxidase axis in CLP aortas.

Weber et al. (1991) identified Anantin as the first competitive antagonist of NPR associated with cGMP formation, i.e., NPR-A and NPR-B receptors. Then, Anantin has been used as non-selective NPR-A/B antagonist. As CNP is the endogenous agonist of NPR-B (Koller and Goeddel, 1992) and has low affinity for NPR-A, Anantin could be considered a NPR-B antagonist for CNP effects. We demonstrated that CNP anticontractile effect was dependent on NPR-B activation, since Anantin reversed this effect. Madhani et al. (2003) showed that CNP relaxation was inhibited by NPR-B antagonist in mice aortas. Rautureau et al. (2010) reported that CNP increased cGMP production independently of sGC in endothelial cells that was accompanied by NO production. Simon et al. (2009) showed CNP hyperpolarizes the endothelial cell membrane in a cGMP-dependent protein kinase (GK) manner. Activation of NPR-B by CNP was evident in Sham and CLP aortas. However, basal NPR-B activity without exogenous CNP was enough to modulate PE-induced vasoconstriction. Anantin at low concentration was able to increase PE maximum contraction only in CLP aortas. The greater effect of Anantin on high concentration suggests an important contribution of endogenous CNP for regulation of vascular tone. Besides, protein expression of NPR-B was not altered in aortas from Sham and CLP rats.

According to Hinder et al. (1997) and Stubbe et al. (2004) described that NPR-A/B antagonist preventing hypotension due to sepsis. This is the first mention that endogenous CNP or NPR-B intrinsic activity can contribute to modulation of vascular tone in aortas from Sham and CLP rats. Moreover, endogenous CNP production can be stimulated by several factors, such as LPS, TNF-α, IL-1α, and IL-1β, acetylcholine and bradykinin (Suga et al., 1993; Anand-Srivastava, 2005). Sepsis is a condition in which plasma levels of CNP are high (Hama et al., 1994). We demonstrated that endogenous CNP on vascular endothelium was lower in CLP than in Sham aortas. However, the peptide expression on VSM cells was greater in CLP than Sham aortas. It suggests a possible release of CNP from the endothelium to VSM in CLP aortas. Although the protein expression of NPR-B was similar in both groups, NPR-B activity might be different in CLP aortas.

Our results showed for the first time the CNP induced anticontractile effect on PE-induced contraction that was greater in CLP than Sham aortas. This effect is dependent on NOS activity, ROS production and NPR-B activation. O<sup>−</sup> 2 and H2O<sup>2</sup> production seems to be part of the same signaling pathway, and PE alone or CNP and PE increase ROS production. In CLP aortas, CNP could be released from the endothelial to VSM cells. NPR-B intrinsic activity or endogenous CNP could modulate the vascular tone. Furthermore, the lower protein expression of NPR-C on VSM cells of CLP aortas and lower relative PKCα mRNA expression could contribute to the reduction of CNPinduced NADPH-dependent ROS production in CLP aortas. If the activation of NOS and NPR-B belongs to a common intracellular signaling remains to be investigated. Moreover, it is evident the contribution of CNP system on vascular hyporeactivity to α1-adrenergic activation in CLP aorta, and it might represents a potential target for future pharmacological studies.

## AUTHOR CONTRIBUTIONS

Conception of the work: LP, BRS, LMB. Design of the work: LP, AFP, BRS, AA, LCP, JETS, LMB. Acquisition and analysis of data: LP, AFP, BRS, AA, LCP. Interpretation of data: LP, AFP, BRS, AA, LCP, JETS, LMB. Drafting or revising of the work: LP, AFP, BRS, AA, LCP, JETS, LMB. Final approval: LP, AFP, BRS, AA, LCP, JETS, LMB. Agreement to be accountable for all aspects of the work: LP, AFP, BRS, AA, LCP, JETS, LMB.

### FUNDING

This study was supported by the Fundação de Amparo a Pesquisa do Estado de São Paulo and Conselho Nacional de Desenvolvimento Científico e Tecnológico.

## ACKNOWLEDGMENTS

The authors are grateful to the Confocal Microscopy Multiuser Laboratory (LMMC, FAPESP #2004/08868-0) at School of Medicine from Ribeirão Preto – University of São Paulo (FMRP-USP).

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fphys. 2016.00226

### REFERENCES


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Pernomian, Prado, Silva, Azevedo, Pinheiro, Tanus-Santos and Bendhack. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Neuronal Nitric Oxide Synthase in Vascular Physiology and Diseases

Eduardo D. Costa<sup>1</sup> , Bruno A. Rezende1, 2, Steyner F. Cortes <sup>3</sup> and Virginia S. Lemos <sup>1</sup> \*

*<sup>1</sup> Department of Physiology and Biophysics, Institute of Biological Sciences, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil, <sup>2</sup> Department of Health Sciences, Post-graduate Institute, Medical Sciences College, Belo Horizonte, Brazil, <sup>3</sup> Department of Pharmacology, Institute of Biological Sciences, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil*

The family of nitric oxide synthases (NOS) has significant importance in various physiological mechanisms and is also involved in many pathological processes. Three NOS isoforms have been identified: neuronal NOS (nNOS or NOS 1), endothelial NOS (eNOS or NOS 3), and an inducible NOS (iNOS or NOS 2). Both nNOS and eNOS are constitutively expressed. Classically, eNOS is considered the main isoform involved in the control of the vascular function. However, more recent studies have shown that nNOS is present in the vascular endothelium and importantly contributes to the maintenance of the homeostasis of the cardiovascular system. In physiological conditions, besides nitric oxide (NO), nNOS also produces hydrogen peroxide (H2O2) and superoxide (O•−) 2 considered as key mediators in non-neuronal cells signaling. This mini-review highlights recent scientific releases on the role of nNOS in vascular homeostasis and cardiovascular disorders such as hypertension and atherosclerosis.

### Edited by:

*Camille M. Balarini, Federal University of Paraíba, Brazil*

### Reviewed by:

*Luciana Venturini Rossoni, University of São Paulo, Brazil Eliana Hiromi Akamine, University of São Paulo, Brazil*

> \*Correspondence: *Virginia S. Lemos vslemos@icb.ufmg.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *19 April 2016* Accepted: *20 May 2016* Published: *02 June 2016*

### Citation:

*Costa ED, Rezende BA, Cortes SF and Lemos VS (2016) Neuronal Nitric Oxide Synthase in Vascular Physiology and Diseases. Front. Physiol. 7:206. doi: 10.3389/fphys.2016.00206* Keywords: neuronal nitric oxide synthase, nitric oxide, hydrogen peroxide, vascular function, hypertension, atherosclerosis

## INTRODUCTION

Since the early 80s, nitric oxide (NO) is considered an essential endothelium-derived molecule, crucial to the maintenance of cardiovascular homeostasis (Furchgott and Zawadzki, 1980). Later on, it became evident that a decrease in the bioavailability of NO participated in several cardiovascular disorders such as atherosclerosis (Napoli et al., 2006) and hypertension (Hermann et al., 2006).

NO is biologically generated by a family of three nitric oxide synthase enzymes (NOS) isoforms: neuronal nitric oxide synthase (nNOS or NOS1), inducible nitric oxide synthase (iNOS or NOS2), and endothelial nitric oxide synthase (eNOS or NOS3). Although nNOS is abundantly expressed in neurons, and associated with the control of neuronal functions (Bredt et al., 1990; Bredt and Snyder, 1992) it is known that this isoform is also expressed in many non-neuronal cells such as in the endothelium and smooth muscle cells of several types of vessels in animals (Boulanger et al., 1998; Loesch et al., 1998; Schwarz et al., 1999) and human (Buchwalow et al., 2002). Recent studies show consistent evidence that this isoform exhibits relevant physiological role in the control of vascular homeostasis (Kurihara et al., 1998; Fleming, 2003; Hagioka et al., 2005; Seddon et al., 2008, 2009).

Besides NO, nNOS also produces H2O<sup>2</sup> in physiological conditions that contributes to endothelium-dependent vascular relaxation (Capettini et al., 2008, 2010). Impairment in endothelial nNOS-derived H2O<sup>2</sup> production has been implicated in the endothelial dysfunction in atherosclerosis (Rabelo et al., 2003; Capettini et al., 2011) and hypertension (Silva et al., 2016). Given the importance of nNOS in health and disease, this mini-review highlights recent scientific releases on the role of nNOS in vascular homeostasis and vascular mal functioning linked to hypertension and atherosclerosis.

### GENE EXPRESSION AND MOLECULAR STRUCTURE OF nNOS

nNOS gene is positioned on chromosome 12 (12q24.2) and distributed over a region greater than 200 kb in human genomic DNA (Hall et al., 1994). It consists of 4299 nucleotides encoding 1434 amino acids (Boissel et al., 1998). nNOS exists as a monomer/dimer mixture, being the dimer the active form. Each monomer consists of two domains: N-terminal (catalytic or oxygenase) and C-terminal (reductase). The N-terminal domain binds to the thiolate-linked heme group, tetrahydrobiopterin (BH4), a redox co-factor; L-arginine the substrate, and the zinc ion. The C-terminal domain has binding sites for flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD), and nicotinamide adenine dinucleotide phosphate (NADPH; Masters et al., 1996; Sagami et al., 2001; Feng et al., 2014).

### nNOS REGULATION

### Intrinsic Factors

### Auto-Inhibitory Domain and C-Terminal Tail

A sequence of 40–50 amino acids inserted in the FMN domain is related to nNOS auto-inhibition by destabilizing calmodulin (CaM) binding to the enzyme and inhibiting intra- and intermodule electron transferring. This interaction occurs in low intracellular Ca2<sup>+</sup> concentration ([Ca2+]i), taking part in the modulation of nNOS activity (Salerno et al., 1997; Daff et al., 1999; Garcin et al., 2004). Similarly, nNOS has a tail sequence of 21–42 amino acids at the C-terminal, related to the enzyme inhibition. Removal of this extension results in increased transference rates of electron flow in the reductase domain (Roman et al., 2000). Deletion of the auto-inhibitory domain and C-terminal tail results in CaM-independent electron transferring through the reductase domain, despite CaM is still required to promote electron transference from the FMN domain to the heme for NO production (Roman and Masters, 2006).

### Dimer Stability

The dimerization maintained by the N-terminal domain is crucial for the catalytic activity of nNOS. Otherwise, the transport of electrons and formation of nNOS products do not exist (Stuehr, 1997). Dimer formation has the participation of residues from the oxygenase domain that form a "hook" which reaches across to the oxygenase domain of the other subunit to coordinate dimer formation (Crane et al., 1998). Zinc binding has a contribution in dimer stabilization (Hemmens et al., 2000). The disulfide bonds formed by cysteine residues along the nNOS molecule and BH<sup>4</sup> binding are also important to stabilize nNOS dimeric form (Hemmens et al., 1998; Kamada et al., 2005).

## Extrinsic Factors

### Phosphorylation

Phosphorylation of nNOS has been shown to be the critical stage in the activation/inactivation of this isoform. Several phosphatases and kinases including protein kinase A, CaM-kinases (CaM-KI and CaMKII), protein kinase C, and phosphatase 1 may regulate the activity of nNOS. For instance, CaM-KI and CaM-KII phosphorylate Ser<sup>741</sup> and Ser<sup>852</sup> , respectively, resulting in reduced activity of the enzyme through inhibition of CaM binding (Song et al., 2004). Phosphorylation on Ser<sup>1412</sup> (in rat) or Ser<sup>1212</sup> (in human) residue is associated with increased activity of nNOS (Chen et al., 2000; Adak et al., 2001).

### nNOS Uncoupling

The deficiency of L-arginine or BH<sup>4</sup> may produce nNOS uncoupling and the enzyme synthesize superoxide instead of NO. Recently, it has been reported impaired NO signaling due to nNOS uncoupling in brain arteries of obese rats and consequent oxidative stress and vasoconstriction (Katakam et al., 2012). Moreover, nNOS uncoupling is associated with penile arteries constriction with erectile dysfunction in a model of metabolic syndrome (Sanchez et al., 2012).

### Protein-Protein Interactions

Protein-protein interaction is one of the key events in controlling the enzymatic activity of NOS. There are numerous proteins that may have physical interaction with nNOS in a variety of roles including activation, inhibition, and trafficking within the cell.

### Ca2+/CaM Complex

The increase in [Ca2+]<sup>i</sup> and its subsequent binding to CaM is the main modulatory event of nNOS activation (Bredt and Snyder, 1990). The first step of nNOS activation consists of binding Ca2<sup>+</sup> in CaM C-terminal domain. In sequence, the CaM C-terminal domain binds to nNOS. Then, in a similar way, Ca2<sup>+</sup> binds to the CaM N-terminal domain, which also binds to nNOS and causes the activation of nNOS by the displacement of the autoinhibitory domain of the enzyme. When the [Ca2+]<sup>i</sup> decrease, CaM dissociates from nNOS, and it becomes inactive again (Weissman et al., 2002).

### Caveolin/Caveolae

Caveolins are scaffolding proteins situated at the caveolae, the flask-shaped non-clathrin-coated invaginations of the plasma membrane (Sowa, 2012). In skeletal muscle, nNOS directly interacts with caveolin-3, involving two distinct and physically separated caveolin scaffolding domains. This interaction inhibits nNOS activity (Venema et al., 1997). In a rat model of myocardial infarction, nNOS upregulation is associated with an increased binding with caveolin-3 (Bendall et al., 2004). Moreover, caveolin-1 interacts with the oxygenase and reductase nNOS domains inhibiting electron transfers (Sato et al., 2004).

### Protein Inhibitor of nNOS (PIN)

The NH2-terminus of nNOS has a binding site for the protein PIN (Jaffrey and Snyder, 1996). This endogenous protein inhibits nNOS by destabilizing the dimer isoform. Curiously, some studies have shown that PIN plays a physiological role in the control of insulin secretion (Lajoix et al., 2006). Moreover, neurogenic erectile dysfunction (NED) may be caused by impairment of nNOS regulation by PIN (Gonzalez-Cadavid and Rajfer, 2004).

### PDZ Domain

The nNOS PDZ domain has 80–120 amino acid residues located in the NH2-terminus. The PDZ domain participates in the formation of active nNOS dimers and interacts with other proteins in different regions of the cell (Roman et al., 2002). A study to assess potential ligands for PDZ domain of nNOS was conducted by screening 13 billion different peptides and had found that this motif binds to peptides ending with Asp-X-Val.

### FORMATION OF nNOS PRODUCTS

NO formation through L-arginine is catalyzed by nNOS in two steps: the hydroxylation of L-arginine to the intermediate Nω-hydroxy-L-arginine (NOHA), which is then oxidized to Lcitrulline and NO (Papale et al., 2012). In the first step, NADPH transfers electrons to FAD and FMN, which have the capacity to reduce molecular oxygen to superoxide (O•− 2 ) (**Figure 1**). At the same time, an electron from flavin-mononucleotide (FMNH) reduces the heme group (Fe3<sup>+</sup> to Fe2+). The reduction of Fe3<sup>+</sup> enables O<sup>2</sup> linking resulting in an O2<sup>−</sup> Fe2<sup>+</sup> complex. The electron from the complex alternates between Fe2<sup>+</sup> and <sup>O</sup>2, resulting in the complex O•− 2 Fe3+. In the deficiency of L-arginine or NOHA, O•− 2 Fe3<sup>+</sup> transfers an electron to O<sup>2</sup> liberating superoxide (O•− 2 ). Studies have revealed that the heme group of nNOS oxidase domain is responsible for 90% of O•− 2 formation by this enzyme (Yoneyama et al., 2001). Alternatively, the intermediate O •− 2 Fe3<sup>+</sup> can receive an electron, forming <sup>O</sup>2<sup>−</sup> Fe3<sup>+</sup> that interacts with H<sup>+</sup> and releases H2O<sup>2</sup> and Fe3+.

In order to make the catalysis of L-arginine possible, BH<sup>4</sup> cofactor must be binding to O•− 2 Fe3<sup>+</sup> present in heme group. Electrons from BH<sup>4</sup> cofactor are responsible for the formation of peroxy complexes (Fe3+-OOH−) with consequent hydroxylation of L-arginine, resulting in the formation of NOHA and regeneration of Fe3<sup>+</sup> from heme group. In the next step, NOHA participates in another oxidation-reduction cycle by binding to Fe3+, which will receive more electrons from the reductase group, resulting in the cleavage of NOHA and release of water, L-citrulline and NO (Abu-Soud et al., 1994, 2000; Rosen et al., 2002).

Therefore, during the enzymatic formation of NO cycle, nNOS also generates H2O<sup>2</sup> and O•− 2 (**Figure 1**). The production of these reactive oxygen species (ROS) by nNOS can occur even at saturating concentrations of L-arginine or NOHA in steps before the formation of NO (Rosen et al., 2002; Tsai et al., 2005; Weaver et al., 2005). At the expense of O•− 2 , the production of H2O<sup>2</sup> is strongly increased by BH<sup>4</sup> (Rosen et al., 2002).

### ROLE OF nNOS IN VASCULAR HOMEOSTASIS

Emerging evidence shows that nNOS has a physiologically relevant role in the control of the cardiovascular system. Here, we outline the recent advances on the role of nNOS in the vascular function.

There are several reports implicating the participation of nNOS in cerebral blood flow (CBF; Pelligrino et al., 1993;

Santizo et al., 2000; Chi et al., 2003). Intraperitoneal injections of the selective nNOS inhibitor 7-nitroindazole (7-NI) depressed baseline CBF in rats (Montécot et al., 1997; Gotoh et al., 2001). Moreover, 7-NI decreased cerebral capillary flow in rats (Hudetz et al., 1998) and global CBF in cats (Hayashi et al., 2002). In rats, during hyperbaric conditions, it was found that the increase in CBF in the cortex prior to the appearance of electrical discharges was completely inhibited by 7-NI (Hagioka et al., 2005).

Aside from cerebral flow, it has been suggested that nNOS-derived NO regulates renal circulation. In the presence of, S-methyl-L-thiocitrulline (SMTC) a nNOS inhibitor, the vasoconstrictor response to angiotensin II is increased in the efferent arteriole (Ichihara et al., 1998). Additional evidence was obtained from nNOS−/<sup>−</sup> mice, where genetic deletion of nNOS decreases medullary blood flow in response to angiotensin II (Mattson and Meister, 2005). In nNOS−/<sup>−</sup> mice Vallon et al. (2001) also found that the feedback control of glomerular vascular tone is attenuated.

Similarly, studies in isolated vessels demonstrate the participation of nNOS in the control of vascular function. In pial arterioles of eNOS−/<sup>−</sup> mice acetylcholine induced an nNOS-cGMP-dependent vasodilation (Meng et al., 1996, 1998). Another work confirmed the presence of nNOS in the endothelium of coronary arteries of eNOS−/<sup>−</sup> mice and showed that shear stress activated endothelial nNOS-derived NO release, compensating the absence of eNOS-derived NO (Huang et al., 2002). In aorta of nNOS−/<sup>−</sup> mice the vasodilator response induced by acetylcholine is reduced (Nangle et al., 2004). In small mesenteric arteries of female rats the inhibition of endothelial nNOS contributes to the decrease in the relaxation induced by estrogen. Furthermore, the same study showed that estrogen rapidly increased the nNOS activity and nNOSmediated NO production in human umbilical vein endothelial cells (Lekontseva et al., 2011). A year later, the same group demonstrated that nNOS contributed to the estrogen-mediated vascular relaxation of mesenteric artery in young, but not in ovariectomized and aging female rats. In the ovariectomized and aging group nNOS functionally became a source of O•− 2 (Lekontseva et al., 2012).

Corroborating the above findings, NO release from nNOS also seems to be important in the control of vascular tone in humans. Expression of nNOS was found in human aorta, carotid, radial and mammary artery (Buchwalow et al., 2002), saphenous vein (Webb et al., 2006), and lung capillary endothelial cells (Lührs et al., 2002).

The first evidence that nNOS had a function in vascular regulation in humans was obtained from children suffering from Duchene muscular dystrophy (DMD). It was shown that nNOSderived NO present in skeletal muscle acts in the blood flow and oxygen transport. nNOS expression is reduced in children with DMD resulting in increased vasoconstrictor response (Sander et al., 2000).

Later on, Seddon et al. (2008)showed the relationship between nNOS and the regulation of blood flow in human. Selective in vivo inhibition of nNOS with SMTC in healthy men promoted a reduction in the brachial artery baseline flow. This effect was eliminated in the presence of L-arginine. A similar reduction was observed with the non-selective inhibitor of NOS (L-NMMA) but required a 20-fold higher dose. This study suggested that nNOSderived NO has a significant role in the physiological regulation of microvascular tone in vivo (Seddon et al., 2008). In another work, the same group investigated the in vivo effects of SMTC in human coronary dilatation. The infusion of SMTC in healthy patients reduced baseline coronary blood flow and coronary artery diameter measured by angiography. They concluded that local nNOS-derived NO is a key physiological regulator of human coronary vascular tone in vivo (Seddon et al., 2009).

All the above works suggesting NO as the mediator of nNOS function in the regulation of vascular tone were based on the assumption that NO was the only physiological vasodilator product of nNOS activation. Our group was the first to show the importance of nNOS-derived H2O<sup>2</sup> in the endothelium-dependent vascular relaxation. We showed that nNOS was constitutively expressed in the endothelium of the mouse aorta and mesenteric resistance artery. Stimulation of those vessels with acetylcholine promoted increase in H2O<sup>2</sup> production. Pharmacological selective nNOS inhibition and nNOS knockdown decreased endothelium-dependent vascular relaxation and H2O<sup>2</sup> production. Finally, incubation of the vessels with catalase, an enzyme that degrades H2O<sup>2</sup> into O<sup>2</sup> and H2O, decreased vascular relaxation (Capettini et al., 2008, 2010; Silva et al., 2016). The participation of nNOS in vascular homeostasis in physiological and pathological conditions is summarized in **Table 1**.

### nNOS IN VASCULAR DISEASES

### Hypertension

Several studies have indicated that the imbalance in nNOS expression and/or activity is involved in the mechanism of pathogenesis of hypertension. In mesenteric arteries from spontaneously hypertensive rats (SHR), nNOS expression was ∼2 times higher than in vessels from control animals (Briones et al., 2000). A similar result showing increased expression of nNOS in vascular smooth muscle cells was found in carotid arteries from SHR. It was shown that activation of nNOS on stimulation by Angiotensin II occurs in hypertensive but not in normotensive animals (Boulanger et al., 1998). Interestingly, in SHR rats the expression and activity of nNOS are decreased in the adrenal gland. Chronic treatment of SHR with antihypertensive drugs, increased the expression and activity of nNOS in the adrenal gland, suggesting that normalization of blood pressure (BP) may be in part related to an increase in nNOS (Qadri et al., 2001).

BP and vascular function were evaluated in normotensive rats chronically treated (6 weeks) with the selective nNOS inhibitor 7- NI. A significant increase in systolic BP was observed in the first 2 weeks of treatment. Corroborating the in vivo study, isolated vessels showed an attenuated relaxant response to acetylcholine in the aorta. These results show that nNOS participates in the regulation of BP and vascular tone (Cacanyiova et al., 2009). In contrast, in SHR, treatment with 7-NI had no effect in blood pressure or acetylcholine-induced vasodilatation in the aorta (Cacanyiova et al., 2009, 2012), suggesting that nNOS function was lost in hypertension.

A recent study revealed that impairment of nNOSderived H2O<sup>2</sup> pathway participates in the endothelial dysfunction and increase in blood pressure in DOCA-salthypertensive mice (Silva et al., 2016). This study showed that 1-(2-trifluoromethylphenyl) imidazole, a selective nNOS inhibitor, and catalase, exhibited a more pronounced reduction of acetylcholine-induced decrease in blood pressure in normotensive than in hypertensive mice. Moreover, selective nNOS inhibition and catalase had a greater inhibitory effect in acetylcholine-induced vasodilatation in control compared to DOCA-salt mice. Also, acetylcholine-induced H2O<sup>2</sup> production and the expression and functioning of nNOS were considerably diminished in the resistance mesenteric arteries of DOCA-salt mice.

### Atherosclerosis

The first evidence that nNOS plays a vasculoprotective role in atherosclerosis came from a work by Wilcox et al. (1997) that showed a correlation between the progression of plaque formation and nNOS mRNA. In 1999, Qian et al. performed experiments with recombinant adenoviruses expressing nNOS transferred to carotid of hypercholesterolemic rabbits and TABLE 1 | Participation of nNOS in the control of vascular function in physiological conditions and during hypertension and atherosclerosis.


*SHR, spontaneously hypertensive rats.*

showed a marked reduction in expression of adhesion molecules and infiltration of inflammatory cells. Additionally, a reduction in lipid deposition was observed after gene transfer. In another work, nNOS−/<sup>−</sup> mice exhibited accelerated neointimal formation and constrictive vascular remodeling caused by blood flow disruption in a model of carotid artery ligation. It was also observed that selective inhibition of nNOS decreased cGMP production, inducing an increase in vasoconstrictor response and accelerating neointimal formation in a rat balloon injury model (Morishita et al., 2002). Using a double knockout mouse (nNOS DKO) that combined genetic deletion of nNOS (nNOS−/−) with a model of atherosclerosis (apoE−/−), Kuhlencordt et al. (2006) showed that the absence of nNOS accelerated the atherosclerotic plaque lesion. After 14 weeks following a "Western-type" atherogenic diet, nNOS DKO animals showed 66% increase of lesion area, compared to apoE−/<sup>−</sup> control mice.

nNOS-derived H2O<sup>2</sup> also seems to participate in endothelial dysfunction in atherosclerosis. Capettini et al. (2011) showed that selective pharmacological inhibition of nNOS, nNOS knockdown and catalase reduced the vasodilator effect of acetylcholine, diminished NO and abolished endothelial-dependent H2O<sup>2</sup> production in wild-type mice, but had no effect in ApoE−/<sup>−</sup> animals. In addition, nNOS functioning was decreased in ApoE−/<sup>−</sup> mice compared to controls.

### CONCLUSIONS

This mini-review summarizes puzzling information on the role of nNOS in the control of vascular homeostasis under physiological

### REFERENCES


and diseases conditions. Recent data indicates that nNOS is constitutively expressed in the endothelial cells of different types of vessels in animals and human. More importantly, nNOSderived products such as NO and H2O<sup>2</sup> play an important role in the control of vascular function and blood pressure. Finally, nNOS participates in the physiopathology of hypertension and atherosclerosis.

### AUTHOR CONTRIBUTIONS

VL defined the research topics and co-wrote the manuscript. EC, BR, and SC co-wrote the manuscript.

### FUNDING

This work was supported by FAPEMIG (Fundação de Apoio à Pesquisa do Estado de Minas Gerais) grant APQ-00683-13 and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) grants 467147/2014-0, 305693/2014-0, and 470860/2012-0.


regulator of glucose-induced insulin secretion. Diabetes 55, 3279–3288. doi: 10.2337/db06-0257


electron flow through the flavin and heme domains and affect modulation by calmodulin. J. Biol. Chem. 275, 29225–29232. doi: 10.1074/jbc.M004766200


nitric oxide synthase-derived hydrogen peroxide. Clin. Sci. (Lond). 130, 895–906. doi: 10.1042/CS20160062


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Costa, Rezende, Cortes and Lemos. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Increased Nitric Oxide Bioavailability and Decreased Sympathetic Modulation Are Involved in Vascular Adjustments Induced by Low-Intensity Resistance Training

Fabrício N. Macedo<sup>1</sup> , Thassio R. R. Mesquita<sup>2</sup> , Vitor U. Melo<sup>1</sup> , Marcelo M. Mota<sup>3</sup> , Tharciano L. T. B. Silva<sup>3</sup> , Michael N. Santana<sup>1</sup> , Larissa R. Oliveira<sup>1</sup> , Robervan V. Santos <sup>1</sup> , Rodrigo Miguel dos Santos <sup>2</sup> , Sandra Lauton-Santos <sup>2</sup> , Marcio R. V. Santos <sup>1</sup> , Andre S. Barreto<sup>1</sup> and Valter J. Santana-Filho<sup>1</sup> \*

*<sup>1</sup> Laboratory of Cardiovascular Pharmacology, Department of Physiology, Federal University of Sergipe, Sao Cristovao, Brazil, <sup>2</sup> Laboratory of Cardiovascular Biology and Oxidative Stress, Department of Physiology, Federal University of Sergipe, Sao Cristovao, Brazil, <sup>3</sup> Department of Healthy Education, Estacio Faculty of Sergipe, Aracaju, Brazil*

### Edited by:

*Valdir Andrade Braga, Federal University of Paraiba, Brazil*

### Reviewed by:

*Marli Cardoso Martins-Pinge, State University of Londrina, Brazil Josiane Campos Cruz, Federal University of Paraiba, Brazil*

> \*Correspondence: *Valter J. Santana-Filho vjsf@infonet.com.br*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *15 May 2016* Accepted: *14 June 2016* Published: *28 June 2016*

### Citation:

*Macedo FN, Mesquita TRR, Melo VU, Mota MM, Silva TLTB, Santana MN, Oliveira LR, Santos RV, Miguel dos Santos R, Lauton-Santos S, Santos MRV, Barreto AS and Santana-Filho VJ (2016) Increased Nitric Oxide Bioavailability and Decreased Sympathetic Modulation Are Involved in Vascular Adjustments Induced by Low-Intensity Resistance Training. Front. Physiol. 7:265. doi: 10.3389/fphys.2016.00265* Resistance training is one of the most common kind of exercise used nowadays. Long-term high-intensity resistance training are associated with deleterious effects on vascular adjustments. On the other hand, is unclear whether low-intensity resistance training (LI-RT) is able to induce systemic changes in vascular tone. Thus, we aimed to evaluate the effects of chronic LI-RT on endothelial nitric oxide (NO) bioavailability of mesenteric artery and cardiovascular autonomic modulation in healthy rats. Wistar animals were divided into two groups: exercised (Ex) and sedentary (SED) rats submitted to the resistance (40% of 1RM) or fictitious training for 8 weeks, respectively. After LI-RT, hemodynamic measurements and cardiovascular autonomic modulation by spectral analysis were evaluated. Vascular reactivity, NO production and protein expression of endothelial and neuronal nitric oxide synthase isoforms (eNOS and nNOS, respectively) were evaluated in mesenteric artery. In addition, cardiac superoxide anion production and ventricle morphological changes were also assessed. *In vivo* measurements revealed a reduction in mean arterial pressure and heart rate after 8 weeks of LI-RT. *In vitro* studies showed an increased acetylcholine (ACh)-induced vasorelaxation and greater NOS dependence in Ex than SED rats. Hence, decreased phenylephrine-induced vasoconstriction was found in Ex rats. Accordingly, LI-RT increased the NO bioavailability under basal and ACh stimulation conditions, associated with upregulation of eNOS and nNOS protein expression in mesenteric artery. Regarding autonomic control, LI-RT increased spontaneous baroreflex sensitivity, which was associated to reduction in both, cardiac and vascular sympathetic modulation. No changes in cardiac superoxide anion or left ventricle morphometric parameters after LI-RT were observed. In summary, these results suggest that RT promotes beneficial vascular adjustments favoring augmented endothelial NO bioavailability and reduction of sympathetic vascular modulation, without evidence of cardiac overload.

Keywords: autonomic nervous system, resistance training, endothelium-dependent relaxing factors, nitric oxide, eNOS enzyme, nNOS enzyme

## INTRODUCTION

Although well known the development of muscular strength, endurance, and mass, resistance exercise it is also related with reduction of cardiovascular disease risk. Currently, the inclusion of resistance exercise has been recommended by the American Heart Association (Pollock et al., 2000) and the American College of Sports Medicine (American College of Sports Medicine, 2009) as an important non-pharmacological strategy for the prevention and treatment of cardiovascularrelated disorders as, hypertension and heart failure. Indeed, it is known that aerobic and resistance training (RT) can promote substantial benefits in physical fitness and health-related factors. Moreover, several studies reported that long-term of RT is able to reduce systolic and diastolic arterial pressure in cardiovascular diseased patients (Kelley and Kelley, 2000; Cornelissen and Smart, 2013).

Although coordinated, autonomic modulation and local signaling pathways, mostly controlled by smooth muscle and endothelial cells, play a pivotal role on the control of vascular tone and consequently, in the regulation of blood pressure (Garland et al., 1995; Thomas, 2011). Despite individual actions, it has been extensively demonstrated the relationship between abnormal endothelial dysfunction and higher incidence of cardiovascular events in diseased patients and in healthy population (Heitzer et al., 2001; Matsuzawa et al., 2013). Indeed, endothelial layer is composed by highly specialized cells which are able to release crucial modulators of vascular tone such as, endothelium-derived hyperpolarizing factors, prostacyclin and nitric oxide (NO) (Ignarro, 1989; Garland et al., 1995). It has been shown that exercise training can improve endothelial function in vascular beds, mainly due to increased shear stress stimulus (Niebauer and Cooke, 1996; Harris et al., 2010). Enhanced endothelium-dependent vasorelaxation has been described after aerobic training due to increased NO bioavailability, reduction in the sensitivity of contractile agonists and vascular sympathetic modulation (Jansakul and Hirunpan, 1999; Higashi and Yoshizumi, 2004). However, the lack of studies dissecting the potential contribution of autonomic modulation and local intracellular mechanisms involved in the decrease of blood pressure induced by RT, encouraged the present study.

Besides of exercise modality, the effects of RT on vascular function seem to be dependent of exercise intensity. Highintensity RT has been associated with deleterious responses on the vascular function and structural remodeling (Cortez-Cooper et al., 2005). On the other hand, recent evidences have been reported that low-intensity RT is able to improve vascular endothelial function and peripheral blood circulation in healthy young and elderly population (Shimizu et al., 2016) and to reduce arterial stiffness (Okamoto et al., 2011, 2014). However, whether low-intensity RT induces a protective or maladaptive effect on vascular function is not yet well understood. Therefore, the aim of this study was to evaluate the participation of endothelial nitric oxide and autonomic modulation in the vascular tone adjustments, and consequently, blood pressure control induced by low-intensity resistance training in healthy rats.

## MATERIALS AND METHODS

### Animals

All experimental procedures were performed in accordance with the Guide for the Care and Use of Laboratory Animals published by National Institutes of Health (NIH, 8th edition) and approved by the Ethics Committee on Animal Research of the Federal University of Sergipe (#01/2013). Twenty-eight male Wistar rats (250–270 g, 12- to 14-weeks old) were obtained from the Animal Care Facility of the Federal University of Sergipe and housed under standard conditions (food and water ad libitum, 12:12 h light/dark cycle and ∼22◦C). The animals were assigned into two experimental groups: sedentary (SED, n = 14) and resistance trained (Ex, n = 14).

### Resistance Training Protocol

SED and Ex animals underwent a 1 week familiarization period (5 days, 5 min per day in rest position) in a customized squat apparatus for RT, as developed by Tamaki et al. (1992). Electrical stimulation (20 V, 0.3 s duration, at 3 s intervals) was applied on the tail of the rat through a surface electrode. After familiarization period, both groups were subjected to a one maximal repetition test (1RM) which consists to determine the maximum weight lifted by the rat in the exercise apparatus. The 1RM test was repeated every 2 weeks in attempt to maintain the desired intensity. Ex group was subjected to a RT protocol which consists in 3 sets of 10 repetitions with intensity defined at 40% of the maximum load established in the 1RM test, the animals were exercised three times per week (alternated days) for 8 weeks. SED group was subjected to a fictitious exercise consisting in a similar procedures and electrical stimulation as Ex group, however, without physical effort.

## In vivo Measurements

The animals were anesthetized with a mixture of ketamine/xylasine (80 mg/kg and 10 mg/kg, respectively, i.p) and polyethylene catheter was implanted (PE-10/PE-50, Intramedic, Becton Dickinson and Company, Sparks, MD, USA) into the femoral artery. The catheter was tunneled and exteriorized in the posterior cervical region and the animals were allowed to recovery for 24 h. Afterwards, the catheter was connected to a pressure transducer (FE221, Bridge Amp, ADInstruments, Bella Vista, NSW, Australia) coupled to a pre-amplifier (Powerlab 8/35, AdInstruments). Values of mean arterial pressure (MAP), systolic arterial pressure (SAP), diastolic arterial pressure (DAP) and heart rate (HR) were obtained and assessed after 8 weeks of experimental procedures.

### Cardiovascular Autonomic Modulation

The baroreflex sensitivity (BRS) was measured in the time domain by the sequence method (Bertinieri et al., 1985). Series beat-to-beat were analyzed by software CardioSeries v2.4 (http://sites.google.com/site/cardioseries). Sequences of at least 4 heart beats with increased SAP followed by PI lengthening or subsequent decrease of SAP with PI shortening with correlation greater than 0.85 were identified as baroreflex sequence. The slope of the linear regression between SAP and PI was considered as a measure of BRS.

Cardiac autonomic balance was evaluated by frequency domain. The PI and SAP variability analysis were performed using the same software previously described. Series beat-to-beat were obtained by pulsatile arterial pressure and converted into points every 100 ms using cubic spline interpolation (10 Hz). The interpolated series were divided into half-overlapping sequential sets of 512 data points (51.2 s). Before calculation of the spectral power density, the segments were visually inspected and the non-stationary data were not taken into consideration. The spectrum was calculated from the Fast Fourier Transformation (FFT) algorithm direct and Hanning window was used to attenuate side effects. The spectrum is composed of bands of low frequency (LF; 0.2–0.75 Hz) and high frequency (HF; 0.75–3 Hz), the results were showed in normalized units, by calculating the percentage of the LF and HF variability with respect to the total power after subtracting the power of the very low frequency (VLF) component (frequencies < 0.20 Hz), namely Low Frequency/High Frequency (LF/HF) ratio.

The LF/HF ratio from pulse interval represents sympathovagal balance. LF and HF components mean cardiac sympathetic and parasympathetic activity. LF from systolic arterial pressure (LFsys) represents sympathetic vascular modulation (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, 1996).

### Vascular Reactivity

Animals were euthanized by an over-dose of thiopental (100 mg/kg) and the superior mesenteric artery was removed, cleaned from connective and fat tissues and sectioned into rings (1–2 mm). Rings were suspended by fine stainless hooks connected to a force transducer (Letica, Model TRI210; Barcelona, Spain) with cotton threads in organ baths containing Tyrode's solution (Composition in mM: NaCl 158.3, KCl 4.0, CaCl<sup>2</sup> 2.0, NaHCO<sup>3</sup> 10.0, C6H12O<sup>6</sup> 5.6, MgCl21.05, and NaH2PO<sup>4</sup> 0.42). The solution was continually gassed with carbogen (95% O<sup>2</sup> and 5% CO2) and maintained at 37◦C under a resting tension of 0.75 g for 60 min (stabilization period). After 1 h of stabilization, endotheliumintact mesenteric rings were pre-contracted with phenylephrine (Phe, 1 µM) and cumulative concentration–response curves to vasodilator agonist, acetylcholine (ACh, 10−9–10−<sup>4</sup> M), were performed in the absence or presence of a NOS inhibitor, Nωnitro-L-arginine methyl ester (L-NAME, 100 µM). Phe-induced vasoconstriction (10−<sup>6</sup> M) was also assessed in the absence or presence of L-NAME. Relaxing responses were plotted as percentage of the contraction induced by Phe. Vasoconstriction induced by Phe was expressed as maximal tension developed (grams).

### Measurements of NO Production

NO production in the mesenteric artery was determined using a fluorescent cell-permeable dye, 4,-amino-5 methylamino-2 ′ ,7′ -diaminofluorescein diacetate (DAF-FM; Molecular Probes), as previously described (Mota et al., 2015). Freshly isolated mesenteric arteries were loaded with 10 µM of the dye for 40 min at 37◦C in Tyrode solution, 20 min after the onset of the probing, some rings were stimulated with 1 µM of ACh and then washed for 40 min. Mesenteric segments were frozen in medium for cryosectioning and cut into 20 µm thick sections. Images were recorded using a fluorescence microscope (Ci-E, Nikon, Japan) under identical settings. The fluorescent intensity was measured using ImageJ software (NIH). A minimum of 10 regions were randomly selected in the endothelial and smooth muscle layers from each mesenteric section. It worthwhile note that smooth muscle exhibits an autofluorescence, therefore, in order to avoid misleading fluorescence measurements, analyses of images were carefully performed selecting the region of interest between the smooth muscle fibers. Fluorescence microscopy images were analyzed according to the intensity of the fluorescence per area and the data are expressed as arbitrary unit (a.u.).

## Measurements of Reactive Oxygen Species Generation

Hearts were quickly removed and frozen in medium for cryosectioning. Frozen tissue blocks were transversely cut into 30 µm thick sections and slides transferred to a recording chamber. Slides were pre-incubated with warmed Tyrode solution for 15 min and then, the sections were loaded with 5 µM of dihydroethidium, a fluorescent cell-permeable dye (DHE; Molecular Probes) for 30 min (Erickson et al., 2008). DHE staining with light fixation was adapted from a published method (Owusu-Ansah et al., 2008). Images were recorded using a fluorescence microscope (Ci-E, Nikon, Japan) and the fluorescence intensity was measured using ImageJ software (NIH).

### Western Blot

Western blot was performed as described previously (Mota et al., 2015). Mesenteric artery was homogenized in ice-cold lysis buffer containing (in mM): 150 NaCl, 50 Tris-HCl, 5 EDTA.2 Na and 1 MgCl2, pH 8.0; 1% Triton X-100, 1% NP-40, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate enriched with a protease inhibitor cocktail (Sigma FAST, Sigma, St. Louis, MO). Homogenates were cleared by centrifugation at 13,000 × g for 15 min at 4◦C and protein content was quantified by Lowry assay. Samples were denaturated in Laemmli's buffer and equal amount of protein (30 µg/lane) was separated on 10% SDS-polyacrylamide gel electrophoresis and then, transferred onto nitrocellulose membrane (2 h at 120 V, Merck-Millipore, Billerica, MA). Membranes were blocked for 2 h in Trisbuffered saline-Tween 20 containing 5% non-fat dry milk at room temperature before incubation with rabbit polyclonal antieNOS, anti-nNOS and goat polyclonal anti-GAPDH antibodies (1:1000, Santa Cruz Biotechnology, Santa Cruz, CA) overnight at 4 ◦C. After washing and incubation for 2 h at room temperature with anti-rabbit or anti-goat IgG-HRP antibody (1:10,000, Sigma-Aldrich, St. Louis, Missouri, USA), immunodetection was performed using enhanced chemiluminescence (Luminata strongTM—Western HRP substrate, Merck-Millipore, Billerica, MA). Digitalized images were analyzed by densitometry using ImageJ software (NIH).

### Cardiac Morphometry

Heart hypertrophy was evaluated by morphometry. After anesthesia, heart beat was stopped in diastole using 10% KCl (i.v.). Hearts were embedded in medium for cryosectioning and frozen at −80◦C. Transversal sections (5 µm) were cut starting from apex to base of the heart and stained with hematoxylineosin for cardiac morphometry. Tissue sections (3–6 for each animal) were examined with (Ci-E, Nikon, Japan) microscope and analyzed with ImageJ software.

### Statistical Analysis

All data are expressed as mean ± S.E.M. Significant differences between groups were determined using Two-way ANOVA followed by Bonferroni's post-hoc test to compare the concentration-response curves obtained in mesenteric rings. One-way ANOVA followed by Bonferroni's post-hoc test was used to compare the Phe-elicited vasoconstriction and NO production. Student's paired t-test was used to compare 1RM and bodyweight intragroup. Student's unpaired t-test was used for all other analysis. All statistical comparisons were made using GraphPad Prism 5.1 (GraphPad Software Inc., San Diego, CA, USA) and values of P < 0.05 were considered to be statistically significant.

## RESULTS

### Exercise Training Efficacy

As shown in the **Table 1**, all groups showed a body weight gain during the study period; however, 8 weeks of low-intensity RT promoted a smaller body weight gain compared to the SED group. Moreover, Ex group showed a significant enhancement on physical fitness assessed by increased 1RM/Body Weight (BW) ratio.

## Hemodynamic Parameters, Morphometry of the Left Ventricle and Oxidative Stress Biomarker

The hemodynamic parameters, morphometric analysis of left ventricular myocardium and reactive oxygen species generation measurements were evaluated at the end of the experimental protocol. As shown in the **Table 2**, Ex group showed a significant reduction in mean (MAP), systolic (SAP) and diastolic (DAP) arterial pressures, associated with decreased heart rate (HR),

TABLE 1 | Bodyweight (BW), one maximal repetition test values (1RM) and 1RM/BW ratio before and after 8 weeks of low-intensity resistance training in healthy rats.


*SED, Sedentary group; Ex, Exercised group; 1RM, One-Repetition Maximum Test. Values expressed as mean* ± *S.E.M. To evaluate differences between groups and intragroup (final vs. initial), it were used unpaired (*\**) and paired (†) T-Test, respectively.* \* , *† p* < *0.05.* when compared with SED group. It is worth mentioning that the hemodynamic measurements were performed in conscious rats. On the other hand, RT neither induced cardiac hypertrophy nor superoxide generation changes compared with SED group, as shown in **Table 2** and **Figure 1**, respectively.

## Vascular Reactivity

Based on hemodynamic findings, we assessed whether RT affects the endothelium-dependent relaxation of mesenteric artery. As expected, ACh induced a concentration-dependent relaxation

TABLE 2 | Hemodynamic parameters and morphology of the left ventricle after 8 weeks of low-intensity resistance training in healthy rats.


*SED, Sedentary group; Ex, Exercised group; MAP, Mean Arterial Pressure; SAP, Systolic Arterial Pressure; DAP, Diastolic Arterial Pressure; HR Heart Rate; LVED, left ventricle external diameter; LVPWT, left ventricle posterior wall thickness; LVID, left ventricle internal diameter. Values expressed as mean* ± *S.E.M. To evaluate differences between groups, it was used unpaired T-test.* \**p* < *0.05 and* \*\**p* < *0.01 when compared SED vs. Ex.*

fluorescence in left ventricle myocardium slides of sedentary (SED) and exercise-trained (Ex) rats. Scale bar = 20 µm. Values are expressed as mean ± S.E.M for 3–5 left ventricle slides analyzed from 3 rats in each group. For data analysis, Student's unpaired *t*-test was used.

in all groups, however, 8 weeks of low-intensity RT induced a significant leftward shift of ACh-induced relaxation in Ex group compared to SED group (**Figure 2A**; pD<sup>2</sup> values: Ex group, 7.1 ± 0.1 and SED group, 6.2 ± 0.1). Maximal relaxant response (Rmax) to ACh was similar between Ex and SED rats (**Figure 2A**). To evaluate the involvement of NO on the enhanced ACh-induced vasorelaxation promoted by RT, inhibition of both constitutive NOS, eNOS and nNOS, markedly reduced the Rmax in SED and Ex group. However, the inhibition was significantly higher in Ex group than SED group. To further explore these results, the difference of the area under the curve (dAUC) demonstrates a greater dependence of NO for ACh-induced vasorelaxation in Ex (81.17 ± 2.7%) than SED group (55.4 ± 5.2%; **Figure 2B**). Furthermore, as shown in the **Figure 3**, the vasoconstriction induced by Phe was decreased in Ex group (0.20 ± 0.05 g) compared with SED group (0.39 ± 0.06 g). In accordance with ACh vasorelaxation response, L-NAME significantly potentiates the vasoconstriction response of Phe in Ex group (0.88 ± 0.07 g) and SED (0.65 ± 0.05 g), however, the developed tension was higher in Ex group than SED group.

## NO Bioavailability and Constitutive NOS Expression

Based on our in vitro findings, we next assessed whether RT affects NO synthesis and protein expression of eNOS and nNOS in mesenteric artery. Interesting, NO production under basal condition was 1.2-fold higher in Ex group than in SED group (**Figure 4**). It is well known that NO production is the result of eNOS and nNOS activities in endothelial cells. Therefore, we indirectly tested the activities of NOS in endotheliumintact mesenteric artery, monitoring the end-product released by active NOS when stimulated for 20 min with ACh (1 µM). It is worth mentioning that the adopted concentration of ACh was determined based on pD<sup>2</sup> values found in our in vitro vascular reactivity data. **Figure 4** shows that ACh stimulation elicited a marked increase on NO production in Ex e SED rats. However, NO production was significantly higher in Ex group than SED group when stimulated with ACh. Consequently, we next investigated whether the expression of both constitutive NOS enzymes were altered after 8 weeks of low-intensity RT in mesenteric artery. In accordance with our findings, western blot analysis revealed increased protein expression levels of both, eNOS (∼2.4-fold) and nNOS (∼1.7-fold), in Ex when compared to SED rats (**Figure 5**).

## Neural Control of Blood Pressure

It is well known that the autonomic nervous system plays a crucial role in blood pressure control. Therefore, we next evaluated whether RT affects arterial baroreflex sensitivity and cardiovascular variability after 8 weeks of low-intensity RT. As shown in the **Figure 6A**, the spontaneous baroreflex sensitivity (BRS) was significantly increased after RT when compared to SED rats. However, RT was able to reduce the power spectral analysis of heart rate (LF/HF ratio; **Figure 6B**) and arterial

pressure variability (LFsys; **Figure 6C**) when compared to SED rats.

## DISCUSSION

In this study, the effect of 8 weeks of low-intensity resistance training (LI-RT) on the hemodynamic parameters, cardiac oxidative stress, left ventricular myocardium remodeling, vascular adjustments and autonomic control were evaluated in normotensive rats. The major findings are: (1) Decrease in both blood pressure and heart rate with no cardiac overload; (2) Improvement of endothelium-dependent vascular function, which is correlated with increased endothelial NO bioavailability and upregulation of eNOS and nNOS in mesenteric artery; (3) Reduction in vasoconstrictor responsiveness to phenylephrine; (4) Enhanced spontaneous baroreflex sensitivity accompanied by reduction in a cardiac and vascular sympathetic modulation.

The model of resistance training used in this work involves a caudal electrical stimulation to accomplish exercise movement (Tamaki et al., 1992). A recent study of our laboratory demonstrated that caudal electrical stimulation isolated is not able to induce changes in vascular response after a single bout of RT (Fontes et al., 2014). In addition, Barauna et al. (2005) showed that stressor parameters, such as plasma catecholamines or adrenal weight did not suffer any change by electrical stimulation isolated. Therefore, the cardiovascular results found in this work are attributed solely to LI-RT.

Muscle strength due a RT protocol are strongly associated with workload. It is well known that muscle hypertrophy and neuromuscular adaptation can enhance this physical valence. Gabriel et al. (2006) demonstrates that RT, independent of the intensity used, develops an important neuromuscular component, however, the size of this adaptation are intensity dependent. LI-RT protocol used in our study were able to increase the maximal weight lifted in Ex group. Even in low intensity, a period of RT was an effective method to increase muscular strength, however there was no difference in body weight between groups. Despite skeletal muscle adaptation, high and moderate intensities of resistance exercise are associated with acute and chronic deleterious cardiovascular effects such as, increase in the arterial stiffness, (Miyachi et al., 2004; Miyachi, 2013) cardiac hypertrophy (Barauna et al., 2005, 2007), arterial baroreflex and vascular sympathetic dysfunction (Niemelä et al., 2008; Collier et al., 2009). On the other hand, LI-RT has been demonstrated as an efficient and safer exercise to be applied to health and pathological populations in humans (Okamoto et al., 2011, 2014) or experimental model (Mostarda et al., 2014).

The LI-RT protocol here adopted consisted in low workload (40% of 1RM), moderated number of repetitions and short sets to minimize pressure overload. It has been shown that pressure overload induces a great production of reactive oxygen species inducing the oxidative stress, which ultimately result in cardiac hypertrophy (Takimoto and Kass, 2007; Brown and Griendling, 2015). In the present study, 8 weeks of LI-RT was not able to induce changes in the superoxide generation nor left ventricular remodeling. High levels of cardiac oxidation are implicated with pathogenesis and aggravation of several cardiometabolic diseases as hypertension, heart failure and diabetes (Lakshmi et al., 2009; Fiorentino et al., 2013). In fact, a number of studies indicated the cardioprotective actions of physical exercise due increased activity and expression of antioxidants enzymes in the heart (Chicco et al., 2006; Azizbeigi et al., 2014; Borges and Lessa, 2015). In contrast, it has been shown left ventricle hypertrophy as consequence of RT performed at high intensity (Huonker et al., 1996; Barauna et al., 2007). Taken together, the absence of oxidative stress and morphological changes in the left ventricular myocardium suggest that the practice of resistance training under low intensity does not develop cardiac risk factors.

It is well known that RT has important impact on hemodynamics parameters reducing the blood pressure in normotensive, hypertensive and diabetics individuals (Kelley and Kelley, 2000; Eves and Plotnikoff, 2006; Cornelissen et al., 2011; Moraes et al., 2012). In humans, recent evidences indicate that acute high intensity resistance exercise seems to be effective in lowering blood pressure in healthy young and hypertensive elderly population (Brito et al., 2014; Duncan et al., 2014). Similarly, Queiroz et al. (2015) showed that low intensity of resistance exercise intensities led to changes in blood pressure in healthy and hypertension population. However, as previously described, maladaptive changes are triggered by high intensity RT which could limit the efficacy–safety ratio of resistance exercise.

In the present study, we observed that 8 weeks of LI-RT reduced mean, systolic and diastolic blood pressure, as well as heart rate. In general, hemodynamic effects of RT have been associated with changes in local signaling pathways and

neurohumoral mechanisms (Katz et al., 1997; Selig et al., 2004; Figueroa et al., 2008; Speretta et al., 2016). Accordingly with our hemodynamic findings and consistent with clinical human evidence (Okamoto et al., 2014), we demonstrated that LI-RT enhances the vasorelaxation response in superior mesenteric artery of trained rats. However, knowing that the vascular tone is resultant of the balance between competing vasodilator and vasoconstrictor factors, we found a decrease in contractile responses to phenylephrine. Based on our data, the mechanism that underlies the increase in responsiveness to ACh, it is suggestive to propose an increased endothelial NO bioavailability in mesenteric artery of trained rats.

As expected, the non-selective NOS inhibitor markedly decreased the endothelium-dependent relaxation induced by ACh and potentiated the Phe-induced vasoconstriction. Surprisingly, the magnitude of these changes was higher in RT animals. As previously demonstrated by our group, acute high intensity RT was also able to activate MAPK/ET-1 pathway, a vasoconstrictor mechanism; however, taking into account the difference between the intensity and acute effect, we suggest that this mechanism might be involved in the potentiated vascular contractile response found in RT rats under NOS inhibition. Moreover, these findings support the notion of the higher dependence of NO bioavailability in mesenteric artery of trained rats.

Confirming our hypothesis, basal production of NO was higher in mesenteric arteries of RT rats. It has been extensively described the pivotal role of eNOS-derived NO from endothelial cells for vascular relaxation (Förstermann and Sessa, 2012). However, although less studied, nNOS-derived relaxing factors such as, NO and H2O2, also contribute to endotheliumdependent relaxation (Capettini et al., 2010). Here, we indirectly test the whole constitutive NOS activities in intact-endothelium mesenteric arteries monitoring the additional synthesis of NO when stimulated with ACh. Hence, a marked increase in AChinduce NO synthesis were found; however, LI-RT potentiated the production of NO. Therefore, to unequivocally demonstrate the involvement of constitutive NOS in the vascular effects induced by LI-RT, protein expression analysis demonstrated an upregulation of eNOS and nNOS in mesenteric artery of RT rats. Previous study has demonstrated the participation of eNOS and nNOS in the control of vascular tone (Nangle et al., 2004), however, the role of nNOS in the endothelial dysfunction and hypertension is extremely limited. Thus, this study provides important insights related to the understanding of vascular adjustments induced by RT and, although not investigated, we suggest that nNOS/NO axis might be involved in the therapeutic actions of resistance exercise.

Importantly, the vascular bed here studied regulates about 20% of total blood flow in the body and changes in its vascular perfusion can represents significant alterations in total vascular peripheral resistance (Blanco-Rivero et al., 2013). Therefore during exercise, mesenteric arteries, suffer a large decrease in the blood flow, described as an intensity-dependent phenomenon known as reactive hyperemia (Joyner and Casey, 2015). Although recommended by several healthcare guidelines, intracellular mechanisms related with resistance exercise response is still largely unexplored. On the other hand, several studies have proposed eNOS/NO, Pi3K/Akt and AMPK signaling as the main intracellular pathways involved the protective cardiovascular actions of the aerobic exercise (McMullen et al., 2007; Wang et al., 2010; Cacicedo et al., 2011). Taken together, the present study added a valuable information about the systemic adaptation induced by LI-RT describing the importance of the expression and functionality of eNOS and nNOS in the vascular adjustments.

Studies have shown that NO can influence autonomic control function by several mechanisms (Patel et al., 1996; Lin et al., 2007; Schultz, 2009). Two important mechanisms postulated are: 1. NO can increase arterial baroreflex sensitivity and 2. NO in nucleus tractus solitarii causes sympathoinhibitory effect, blocking the rostral ventrolateral medulla (Schultz, 2009). In addition, it is known that arterial baroreflex and vascular sympathetic activity are associated with increased NO bioavailability in vascular beds (Kolo et al., 2004; Gamboa et al., 2007; Conti et al., 2013). Although we utilized indirect measurements of autonomic control, our results showed a higher spontaneuous baroreflex sensitivity and lower cardiac and vascular sympathetic modulation in LI-RT rats. Interestingly, these mechanisms may contribute to a high NO bioavailability, inhibiting the metabolization and release of noradrenergics neurotransmitters on vascular sympathetic nerve endings (Greenberg et al., 1990; Daveu et al., 1997; Kolo et al., 2004). Gmitrov (2015), observed other important effect of NO-baroreflex axis, demonstrated by the baroreflex-mediated increament in sensitivity to NO. Taken all together, we belive that baroreflex-NO axis works in a state of positive feedback promoting a decrease in heart and vascular sympathetic modulation and increasing cardiovascular safety.

In summary, our results suggest that LI-RT promotes a decrease in the blood pressure associated with enhanced Achinduced vasorelaxation response in rats mesenteric artery. More importantly, our results clearly show that these changes are followed by an upregulation of eNOS and nNOS leading to elevated NO bioavailability in vascular tissue. Besides of that, spontaneuous baroreflex sensitivity improvement followed

### REFERENCES


by reduction of cardiovascular sympathetic modulation also contribute to lower values of basal blood pressure and bradycardia. Such evidence confirm that LI-RT is able to promote systemic cardiovascular adjustments with no evidences of cardiovascular overload, being considered safer to be applied as a non-pharmacological strategy to health maintenance.

### AUTHOR CONTRIBUTIONS

FM, TM, VM, MM, TS, MS, LO, RS, and Rd performed experiments, data analyses and drafted the manuscript. SL, MRVS, AB, and VS designed the study and contributed to the manuscript. All authors approved the final version of the manuscript.

### FUNDING

This study was supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil)— Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Brazil) and Fundação de Apoio à Equipe e à Inovação Tecnológica do Estado de Sergipe (FAPITEC/SE—Brazil).

mouse aorta. Am. J. Physiol. Heart Circ. Physiol. 301, H1255–H1265. doi: 10.1152/ajpheart.01279.2010


measurement, physiological interpretation, and clinical use. Circulation 93, 1043–1065. doi: 10.1161/01.CIR.93.5.1043


mesenteric arterial bed. Am. J. Physiol. Heart Circ. Physiol. 286, H296–H303. doi: 10.1152/ajpheart.00668.2003


comparison between normotensive and hypertensive men. Scand. J. Med. Sci. Sports 25, 486–494. doi: 10.1111/sms.12280


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer JC and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review

Copyright © 2016 Macedo, Mesquita, Melo, Mota, Silva, Santana, Oliveira, Santos, Miguel dos Santos, Lauton-Santos, Santos, Barreto and Santana-Filho. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

## Different Anti-Contractile Function and Nitric Oxide Production of Thoracic and Abdominal Perivascular Adipose Tissues

### Jamaira A. Victorio<sup>1</sup> † , Milene T. Fontes 2 †, Luciana V. Rossoni <sup>2</sup> \* and Ana P. Davel <sup>1</sup> \*

*<sup>1</sup> Department of Structural and Functional Biology, Institute of Biology, University of Campinas, Campinas, Brazil, <sup>2</sup> Vascular Physiology Lab, Department of Physiology and Biophysics, Institute of Biomedical Sciences, University of São Paulo, São Paulo, Brazil*

### Edited by:

*Camille M. Balarini, Federal University of Paraíba, Brazil*

### Reviewed by:

*Zhihong Yang, University of Fribourg, Switzerland Agata Lages Gava, Universidade Federal do Espírito Santo, Brazil*

### \*Correspondence:

*Luciana V. Rossoni lrossoni@icb.usp.br Ana P. Davel anadavel@unicamp.br*

*† These authors have contributed equally to this work.*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *11 May 2016* Accepted: *27 June 2016* Published: *12 July 2016*

### Citation:

*Victorio JA, Fontes MT, Rossoni LV and Davel AP (2016) Different Anti-Contractile Function and Nitric Oxide Production of Thoracic and Abdominal Perivascular Adipose Tissues. Front. Physiol. 7:295. doi: 10.3389/fphys.2016.00295* Divergent phenotypes between the perivascular adipose tissue (PVAT) surrounding the abdominal and the thoracic aorta might be implicated in regional aortic differences, such as susceptibility to atherosclerosis. Although PVAT of the thoracic aorta exhibits anti-contractile function, the role of PVAT in the regulation of the vascular tone of the abdominal aorta is not well defined. In the present study, we compared the anti-contractile function, nitric oxide (NO) availability, and reactive oxygen species (ROS) formation in PVAT and vessel walls of abdominal and thoracic aorta. Abdominal and thoracic aortic tissue from male Wistar rats were used to perform functional and molecular experiments. PVAT reduced the contraction evoked by phenylephrine in the absence and presence of endothelium in the thoracic aorta, whereas this anti-contractile effect was not observed in the abdominal aorta. Abdominal PVAT exhibited a reduction in endothelial NO synthase (eNOS) expression compared with thoracic PVAT, without differences in eNOS expression in the vessel walls. In agreement with this result, NO production evaluated *in situ* using 4,5-diaminofluorescein was less pronounced in abdominal compared with thoracic aortic PVAT, whereas no significant difference was observed for endothelial NO production. Moreover, NOS inhibition with L-NAME enhanced the phenylephrine-induced contraction in endothelial-denuded rings with PVAT from thoracic but not abdominal aorta. ROS formation and lipid peroxidation products evaluated through the quantification of hydroethidine fluorescence and 4-hydroxynonenal adducts, respectively, were similar between PVAT and vessel walls from the abdominal and thoracic aorta. Extracellular superoxide dismutase (SOD) expression was similar between the vessel walls and PVAT of the abdominal and thoracic aorta. However, Mn-SOD levels were reduced, while CuZn-SOD levels were increased in abdominal PVAT compared with thoracic aortic PVAT. In conclusion, our results demonstrate that the anti-contractile function of PVAT is lost in the abdominal portion of the aorta through a reduction in eNOS-derived NO production compared with the thoracic aorta. Although relative SOD isoforms are different along the aorta, ROS formation, and lipid peroxidation seem to be similar. These findings highlight the specific regional roles of PVAT depots in the control of vascular function that can drive differences in susceptibility to vascular injury.

Keywords: thoracic aorta, abdominal aorta, nitric oxide, oxidative stress, perivascular adipose tissue, endothelium

## INTRODUCTION

Aortic atherosclerotic lesion and aneurysm are predominant in the abdominal rather than the thoracic portion of the aorta, suggesting inherent properties influencing regional aortic susceptibility to injury; however, the mechanisms involved in this susceptibility are not well understood. Despite being segments of the same artery, elastic features, and collagen contents vary along the aorta, which contribute to the physiological characteristics of pulse wave and blood distribution (Sokolis et al., 2008; Tsamis et al., 2013). Moreover, although the levels of endothelium-dependent relaxation induced by acetylcholine appear to be similar (Oloyo et al., 2012), the mechanisms involved in the alpha-adrenergic-mediated contraction might differ in the thoracic vs. the abdominal aorta (Lamb et al., 1994; Asbún-Bojalil et al., 2002).

It has been demonstrated that perivascular adipose tissue (PVAT), an adipose depot surrounding most arteries, plays an important role in vascular homeostasis (Szasz and Webb, 2012). PVAT of murine thoracic aorta secretes vasoactive substances, including adiponectin (Fésus et al., 2007), angiotensin 1–7 (Lee et al., 2009), leptin (Gálvez-Prieto et al., 2012), H2S (Fang et al., 2009), and a still unidentified adipocyte-derived relaxing factor(s) (ADRFs; Fésus et al., 2007; Schleifenbaum et al., 2010; Oriowo, 2015). These vasoactive factors can induce an anti-contractile effect through endothelial nitric oxide (NO) synthesis and release (Gálvez-Prieto et al., 2012) and/or through activation of vascular smooth muscle K<sup>+</sup> channels (Gao et al., 2007). More recently, it was also demonstrated that PVAT from the thoracic aorta expresses the endothelial isoform of NO synthase (eNOS; Araujo et al., 2015; Xia et al., 2016) and produces NO. PVAT-derived NO mediates relaxation of the adjacent thoracic aortic wall, indicating NO as a potential ADRF in this vessel (Xia et al., 2016).

Along the aorta, the phenotype of the PVAT is distinguished. Abdominal aorta PVAT resembles the phenotype of white adipose tissue (WAT), while fat surrounding the thoracic aorta shares the characteristics of brown adipose tissue (BAT; Police et al., 2009; Fitzgibbons et al., 2011). In addition, the expression levels of inflammatory genes and markers of immune cell infiltration are greater in abdominal PVAT than in thoracic PVAT, suggesting that the WAT phenotype is more pro-inflammatory and atherogenic than the BAT phenotype in PVAT (Padilla et al., 2013). Accordingly, PVAT from human coronary arteries with the WAT phenotype exhibits a pro-inflammatory profile with reduced adiponectin levels compared to BAT subcutaneous and peri-renal adipocyte depots (Chatterjee et al., 2009).

Although the anti-contractile and anti-inflammatory effects of thoracic aortic PVAT are well-known, the possible role of PVAT regulating the vascular reactivity and redox status of the abdominal aorta is still poorly understood. A previous study demonstrated that the anti-contractile effect of abdominal aortic PVAT is less pronounced when compared to the thoracic portion of the aorta (Watts et al., 2011). However, the mechanism involved in this regional difference remains unclear. This study aimed at investigating the comparative effects of thoracic vs. abdominal aortic PVAT on anti-contractile function, the oxidative profile, and NO synthesis and availability.

### Animals

All animal procedures were in accordance with the ethical principles for animal experimentation adopted by the Brazilian Society of Laboratory Animal Science (SBCAL/COBEA) and were approved by the Ethics Committee on Animal Use of the University of Campinas—UNICAMP (protocol number: 3523-1) and of the Institute of Biomedical Science at the University of Sao Paulo (protocol number 53, sheet 19, book 03).

Experiments were conducted in 3- to 4-month-old male Wistar rats (Multidisciplinary Center for Biological Research, UNICAMP) maintained at a constant room temperature (22– 24◦C) and light cycle (12:12 h light:dark) with food and water allowed ad libitum to all animals. At the time of the experiments, animals were euthanized under anesthesia (ketamine 80 mg/kg and xylazine 5 mg/kg; i.p.).

### Vascular Reactivity Study

Thoracic and abdominal aorta were isolated, placed in Petri dishes with cold Krebs-Henseleit solution (in mM: 118 NaCl, 4.7 KCl, 25 NaHCO3, 2.5 CaCl2-2H2O, 1.2 KH2PO4, 1.2 MgSO4- 7H2O, 11 glucose and 0.01 EDTA) and sectioned into 3-mm rings with or without PVAT. In some experiments, endothelial cells were removed by gently rolling the preparation back and forth with a needle. Rings were mounted at a resting tension of 1 g in an organ chamber bath (Panlab Harvard Apparatus, Cornellà - Barcelona, Spain) containing Krebs– Henseleit solution continuously aerated with 95% O2, 5% CO<sup>2</sup> (pH = 7.4; 37◦C) as previously described (Davel et al., 2006, 2008). Isometric tension was recorded using an isometric force transducer (MLT0420, AdInstruments) connected to an acquisition PowerLab 8/30 system for tension recording (LabChart 7, AdInstruments).

After a 30-min equilibration period, all aortic rings were initially exposed twice to 75 mM KCl, the first to check their functional integrity and the second to assess the maximum contractility. Following the wash, vascular reactivity was investigated with cumulative concentration-response curves to acetylcholine (0.1 nM – 10 µM) in vessels contracted with phenylephrine (1 µM). Endothelial denudation was confirmed by acetylcholine-induced relaxation <10%. After the samples

TABLE 1 | Potency (−LogEC50) and maximal response (Rmax) values to phenylephrine-induced contraction in thoracic and abdominal aortas with (+) or without (−) endothelium (E) and perivascular adipose tissue (PVAT).


*Data are expressed as the means* ± *SEM; the number of animals is indicated in the parenthesis. Student's t-test, P* < *0.05:* \**PVAT*+*/E*+ *vs. PVAT*−*/E*+*;* #*PVAT*+*/E*− *vs. PVAT*−*/E*−*.*

were washed, the contractile response to phenylephrine (1 nM – 10 µM for thoracic aorta and 1 nM – 100 µM for abdominal aorta) was evaluated. Some aortic rings with PVAT and without endothelium were incubated with NO synthase inhibitor L-NAME (100 µM) for 30 min prior to phenylephrine concentration-response curves.

At the end of the experiments, the length of each ring was measured. Vasoconstrictor responses to phenylephrine are expressed as mN/mm. Relaxation responses to acetylcholine are expressed as the percentage of relaxation of the contractile response induced by phenylephrine (1 µM).

## Nitric Oxide (NO) and Reactive Oxygen Species (ROS) Evaluation In situ

Thoracic and abdominal aortic tissues with their respective coated PVAT were isolated, placed in Petri dishes with cold Krebs-Henseleit solution, cleaned and sectioned into 3-mm rings. Next, in a dark chamber, aortic segments were incubated for 30 min with Krebs-Henseleit solution (pH = 7.4, 37◦C) plus 4,5-diaminofluorescein diacetate (DAF-2, 10 µM) or with dihydroethidium (DHE, 2 µM) for NO and ROS measurement, respectively, as previously described (Gil-Ortega et al., 2014). Some aortic segments were co-incubated with DAF-2 plus

### FIGURE 2 | Continued

figure and densitometric analysis is expressed as a fold change of THO expression at the bottom panel. (B) Top panels-Representative fluorographs of 4,5-diaminofluorescein diacetate (DAF-2) signal obtained in transverse sections of vessel walls and PVAT from THO and ABD aortic tissues in the absence (upper panel) and presence (lower panel) of L-NAME. Scale bar = 100 µm (20X objective). Bottom panel-Quantified NO availability, measured as DAF-2 fluorescence intensity in the endothelial layer and PVAT of THO and ABD aorta. (C) Concentration-response curves to phenylephrine in rat thoracic (left panel) and abdominal (right panel) aorta without (−) endothelium (E) and with (+) PVAT in the absence or presence of L-NAME. Data are expressed as the means ± SEM; the number of animals is indicated in the bars or in parenthesis. Student's *t*-test, *P* < 0.05: \*ABD vs. THO PVAT; Two-way ANOVA, *P* < 0.05: & PVAT+E− vs. PVAT+E− L-NAME.

L-NAME (1 mM) to evaluate the specific generation of NO from NO synthase or with DHE plus MnTMPyP (25 µM), a cell-permeant superoxide dismutase (SOD) mimetic, to evaluate superoxide anion formation. Subsequently, the aortic segments with PVAT were fixed in 4% paraformaldehyde for 4 h and then embedded in freezing medium (Tissue-Tek, Sakura Finetek, Torrance, CA). Transverse sections (20 µm thick) of frozen arteries were obtained on a cryostat. Digital images were collected on a microscope (Nikon, Chiyoda-ku, Tokyo, Japan) equipped with epifluorescence and fluorescein/rhodamine filters using a 20X objective. The images were analyzed using ImageJ software (NIH, Bethesda, MD, USA). NO availability was evaluated by DAF-2 mean optical density of the fluorescence in the endothelium and PVAT, and ROS production were analyzed based on the integrated density of the DHE fluorescence normalized by the number of nuclei labeled with ethidium bromide (EB-positive nuclei) in the vascular wall segment of the aorta and its respective PVAT.

### Western Blotting

Total protein extracts were obtained from abdominal and thoracic aortic samples and their respective PVAT depots. Tissues were homogenized in cold RIPA lysis buffer (Merck Millipore, Billerica, MA, USA) containing phenylmethylsulfonyl fluoride (1 mM PMSF), Na3VO<sup>4</sup> (1 mM) and protease inhibitor cocktail (2 µL/mL PIC, Sigma-Aldrich).

Protein extracts (75 µg) were separated by SDS–PAGE, and proteins were transferred to PVDF membranes (GE HealthCare, Little Chalfont-Buckinghamshire, UK) using a Mini Trans-Blot Cell system (Bio-Rad, Hercules, CA, USA) containing 25 mM Tris, 190 mM glycine, 20% methanol and 0.05% SDS. Membranes were blocked for 90 min at room temperature with 5% albumin in Tris buffer (10 mM Tris, 100 mM NaCl and 0.1% Tween 20). Membranes were then incubated overnight at 4◦C with the primary antibodies anti-eNOS (1:750; BD Transduction, Franklin Lakes, NJ, USA), anti-EC-SOD (1:1,000; Enzo Life Science, Farmingdale, New York, USA), anti-Mn-SOD (1:1,000; Enzo Life Science), anti-CuZn-SOD (1:5,000; Sigma Aldrich), and anti-4-hydroxynonenal (4-HNE; 1:2,000; Abcam, Cambridge, UK).

After washing, membranes were incubated for 90 min with a peroxidase-conjugated IgG antibody specific for the primary antibody used. Protein expression was detected with Pierce ECL Western Blotting Substrate (Thermo Scientific, Rockford, IL, USA), and the membranes were subjected to autoradiography (Amersham Hyperfilm ECL, GE Healthcare). The blots were digitized, and intensity was quantified using ImageJ 1.46p software (National Institutes of Health). Ponceau staining was used to normalize expression of the evaluated proteins in each sample.

### Drugs

Acetylcholine chloride, phenylephrine hydrochloride, L-NAME, and DAF-2 were purchased from Sigma-Aldrich CO (Saint Louis, MO, USA). MnTMPyP was purchased from Calbiochem (Merck Millipore). DHE was purchased from Invitrogen (Grand Island, NY, USA).

### Statistical Analysis

Results are expressed as the means ± SEM. Data were analyzed using GraphPad Prism 5.0 software (GraphPad Software Corp., USA). Concentration–response curves were analyzed using two-way ANOVA followed by the Bonferroni's post-test. For comparisons between abdominal and thoracic aortic samples in the same condition, the Student t-test was used. Values of P < 0.05 were considered significantly different.

## RESULTS

### PVAT Exerts an Anti-Contractile Effect in the Thoracic but Not the Abdominal Aorta

To determine the anti-contractile effects of PVAT in thoracic and abdominal aortic tissues, we performed concentration-response curves to phenylephrine in rings with (open symbols) or without (filled symbols) PVAT in intact (circle) or denuded (triangle) endothelium. Thoracic aortic rings with PVAT and intact endothelium (PVAT+E+) presented a significant reduction in potency and maximal response to phenylephrine when compared to rings without PVAT (PVAT−E+; **Figure 1A** and **Table 1**). Although the endothelium damage increased the phenylephrineinduced contraction (compare PVAT−/E+ vs. PVAT−/E−, **Figure 1A**), the anti-contractile effect of PVAT was still observed in endothelium-denuded rings. Thus, the presence of PVAT (PVAT+E−) in endothelium-denuded rings also reduced both the potency and maximal response to phenylephrine when compared to rings without PVAT and endothelium (PVAT−E−; **Figure 1A** and **Table 1**). In contrast, the presence of PVAT did not alter the phenylephrine-induced contraction in either intact or denuded endothelium abdominal aortic rings (**Figure 1B** and **Table 1**).

KCl-induced contractions were similar in both thoracic and abdominal aortic segments without (PVAT−E+; THO: 8.0 ± 0.5 vs. ABD: 8.7 ± 0.4 mN/mm) or with PVAT (PVAT+E+; THO: 9.4 ± 0.5 vs. ABD: 7.7 ± 0.5 mN/mm).

We also assessed the endothelium-dependent relaxation response to acetylcholine in thoracic and abdominal aorta.

### FIGURE 3 | Continued

evaluated at the basal level and in the presence of the SOD mimetic MnTMPyP. Scale bar = 100 µm (20X objective). Values of the integrated density of hydroethidine-positive (EB-positive) nuclei fluorescence were normalized to nuclei number, which was analyzed by DAPI staining in each sample. (B) Expression of 4-hydroxynonenal (4-HNE) adducts in vessel walls and PVAT from ABD and THO aorta. Representative blots and Ponceau S staining were demonstrated at the right panel and densitometric analysis is expressed as the fold change of THO expression (left panel). Data are expressed as the means ± SEM; the number of animals is indicated in the bars of the graph. Student's *t-*test.

As expected, endothelium damage blocked the vasodilatation induced by acetylcholine in both thoracic and abdominal aorta (**Figures 1C,D**). However, no effects of PVAT on the acetylcholine-induced relaxation were observed in either thoracic or abdominal aorta (**Figures 1C,D**).

### eNOS Expression and NO Availability Is Impaired in PVAT of the Abdominal Aorta

There was a non-significant trend (p < 0.07) toward reduced eNOS protein expression in the abdominal when compared to the thoracic aorta (**Figure 2A**), whereas abdominal PVAT showed a 60% reduction in eNOS expression compared with thoracic PVAT (**Figure 2A**). In accordance with these data, NO availability evaluated based on DAF-2 fluorescence was not significantly altered in the endothelium of abdominal vs. thoracic aorta (p < 0.053), but abdominal PVAT showed a decrease in NO availability of 34% compared with thoracic PVAT (**Figure 2B**). L-NAME incubation significantly reduced NO bioavailability in both the endothelium and PVAT of abdominal and thoracic portions of aorta (**Figure 2B**).

The role of NO-derived from PVAT on phenylephrineinduced contraction was evaluated by the L-NAME incubation in thoracic and abdominal PVAT+E− rings. L-NAME increased the potency and maximal response to phenylephrine in thoracic (Rmax: PVAT+/E− = 8.7 ± 0.3 vs. PVAT+/E− L-NAME = 10.5 ± 0.4 mN/mm, p < 0.05; −LogEC50: PVAT+/E− = 6.6 ± 0.12 vs. PVAT+/E− L-NAME = 7.0 ± 0.07, p < 0.05) but not in abdominal aorta (Rmax: PVAT+/E− = 10.9 ± 0.1 vs. PVAT+/E− L-NAME = 11.6 ± 0.4 mN/mm, p>0.05; −LogEC50: PVAT+/E− = 6.7 ± 0.15 vs. PVAT+/E− L-NAME = 6.9 ± 0.08, p > 0.05; **Figure 2C**).

### ROS Production and Lipid Peroxidation Did Not Vary along the Aorta

ROS production was detected based on DHE fluorescence (**Figure 3A**), and lipid peroxidation was evaluated based on the expression of 4-HNE adducts (**Figure 3B**). ROS production was almost fully inhibited by MnTMPyP in abdominal and thoracic aorta, suggesting superoxide as the main ROS evaluated in situ by DHE fluorescence in both the vascular wall and PVAT (**Figure 3A**). Both ROS production and lipid peroxidation were similar in abdominal and thoracic aortic tissues and PVAT (**Figures 3A,B**).

## Anti-Oxidative Profiles of Abdominal and Thoracic Aortic Tissue and PVAT

The protein expression levels of SOD isoforms were investigated in abdominal and thoracic tissues. EC-SOD did not differ between abdominal and thoracic aortic tissues and PVAT (**Figure 4A**), while reduced Mn-SOD expression was detected in abdominal compared with thoracic PVAT, without changes in the vascular wall (**Figure 4B**). In contrast, CuZn-SOD expression was increased in abdominal PVAT compared to thoracic PVAT, with no regional differences in the aortic wall (**Figure 4C**).

## DISCUSSION

The present results showed that: (1) the anti-contractile effect of PVAT on alpha-adrenergic-induced contraction observed in thoracic aorta is completely lost in the abdominal section, (2) there is a significant reduction in eNOS-derived NO production in PVAT but not in the endothelium of the abdominal vs. the thoracic aorta, and (3) ROS production and lipid peroxidation levels appear to be similar along the aorta, although the relative expression levels of SOD isoforms are different. Therefore, a minor contribution of NO in the abdominal vs. the thoracic adipose depot, rather than changes in redox status, might be involved in the functional regional differences along the aorta.

Anti-contractile effects of murine thoracic periaortic fat have been demonstrated in response to several contractile agonists, including phenylephrine, serotonin and angiotensin II (Löhn et al., 2002; Fésus et al., 2007; Gao et al., 2007; Lee et al., 2011; Sun et al., 2013; Araujo et al., 2015). Importantly, the endothelium contributes to the anti-contractile effect of PVAT in the response to angiotensin II (Gálvez-Prieto et al., 2012), while it only attenuates or does not alter its effect on the alphaadrenergic mediated contraction in the thoracic aorta (Gao et al., 2007; Lee et al., 2009, 2011). In agreement with these data, both endothelium and PVAT, independently of each other, restrain the contractile response to phenylephrine in the thoracic aorta. In contrast, in the abdominal aorta, only the endothelium exerted an anti-contractile effect in the phenylephrine-induced contraction. These data are consistent with previous data demonstrating anticontractile effects of endothelium-derived NO in both thoracic and abdominal aortic tissues (Kleinbongard et al., 2013), while the anti-contractile function of periaortic fat on the angiotensin II-induced contraction is impaired in the abdominal section (Watts et al., 2011).

Although the contractile response to phenylephrine varies along the aorta, endothelium-dependent relaxation levels induced by acetylcholine were similar in abdominal and thoracic aorta. Similar relaxation responses to acetylcholine were also observed in abdominal vs. thoracic sections of aorta in male Sprague-Dawley rats (Oloyo et al., 2012). In addition, PVAT did not play a role in the aortic relaxation response to acetylcholine, as previously demonstrated (Ketonen et al., 2010; Gálvez-Prieto et al., 2012).

### FIGURE 4 | Continued

and cytoplasmatic CuZn-SOD (C) in vessel walls and perivascular adipose tissue (PVAT) from THO and ABD aorta. Representative blots and Ponceau S staining were demonstrated at the upper panels of the figures and densitometric analysis is expressed as the fold change of THO expression at the bottom panel. Data are expressed as the means ± SEM; the number of animals is indicated in the bars of the graph. Student's *t-*test, *P* < 0.05: \*ABD vs. THO PVAT.

NO availability using DAF-2 fluorescence was previously demonstrated in PVAT of mouse mesenteric arteries (Gil-Ortega et al., 2010) and thoracic aorta (Xia et al., 2016). Here, for the first time, we demonstrated that NO availability in PVAT was significantly reduced in abdominal vs. thoracic aorta, whereas endothelium-derived NO levels were similar between the two portions of the aorta. The reduction in DAF-2 fluorescence in the abdominal periaortic fat was accompanied by a similar magnitude of reduction in eNOS expression in this tissue, while eNOS expression in the endothelium remained the same along the aorta. Reinforcing these results, it is interesting to observe that in PVAT preserved but endothelium denuded abdominal aortic rings the inhibition of NO synthesis did not change the phenylephrine-induced contraction, while it was increased in thoracic aortic rings. The cell-specific localization of the higher eNOS protein expression in thoracic aortic PVAT remains to be elucidated. Because microvessels are present throughout the thoracic PVAT, we cannot rule out endothelial cells from these vessels as a source of NO in PVAT, albeit the vast majority of eNOS positive cells in PVAT are adipocytes (Xia et al., 2016).

It is known that the abdominal aorta exhibits a different PVAT phenotype compared to the thoracic aorta: while abdominal PVAT exhibits a phenotype of pro-inflammatory fat, thoracic periaortic fat shares characteristics of BAT (Padilla et al., 2013). Interestingly, enlargement of lipid droplet morphology has been observed in thoracic periaortic fat from obese mice (Fitzgibbons et al., 2011). This change in phenotype induced by a high-fat diet is accompanied by a significant reduction in PVAT-derived NO production related to decreased eNOS phosphorylation (Xia et al., 2016). However, aerobic exercise training can overbrowning and enhances the eNOS expression in thoracic PVAT (Araujo et al., 2015). These studies support the hypothesis that physiological white fat depots surrounding the abdominal aorta may not exhibit a significant anti-contractile function due to less pronounced eNOS expression and NO production and/availability.

In the absence of PVAT and endothelium, abdominal and thoracic portions of the aorta exhibit similar raw forces in response to phenylephrine, suggesting both as major mechanisms involved in the regional contractile differences along the aorta. However, we cannot exclude other intrinsic factors. Although contraction in response to phenylephrine of rat abdominal and thoracic aorta is mediated via alpha1D-adrenergic receptor subtype (Asbún-Bojalil et al., 2002), prostanoids derived from smooth muscle cells contribute to contraction only in the abdominal aorta (Lamb et al., 1994). In addition, distal aortic segments are stiffer than proximal ones (Sokolis et al., 2008; Devos et al., 2015), although the mechanisms underlying this regional difference within the aorta is not fully established. Previous study has shown that impairment of eNOS activity is an important mechanism inducing vascular extracellular matrix remodeling in a model of abdominal aortic aneurism (Gao et al., 2012). Therefore, the substantial reduction in NO availability observed in the present study at the abdominal aorta might be a relevant factor implicated in the susceptibility of this part of the aorta to vascular injury.

It is well-known that NO availability is impaired by the spontaneous reaction of NO with superoxide anion causing the generation of peroxynitrite, a potent oxidant with potential cytotoxicity. Gil-Ortega et al. (2014) demonstrated that an enhanced superoxide generation in PVAT surrounding mesenteric arteries of obese mice is associated with a loss of the anti-contractile effect of PVAT on the contractile response of mesenteric arteries to noradrenaline. Therefore, we attempted to investigate whether oxidative differences along the aorta could be involved in the impaired anti-contractile function and NO availability noted in the abdominal aorta. Interestingly, no regional differences were noted in either superoxide generation, based on DHE fluorescence, or lipid peroxidation, assessed based on the presence of 4-HNE-protein adducts between the vascular wall and PVAT of abdominal and thoracic portions of the aorta. Early dysfunction of thoracic aortic PVAT was associated with increased lipid peroxidation, evaluated based on the presence of thiobarbituric acid reactive species (TBARS), in response to high fructose diet (Rebolledo et al., 2010). However, this change reflected those reported in abdominal adipose depots (Alzamendi et al., 2009), suggesting that they are the consequence of general rather than local oxidative stress involved in the PVAT control of vascular function.

The major enzymatic control of superoxide anion levels in the vessel wall is exerted by SOD isoforms (Faraci and Didion, 2004). SOD activity was also detected in PVAT of murine mesenteric arteries and thoracic aorta, playing an important role in the redox status of vascular wall (Rebolledo et al., 2010; Gil-Ortega et al., 2014). SODs convert superoxide anion into hydrogen peroxide, thereby protecting NO availability and signaling. Although the three isoforms of SOD catalyze the same reaction, they differ in both localization and their importance for vascular function. Here, we noted similar patterns of SOD expression between the abdominal and thoracic aortic walls. However, in comparison

### REFERENCES


with the thoracic PVAT, abdominal PVAT exhibited reduced protein levels of Mn-SOD and enhanced cytosolic CuZn-SOD. Interestingly, CuZn-SOD overexpression protects the aorta from lipid peroxidation and DNA fragmentation, whereas Mn-SOD heterozygous-deficient mice exhibited enhanced aortic lipid peroxidation and apoptosis, mechanisms of importance for the development of atherosclerosis (Guo et al., 2001). However, the functional significance of differences in vascular expression or activity in PVAT is still poorly understood; future studies will be necessary to address this question.

Taken together, the present results demonstrate that, compared with the thoracic aorta, the anti-contractile function of PVAT is impaired in the abdominal portion of the aorta through a reduction in eNOS-derived NO production without changes in ROS generation and lipid peroxidation. Knowing that aortic atherosclerotic lesions and aneurysms are predominant in abdominal rather than thoracic portions of the aorta, these findings highlight a new specific regional role of PVAT depots in the control of vascular function that may drive differences in susceptibility to vascular injury between these two portions of the aorta.

### AUTHOR CONTRIBUTIONS

Conceived and designed the study: LR, AD. Performed the experiments: JV, MF. Analyzed and interpreted the data: JV, MF, LR, AD. Wrote the manuscript: JV, MF, LR, AD.

### FUNDING

This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP grants 14/07947-6 and 14/20303-0) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq grant #447507/2014-1). AD and LR are research fellows from Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil).

### ACKNOWLEDGMENTS

The authors thank Gisele K. Couto for the excellent technical assistance in the experiments evaluating DAF-2 and DHE fluorescence.


aorta than in the abdominal aorta. J. Magn. Reson. Imaging 41, 765–772. doi: 10.1002/jmri.24592


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Victorio, Fontes, Rossoni and Davel. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Comparative mRNA and MicroRNA Profiling during Acute Myocardial Infarction Induced by Coronary Occlusion and Ablation Radio-Frequency Currents

Eduardo T. Santana<sup>1</sup> , Regiane dos Santos Feliciano2, 3, Andrey J. Serra<sup>2</sup> , Eduardo Brigidio<sup>3</sup> , Ednei L. Antonio<sup>4</sup> , Paulo J. F. Tucci <sup>4</sup> , Lubov Nathanson<sup>5</sup> , Mariana Morris <sup>5</sup> and José A. Silva Jr. <sup>3</sup> \*

*<sup>1</sup> Rehabilitation Department, Universidade Nove de Julho, São Paulo, Brazil, <sup>2</sup> Biophotonics Department, Universidade Nove de Julho, São Paulo, Brazil, <sup>3</sup> Medicine Department, Universidade Nove de Julho, São Paulo, Brazil, <sup>4</sup> Cardiac Physiology Department, Universidade Federal de São Paulo, São Paulo, Brazil, <sup>5</sup> Institute for Neuro-Immune Medicine, Nova Southeastern University, Fort Lauderdale, FL, USA*

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Deborah A. Scheuer, University of Florida, USA Maike Krenz, University of Missouri, USA*

> \*Correspondence: *José Antonio Silva jasjr2020@gmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *16 August 2016* Accepted: *07 November 2016* Published: *25 November 2016*

### Citation:

*Santana ET, Feliciano RdS, Serra AJ, Brigidio E, Antonio EL, Tucci PJF, Nathanson L, Morris M and Silva JA Jr. (2016) Comparative mRNA and MicroRNA Profiling during Acute Myocardial Infarction Induced by Coronary Occlusion and Ablation Radio-Frequency Currents. Front. Physiol. 7:565. doi: 10.3389/fphys.2016.00565* The ligation of the left anterior descending coronary artery is the most commonly used experimental model to induce myocardial infarction (MI) in rodents. A high mortality in the acute phase and the heterogeneity of the size of the MI obtained are drawbacks recognized in this model. In an attempt to solve the problem, our group recently developed a new MI experimental model which is based on application of myocardial ablation radio-frequency currents (AB-RF) that yielded MI with homogeneous sizes and significantly reduce acute mortality. In addition, cardiac structural, and functional changes aroused by AB-RF were similar to those seen in animals with MI induced by coronary artery ligation. Herein, we compared mRNA expression of genes that govern post-MI milieu in occlusion and ablation models. We analyzed 48 mRNAs expressions of nine different signal transduction pathways (cell survival and metabolism signs, matrix extracellular, cell cycle, oxidative stress, apoptosis, calcium signaling, hypertrophy markers, angiogenesis, and inflammation) in rat left ventricle 1 week after MI generated by both coronary occlusion and AB-RF. Furthermore, high-throughput miRNA analysis was also assessed in both MI procedures. Interestingly, mRNA expression levels and miRNA expressions showed strong similarities between both models after MI, with few specificities in each model, activating similar signal transduction pathways. To our knowledge, this is the first comparison of genomic alterations of mRNA and miRNA contents after two different MI procedures and identifies key signaling regulators modulating the pathophysiology of these two models that might culminate in heart failure. Furthermore, these analyses may contribute with the current knowledge concerning transcriptional and post-transcriptional changes of AB-RF protocol, arising as an alternative and effective MI method that reproduces most changes seem in coronary occlusion.

Keywords: myocardial infarction, RNA, microRNAs, rats, Wistar, methods

## INTRODUCTION

The ligation of the left anterior descending coronary artery is a classical experimental model to induce myocardial infarction (MI) in rodents. However, this method showed a high mortality rate in the acute phase post-MI and large heterogeneity of the MI size that are recognized drawbacks of this model. Antonio et al. (2009) developed a new MI experimental protocol, which is based on application of ablation radio-frequency currents (AB-RF) in the myocardium. This technique, initially used as a heart failure (HF) method, yielded MI with homogeneous sizes, and significantly reduces acute mortality rate. Although cardiac structural and functional changes aroused by AB-RF were similar to those seen in animals with MI induced by coronary artery ligation (Antonio et al., 2009), extensive comparison of molecular signature of these models is yet unknown.

There are strong evidences of several upregulated, downregulated, or stable mRNA expression during MI and HF progression. Many authors assessed an expression pattern in different MI phases and issues as cardiomyocytes responses to remodeling were addressed in different studies, focusing in differentially expressed genes (Dhalla et al., 2012; Sofia et al., 2014; Katz et al., 2016; Kirk and Cingolani, 2016). These changes may occur at multiple levels, including in transcription regulation, epigenetic processes, mRNA stability, and more. MicroRNAs (miRs), a specific group of 18–24 nucleotides non-coding RNAs are compelling regulators that can both destabilizing and inhibiting translation of mRNAs (Kukreja et al., 2011).

To compare these two MI protocols, we performed the evaluation of mRNA expression of nine different signal transduction pathways (cell survival, cell cycle, oxidative stress, apoptosis, molecules related to calcium signaling, cell growth factors, transcription factors, angiogenesis, and inflammation) in rat left ventricle with MI generated by experimental protocols of coronary occlusion and AB-RF. Therefore, along with identification of differentially expressed genes, we performed a large-scale microRNA identification in these MI protocols. The mRNA and microRNA transcriptomes analyses may help to draw feasible molecular mechanisms associated with a particular MI model that seldom would have functional significance alone. Thus, AB-RF due its lack of complexity might be an effective and fast method to achieve MI.

## METHODS

### Animals

Sixty-five male Wistar rats (∼230 g) were used in the present work. The animals were cared for in compliance with the Principles of Laboratory Animal Care, as formulated by the National Institutes of Health (National Institutes of Health publication no. 96-23, revised in 1996). The protocol was approved by the Ethics Committee for Animal Care and Use at the Universidade Nove de Julho, in Sao Paulo, Brazil (034/2012). The animals were randomized into three groups: animals that undergo to left coronary descendent artery ligation (Occlusion, n = 37); animals that undergo to left ventricle RF ablation

## Anterior Descending Coronary Artery Ligation

This MI method was based on work by Johns and Olson (1954), with minor adaptations. This technique is being routinely applied to rat in our laboratory (Antonio et al., 2009; Manchini et al., 2014). Rats were anesthetized with 4% halothane inhalation, intubated and mechanically ventilated with positive pressure in rodent ventilator (Harvard Model 683, Holliston, MA, USA). After trichotomy, lateral thoracotomy was performed at the place where the heart impacts on palpation. With the animal in the supine position was made 2 cm incision in the skin and dilatation of the pectoral and intercostal muscles with the help of Kelly curved forceps. After dilatation of the intercostal muscles, the ribs were isolated with the help of Kelly forceps and retractor Stevenson adapted. Then it was held a pericardiotomy and exposure of the heart to visualize the anterior descending coronary artery (ADCA).

For "MI generation, ADCA was occluded to ∼3 mm from the origin of the aorta through 5.0 nylon suture. After being checked the results of the suture, the retractor was removed, lung hyperinflation is promoted and the thorax was closed by purse string suture previously prepared around the incision edges. Postoperative care as analgesia (meperidine, 20 mg/kg, SC) and search for signs of anorexia, fever, vomiting, or abnormal respiration were conducted in all experimental animals.

### Radio-Frequency Ablation

Under 4% halothane anesthesia and immobilization in the supine decubitus position, the left thoracotomy was performed in the fourth intercostal space. The ribs were separated by retractors. After pericardium opening, the electrode (forceps) was placed in position to gently embrace the heart, and the catheter tip was placed on the LV anterolateral wall, perpendicularly to the tissue. AB-RF lesions (one ablation/rat) were achieved using a modified unipolar mode, following the procedure used by Antonio et al. (2009) and Dos Santos et al. (2013). Briefly, a custom-made catheter with a single electrode located at its tip was used to deliver RF energy against an indifferent electrode with a large area. The catheter tip was a single aluminum domeshaped electrode (similar to a round domed screw head), 4.5 mm in diameter and 4.0 mm in length. This electrode was connected to an electrically-insulated flexible coaxial cable that was able to deliver very high frequency currents. A copper plate (14.6 mm) was located at the posterior aspect of the heart. Special steel forceps designed to support the heart during AB-RF was used as the indifferent electrode. The distal end of the forceps took the shape of 2 small shells (0.9 cm in diameter). Given the large surface of these shells compared with the rat heart, energy could be delivered to the myocardium without a significant rise in impedance. Thus, the cable was connected to the proximal end of the forceps.

A commercially available RF generator (model TEB RF10; Tecnologia Eletronica Brasileira Ltda, Sao Paulo, Brazil) was used to create RF-Ab lesions. RF current (1000 KHz) was delivered at constant power (12 watts) for 12 s. Power and impedance was monitored during each application and the mean values were recorded. An automatic power output shut down was triggered if impedance exceeded 200 ohms. The damaged tissue was characterized by the presence of a clear white disk-shaped region of coagulation necrosis that appeared immediately after current application. Then, the heart was instantly returned to the thorax, pulmonary hyperinsufflation was performed, and a previously made pursestring suture was used to close the chest. A continuously monitored D2 electrocardiogram was performed during the procedure. Ventricular Ab protocol was always lower than 2 min and postoperative care was performed as stated previously.

### Transthoracic Doppler Echocardiograms

Animals from occlusion and ablation groups undergo to ECHO examination to evaluate the infarct size 3 days after MI induction using a HP SONOS 5500 (Philips Medical System, Andover, MA) with a 12-MHz transducer at a depth of 2 cm, according to previous reports (Antonio et al., 2009; Silva et al., 2014). Briefly, under 4% halothane anesthesia, 2-dimensional and M-mode images from the parasternal longitudinal, transverse, and apical views were obtained and recorded on a 0.5-inch videotape. The imaging analysis and measurements were performed offline. MI was detected by ECHO on the basis of subjective identification of akinesis or dyskinesis. Measurements of end-diastolic (LVAd) and end-systolic (LVAs) LV transverse areas were performed in the 3 transverse planes (basal, medium, and apical) and LV systolic function was estimated by the fractional area change (FAC: LVAd—LVAs/LVAd × 100). Diastolic function was assessed by calculating the peak E and A blood flow mitral velocities and the E/A ratio. Sample volume of the pulsed wave Doppler was positioned at the tips of the mitral valve leaflets in an apical 4-chamber view. Only animals bearing infarct sizes higher than 40% of LV in both procedures were included in the experimental groups. An experiment cardiologist made a blind analysis of all echocardiograms.

### mRNA Quantification Analysis

Animals were euthanized by decapitation 1 week after each MI protocols. An area from a remote MI site (septum) was collected from each animal and used for comparative messenger RNA expression using TaqMan microarray plates to identify mRNAs that were differentially expressed among the different groups of rat heart tissues; control, occlusion, and ablation MI samples.

Samples weighing between 0.2 and 0.5 g, were homogenized in Trizol <sup>R</sup> Reagent for extracting RNA according to manufacturer's instructions. Then, extraction was performed with equal volume mixture containing phenol-chloroform-isoamyl alcohol in a ratio of 25:24:1, followed by precipitation with 0.2 M sodium acetate and 2 volumes of absolute ethanol. The precipitated RNA was washed with 70% ethanol to eliminate the phenol and salt residues, and solubilized in DEPC-water. The total RNA extracted was treated with deoxyribonuclease 10 U RNase-free for 1 h at 37◦C. The concentration of total RNA samples was determined spectrophotometrically at a wavelength of 260 nm. The RNA integrity was verified after gel electrophoresis in 1% agarose containing 0.5µg/ml ethidium bromide, irradiation with ultraviolet light. The quantification of the total RNA samples was made using the NanoDrop ND-2000 spectrophotometer apparatus (NanoDrop Products, Wilmington, DE, USA) where 1U A260 RNA corresponds to 40µg/ml. Samples were only used free of contaminants (A260/A230 ∼ 1.8) and protein (A260/A280 = 1.8–2.0). The integrity of total RNA was assessed by observing the proportion of bands related to 18S and 28S rRNA on agarose gel electrophoresis 1% stained with SYBR <sup>R</sup> Safe (Life technologies). To eliminate genomic DNA contamination of the samples, 1µg of total RNA (8 µL) was incubated with 1 unit (1 µL) of DNase I / RNase Free—(Invitrogen, USA) in the presence of one solution containing 20 mM Tris HCl, pH 8.4 and 2 mM MgCl2 for 15 min at 37◦C, followed by incubation at 65◦C for 10 min to inactivate the DNAse I.

After the above treatment, reverse transcription reaction was performed (RT-PCR) for cDNA synthesis. At 1µg total RNA treated were added 2µL of incubation buffer (50 mM KCl, Tris-HCl pH 8.4, 20 mM MgCl2, 2.5 mM), 1 unit of reverse transcriptase (1µl) (Invitrogen) 2µl Randon Primer (Invitrogen) 0.8µL of oligonucleotides (dNTPs, 100 mM) and 4.2µL of ultrapure H2O to a final reaction of 20µL. The samples were then subjected to the following incubations: 25◦C for 10 min, 37◦C for 120 min, 85◦C for 5 min. After the reaction, the cDNA samples were kept at −20◦C for further Real-time PCR analysis.

The reaction of polymerization chain in real time (Real-Time PCR) combines PCR amplification with automated fluorescent detection. Amplification and data acquisition were performed with TaqMan probe using Abi Prism 7500 Fast equipment (Applied Biosystems) as previously described (Silva et al., 2014). The fluorescence excitation capture was performed on each PCR amplification cycle, providing a real-time quantification of the sequences of genes of interest. The protocol used for the Real time PCR reactions was: 1.0µl of cDNA was added 5µL of Solution TaqMan Fast Universal Master Mix 2X (Applied Biosystems, USA) and sufficient water to 10µL reaction in each well of the 96 well-plate. The samples were applied in duplicate, and then incubated at 95◦C for 20 s, and passed through 40 thermal cycles at 95◦C for 3 s, 60◦C for 30 s. TaqMan Array plates 96— Well-Plate FAST been customized by Applied Biosystems/Life Technologies according to the chosen genes.

All reactions were subjected to the same conditions of analysis and normalized by ROX passive reference dye signal for correction of fluctuations in reading due to changes in volume and evaporation over the reaction. The results, expressed in Ct value refer to the number of PCR cycles required for the fluorescent signal reaches the detection threshold. The differentially expressed genes were normalized by the expression level of the housekeeping genes GAPDH or 18S subunit ribosomal RNA, which expression was shown to remain unchanged under experimental conditions. Fast SDS 1.4 software (Applied Biosystems) was used for data processing. The 1Ct values of the samples were determined by subtracting the average Ct value of the mRNA of the target gene from the average Ct value of the GAPDH housekeeping gene or 18S rRNA. The 2−11Ct parameter was used to express the relative expression data.

## NanoString nCounter Assay for miRNA Profiling

We profiled miRNAs using NanoString nCounter-miRNA expression analyses (NanoString Technologies, Seattle, Washington) that uses molecular barcodes and single-molecule imaging to detect and count RNAs without PCR amplification to analyze the global expression of miRNAs in the same total RNA samples used for mRNA quantification. Those samples were isolated from septum 1 week after both MI procedures of occlusion and ablation. The method quantifies 423 endogenous rat miRNAs (based in miRBase, version 17) of each septal specimen remote to MI area in two plates that accommodated the 24 samples (n = 8/group).

Raw data were processed using the NanoStringNorm R package (Waggott et al., 2012). The raw data were log2 transformed and normalized using the mean of the six positive controls, which were used to calculate a scaling factor in each column (lane/sample) as suggested by NanoString. The internal positive spike controls were present in each reaction to account for minor differences in hybridization, purification, or binding efficiencies. The data were further background corrected by subtracting the mean of the six negative controls followed by quantile normalization.

Total RNA (100 ng) was used as input for nCounter miRNA sample preparation reactions and the reactions were performed, as per the manufacturer's instructions (NanoString Technologies). Small RNA sample preparation involves the ligation of a specific DNA tag onto the 3′ end of each mature miRNA. These tags normalize the melting temperatures of the miRNAs and provide a unique identification for each miRNA species in the sample. Excess tags were then removed, and the resulting material was hybridized with a panel of miRNA: tag-specific nCounter capture and barcoded reporter probes. Hybridized probes were then purified and immobilized on a streptavidin-coated cartridge using the nCounter Prep Station (NanoString Technologies). Data collection was carried out on the nCounter Digital Analyzer (NanoString Technologies) following manufacturer's instructions to count individual fluorescent barcodes and quantify target RNA molecules present in each sample. For each assay, a high-density scan (600 fields of view) was performed.

The nCounter miRNA data was also confirmed through cross-platform validation in 10 randomly selected study samples using the TaqMan probes on the 7500 Fast Real Time PCR System. Furthermore, differentially expressed miRNAs were also quantified by Real-time PCR (data not shown). The average Pearson correlation coefficient was 0.70 (0.60–0.80) between the two platforms, thus confirming the robustness of the nCounter platform.

### Statistical Analysis

Data were analyzed with GraphPad Prism software (La Jolla, CA, USA). The Shapiro-Wilk test was used to verify normality and error variances. Results were evaluated using two-way ANOVA complemented by Tukey test was used to detect differences between three groups at sample with normal distribution. A p ≤ 0.05 was considered significant and the results are expressed as mean ± standard error of the mean (SEM).

## RESULTS

### Mortality Rate Comparison

Twelve of 37 rats (32%) died post occlusion-induced MI, mainly because large infarction sizes. In contrast, only one rat died after RF-Ab of unspecific cause. This animal showed the commitment of 42% of the LV area, dismissing the idea of extensive Ml above average as a cause of death.

## TaqMan Microrrays Assays

Addressing to changes in remote MI mRNA expression after coronary occlusion compared to control, we observed 38 differentially expressed mRNAs involved in extracellular matrix remodeling [collagen III 1a (COLIII1a):1.59-fold; collagen I 1a (COLI1a): 3.09-fold; tenascin: 3.32-fold and matrix metallopeptidase 9: 2.41-fold], inflammation mediators [interleukin 6 (IL6): 1.06-fold; tumor necrosis factor receptor 1a (TNFRSFLA): 1.40-fold and tumor necrosis factor, alpha (TNFa): 1.13-fold] and a survival marker (AKT1: 1.65-fold) that were found to be upregulated after coronary occlusion compared to control.

Genes involved in apoptosis signaling [Bcl2 associated X protein (Bax): 1.10-fold and p 53: 1.11-fold] were found to be upregulated, whereas mitogem activated protein kinase 14 (MAPK14) and mitogem activated protein kinase 1 (MAPK1) were found to be downregulated (−1.07- and −1.09-fold, respectively) after coronary occlusion compared to control. Among oxidative stress genes, two were found to be upregulated after coronary occlusion [glutathione peroxidase 4 (GPX4): 1.34-fold and heat shock protein 12b (HSPA12B): 1.34-fold], whereas catalase and superoxide dismutase 1 were found to be downregulated by −1.03- and −1.04-fold, respectively.

Expressions of hypertrophy biomarkers [endothelin: 1.78-fold; natriuretic peptide A and B: 4.48- and 2.18-fold, respectively; protein kinase C, alpha: 1.13-fold and nuclear factor of activate T-cells cytoplasmic, calcineurin—dependent 3 (NFATC3): 1.10-fold] were found to be upregulated after coronary occlusion compared to control. Interestingly, calcineurin like EF hand protein 2, angiotensin converting enzyme (ACE1) and angiotensin II receptor type 1a (−1.77-, −1.05- and −1.19-fold, respectively) were found to be downregulated after coronary occlusion.

Two myofilamentar protein expressions were analyzed, myosin, heavy chain beta was found to be upregulated (1.59-fold) whereas myosin, heavy chain alpha was found to be downregulated (−1.52-fold) compared to control. The genes involved in calcium kinetics [phospholamban: 1.10 fold and solute carrier family 8 (sodium/calcium exchange), member 1: 1.38-fold] were found to be upregulated after coronary occlusion compared to control. However, also involved in calcium kinetics, ryanodine receptor 2, ATPase, calcium transporting cardiac muscle, slow twitch 2 (ATP2A2) and calsequestrin 2 (−1.25, −1.06-, −1.05-fold, respectively) were found to be downregulated after MI occlusion-induced. Vascular endothelial growth factor A (VEGFA, −1.50-fold) was found to be downregulated after occlusion when compared to control.

Among metabolism-related genes [hexoquinase I: 1.59-fold; uncopling protein 2: 1.56-fold; solute carrier family 2 (facilited glucose transporter) member 1: 1.08-fold and taffazin: 1.10-fold] were found to be upregulated after coronary occlusion. Apart from these, phosphofructokinase expression (−1.53-fold) was found to be downregulated after coronary occlusion compared to control.

Comparing remote MI mRNA expression after ablation protocol to control, we found 38 differentially expressed genes. Extracellular matrix remodeling genes were found to be upregulated [collagen III 1a (COLIII1a):1.65-fold; collagen I 1a (COLI1a): 3.87-fold; tenascin: 6.68-fold and transforming growth factor, beta 1 (TGF1b): 1.57-fold] when compared to control. Also, inflammation mediators genes [interleukin 6 (IL6): 1.17 fold and tumor necrosis factor, alpha (TNFa): 1.83-fold] showed increased expression after myocardial ablation in comparison with control.

Furthermore, mRNA expressions of apoptosis signaling genes [Bcl2 associated × protein (Bax): 1.30-fold, p 53: 1.82 fold, mitogem activated protein kinase 1 (MAPK1): 1.92-fold], survival effector (AKT1): 2.25-fold, calcium kinetics [ryanodine receptor 2: 2.29-fold; ATPase, calcium transporting cardiac muscle, slow twitch 2 (ATP2A2): 1.92-fold; calsequestrin 2: 1.93-fold; phospholamban: 3.36-fold and solute carrier family 8 (sodium/calcium exchange), member 1: 3.29-fold], metabolismrelated genes [hexoquinase I: 2.74-fold; uncopling protein 2: 2.35-fold; solute carrier family 2 (facilited glucose transporter) member 1: 1.90-fold; phosphofructokinase: 1.18-fold; NADH dehydrogenase (ubiquinone) 1 alpha subcomplex 3 (NDUFA3): 2.27-fold and taffazin: 1.58-fold] were also found to be upregulated after ablation. In addition, myosin heavy chain beta (2.73-fold)] was found to be upregulated after ablation compared to control, however, the expression of the alpha isoform remained unchanged after ablation.

Among oxidative stress genes analyzed, two were found to be upregulated after ablation [glutathione peroxidase 4 (GPX4): 2.27-fold and catalase: 2.29-fold], whereas superoxide dismutase 1 was found to be downregulated by −1.22-fold when compared to control. In addition, expression of vascular endothelial growth factor A (VEGFA, −1.50-fold) was found to be downregulated after ablation when compared to control.

Most of the hypertrophy biomarkers [angiotensin converting enzyme (ACE1): 1.12-fold; angiotensin converting enzyme 2 (ACE2): 1.58-fold; angiotensin II receptor type 1a: 1.69-fold; endothelin: 2.22-fold; natriuretic peptide A and B: 1.91- and 2.21-fold, respectively; protein kinase C, alpha: 2.29-fold; nuclear factor of activate T-cells cytoplasmic, calcineurin—dependent 3 (NFATC3): 2.22-fold; insulin like growth factor (IGF1): 1.55 fold] were found to be upregulated after ablation compared to control. Among these protein kinase C, beta (−1.25-fold) was found to be downregulated after ablation.

Analyzing changes in cardiac mRNA expression in occlusion compared to ablation, we observed a smaller number of differentially expressed mRNA (nineteen genes) that were found to be differently expressed after occlusion than when comparison was performed against control). Genes involved in extracellular matrix remodeling [collagen III 1a (COLIII1a):1.25 fold; tenascin: 2.01-fold; matrix metallopeptidase 9: 2.03 fold and transforming growth factor, beta 1 (TGF1b): 1.42 fold], survival marker (AKT1: 1.37-fold), apoptosis signaling [mitogem activated protein kinase 14 (MAPK14): 1.21-fold], calcium kinetics [phospholamban: 6.16-fold; solute carrier family 8 (sodium/calcium exchange), member 1 (SLC8A1): 2.39-fold; calsequestrin 2: 2.04-fold and ATPase, calcium transporting cardiac muscle, slow twitch 2 (ATP2A2): 2.05 fold], metabolism-related genes [hexoquinase I: 1.72-fold; uncopling protein 2: 1.51-fold; solute carrier family 2 (facilited glucose transporter) member 1: 1.08-fold and taffazin: 1.75 fold and phosphofructokinase: 1.80-fold] were found to be upregulated after coronary occlusion when compared to ablation protocol.

None oxidative stress genes were differently expressed comparing both MI procedures. Regarding to inflammatory response genes, only tumor necrosis factor receptor, member 1a mRNA (TNFRSFLA, −1.05-fold) was found to be downregulated in occlusion when compared to ablation. Expressions of hypertrophy biomarkers [calcineurin like EF hand protein 2: 1.72-fold; angiotensin converting enzyme 2 (ACE2): 1.99-fold and insulin like growth factor 1 (IGF1):1.68-fold] were found to be upregulated after occlusion compared to ablation. Protein kinase C, gamma (−1.23-fold) was found to be downregulated after occlusion when compared to ablation. All data referring to Taqman Realtime PCR are summarized on **Table 1**.

### High Throughput miRNA Analysis

From 423 miRNAs, we only found eight differentially expressed miRNAs among experimental groups. This low detection support the idea that, besides particular differences in miRNA expression between occlusion and ablation procedures, only few miRNAs showed differentially expressed among total miRNA (<2% of total miRNA). Following, we highlighted the differentially expressed miRNAs detected on NanoString platform.

The data show that mir-221 presented a method-specific switch in expression; occlusion had higher levels of these miRNAs than ablation. Similarly, miR-34c and mir-93 expressions were higher expressed in occlusion at 1 week compared to control, and also in occlusion compared to ablation.

We also identified miRNAs with changes in expression only after occlusion. Specifically, mir-301 and mir-17-5p showed different expression levels after occlusion, while mir-301 expression decreased, we observed an increased in mir-17-5p expression compared to control.

The mir-9 expression detected was similar between occlusion and ablation and different when compared to control, whereas both diminished. Ablation showed diminished expressions of mir-542-5p and mir-1949 compared to control and occlusion. Together, these results show that there are few changes in miRNAs content between MI methods 1 week after infarction. All data regarding to miRNA quantification are presented on **Tables 2**–**4**.



*(Continued)*

### TABLE 1 | Continued


*Gene ID refers to NCBI database numbers. OC* × *SH refers to Occlusion group vs. Sham animals, AB* × *SH refers to Ablation group compared to Sham animals, OC* × *AB refers to the comparison of both MI protocols.* \**p* ≤ *0.05,* \*\*\**p* ≤ *0.001, unaltered refers to p* > *0.05. More details are provided in Methods Section.*

### DISCUSSION

Data from transcriptomes have opened a window for more integrated analysis of both qualitative and quantitative changes in global mRNA expression at a particular time point of any biological system (Agnetti et al., 2007). In the last decade, there have been several efforts to dissect out cardiac mRNA and proteome in a wide array of cardiovascular diseases (McGregor and Dunn, 2006) using different animal models or plasma samples collected from human patients (Seenarain et al., 2010; Haas et al., 2011; Silbiger et al., 2011; Drastichova et al., 2012; Chowdhury et al., 2013; Marshall et al., 2014; Petriz and Franco, 2014). While novel biomarkers were identified from some of these studies (Haas et al., 2011; Silbiger et al., 2011; Chowdhury et al., 2013), authors reported general alterations in the cardiac transcriptome and proteome profile that suggested changes in Ca2+ handling proteins (Seenarain et al., 2010), energy metabolism proteins (Jin et al., 2006; Meng et al., 2009), and mediators of apoptotic signaling (Drastichova et al., 2012; Marshall et al., 2014; Petriz and Franco, 2014).

Here, we attempted to achieve differentially expressed mRNA of remote area of MI between coronary occlusion and by radio frequency ablation in rats. Heretofore, we believe our study is the first report that compares differentially mRNA expression profile and large-scale miRNA analysis of these different MI procedures, in order to establish ablation-induced MI as an alternative experimental MI method, with the advantages of uniformity of MI sizes and low mortality rates. Surprisingly, we observed more mRNA expression similarities than differences between these protocols.

Extracellular matrix (ECM) components are important for mechanical support and effective functioning of the cardiovascular system. Therefore, alteration of the ECM may directly result in changes of mechanical properties and functional impairment. The ECM also plays significant roles in tissue remodeling in stress responses. Several extracellular matrix genes take part of the cardiac remodeling after MI. In tune with earlier reports (Roy et al., 2006; Manchini et al., 2014), mRNA modifications regarding to cardiac remodeling were observed in our experimental groups. mRNA expression analysis revealed significant upregulation of most genes after both occlusion and ablation compared to control. We found an increased tenascin (TNC) mRNA expression in both occlusion


TABLE 3 | MiRNAs that were altered significantly (p ≤ 0.05) after myocardial infarction induced by ablation compared to Sham.


TABLE 4 | MiRNAs that were altered significantly (p ≤ 0.05) after myocardial infarction induced by occlusion compared to ablation.


and ablation. Tenascin is sparsely detected in the normal adult myocardium, but reappears when the heart remodels its structure in response to pathologic insults, such as acute MI (Willems et al., 1996; Imanaka-Yoshida et al., 2001; Sato et al., 2006; Odaka et al., 2008), myocarditis (Imanaka-Yoshida et al., 2002; Sato et al., 2002; Morimoto et al., 2005), hibernation (Frangogiannis et al., 2002), ischemia-reperfusion (Taki et al., 2010), hypertensive cardiac fibrosis (Nishioka et al., 2007), chronic cardiac rejection (Franz et al., 2010), and some cases of dilated cardiomyopathy (DCM) (Tamura et al., 1996; Tsukada et al., 2009) closely associated with inflammation. Several studies suggest that TNC could help tissue reconstruction of the edge of the residual myocardium as a de-adhesion protein. TNC could loosen strong adhesion of cardiomyocytes (Imanaka-Yoshida et al., 2001, 2004) and upregulate the expression and activity of matrix metalloproteinases (MMPs) (Collins et al., 2004). In fact, expression of matrix metalloproteinase 9 was found to be induced after occlusion but remained unaltered after ablation, indicating that this particular increased expression may be consequent to TNC stimulation. Most matrix cellular proteins are minimally expressed in normal young adult hearts, but are markedly upregulated following cardiac injury. Matrix cellular proteins induced in the infarcted heart appear to serve as transducers of key molecular signals in cardiac repair and act as modulators of cell migration, proliferation, and adhesion (Frangogiannis, 2012).

Fibrosis is often considered to be the end inflammatory reaction, and both stimuli generated by occlusion and ablation were able to increase collagen I mRNA content. Regarding to collagen III, the mRNA expression remained unchanged. This data is in consonance with other reports that indicated an augmentation in expression of collagen fibers type I as the end product when tissue is repaired during heart failure in animal models and human patients (Stefanon et al., 2013; Yabluchanskiy et al., 2013), while collagen III is quickly synthetized and gradually switched to collagen I fibers. A high expression of TGF1b in occlusion group compared to ablation suggested that there would be different pathways activated after this procedure. TGF-β1 is a persistent stimulus in the chronic and inappropriate wound healing phase that is marked by hypertrophic scarring and eventual stiffening of the entire myocardium, ultimately leading to the pathogenesis of heart failure following MI (Zeglinski et al., 2016). Moreover, TGFβ1 is a key pro-fibrotic cytokine that is markedly elevated in experimental MI, and anti-TGF gene therapy mitigates cardiac remodeling by affecting cardiac fibrosis and infarct tissue dynamics (Okada et al., 2005).

In the infarcted myocardium, necrotic cardiomyocytes release danger signals, activating an intense inflammatory response. Inflammatory pathways play a crucial role in regulation of a wide range of cellular processes involved in injury, repair, and remodeling of the infarcted heart. Pro-inflammatory cytokines,

such as tumor necrosis factor α and interleukin 1, are markedly upregulated in the infarcted myocardium and promote adhesive interactions between endothelial cells and leukocytes by stimulating chemokine and adhesion molecule expression (Saxena et al., 2016).

In fact, we found that some genes involved in inflammation were upregulated after both occlusion and ablation. Upregulation of all these mRNAs after MI procedures could be due to an establishment of a secondary inflammation that accomplishes MI by necrosis and other cell death mechanisms as reported by earlier studies (Manchini et al., 2014; Saxena et al., 2016). Importantly, inflammation mediator expressions were mostly unaltered when occlusion and ablation were compared, indicating the same activation pathways, there was an increased mRNA expression of most interleukin analyzed in both procedures when compared to control. TNF receptor mRNA, for instance, showed increased expression in both procedures when compared to control, and presented a downregulation after occlusion when compared to ablation. This difference might indicate a variation in the time course and extension of stimulus that leads to inflammation. Several reports on the expression status of these mRNAs in occlusion-induced MI are available and there is a consensus that inflammation plays an important role after cardiomyocytes death (Bao et al., 2008; Zhu et al., 2015).

Our study also revealed an upregulation of Bax and p53 mRNA after both procedures compared to control, although both genes remained unaltered compared among procedures, corroborating that the MI protocols show the same mRNA expression pattern. It has been reported earlier that MI presented an important upregulation of mitochondria-related pro-apoptotic members to meet increased stimulation after MI procedures (Xu et al., 2015). In occlusion, we observed a downregulation of mitogen activated protein kinases 1 and 14 when compared to control. These two kinases have several functions, while MAPK1 act as an integration point for multiple biochemical signals, and are involved in a wide variety of cellular processes such as proliferation, differentiation, transcription regulation and development, the MAPK14 is member of p38 MAPK family, responsive to stress stimuli, such as cytokines, ultraviolet irradiation, heat shock, and osmotic shock, and are involved in cell apoptosis, differentiation and autophagy. These intracellular mitogen-activated protein kinases (MAPKs) signaling cascades play a key role in the pathogenesis of cardiac and vascular disease. ERK1/2, a member of the MAPKs, is concerned with the regulation of cell proliferation, differentiation, survival, and apoptosis. Although, most researchers have maintained that the phosphorylation of ERK1/2 played an important role in protecting against myocardial I/R injury (Clark et al., 2007), others have shown that ERK1/2 phosphorylation may aggravate myocardial cell injury (El-Mahdy et al., 2016). There is strong in vivo evidence that activation of the p38 mitogen-activated protein kinase (MAPK) family of stress-activated kinases exacerbates myocardial injury following prolonged ischemia (Clark et al., 2007). Pro-inflammatory cytokines such as IL-1 and tumor necrosis factor alpha (TNFα), which are potent stimuli for the p38 MAPK pathway, are elevated in the infarcted heart and appear to be detrimental for post-MI myocardial remodeling and progression to heart failure (Aukrust et al., 2005).

Oxidative stress-related mRNA presented markedly expression in occlusion and ablation when compared to control, although showed similar expression in comparison between MI procedures. Reactive oxygen species (ROS) such as superoxide anions (·O − 2 ) and hydroxyl radicals (·OH) cause the oxidation of membrane phospholipids, proteins, and DNA (McCord, 1985) and have been implicated in a wide range of pathological conditions including ischemia-reperfusion injury (Chen et al., 2001), neurodegenerative diseases (Mizuno et al., 1998) and aging (Trifunovic et al., 2004).

Under physiological conditions, their toxic effects can be prevented by scavenging enzymes such as superoxide dismutase (SOD), mitochondrial phospholipid hydroperoxide glutathione peroxidase (GPX4), and catalase as well as by other nonenzymatic antioxidants. However, when the production of ROS exceeds the capacity of antioxidant defenses, oxidative stress might have a harmful effect on the functional and structural integrity of biological tissue. ROS cause contractile failure and structural damage in the myocardium (Toussaint et al., 1993). The importance of oxidative stress is increasingly emerging with respect to a pathophysiological mechanism of LV remodeling responsible for heart failure progression.

We identified GPX4 as the only oxidative stress molecule that showed increased mRNA expression after both procedures compared to control. GPX gene overexpression inhibited the development of LV remodeling and failure after MI, which might contribute to the improved survival (Shiomi et al., 2004). These findings not only extended the previous observation that employed antioxidants, but also revealed the major role of ROS in the pathophysiology of myocardial remodeling. These effects were associated with the attenuation of myocyte hypertrophy, apoptosis, and interstitial fibrosis (Shiomi et al., 2004). Similarly, overexpression of the GPX gene attenuated myocardial remodeling and preserved diastolic function in diabetic heart (Matsushima et al., 2006). Therefore, therapies designed to interfere with oxidative stress by using GPX could be beneficial to prevent myocardial remodeling and failure.

Superoxide dismutase 1 is the primary mitochondrial antioxidant enzyme and is essential for maintaining normal cell development and function. Overexpression of the SOD1 gene has been shown to be beneficial in various animal models of cardiac diseases (Yen et al., 1996; Chen et al., 1998). Interestingly, we found to be a downregulation of SOD1 mRNA after MI procedures compared to control. SOD1 gene overexpression also elevated levels of myocyte catalase and mitochondrial GSH, which might also act together with SOD1 against oxidative stress (Khaper et al., 2003). Catalase gene was differently expressed in each procedure compared to control. The expression increase in ablation could be related to the lesion extension after MI induction (Tsutsui et al., 2009).

Upregulation of hypertrophic markers after MI suggests an activation of a mechanism that promotes enlargement of the heart chambers as a compensative strategy to pathophysiologic state. Hence, it could be suggested that differences concerning hypertrophy in ablation when compared to control is an indicative of severity of stimulus triggered by this MI procedure (Zannad et al., 2010; Gaggin and Januzzi, 2013). Ligation of coronary artery provokes an augmentation on hypertrophic markers, such as the cardiac natriuretic peptide A content, that showed significant upregulation when compared to ablation. The natriuretic peptides represent the gold standard for biomarkers in HF, and the understanding about their biology and their clinical use have both grown exponentially since their introduction. Structurally conserved across multiple species, a number of structurally similar natriuretic peptides have been identified: atrial natriuretic peptide (ANP), urodilantin (an isoform of ANP), B-type natriuretic peptide (BNP), C-type natriuretic peptide and Dendroaspis natriuretic peptide (Cea, 2005). Of these, ANP and BNP are transcribed and primarily produced in the myocytes of atria and ventricles, respectively (Mukoyama et al., 1991), both are produced in response to myocardial stretch due to pressure or volume overload (Kinnunen et al., 1993), conditions commonly found in HF. The biological functions of ANP and BNP include various compensatory mechanisms such as natriuresis, diuresis, and vasodilation (Cody et al., 1986; Marcus et al., 1996).

Myofilamentar fibers also are important markers of hypertrophy. Myosin isoform beta was upregulated after MI procedures compared to control, which is in agreement to previous report where myosin over-expression is related as a molecular sign of overloaded myocardium (Gupta, 2007). Although an augmentation of mRNA expression was observed when experimental groups were compared to control, there were no expression changes between both MI procedures. A major shift in the myosin isoform distribution occurs during pathologic hypertrophy mediated by pressure and volume overload. This hypertrophy is associated with induction of βMHC at the expense of αMHC. This change from the α- to β-MHC phenotype is taken as a hallmark of pathologic hypertrophy, which is more intense in pressure-overload than in volume-overload hypertrophy (Gupta, 2007).

Particularly, all five genes related to calcium dynamic showed altered mRNA expression in our study. Drastic increase in the level of phospholamban and sodium calcium exchange SLC8A1 mRNAs were observed after MI generated by ablation compared to both occlusion and control. Failing heart muscle generally exhibits distinct changes in intracellular calcium (Ca2+) handling, including impaired removal of cytosolic Ca2+; reduced Ca2<sup>+</sup> loading of the cardiac sarcoplasmic reticulum (SR) with downregulation of SR Ca2+-ATPase 2 (SERCA2); and defects in SR Ca2<sup>+</sup> release accompanied by impairment of cardiac relaxation and systolic function (Morgan et al., 1990; Marx et al., 2000). Sodium-calcium exchange (NCX) is the major Ca2<sup>+</sup> efflux mechanism of ventricular cardiomyocytes. Consequently, the exchanger plays a critical role in the regulation of cellular Ca2<sup>+</sup> content and hence contractility (Ottolia et al., 2013).

On the other hand, calsequestrin 2, cardiac ryanodine receptor and ATPase, calcium transporter were downregulated after coronary occlusion compared to control. Calsequestrin (CSQ2), as the major Ca2<sup>+</sup> binding protein in the sarcoplasmic reticulum of cardiac myocytes, communicates changes in the luminal Ca2<sup>+</sup> concentration to the cardiac ryanodine receptor (RYR2) channel (Gaburjakova et al., 2013). Combined with biochemical observations highlighting the Ca2<sup>+</sup> dependence of the CSQ2- RYR2 interaction, these results indicate that the changes in RYR2 activity caused by luminal Ca2<sup>+</sup> may be presumably attributed to the CSQ2 dissociation from the channel complex; and thus, it is highly likely that CSQ2 plays an active role in communicating changes in [Ca2+] to the RYR2 channel (Gaburjakova et al., 2013).

The metabolic adaptation of heart glycolysis to substrate availability, workload, and hormones has been known for many years (Taegtmeyer, 1994; Stanley et al., 1997) and the clinical relevance of glucose metabolism to heart diseases has been reviewed (Taegtmeyer, 1994; Depré et al., 1998).Six mRNA implied to the metabolism of cardiomyocytes were relevant in our expression investigation. All six genes were found to be upregulated after MI induced by occlusion and ablation compared to control, a common expression pattern seem before in our mRNA quantification. Phosphosfructokinase 1, a tetrameric enzyme that phosphorylates fructose-6-phosphate to fructose-1,6-bisphosphate, committing glucose to glycolysis (Opie, 1978), was decreased after occlusion, despite the fact that it was more marked after ablation, both compared to control. An augmentation of phosphofructokinase mRNA in occlusion compared to ablation might indicate a more intense metabolic rate after hypoxia, to reverse cell damage and death.

Two detected mRNA in this group, NAD dehydrogenase 1 alpha 3 and taffazzin (TAZ), increased after both MI procedures compared to control. So far, there are no clear reports on taffazin expression; while some authors observed a remarkable over 90% decrease in TAZ mRNA expression during the development of HF in rats and humans, others reported that its expression remains unchanged after MI (Athéa et al., 2007; Harjot et al., 2009). The NADPH oxidase family (Nox) is a major and dedicated cellular ROS generation system in cardiac myocytes and many other cell types, including neurons (Lambeth, 2004; Bedard and Krause, 2007). It has been suggested that decreases in NAD+/NADH ratio, which have been observed in mitochondrial diseases or metabolic diseases such as diabetes, promote reductive stress (Ido, 2007; Ying, 2008). Increased NADPH level facilitates the reduction of glutathione, which serves as an important antioxidant mechanism. It should be mentioned that ischemia elevates NADH level (NAD+/NADH ratio declines; Ying, 2008).

VEGFa, usually referred to as VEGF, acts as a key player in vasculogenesis and angiogenesis. VEGFa mRNA was found to be downregulated after both MI procedures compared to control. Zhao et al. (2010) found that VEGFa levels within the infarcted myocardium were persistently suppressed post MI. VEGFR expression was significantly increased only at the border zone at day 1, but not in the later stages. The expression of VEGFa/VEGFR remained unchanged in the noninfarcted myocardium.

On the other hand, Akt1 was upregulated exclusively after MI procedures compared to control. Multiple signaling pathways downstream of Akt1 control cell survival, growth, metabolism, cell cycle progression, as well as motility of vascular cells (Somanath et al., 2006). Akt1 signaling might be involved in the regulation of several aspects of cardiac function and repair following an ischemic injury. Akt1 signaling seems to promote fibrosis in post-MI hearts. Akt1 function in cardiac fibrosis and hypertrophy is supported by multiple studies using both overexpression of Akt1 in heart (Nagoshi et al., 2005; Shiojima et al., 2005; Shiojima and Walsh, 2006) and Akt1 deletion (Shimizu et al., 2010) models. Although lack of Akt1 enhances LV damage and apoptosis in cardiomyocytes immediately after I/R injury, in a longer term, it improves cardiac function and LV remodeling after MI (Shimizu et al., 2010).

### miRNA Analysis

In the current study, we found only few microRNAs are differentially expressed between both MI models or compared to control. All differentially expressed miRNA expression were validated with Real time PCR and presented decreased expressions after ablation compared to occlusion as observed in nCounter analyzes. In response to ischemia or I/R, cardiac cells including cardiomyocytes, cardiac fibroblasts, and endothelial cells undergo variable miRNA dysregulation. These changes can be apparent as early as 15 min after coronary artery ligation and may last for hours, days, or months (Zhou et al., 2016). This dysregulation may comprise a mix of protective and deleterious effects. Following very brief ischemia or reperfusion, such changes in miRNA expression might be relevant to early cell survival. A prolonged ischemia and/or reperfusion will cause permanent injury to the heart, and changes in miRNA expression may be pertinent to apoptotic status and other chronic processes including cardiac fibrosis and remodeling (Zhou et al., 2016).

In our occlusion protocol, we observed an enhanced expression of miR-221. Notwithstanding, we detected an increased ANP and BNP mRNA expression levels after 1 week. These mRNA increments observed in MI generated by occlusion is well established. Upregulation of miRNA-221 is also related to antiapoptotic effects in cancer cells (Zhou et al., 2016) and showed pro-proliferative, pro-migration, and antiapoptotic effects in vascular smooth muscle cells (Zhou et al., 2016). The miRNA-221 is significantly upregulated in patients with hypertrophic cardiomyopathy (HCM) and in a mouse model of cardiac hypertrophy and heart failure induced by pressure overload (Su et al., 2015). In vitro overexpression of miR-221 alone is sufficient to increase the size of cardiomyocytes, accompanied by enhanced expression levels of atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) (Wang et al., 2012). However, the in vivo roles and molecular mechanisms of miR-221 in the regulation of cardiac remodeling remain unclear.

Our study identified an upregulation of miR-34c after occlusion compared to control, and its upregulation compared to ablation. Reports showed that miR-34c is induced after acute ischemic damage (Greco et al., 2009), miR-34b and miR-34c belong to an evolutionary conserved miRNA family that plays a fundamental role in the p53 tumor suppressor network (He et al., 2007). Since MI procedures promotes p53-dependent apoptosis (Fiordaliso et al., 2001), it is tempting to speculate that the p53/miR-34 axis may be involved in a MI-induced death pathway.

Panguluri et al. (2013) identified a significant elevation of miR-301a as a key modulator in diabetic condition. Loss of function approach in in vitro model (H9C2 cells) using miR-301a inhibitor was used to test if the increase in miR-301 has a role in regulation of Kv4.2 gene expression, a voltage-gated K<sup>+</sup> channel. We observed a decreased miR-301 and miR-93 expressions in occlusion group compared to ablation. Computational analysis of the Homo sapiens and Mus musculus RyR2-3′UTR revealed conserved binding sites for miR-93. These results demonstrate that miR-93 negatively regulates RyR2-3′UTR, which could lead to a decreased RyR2 protein expression (Chiang et al., 2014). In fact, we, analyzing downstream, observed a decrease in RyR2 mRNA expression in occlusion model compared to control.

MiR-17 was identified as a novel Apaf-1-targeting miRNA (Song et al., 2015). The delivery of exogenous miR-17 suppressed the apoptotic protease activating factor 1 (Apaf-1) expression and consequently attenuated formation of the apoptosome complex containing caspase-9, as demonstrated by coimmunoprecipitation and immunocytochemistry. Furthermore, miR-17 suppressed the cleavage of procaspase-9 and the subsequent activation of caspase-3, which is downstream of activated caspase-9. Together, these results demonstrated the potential of miR-17 as an effective anti-apoptotic agent (Song et al., 2015). We detected an increased expression of miR-17-5p only after occlusion procedure. Moreover, Li et al. (2013) investigated the role of miR-17 in cardiac matrix remodeling following myocardial infarction (MI). Using real-time PCR, miR-17 was up-regulated most dramatically: 3.7-fold and 2.4 fold in the infarct region 3 and 7 d post-MI, respectively, and 2.4-fold in the border zone at d 3 compared to sham control (P < 0.01). Thus, it suggests that miR-17 participates in the regulation of cardiac matrix remodeling and provides a novel therapeutic approach using miR-17 inhibitors to prevent remodeling and heart failure after MI (Li et al., 2013).

Wang et al. (2010) evidenced that miR-9 can suppress myocardin expression, a transcriptional cofactor of NFATc3 expressed at a relatively low level in cardiomyocytes under physiological conditions. However, it can be up-regulated by hypertrophic stimulation and consequently mediate hypertrophic signals. Administration of miR-9 could attenuate cardiac hypertrophy and ameliorate cardiac function. We observed a diminished miR-9 expression on both MI procedures compared to control, corroborating the data presented by Wang et al. (2010), whereas miR-9 suppression might increase cardiac hypertrophy by NFATc3 stimulation. In fact, we also observed an increased NFATc3 mRNA expression after occlusion and ablation 1 week-post MI compared to sham.

To the best of our knowledge, reports of miR-542 and miR-1949 expressions in health or damaged cardiac tissue were not yet described. We found a downregulation of both miRNAs in ablation compared to occlusion and control. MiR-542-5p may induce double-strand DNA breaks and reactive oxygen species accumulation in transfected cells (Faraonio et al., 2012). In addition, Yoon et al. (2010) suggested that survivin is a direct target of miR-542-3p and growth inhibition by miR-542-3p may have a potential utility as an anti-cancer therapy.

The expression of miR-1949 was found to be deregulated and abundant in the rat bladder following spinal cord injury. Bioinformatics demonstrated that retinoblastoma 1, which is involved in tumorigenesis, is a target gene of miR-1949 (Wang et al., 2015), but still there was not a clear relation to the myocardium.

### CONCLUSION

We have identified transcriptional modifications in remote area of rat myocardium after MI induced by occlusion and ablation. Our findings pointed out that myocardium reacts, besides different insults, with more similarities than differences in areas remote to MI. Cardiac mRNA signatures after MI generated by both protocols were found to cause alterations in almost the same genes. This indicates a pattern in the way the tissue deals with the injury, although the response magnitude varies from one procedure to another. Notwithstanding, few miRNAs were detected to be differently expressed between occlusion and ablation experimental groups. Evaluation of mRNA expression of an increased number of genes and proteome would lead to better understanding of the miRNA targets associated with pathophysiological states seen in both MI procedures. Considering all aspects addressed in this study, we compared a genic and post-transcriptional modifications signatures in both methods and concluded that RF-Ab might be an alternative protocol to achieve MI with lower rates of animals' mortality and with a simple and reproducible method and mostly importantly,

### REFERENCES


with activation of resembling signal pathways. However, we hope to address to more detailed analyses and pathway intersections in others studies in current development in our laboratory.

### ETHICS STATEMENT

Ethics Committee for Animal Care and Use at the Nine of July University, UNINOVE. Consent number 034/2012.

### AUTHOR CONTRIBUTIONS

JS: Idealization, execution, and writing of the manuscript. ES: Execution and analysis of the data. RF: Execution and analysis of the data. AS: Idealization and analysis of the data. EB: Execution and analysis of the data. EA: Execution of protocols. PT: Idealization and analysis of the data. LN: Idealization and analysis of the data. MM: Idealization and analysis of the data.

### ACKNOWLEDGMENTS

This work was supported by grants number 2009/54225-8, 2012/15808-0, and 2015/11028-9 from São Paulo Research Foundation (FAPESP) and National Council for Scientific and Technological (CNPq grant numbers 477458/2009-2 and 479395/2012-8). The authors would like to thank FAPESP, CNPq, UNINOVE, and UNIFESP for all support.

ischemia/reperfusion injury in transgenic mice. J. Mol. Cell. Cardiol. 30, 2281–2289. doi: 10.1006/jmcc.1998.0789


the heart during ischemia/reperfusion injury. J. Clin. Invest. 115, 2128–2138. doi: 10.1172/JCI23073


of kallikrein-kinin pathway and angiogenesis. PLoS ONE 9:e91017. doi: 10.1371/journal.pone.0091017


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Santana, Feliciano, Serra, Brigidio, Antonio, Tucci, Nathanson, Morris and Silva. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Low-Level Laser Application in the Early Myocardial Infarction Stage Has No Beneficial Role in Heart Failure

Martha T. Manchini 1, 2 \*, Ednei L. Antônio<sup>2</sup> , José Antônio Silva Junior <sup>3</sup> , Paulo de Tarso C. de Carvalho<sup>1</sup> , Regiane Albertini <sup>2</sup> , Fernando C. Pereira<sup>1</sup> , Regiane Feliciano<sup>1</sup> , Jairo Montemor <sup>2</sup> , Stella S. Vieira1, 2, Vanessa Grandinetti <sup>1</sup> , Amanda Yoshizaki <sup>2</sup> , Marcio Chaves <sup>1</sup> , Móises P. da Silva<sup>1</sup> , Rafael do Nascimento de Lima<sup>1</sup> , Danilo S. Bocalini <sup>4</sup> , Bruno L. de Melo<sup>2</sup> , Paulo J. F. Tucci <sup>2</sup> and Andrey J. Serra1, 2

*<sup>1</sup> Laboratory of Biophotonic, Nove de Julho University, São Paulo, São Paulo, Brazil, <sup>2</sup> Laboratory of Cardiac Physiology, Federal University of São Paulo, São Paulo, Brazil, <sup>3</sup> Medicine Program, Nove de Julho University, São Paulo, Brazil, <sup>4</sup> Translational Physiology Laboratory, Brazil Physical Education and Aging Science Program, S*ã*o Judas Tadeu University, São Paulo, Brazil*

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Daryl Schwenke, University of Otago, New Zealand Paul Kenneth Witting, University of Sydney, Australia*

> \*Correspondence: *Martha T. Manchini*

*martha.manchini@gmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *15 August 2016* Accepted: *11 January 2017* Published: *30 January 2017*

### Citation:

*Manchini MT, Antônio EL, Silva Junior JA, de Carvalho PdTC, Albertini R, Pereira FC, Feliciano R, Montemor J, Vieira SS, Grandinetti V, Yoshizaki A, Chaves M, da Silva MP, de Lima RdN, Bocalini DS, de Melo BL, Tucci PJF and Serra AJ (2017) Low-Level Laser Application in the Early Myocardial Infarction Stage Has No Beneficial Role in Heart Failure. Front. Physiol. 8:23. doi: 10.3389/fphys.2017.00023* Low-level laser therapy (LLLT) has been targeted as a promising approach that can mitigate post-infarction cardiac remodeling. There is some interesting evidence showing that the beneficial role of the LLLT could persist long-term even after the end of the application, but it remains to be systematically evaluated. Therefore, the present study aimed to test the hypothesis that LLLT beneficial effects in the early post-infarction cardiac remodeling could remain in overt heart failure even with the disruption of irradiations. Female Wistar rats were subjected to the coronary occlusion to induce myocardial infarction or Sham operation. A single LLLT application was carried out after 60 s and 3 days post-coronary occlusion, respectively. Echocardiography was performed 3 days and at the end of the experiment (5 weeks) to evaluate cardiac function. After the last echocardiographic examination, LV hemodynamic evaluation was performed at baseline and on sudden afterload increases. Compared with the Sham group, infarcted rats showed increased systolic and diastolic internal diameter as well as a depressed shortening fraction of LV. The only benefit of the LLLT was a higher shortening fraction after 3 days of infarction. However, treated-LLLT rats show a lower shortening fraction in the 5th week of study when compared with Sham and non-irradiated rats. A worsening of cardiac function was confirmed in the hemodynamic analysis as evidenced by the higher LV end-diastolic pressure and lower +dP/dt and −dP/dt with five weeks of study. Cardiac functional reserve was also impaired by infarction as evidenced by an attenuated response of stroke work index and cardiac output to a sudden afterload stress, without LLLT repercussions. No significant differences were found in the myocardial expression of Akt1/VEGF pathway. Collectively, these findings illustrate that LLLT improves LV systolic function in the early post-infarction cardiac remodeling. However, this beneficial effect may be dependent on the maintenance of phototherapy. Long-term studies with LLLT application are needed to establish whether these effects ultimately translate into improved cardiac remodeling.

Keywords: angiogenesis, cardiac remodeling, cardiac performance, low-level laser therapy, myocardial infarction

## INTRODUCTION

Myocardial infarction (MI) is a major cause for heart failure (HF) development (Yancy et al., 2013). Data are showing that three million people are affected by MI in the USA, and more than 400.000 new cases are reported for each year. In fact, ∼50% of patients will die within 5 years, and 40% die 12 months after the first HF hospitalization (Kolseth et al., 2014).

The acute MI triggers an adverse process known as cardiac remodeling, in which there is left ventricular (LV) dilation and enlargement of the ischemic tissue (Serra and Tucci, 2016). Moreover, an impaired LV systolic and diastolic function and a reduced myocardial inotropism are well-documented findings (dos Santos et al., 2013; Antonio et al., 2015). Several mechanisms are shown to be implicated in cardiac remodeling, including adrenergic hyperactivity, renin-angiotensin-aldosterone system, apoptosis, autophagy, fibrosis, inflammation, oxidative stress, calcium handling abnormalities, and metabolic dysfunction (Whelan et al., 2010; Carlos et al., 2016; Ziff et al., 2016). Moreover, post-infarction cardiac remodeling is associated with a higher prevalence of cardiac rupture, arrhythmias, and formation of aneurysms. In the long term, there is the development of HF and sudden death (Whelan et al., 2010; Ziff et al., 2016).

Several interventions have been proposed to alleviate cardiac remodeling to prolong or prevent the development of HF (Carlos et al., 2016). However, current therapies have shown only modest results in survival or potential adverse properties (Yancy et al., 2013; Grosman-Rimon et al., 2016). In latest years, experimental studies have punctuated that the low-level laser therapy (LLLT) may be a promising approach to modulate various biological processes (Albertini et al., 2008; Pires et al., 2011). The LLLT stimulates photoreceptors in the mitochondrial respiratory chain, resulting in increased ATP, increased growth factor secretion and tissue healing (Tuby et al., 2006; Huang et al., 2011; Peplow et al., 2012). A cardiac LLLT effect has been reported for over 10 years, in which infarcted rats showed a lower myocardial necrosis (Oron et al., 2001b), LV dilatation (Ad and Oron, 2001; Yaakobi et al., 2001), and most favorable milieu to prevent scar disruptions (Whittaker and Patterson, 2000) with LLLT. More recently, our group demonstrated reduced infarct size, attenuated the systolic dysfunction and beneficial modulates inflammation and expression of vasoactive peptides in rats submitted to LLLT (Manchini et al., 2014).

In a recent systematic review, we have reported that many studies have only assessed the LLLT role at MI early stage, in which data reporting effects on the progression to HF are limited (Carlos et al., 2016). Moreover, an intriguing is issue shown to be an attenuated cardiac remodeling in animals submitted to LLLT only at the initial phase of injury. In this regard, it has been shown benefits of LLLT after several weeks post-MI, e.g., decreased infarct size and cardiac dilation (Oron et al., 2001b; Yaakobi et al., 2001). Although these data indicate that the benefits of LLLT in the acute phase of MI may persist in overt HF, there are some limitations that should be considered: (i) there is no blinding for the experimental group or outcomes. A more suitable method would be to blind the infarct size and LLLT; (ii) inclusion/exclusion criteria has not been stated (e.g., animals with similar infarct sizes). The control of infarct size it seems to be a key issue because the remodeling is intensified on larger infarctions. Thus, it is doubtful to consider a beneficial cardiac remodeling LLLT effect because of the intragroup infarct size variability; (iii) there is only cross-sectional design studies, and the causality results cannot be determined. Therefore, this study was designed to determine whether LLLT application benefits at the MI early stage remains in overt HF same with disruption of treatment.

### MATERIALS AND METHODS

### Animals and Experimental Design

This study was carried out in accordance with the recommendations of Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NHI, no. 85-23, revised 1996). The protocol was approved by the Institutional Research Ethics Committee of the Nove de Julho University, São Paulo, Brazil (number: 0015–2012). Experiments were performed under ketamine (50 mg/kg), and Xylazine (10 mg/kg) mixture anesthesia, and efforts were made to minimize the suffering of animals.

**Figure 1** illustrates the experimental design. Forty-seven female Wistar rats weighing 250–280 g were assigned to LLLT or non-LLLT. The MI was produced by permanent arterial coronary occlusion, and then rats were randomized to one of the following groups: infarcted rats non-treated with LLLT (MI-N, n = 7); infarcted rats submitted to LLLT (MI-LLLT, n = 6). Sham rats (n = 8) were operated upon similarly, although the coronary occlusion was avoided. The Sham and MI-N groups were submitted to a similar LLLT procedure, yet the device was kept off (placebo). Echocardiographic analyses were carried out on 3 days and 5 weeks post-infarction. We have included in the study only rats with large infarcts, which showed be defined on the 3rd-day post-infarction as a size ≥ 37% of LV (dos Santos et al., 2013). At the end of the 5-week, rats were euthanized by decapitation according to a protocol detailed elsewhere (AVMA Panel on Euthanasia. American Veterinary Medical Association, 2001). The infarct scar was removed from LV, and remote myocardial tissue was immediately stored in a cryogenic tube and kept frozen in liquid nitrogen for molecular analysis.

To date, nine rats died in coronary occlusion surgery, four during the peri-operative period, and two in the hemodynamic evaluation. We excluded 11 rats because they had infarct sizes < 37%.

### MI Model

The MI was induced according to a well-established technique (Antonio et al., 2015). Briefly, under anesthesia and artificial

**Abbreviations:** MI, Myocardial infarction; HF, Heart failure; LLLT, Low-level laser therapy; LV, Left ventricular; ATP, Adenosine triphosphate; NHI, National Institutes of Health; MI-LLLT, Infarcted rats submitted with laser; MI-N, Infarcted rats non-treated with laser; AlGaInP, Aluminum Indium Gallium Phosphorus; EDP, End diastolic pressure; +dP/dt, Positive derivatives of the developed pressure; -dP/dt, Negative derivatives of the developed pressure; SV, Stroke volume; SP, Systolic pressure; SWI, Stroke work index; CO, Cardiac output; VEGF, Vascular endothelial grow factor.

ventilation (Harvard Rodent Ventilator, Model 863; Harvard Apparatus, Holliston, MA, USA), a left thoracotomy was performed. The heart was exteriorized, and the left anterior descending coronary artery was occluded near its origin with 6- 0 polypropylene. The heart was rapidly returned to its original position and the thorax closed.

## Phototherapy

Aluminum Indium Gallium Phosphorus—AlGaInP (Twin Laser—MM Optics, São Carlos, SP, Brazil) was used for irradiation under the parameters in **Table 1**. After thoracotomy, the coronary occlusion was carried out as describe above, and the heart was put in the chest to recover itself for 60 s and then the organs was externalized. The laser/placebo was applied directly to the myocardial tissue targeting infarcted area. In the 3rd day, rats were anesthetized, and a new thoracotomy was performed at the same surgical site to heart exteriorization and laser/placebo application. Sham group was exposed to all experimental procedures, though the LLLT device was off.

## Echocardiography

Rats were anesthetized as described above (K-X mixture) and LV echocardiography was performed using a 12-MHz transducer connected to an HP Sonos-5500 (Hewlett–Packard, Palo Alto, CA, USA). The infarct size was evaluated on transverse 2 dimensional view and reported as percent of the LV perimeter on the basal, mid transversal, and apical planes (Sofia et al., 2014). The MI was defined as the presence of a segment with increased echogenicity and modification in myocardial thickening or systolic movement (hypokinesia, akinesia, or dyskinesia). Systolic function was analyzed by the fractional shortening (Serra et al., 2010). Diastolic function was not evaluated owing to the fusion of the A and E waves.

## LV Hemodynamic Study and Afterload Stress

Immediately after echocardiography, baseline hemodynamic evaluation was performed under adjusted anesthesia (K-X mixture) and oxygen-enriched ventilation with a closed chest.

### TABLE 1 | Protocol of LLLT irradiation.


The left femoral vein was accessed for drug administration, and a 2-F gauge Millar catheter-tip micromanometer (model SPR-320, Millar Instruments, Houston, TX, USA) was inserted into the right carotid artery into the LV cavity. Moreover, an ultrasound flow probe (Transonic System Inc., Ithaca, NY, USA) was positioned in the ascending aorta. The following data were analyzed (Acknowledge software, Biopac System, Santa Barbara, CA, USA): LV systolic (SP) and end-diastolic pressures (EDP), rate of change of LV pressure (+dP/dt and −dP/dt), heart rate, and cardiac output (CO), and stroke volume (SV). Stroke work index (SWI) was stated as previously described (dos Santos et al., 2010). Thereafter, sudden LV afterload increases were achieved using a single phenylephrine in bolus injection (15–25 µg/kg, i.v.) (dos Santos et al., 2010).

### Biometric Data

After hemodynamic analysis, hearts were quickly removed and weighed. Myocardial mass was indexed by body weight and used as a hypertrophy marker.

## Myocardial Fibrosis

Hearts were removed in 3 days and 5 weeks after infarction or sham surgery and fixed in 4% buffered formaldehyde overnight. The LV fragments were washed with PBS, dehydrated through a graded series of ethanol, diaphonized with Xylol and embedded with paraplast. Samples were cut into 3 mm thick sections and stained with Masson's trichome. The fibrous tissue was evaluated in 6 randomized 40 x magnification using a Nikon Eclipse E200 microscope and Nikon Infinity Optical System (Kurobane Nikon Co., Tochigi, Japan), and Image Pro-Plus software, version 4.0 (Media Cybernetics Inc., Rockville, MD, USA).

### Western Blot

Proteins were extracted from the LV remote area as previously described by us (Silva et al., 2014). Homogenate protein samples of 30 µg were subjected to SDS-PAGE in 10% polyacrylamide gel. Separated proteins were transferred onto hydrophobic polyvinylidene difluoride membranes (Hybond-P, Amersham Biosciences; Piscataway, J, USA), and the transfer efficiency was examined with 0.5% Ponceau S. The membranes were soaked in a blocking buffer (5% nonfat dry milk and 0.1% Tween 20 in PBS, pH 7.5) for 1 h at room temperature and then incubated overnight at 4◦C with primary antibodies: rabbit anti-Akt1 (1:5000 dilution; Abcam, Cambridge, MA, USA); rabbit anti-phosphoSer473Akt1 (1:5000 dilution; Abcam, Cambridge, USA); goat anti-VEGF (1:1000; Abcam, Cambridge, MA, USA); anti-GAPDH (1:500; Santa Cruz Biotechnology, Santa Cruz, CA, USA). After overnight incubation, membranes were washed five times and then incubated for 1 h with horseradish peroxidase-conjugated goat anti-rabbit and rabbit anti-goat secondary antibodies (1:2000; Invitrogen, San Diego, CA, USA). Membranes were finally washed five times with blocking buffer and then rinsed twice in PBS. Bound antibody was detected by using chemiluminescence reagent for 1 min. The bands were imaged by using Amersham Imager 600 system (GE Health Care, Little Chalfont, UK). UK).

### Statistical Analysis

Data were analyzed using GraphPad Prism software 5.0 (La Jolla, CA, USA). Shapiro-Wilk test was used to verify normality data. Levene test was applied to assess the equality of variances. One-way ANOVA complemented by Newman–Keuls post hoc was applied to detect differences between groups in crosssection analysis. Two-way repeated ANOVA complemented by Bonferroni post hoc was applied to paired data. Kruskal-Wallis followed by Dunn's multiple comparison tests were applied to

### TABLE 2 | Biometric data.

non-normality data. Statistical significance was set at p ≤ 0.05. Data are expressed as mean ± SD.

## RESULTS

### LLLT Does Not Affect Structural and Functional Abnormalities of LV

The biometric data are shown in **Table 2**. Average body weight was similar between the three experimental groups on the 3rd day and 5 weeks of study. Infarcted rats showed a similar heart mass as well as heart mass-to-body weight ratio when compared with Sham rats, and phototherapy had no repercussion on heart mass. Besides, phototherapy also had no effect on infarct size. As evidenced in **Figure 2**, quantitative analysis for Masson trichome staining indicated no significant differences in the collagen content between experimental groups with 3 days post-MI. However, infarcted rats showed a significant increase of fibrosis over 5 weeks post-MI, in which LLLT had no significant effect.

As seen in **Figure 3**, there was LV dilatation with only 3 days post-infarction, in which diastolic diameter was significantly higher in MI+LLLT group while the systolic diameter was higher in all infarcted groups when compared with Sham group. At the end of the 5-week experimental period, both diastolic and systolic diameters were shown to be significantly increased in all infarcted groups when compared with Sham group. The LV systolic dysfunction was apparent early as the 3rd-day post-infarction, as evidenced by a minor fractional shortening. The beneficial role of LLLT was only noticed in the early (3 days) post-infarction cardiac remodeling, in which the fractional shortening of the MI-LLLT group was significantly higher than MI-N group. On the other hand, treated-LLLT rats show a lower LV performance in the 5th week of study when compared with Sham and MI-N rats.

Afterward second echocardiographic analysis, an invasive hemodynamic evaluation was carried out to determine LV ejection performance. As reported in **Figure 4**, data also indicate deteriorating LV function, in which +dP/dt and −dP/dt values were significantly lower in MI-N and MI-LLLT groups compared with Sham group under basal conditions. In addition, a higher EDP was reported only for MI-LLLT group, which also showed a more marked reduction on −dP/dt when compared to MI-N and Sham group. The LV ejection parameters from all infarcted groups did not differ significantly from those of the Sham group when evaluated under basal conditions, as evidenced by SWI


*BW, body weight. Two-way repeated ANOVA was applied for body weight and infarct size analysis. ANOVA one way was applied for heart mass and heart mass index. There were no significant differences as a result of time or phototherapy. Data are expressed as mean* ± *SD.*

and CO. These findings led to analyze the cardiac functional reserve during sudden afterload stress as a result of in bolus phenylephrine injection. For suitable homogenization, we carried out experiments to raise the blood pressure of 50–70% over the baseline level (dos Santos et al., 2010). This afterload range was accompanied by a higher increase of +dP/dt and −dP/dt in Sham group than in all infarcted groups. Furthermore, CO decreased more dramatically in all infarcted groups when compared with Sham group. Ultimately, Sham rats showed SW increase, whereas the SW was remarkably reduced in all infarcted rats.

## Survival/Angiogenesis Factors Are Not Affected by MI or LLLT

It has been postulated that the cardioprotective effects of LLLT shown to be associated with increased angiogenesis, and this action is linked to modulation of vascular endothelial growth factor (VEGF) (Tuby et al., 2006, 2008). Thus, we have investigated the Akt1/VEGF pathway in the remote myocardial after 5 weeks following injury. Data in **Figure 5** indicate that the MI and phototherapy did not affect the expression of the total Akt1, Akt<sup>1</sup> phosphorylated at Serine 473 and Akt1/pAkt<sup>1</sup> ratio, which is a marker of its activity. Notwithstanding, VEGF expression was also not significantly different between the experimental groups.

## DISCUSSION

Data showing that LLLT application only at the MI early stage could result in a long-term beneficial effect on cardiac remodeling are intriguing. In fact, LLLT action has been achieved until several weeks after discontinuing of the irradiation, e.g., a minor infarct size (Oron et al., 2001b).

We showed here that the LLLT improved LV systolic function only 3 days post-infarction, which confirms previous data from our lab (Manchini et al., 2014). On the other hand, we have not reported a cardioprotective LLLT role

FIGURE 4 | Repercussion on baseline hemodynamic and sudden afterload after 5 weeks of MI (Sham, n = 8; MI-N, n = 7; MI+LLLT, n = 6). Data are means ± SD. *P*-values were determined by one-way ANOVA and *post hoc* Newman-Keuls test. LVEDP: left ventricular end-diastolic pressure; +dP/dt: maximum positive time derivative of developed pressure; −dP/dt: maximum negative derivative of developed pressure; SWI: stroke work index CO: cardiac output.

during evolution to overt HF, as illustrated by the no effect on infarct size, cavity dilation and LV systolic performance at the end of the study. Yang et al. (2011) have published similar findings in rats subjected to LLLT with up to 72 h post-infarction. Accordingly our data, these authors also reported no beneficial LLLT effect in LV diastolic and systolic diameter as well as LV performance on echocardiographic analysis. We advance these findings to explore whether the LLLT could increase functional heart reserve for an increased LV afterload. In fact, a minor functional LV reserve shown be a marker for cardiac remodeling progression (Fletcher et al., 1981), in which it can be the result of a decreased myocardial inotropism at a given loading level (Francis et al., 2001). As illustrated in **Figure 3**, infarcted rats had exacerbated LVEDP and decreased +dP/dt, −dP/dt, SW, and CO as a response to sudden afterload increases. Mechanisms associated with changes in cardiac performance are not fully clarified, but they may be linked to an altered handling Ca2<sup>+</sup> and myofilament Ca2<sup>+</sup> sensitivity (Pfeffer and Braunwald, 1990). Moreover, post-infarct ventricular dilatation is shown to be limiting for the intracavitary pressure development, as defined by the Laplace (Pfeffer and Braunwald, 1990; dos Santos et al., 2013). Importantly, LLLT had no effect on the functional cardiac abnormalities.

To our knowledge, infarct size has been the main variable affected by LLLT, and many studies have shown a minor injury size with several weeks post-LLLT application (Oron et al., 2001a,b; Yaakobi et al., 2001; Yang et al., 2011). It is hard to understand the differences of our findings to previously studies. A key reason may be the randomization, in which we have only included animals with large infarcts. The comparison of experimental groups that have a similar infarct size at baseline is a critical issue to avoid the causality of results and has not been controlled in previous studies. Thus, while it may be understood that the LLLT lead to a lower infarct size, it is possible also that rats with lower infarct sizes have been included in the LLLT-treated group. Moreover, previous investigations have only carried out a cross-sectional analysis (Oron et al., 2001a; Yaakobi et al., 2001; Yang et al., 2011), in which the causality cannot be determined. In this regard, we have analyzed the longitudinal repercussion of LLLT (**Table 2**) to clarify whether infarct size at baseline (3 days) changed over time as an effect of phototherapy. Other issues that show be investigated to understand the differences in our findings for previously studies are (i) the differences in irradiation parameters and (ii) approach to analyzing the infarct size (e.g., histomorphometric or echocardiographic).

Cardioprotective effects of LLLT are often attributed to angiogenic factors in a wide range of tissues (Dourado et al., 2011; Feng et al., 2012; Cury et al., 2013), including the ischemic myocardium (Tuby et al., 2006). Thus, there are findings showing greater pro-angiogenic stimuli (e.g., VEGF expression) in infarcted hearts that received LLLT only in the early MI (Mirsky et al., 2002; Zhang et al., 2010). In our study, there was no increased VEGF expression and its well-known downstream— Akt with 3 days and 5 weeks post-MI. It is shown to be reported that time of analysis of VEGF post-infarction may be a reason for our findings. Zhao et al. (2010) conducted experiments on infarcted rats to investigate the temporal expression of angiogenic factors. The authors observed a significant increase in the VEGF protein levels at the border zone only during day one post-MI and with subsequent decline in 28 days. Consequently, we cannot exclude an effect of LLLT on angiogenic VEGF signaling because our analysis may have been influenced by the timeline.

In summary, our findings illustrate that LLLT improves LV systolic function in the early post-infarction cardiac remodeling.

### REFERENCES


However, this beneficial effect may be dependent on the maintenance of phototherapy. Long-term studies with LLLT application are required to establish whether these effects ultimately translate into improved cardiac remodeling.

### AUTHOR CONTRIBUTIONS

MM, drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted all data. EA, drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted all data. JS, drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted the all data s and protein expression protocols. Pd, laser protocol and dosage. RA, laser protocol and dosage. FP, drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted all data. RF, performed experiments and protein expression protocols. JM, Echocardiogram analysis. SV, drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted all data. VG, drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted all data. Md, performed experiments and protein expression protocols. AY, performed experiments and histological analysis. MC, performed experiments and histological analysis. Rd, drafted the work and substantially contributed to work design, as well as, acquired, analyzed and interpreted all data. DB, performed experiments and protein expression protocols. Bd, performed experiments and protein expression protocols. PT, oversaw the design and performance of the experiments, analyzed data, interpreted the results of the experiments and edited the final format of the manuscript. AS, oversaw the design and performance of the experiments, analyzed data, interpreted the results of the experiments, edited and revised manuscript. All authors revised the work critically, approved the final version to be published and declared accountable for all aspects of the work.

### FUNDING

This study was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (grant #400851/2014-8) and FAPESP (Grants: 09-54225/8; 15/11028-9).


performance in rats with heart failure. Can. J. Physiol. Pharmacol. 88, 724–732. doi: 10.1139/y10-062


than prevents myocardial dysfunction in rats with sustained beta-adrenergic hyperactivity. J. Physiol. 588, 2431–2442. doi: 10.1113/jphysiol.2010.187310


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2017 Manchini, Antônio, Silva Junior, de Carvalho, Albertini, Pereira, Feliciano, Montemor, Vieira, Grandinetti, Yoshizaki, Chaves, da Silva, de Lima, Bocalini, de Melo, Tucci and Serra. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Exercise Training Attenuates Right Ventricular Remodeling in Rats with Pulmonary Arterial Stenosis

Brunno Lemes de Melo<sup>1</sup> , Stella S. Vieira<sup>1</sup> , Ednei L. Antônio<sup>1</sup> , Luís F. N. dos Santos <sup>1</sup> , Leslie A. Portes <sup>1</sup> , Regiane S. Feliciano<sup>2</sup> , Helenita A. de Oliveira<sup>2</sup> , José A. Silva Jr. <sup>2</sup> , Paulo de Tarso C. de Carvalho<sup>2</sup> , Paulo J. F. Tucci <sup>1</sup> and Andrey J. Serra1, 2 \*

*<sup>1</sup> Cardiac Physiology Laboratory, Federal University of São Paulo, São Paulo, Brazil, <sup>2</sup> Biophotonic Laboratory, Nove de Julho University, São Paulo, Brazil*

Introduction: Pulmonary arterial stenosis (PAS) is a congenital defect that causes outflow tract obstruction of the right ventricle (RV). Currently, negative issues are reported in the PAS management: not all patients may be eligible to surgeries; there is often the need for another surgery during passage to adulthood; patients with mild stenosis may have later cardiac adverse repercussions. Thus, the search for approaches to counteract the long-term PAS effects showed to be a current target. At the study herein, we evaluated the cardioprotective role of exercise training in rats submitted to PAS for 9 weeks.

### Edited by:

*Camille M. Balarini, Federal University of Paraíba, Brazil*

### Reviewed by:

*Georgina May Ellison-Hughes, Kings College London, UK Marcelo Perim Baldo, Unimontes, Brazil*

> \*Correspondence: *Andrey J. Serra andreyserra@gmail.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *09 August 2016* Accepted: *27 October 2016* Published: *05 December 2016*

### Citation:

*de Melo BL, Vieira SS, Antônio EL, dos Santos LFN, Portes LA, Feliciano RS, de Oliveira HA, Silva JA Jr., de Carvalho PdTC, Tucci PJF and Serra AJ (2016) Exercise Training Attenuates Right Ventricular Remodeling in Rats with Pulmonary Arterial Stenosis. Front. Physiol. 7:541. doi: 10.3389/fphys.2016.00541* Methods and Results: Exercise resulted in improved physical fitness and systolic RV function. Exercise also blunted concentric cavity changes, diastolic dysfunction, and fibrosis induced by PAS. Exercise additional benefits were also reported in a pro-survival signal, in which there were increased Akt<sup>1</sup> activity and normalized myocardial apoptosis. These findings were accompanied by microRNA-1 downregulation and microRNA-21 upregulation. Moreover, exercise was associated with a higher myocardial abundance of the sarcomeric protein α-MHC and proteins that modulate calcium handling—ryanodine receptor and Serca 2, supporting the potential role of exercise in improving myocardial performance.

Conclusion: Our results represent the first demonstration that exercise can attenuate the RV remodeling in an experimental PAS. The cardioprotective effects were associated with positive modulation of RV function, survival signaling pathway, apoptosis, and proteins involved in the regulation of myocardial contractility.

Keywords: artery pulmonary stenosis, cardiac hypertrophy, cardiac remodeling, exercise training, right ventricular hypertrophy

## INTRODUCTION

Pulmonary arterial stenosis (PAS) is a common congenital heart disease that affects the RV outflow tract (Bonow et al., 2008; Tarasoutchi et al., 2011; Ananthakrishna et al., 2014), in which has been highly prevalent in women (Egbe et al., 2014). There is a consensus that patients with severe PAS are eligible for surgery intervention (e.g., balloon valvuloplasty), in which it shows to be implemented at an early age to prevent adverse RV remodeling. On the other hand, a mild RV obstruction is often only clinically supervised (Kan et al., 1982; Rey et al., 1988; Stanger et al., 1990). In this regard, a possible hypertrophy and functional RV abnormality is monitored in the long-term as a means to prevent advance to heart failure (Pokreisz et al., 2007; Kittipovanonth et al., 2008; Baumgartner et al., 2010). Although the current PAS clinical management has managed to correct the congenital defect and improve physiological RV variables, some weaknesses were reported. The surgical stenosis correction in early life is an important limitation in developing countries, in which late intervention may underlie the impaired RV remodeling, loss of functional capacity and reduced quality of life (Tchoumi et al., 2011; Romeih et al., 2013). It is common that rectifications of malformations affecting the RV outflow tract can result in valve regurgitation, restenosis, and later on it will need valve implantation (Ananthakrishna et al., 2014; Sizarov and Boudjemline, 2016). Moreover, data are showing that a maintenance of cardiac function is observed after invasive intervention without following evidence functional gain beyond the acute stage (Lurz et al., 2011). Further, an impaired RV systolic function can appear even after surgery repair in conditions of complex congenital defects (Khraiche and Ben Moussa, 2016). Ultimately, patients with mild PAS can have later RV remodeling, i.e., stenosis is exacerbated by body growth because of the limited size of the conduit. In the knowledge of the natural history of PAS, data are showing that mild cases can become severe at a later date (Anand and Mehta, 1997; Baumgartner et al., 2010).

Based on the limitations mentioned above, new approaches have been evaluated to counteract the long-term PAS effects. In an animal PAS model, Borgdorff and coworkers have shown that pharmacological treatment with Sildenafil was effective in attenuating RV dysfunction (Borgdorff et al., 2012). Considering that a lower exercise capacity is commonly seen in patients with RV hypertrophy (Meadows et al., 2007) or who had surgical repair during childhood (Roos-Hesselink et al., 2006), exercise training can be an attractive approach. Moreover, exercise has shown beneficial cardiac effects on models of left ventricular pressure-overload. In a recent study by Souza and coworkers, exercise training attenuated cardiac remodeling and preserved systolic and diastolic function in rats submitted to ascending aortic stenosis (Souza et al., 2015). In contrast, we found no data about the effects of exercise on RV pressure-overload hypertrophy because of mechanical obstruction to outflow. The only evidence comes from experimental studies carried out with RV pressure overload on pulmonary hypertension induced by monocrotaline. In this regard, exercise training was able to improve survival and cardiac function (Handoko et al., 2009).

In this study, we evaluated whether an aerobic exercise training would improve functional capacity and attenuate the RV remodeling in an experimental PAS model. We specifically hypothesized (i) that exercise would increase physical fitness as a safety approach, in other words, without adverse effects, and (ii) the findings obtained could be accomplished by attenuation of pathological RV remodeling.

### MATERIALS AND METHODS

### Animals and Experimental Design

The investigation complies with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication no. 85-23, revised 1996). The experimental protocol was approved by the Institutional Research Ethics Committee from the Federal University of São Paulo, São Paulo, Brazil (process: 1673020414). Sixty-three female Wistar rats, weighing 180–220 g and aged 8–9 weeks, were assigned to one of the following three groups: No-trained Sham (n = 21); no-trained stenosis (SS, n = 21); and trained stenosis (TS, n = 21). **Figure 1** illustrates the experimental design, in which at first step the animals were submitted to stenosis or Sham procedure. A week later, echocardiographic examination was performed in all animals, and stenosis rats were then randomly assigned to SS or TS group. Following 3 days, the animals were submitted to functional fitness test on a motor-driven treadmill. Then, animals were followed for over 8 weeks.

### PAS Surgery

Rats were anaesthetized with ketamine plus xylazine (50 mg/Kg, Dopalen <sup>R</sup> , Vertbrands, Paulínia, SP, BRA, i.p.) mixture, intubated, and mechanically ventilated (Rodent ventilator, model 683, Harvard Apparatus, Holliston, MA, USA). A left thoracotomy was performed, and the pulmonary artery (PA) was carefully dissected free from the aorta. A silk thread was positioned under the PA, and an 18-gauge needle was placed alongside the PA. A suture was tied tightly around the needle, and the needle was rapidly removed to produce a fixed constricted opening in the lumen equal to the needle diameter. The combination of a fixed banding around the PA and the animal growth resulted in a high RV afterload. The Sham animals underwent the same procedure except PAS (Faber et al., 2006).

## Functional Fitness Assessment (Maximal Oxygen Uptake, VO2max)

The functional fitness was evaluated using a motorized treadmill coupled to a gas analyzer (Panlab, Harvard Bioscience Company, MA, USA). Prior to the physical test, rats were introduced to running as previously described by our group (Amadio et al., 2015). Each rat had to undergo a 2-min warm-up period at 25 cm/s, and the running speed was increased by 9 cm/s every 2 min to reach exhaustion. An oxygen uptake steady state on progressive increases in running speed and a respiratory exchange ratio of ≥1.05 were taken into account to define the VO2max.

### Echocardiography

After 48 h of stenosis or Sham surgery, rats were anaesthetized with ketamine and xylazine mixture as reported above, and transthoracic echocardiography was performed to determine

**Abbreviations:** RV, Right ventricle; PAS, Pulmonary arterial stenosis; SS, Notrained stenosis; TS, Trained stenosis; PA, Pulmonary artery; RVDA, Right ventricle diastolic transverse area; RVSA, Right ventricle systolic transverse area; FAC, Fraction area change; MHC, Myosin heavy chain.

pressure gradient using a 12 MHz transducer Sonos-5500 (Hewlett-Packard, Andover, MA, USA). The echocardiography was also carried out 9 weeks later to RV morphofunctional study. Measurements were performed according to the echocardiographic RV guidelines. The 2D modality was used to measure the RV. The right parasternal short-axis view at the level of the papillary muscles was used for determination of the diastolic and systolic RV area. Measurements of the RV outflow tract were obtained from the parasternal short-axis view at the level of the aortic valve during end-diastole (Egemnazarov et al., 2015). We included only animals with pressure gradients between 25–55 mmHg (Manchini et al., 2014). The diastolic (RVDA) and systolic (RVSA) transverse areas of the RV were measured at the basal, middle, and apical view. The final value was the arithmetic mean of the measures of the three views. Systolic function was analyzed by the fractional area change (FAC) as a function for the following equation: FAC = RVDA-RVSA/RVDA × 100. Pulsed Doppler at the tricuspid valve level provided the flow velocity curve to analyze the diastolic function (E/A wave ratio).

### Exercise Training Protocol

Rats were subjected to running training on a motor-driven treadmill (CL4002, Caloi, São Paulo, Brazil) after 7 days of surgical procedures for 8 weeks. Rats ran 6 times a week, and each session lasted up to 60 min. For each session, the treadmill speed was 18 m/min for 30 min and 22 m/min for the remaining 30 min. This research protocol has shown to be effective in alleviating the pathological cardiac remodeling (Serra et al., 2008, 2010).

### Myocardial Growth Markers and Fibrosis

Twenty-four hours after the last training session or sedentary status the rats were anesthetized with urethane overdose (4.8 g kg kg−<sup>1</sup> i.p.) and the hearts were quickly removed. The RV chamber was cut, separated from the rest of the heart and weighted to be used as a myocardial hypertrophy indicator. Afterwards, RV tissue was transversally sectioned at the mid-RV level and fixed in 10% formalin buffered solution. Tissue samples were sectioned in 7 µm thickness and stained with hematoxylin-eosin for cross-sectional cardiomyocyte area analysis at approximately 20 visual fields for each animal. Myocytes with visible nuclei and intact cellular membranes were chosen for evaluation. The analyses were performed at 40× magnification using an Olympus image acquisition system (Waltham, MA, USA) (Soci et al., 2011). The RV fibrosis was evaluated in tissue samples stained with picro-sirius red. The collagen content was assessed at 40× magnification using an Olympus image acquisition system (Egemnazarov et al., 2015). The percentage fibrosis was calculated by dividing the total area of collagen by the total analyzed tissue area multiplied by 100.

## mRNA Assay

Total RNA was extracted from frozen RV with 1 ml of TRIzol reagent (Gibco BRL, MD, USA). RNA quantification was determined using a SpectraMax M5 spectrophotometer system (Molecular Devices, CA, USA). One microgram of RNA was used for cDNA synthesis and Real-Time PCR gene expression analysis. First, contaminating DNA was removed using DNase I (Invitrogen, CA, USA) at a concentration of 1 unit/µg RNA in the presence of 20 mM Tris-HCl, pH 8.4, containing 2 mM MgCl<sup>2</sup> for 15 min at 37◦C, followed by incubation at 95◦C for 5 min. Then, the reverse transcription was performed with 200 µl reaction in the presence of 50 Mm Tris-HCl, pH 8.3, 3 mM MgCl2, 10 mM dithiothreitol, 0.5 mM dNTPs, and 50 ng of primers with 200 units of Moloney murine leukemia virus-reverse transcriptase (Invitrogen). The conditions were: 20◦C for 10 min; 42◦C for 45 min; 95◦C for 5 min. One microliter of reverse transcription reaction was used for real-Time PCR at Applied Biosystems 7500 Fast PCR (ABI Prism, Applied Biosystems, CA, USA) using the SYBR Green supermix (ABI Prism). Experiments were performed in triplicates for each target genes: β-MHC (primers forward 5′ -AAGTGGGCAGCATCACCTAC-3′ reverse 5 ′ -GCCGGCTCTGTAACTTCCTT-3′ (GenBank: NM\_021838); α-MHC (primers forward 5′ -TCAGGCTTGGGTCTTGTTAGC GAAGAGAAACTTCCAGGGGCA-3′ reverse 5′ -AGGCTC TTTCTGCTGGACA-3′ (GenBank: NM\_012611.3). Target gene

de Melo et al. Exercise and Arterial Stenosis

abundance was quantified as a relative value for internal control β-Actin: forward primer 5′ -AGAGGGAAATCGTGCGTGAC-3′ and reverse primer 5′ -AGGAAGGAAGGCTGGAAGAGA -3′ (GenBank: NM\_031144.3).

### MicroRNA Assay

The cDNA for miRNA analysis was synthesized from total RNA using specific primers according to the TaqMan microRNA assay. The 15µl reactions obtained by the TaqMan MicroRNA Reverse Transcription Kit protocol (Applied Biosystems, CA, USA) were incubated in a Thermal Cycler (Applied Biosystems, CA, USA) for 30 min at 16◦C, 30 min at 42◦C, and 5 min at 85◦C. They were then maintained steadily at 4◦C. The real-time PCR quantification was performed by using the TaqMan MicroRNA Assay protocol (Applied Biosystems, CA, USA). The 20µl PCR reaction solution contained 10µl TaqMan Universal PCR master mix II (2x), 1.33µl RT product, 7.67 µl nuclease-free water, and 1µl of primers and probe mix from the TaqMan MicroRNA Assay protocol for microRNA-1 and 21. The reactions were performed at 95◦C for 10 min, and then in 40 cycles of 95◦C for 15 s and 60◦C for 1 min. Samples were normalized by evaluating the U6 gene.

### Western Blot

Frozen RV was homogenized using ice-cold lysis buffer and proteinase inhibitor cocktail (Manchini et al., 2014). Lysates corresponding to 30µg of protein were subjected to 10% SDS-PAGE. Separated proteins were transferred to PVDF membrane (Amersham Biosciences, NJ, USA) and transfer effectiveness was examined with 0.5% Ponceau S. After blocking with 5% non-fat dry milk for 2 h at room temperature, PVDF membranes were probed with Abcam (Cambridge, MA, USA) primary antibodies for rabbit ant-Akt<sup>1</sup> (1:5000), rabbit anti-p-Akt<sup>1</sup> (1:2500), rabbit anti-Caspase3; rabbit anti-Bax (1:1000), rabbit anti-Bcl-2 (1:1000), rabbit anti-Bcl-xL (1:500), rabbit antiβ-MHC (1:5000), rabbit anti-α-MHC (1:5000), rabbit anti-L-type Ca++ (1:500), rabbit anti-ryanodine receptor (1:1000), rabbit anti-Serca 2 (1:1000), and rabbit anti-Na+/Ca++ exchanger (1:100) in overnight incubation. Membranes were then washed five times with PBS and incubated for 1 h with horseradish peroxidase-conjugated goat anti-rabbit (1:20,000; Zymed, CA, USA). Membranes were again washed five times with blocking buffer and then rinsed twice with PBS. Antibodies binding were detected by chemiluminescence reagents (Amersham Biosciences, NJ, USA), and images were captured using an Amersham Imager 600 system. Quantification of target proteins was normalized for the internal control glyceraldehyde 3 phosphate dehydrogenase.

### Statistical Analysis

Data were analyzed using GraphPad Prism software version 5.0 (La Jolla, CA, USA) by analysis normality with Shapiro-Wilk test. One-way ANOVA complemented by Newman–Keuls posthoc was used to detect differences between groups for crosssection analysis. Two-way repeated ANOVA complemented by Bonferroni post-hoc were applied to paired data. Kruskal-Wallis followed by Dunn's multiple comparison tests were applied to non-normality data. Statistical significance was set at p-value ≤0.05. Data are expressed as mean ± standard error of the mean.

## RESULTS

## Overall Characteristics and Functional Fitness of Animals

All rats survived the stenosis surgery or sham procedure. During the stenosis period, the rats did not show signs of cyanosis or congestion, as demonstrated by the similar liver (SHAM: 68.1 ± 3.5; SS: 66.1 ± 2.9; TS: 67.1 ± 3.6; %, p = 0.3) and lung (SHAM: 71.4 ± 4.7; SS: 71.7 ± 5.2; TS: 71.6 ± 8.6; %, p = 0.9) weights. We performed the experiments in female rats because it allowed a strict control body weight. It is true that the stenosis degree could result in very dissimilar RV afterload during the study, thereby a bias would be set. As illustrated in **Figure 2A**, all experimental groups showed a similar average body weight

FIGURE 2 | Exercise training improves functional fitness in animals with PAS. Data are representative of eight samples from each group in the baseline (filled columns) and at the end of the study (open columns). (A) Body weight. (B) Pressure gradient; (C) maximal oxygen uptake (VO2max). #*<sup>p</sup>* <sup>&</sup>lt; 0.001 vs. SHAM group for the respective time.

at baseline and at the end of the study. The increase in body weight was associated with significant elevation in pulmonary transvalvular gradient only in stenosis groups (**Figure 2B**). This finding shows that the construction of the outflow tract of the RV was kept during the observation period, and echocardiography showed that the stenosis degree was similar between SS and TS groups. Moreover, mean valvar gradient values are within a range of mild to moderate stenosis definition (Hirth et al., 2006), which usually receives only clinical monitoring of RV remodeling. This issue emerges as a main way of analyzing the impact of exercise because critical stenosis requires a more invasive intervention (Godart et al., 2014). The **Figure 2** also shows the functional fitness data. The VO2max was not significantly different between groups at the baseline and at the end of the study, but there was a slight decrease in the SHAM and SS groups after 8 weeks (**Figure 2C**). A single functionality difference was evident only in the TT rats, in which VO2max was significantly increased with training.

## Exercise Prevents Concentric RV Remodeling and Fibrosis

On transthoracic echocardiography (**Figure 3A**), we have confirmed previous findings from the literature that the RVpressure overload results in a concentric cardiac remodeling (Faber et al., 2006). From the baseline to end of the study, non-trained stenosis rats (SS group) had a significant reduction in diastolic and systolic RV transverse areas, characterizing a concentric RV remodeling. On the other hand, exercise prevented these concentric changes in the cavity. The echocardiographic analysis also revealed reduced RV systolic performance in the first analyzes for animals submitted to stenosis. An evaluation 8 weeks after operation illustrated that systolic dysfunction was stable in SS group, but a significant improvement was noted in the animals submitted to exercise (TS group). The stenosis rats exhibited an increased ratio of RV weight to body weight and cross-section cardiomyocyte area compared with SHAM rats, in which trained animals showed a lower cross-section cardiomyocyte area compared to SS group (**Figure 3B**). Our structural analysis corroborates a well-established finding of myocardial fibrosis associated with RV pressure overload (Baicu et al., 2012; Egemnazarov et al., 2015). Thereby, stenosis resulted in a significant increase in collagen content (**Figure 3C**). Notably, exercise prevents the collagen deposition, in which TS group had values that did not differ significantly from the SHAM group.

## Exercise Triggers Survival Pathway and Inhibits Myocardial Apoptosis

We analyzed the Akt<sup>1</sup> and p-Akt<sup>1</sup> expression, as altering its activity provides a potent pro-survival signal (Matsui et al., 2003). As illustrated in **Figure 4A**, there was a significant upregulation of the activated Akt<sup>1</sup> form only in trained animals. We further investigated whether the increase in Akt<sup>1</sup> activity could be associated with a beneficial role of exercise in cell death. Thus, RV apoptosis was assessed by cleaved Caspase-3 levels, which was significantly increased in PAS animals (**Figure 4B**). These findings were accomplished by significant changes in upstream components of myocardial mitochondrial-dependent apoptotic signaling pathways. Compared to SHAM group, the protein pro-apoptotic Bax level significantly increased while the antiapoptotic factors Bcl-2 and Bcl-xL decreased in the SS and TS groups. Intriguingly, exercise restored the cleaved Caspase-3 to a similar level of SHAM animals. These findings were accomplished by a significant increase of anti-apoptotic Bcl-xL. Hence, exercise training reversed to normal the Bcl-xL/Bax ratio in myocardial.

## Myocardial MicroRNA-1 and -21 Levels are Regulated by Exercise

MicroRNAs are emerging as key regulators of gene expression, and their role in cardiac hypertrophy is becoming increasingly apparent (Harada et al., 2014). When looking at the expression of microRNA-1 we found no effect of pulmonary stenosis per se, but there was a significant reduction with exercise (**Figure 5A**). We observed a strong effect of stenosis on the microRNA-21 expression, in which was down-regulated in the SS group (**Figure 5B**). On the other hand, a restored microRNA-21 level was reported with exercise.

## Exercise Modulates Myocardial Gene/Protein Isoform Switches

We assessed how the pathophysiologic modifications in the RV were reflected by gene/protein expression changes of two distinct types of myosin heavy chain (MHC), referred as β-MHC and α-MHC. These analyses were carried out because the relative distribution of β- and α-MHC is altered in cardiac hypertrophy and showed to be directly correlated with the myocardial dysfunction (Nadal-Ginard and Mahdavi, 1989). Consistent with myocardial remodeling, there was increased β-MHC and decreased α-MHC gene expression in stenosis animals (**Figure 6A**). However, the combination of the stenosis with the exercise resulted in additional expression of the β-MHC and attenuation in downregulation of α-MHC. Thereby, the β/α-MHC ratio was increased in SS group whereas this was attenuated by exercise. The **Figure 6B** illustrates that β-MHC and α-MHC protein expression was upregulated for both SS and TS groups. Moreover, stenosis-induced overexpression of α-MHC protein was significantly enhanced by exercise. Then, the increased β/α-MHC ratio induced by stenosis was completely abrogated in trained animals.

## Exercise Modulates Expression of Ca++-Regulating Proteins

Quantitative changes in expression of the proteins that modulate calcium handling with correlations to functional alterations were reported in several experimental models of cardiac disease (Wankerl and Schwartz, 1955). Here, pulmonary stenosis induced a significant increase in L-type Ca++ channel and as Na+/Ca++ exchanger expression (**Figure 7**). Furthermore, SS animals exhibited lower ryanodine receptor content and a no significant reduction of Serca 2. Exercise protocol restored both ryanodine receptor and Serca 2 expression to basal levels. In addition, our results demonstrate an imbalance of Na+/Ca++

exchanger and Serca 2 expression in SS animals, while TS group has not reported this finding.

### DISCUSSION

In this study, we showed for the first time a useful role of exercise in a major congenital malformation affecting the RV outflow, i.e., PAS. Previous studies have shown only beneficial effects of exercise on RV remodeling in pulmonary hypertension model (Souza-Rabbo et al., 2008; Colombo et al., 2013). We have wellordered the body weight of the animals at baseline to ensure a comparable stenosis level between experimental groups, thereby imposing similar RV overload (**Figure 2B**).

We have observed RV pressure gradients ranging ∼40 to 50 mmHg/s, and this led to RV hypertrophy and remodeling, as reflected by increased RV mass, increased cardiomyocyte cross-sectional area, a concentric pattern of the RV, and fibrosis.

As previously described by other authors (Egemnazarov et al., 2015), myocardial remodeling was associated with reactivation of the "fetal gene program," as shown by the shift from α-MHC to β-MHC expression. Thus, as seen in **Figure 4**, the β-MHC to α-MHC ratio was shifted upwards in transcriptional and translational level. To date, although systolic and diastolic dysfunction has been reported in our PAS model, we did not observe signs of cardiac failure, as shown by the lack of cyanosis and liver or lung congestion. These findings are in line with interpretations from other groups (Johnson et al., 2011; Egemnazarov et al., 2015) and emerged as a possibility to explain because no significant reduction in functional capacity was noticed with PAS (**Figure 2C**).

To our knowledge, no published studies have directly assessed the role of exercise on RV hypertrophy induced by PAS. There is only information showing that an exercise program may be beneficial in other RV pressure overloads. In a pulmonary hypertension model induced by monocrotaline, Colombo et al. (2013) showed that hypertensive-exercised animals exhibited reduction in collagen, increase in vessels, and attenuation of diastolic dysfunction when compared with the hypertensiveuntrained animals. A more recent study showed that exercise was capable of attenuating RV hypertrophy, improving cardiac function, and increasing exercise tolerance and survival in rats with pulmonary hypertension (Moreira-Gonçalves et al., 2015). Here, we have reported that exercise exerts several beneficial effects on RV remodeling induced by PAS. Our histomorphometric data show that exercised-stenosis animals had lower cellular hypertrophy and fibrosis. Moreover, the echocardiographic analysis revealed a cardioprotective action of exercise on RV architecture and function, i.e., there was prevention of concentric cavity pattern, improved systolic performance and conserved diastolic function. Thus, it may be suggested that TS animals exhibited a prototype of physiological myocardial stimulus associated to preserved cavity geometry and improved cardiac function (Ellison et al., 2012). These benefits occurred regardless of a reduction in gradient pressure, which supports the hypothesis that the exercise can attenuate myocardial remodeling independent of the cardiac overload lowering the effects of exercise (Libonati, 2013). Overall, benefits

from the exercise may explain the improved functional capacity in trained animals, as noted for the increase in VO2max—a gold standard to assess exercise tolerance (Carlson, 1955). Therefore, considering that, a poor exercise tolerance is usually seen in patients with RV hypertrophy or who had surgical repair (Roos-Hesselink et al., 2006; Meadows et al., 2007), exercise can be an attractive approach.

We examined RV histological and molecular features to provide insights into the exercise-induced cardioprotection. Exercise animals showed minor cardiomyocyte hypertrophy with fibrosis inhibition, which can be linked to restoring the diastolic function (E/A ratio) as a result of myocardial stiffness normalization. Moreover, maintenance of RV cavity dimension and minor cross-sectional area of cardiomyocytes can be related to increased activity of Akt1. In fact, some studies demonstrate that the Akt activation has a favorable impact on cardiac pressure overload by inhibiting or preventing pathological processes (Faber et al., 2006; McMullen et al., 2007; Harston et al., 2011). Considering that the Akt is one of the major upstream signals of the Bcl-2 family, a higher activity of Akt can extend to the inhibition of broader cascade of apoptosis in trained animals. In fact, exercise training inhibited the increase of key downstream effector**—**cleaved Caspase-3, and more than normalized levels of anti-apoptotic factor Bcl-xL. This information is in line with current literature, in which exercise is shown to be a useful approach to prevent heart apoptosis caused by an increased afterload (Huang et al., 2012).

We also tried to explore the expression of two welldocumented microRNAs to respond to hypertrophic stimuli. It can be observed that mirRNA-1 was unchanged in the untrained animals. These findings are surprising because it was presumed that the microRNA-1 levels could be reduced with PAS, as previously reported in rodents subjected to aortic binding (Wang et al., 2009). However, microRNA-1 was reduced only in trained animals, showing that the physiological exercise stimulus persists even with pathological PAS insult. Previous studies have identified similar data (Wang et al., 2009; Ellison et al., 2012), but the physiological impact of altered mirRNA-1 level still needs clarification in healthy and diseased heart. Further, microRNA-21 expression was found to be significantly lower in SS animals, and these findings against previous studies showing upregulation of microRNA-21 in response to aortic binding. In this regard, microRNA-21 ablation has been advocated as a beneficial therapeutic to cardiac remodeling. Thum et al. (2008) reported that microRNA-21 knockdown prevented cardiac hypertrophy and fibrosis in response to pressure overload. The simple interpretation of these findings could indicate that the increase in microRNA-21 with exercise would not provide for myocardial remodeling. Nevertheless, some issues show to be clarified: (i) Is microRNA-21 upregulated in both ventricles in response to chronic elevated afterload? (ii) How can microRNA-21 ablation caused by exercise affect the RV remodeling? (iii) The maintenance of myocardial microRNA-21 level in SHAM and TS groups leads us to believe that the microRNA-21 works to promote cell survival through inhibition of PTEN (phosphatase and tensin homolog deleted on chromosome ten). The PTEN degrades phosphatidylinositol-3,4,5-trisphosphate, which is produced by phosphoinositide 3-kinase and is essential for activation of the pro-survival Akt pathway (Kukreja et al., 2011). This issue was well reported in studies showing that the PI3K inhibitor abrogated the protective miR-21 effect on myocardial apoptosis (Tu et al., 2013). Thus, it is possible that steady-state microRNA-21 level can mediate inhibition of apoptosis induced by PAS in our trained animals. This assumption has not been explored.

It has also been indicated that exercise more than restored systolic RV function in PAS animals. It is reasonable to admit that apoptosis inhibition resulted in conserved cardiomyocyte amount in the TS animals, contributing to preserving RV function. We have reported changes in the α-MHC and β-MHC composition in transcriptional e translational level. Considering that, an increased α-MHC expression, which has high Ca++ and actin-activated ATPase activity (Nadal-Ginard and Mahdavi, 1989), it is possible that the increase in α-MHC mediates improved systolic function in the TS group. Moreover, the slight increase of α-MHC can counteract

FIGURE 6 | Exercise training modulates β-MHC and α-MHC isoform expression in myocardial. (A) Gene expression was evaluated by real-time PCR (SHAM, *n* = 4; SS, *n* = 6; TS, *n* = 6). (B) Protein expression was evaluated by Western blot (SHAM, *n* = 7; SS, *n* = 6; TS, *n* = 7). Values were normalized for levels of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). #*p* < 0.05 vs. SHAM group. &*p* < 0.01 vs. SS group.

the increased β-MHC expression, which has lower ATPase activity. Besides, exercise showed to modulate the altered Ca++ regulating proteins expression induced by PAS. Although the SS and TS groups have shown similar increase in L-type Ca++ channel density (**Figure 7**), which can trigger a high amount of Ca++ from the Serca 2 (Balke and Shorofsky, 1998), the reduced ryanodine density could decrease the gain of Ca ++-induced Ca++ in the SS group. This finding was well documented in studies carried out mice with reduced ryanodine receptor, in which Ca++ release from sarcoplasmic reticulum was impaired in isolated cardiomyocytes (Zou et al., 2011). In addition, Serca 2 tended to be depressed and Na+/Ca++

exchanger/Serca 2 ratio was increased in RV from SS animals. We did not directly assess the Ca++ uptake function of the Serca 2 or the impact of increased Na+/Ca++ exchanger, but it can be assumed that these findings will slow the decline in [Ca++]i during relaxation and/or cause a net loss of Ca++ due to alternative removal via NCX (Bers et al., 2002). On the other hand, this condition could not exist in the trained animals, in which ryanodine receptor and Serca 2 densities as well as Na+/Ca++ exchanger/Serca 2 ratio were restored.

In summary, the major outcome of this study was to show that exercise training attenuates myocardial remodeling and improves RV function as well as functional fitness in rats with PAS. These changes were found to be associated with preserved collagen content and apoptosis in myocardial. The cardioprotective role of exercise may be caused by positive modulation of sarcomeric protein and calcium handling protein levels. For clinical purposes, exercise training can be considered as a useful approach for several patients with increased RV afterload, whether from PAS, pulmonary arterial hypertension, or other cardiac disorders, without a welldefined RV failure until delayed clinical course. Therefore, exercise may be a key tool to the development of overt RV failure. Moreover, given that an increased RV afterload

### REFERENCES


can result in biventricular injury and dysfunction (Okumura et al., 2015), it is possible that exercise benefits also exist in the left ventricle. Unfortunately, we have not carried out analysis of left ventricle and this will be evaluated in the future.

### AUTHOR CONTRIBUTIONS

BLM, SSV, ELA, LFNS, LP, and RL contributed to the design of the study, acquisition of data, and result analyses. HAO performed statistical analyses. JASJ and PTCC revised the manuscript, and approved the final version. PJFT and AJS raised grant funding, contributed to the design of the study, revised the manuscript, and approved the final version.

### FUNDING

This work was supported by the FAPESP on grant numbers 09-54225/8 and 2015/11028-9.

### ACKNOWLEDGMENTS

We thank Pontual Translation for critical reviewing of English language.

Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, and Society of Thoracic Surgeons. Circulation 118, e523–e661. doi: 10.1161/CIRCULATIONAHA.108.190748


ischemia/reperfusion injury via PTEN/Akt pathway. PLoS ONE 8:e75872. doi: 10.1371/journal.pone.0075872


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 de Melo, Vieira, Antônio, dos Santos, Portes, Feliciano, de Oliveira, Silva, de Carvalho, Tucci and Serra. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# Simulated Microgravity and Recovery-Induced Remodeling of the Left and Right Ventricle

Guohui Zhong<sup>1</sup> , Yuheng Li <sup>1</sup> , Hongxing Li <sup>2</sup> , Weijia Sun<sup>1</sup> , Dengchao Cao<sup>3</sup> , Jianwei Li <sup>1</sup> , Dingsheng Zhao<sup>1</sup> , Jinping Song<sup>1</sup> , Xiaoyan Jin<sup>1</sup> , Hailin Song<sup>2</sup> , Xinxin Yuan<sup>3</sup> , Xiaorui Wu<sup>1</sup> , Qi Li <sup>1</sup> , Qing Xu<sup>4</sup> , Guanghan Kan<sup>1</sup> , Hongqing Cao<sup>1</sup> , Shukuan Ling<sup>1</sup> \* and Yingxian Li <sup>1</sup> \*

*<sup>1</sup> State Key Laboratory of Space Medicine Fundamentals and Application, China Astronaut Research and Training Center, Beijing, China, <sup>2</sup> Key Laboratory of Molecular and Cellular Biology of Ministry of Education, College of Life Science, Hebei Normal University, Shijiazhuang, China, <sup>3</sup> State Key Laboratory of Agrobiotechnology, College of Life Sciences, China Agricultural University, Beijing, China, <sup>4</sup> Medical Experiment and Test Center, Capital Medical University, Beijing, China*

### Edited by:

*Valdir Andrade Braga, Federal University of Paraíba, Brazil*

### Reviewed by:

*Eugene Nalivaiko, University of Newcastle, Australia Jose Luiz De Brito Alves, Federal University of Paraíba, Brazil*

### \*Correspondence:

*Shukuan Ling sh2ling@126.com; Yingxian Li yingxianli@aliyun.com*

### Specialty section:

*This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology*

Received: *27 April 2016* Accepted: *17 June 2016* Published: *29 June 2016*

### Citation:

*Zhong G, Li Y, Li H, Sun W, Cao D, Li J, Zhao D, Song J, Jin X, Song H, Yuan X, Wu X, Li Q, Xu Q, Kan G, Cao H, Ling S and Li Y (2016) Simulated Microgravity and Recovery-Induced Remodeling of the Left and Right Ventricle. Front. Physiol. 7:274. doi: 10.3389/fphys.2016.00274* Physiological adaptations to microgravity involve alterations in cardiovascular systems. These adaptations result in cardiac remodeling and orthostatic hypotension. However, the response of the left ventricle (LV) and right ventricle (RV) following hindlimb unloading (HU) and hindlimb reloading (HR) is not clear and the underlying mechanism remains to be understood. In this study, three groups of mice were subjected to HU by tail suspension for 28 days. Following this, two groups were allowed to recover for 7 or 14 days. The control group was treated equally, with the exception of tail suspension. Echocardiography was performed to detect the structure and function changes of heart. Compared with the control, the HU group of mice showed reduced LV-EF (ejection fraction), and LV-FS (fractional shortening). However, mice that were allowed to recover for 7 days after HU (HR-7d) showed increased LVIDs (systolic LV internal diameter) and LV Vols (systolic LV volume). Mice that recovered for 14 days (HR-14d) returned to the normal state. In comparison, RV-EF and RV-FS didn't recover to the normal conditions till being reloaded for 14 days. Compared with the control, RVIDd (diastolic RV internal diameter), and RV Vold (diastolic RV volume) were reduced in HU group and recovered to the normal conditions in HR-7d and HR-14d groups, in which groups RVIDs (systolic RV internal diameter) and RV Vols (systolic RV volume) were increased. Histological analysis and cardiac remodeling gene expression results indicated that HU induces left and right ventricular remodeling. Western blot demonstrated that the phosphorylation of HDAC4 and ERK1/2 and the ratio of LC3-II / LC3-I, were increased following HU and recovered following HR in both LV and RV, and the phosphorylation of AMPK was inhibited in both LV and RV following HU, but only restored in LV following HR for 14 days. These results indicate that simulated microgravity leads to cardiac remodeling, and the remodeling changes can be reversed. Furthermore, in the early stages of recovery, cardiac remodeling may be intensified. Finally, compared with the LV, the RV is not as easily reversed. Cardiac remodeling pathways, such as, HDAC4, ERK1/2, LC3-II, and AMPK were involved in the process.

Keywords: simulated microgravity, cardiac remodeling, declining function, recovery, remodeling pathways

## INTRODUCTION

There are various changes in the human cardiovascular system, including a cephalic fluid shift (Thornton et al., 1987), changes in cardiac systolic volume (Bungo et al., 1987; Caiani et al., 2006), and over time, a loss of left ventricular mass due to microgravity during space flight (Perhonen et al., 2001; Summers et al., 2005). The adaptation and adjustments that characterize the responses to the metabolic demands of activity and gravitational loading on Earth are changed dramatically under conditions of microgravity. Chronic reduction in metabolic demand and oxygen uptake reduces the demand on cardiac output and tissue perfusion, resulting in cardiac atrophy and a decline in function, and further leads to orthostatic intolerance upon return to full gravity with the potential risk of irreversible structural changes that may become pathological (Marcus et al., 1977; Zile et al., 1993; Perhonen et al., 2001). Because of this, it is essential to determine the severity of cardiac changes upon return to the ground. Many studies have demonstrated that cardiac remodeling induced by microgravity and/or simulated microgravity is associated with a decline in cardiac function. However, the changes in heart structure and function during reloading following simulated microgravity are not wellunderstood.

An abundance of data has provided insight into the changes that occur in the left ventricle (LV; Summers et al., 2005; Westby et al., 2016). There are no data, however, on remodeling in the right ventricle (RV) under weightlessness due to reduced gravitational loading. The changes in the LV and RV that occur in astronauts during space flight and their subsequent return to the ground are poorly understood. On the cellular level, the remodeling responses of the LV and RV to pressure overload are largely similar. There are several major signaling molecules involved in cardiac remodeling induced by external or intrinsic stimuli, including HDAC4, AMP-activated protein kinase (AMPK), ERK1/2, and LC3-II. However, there is a divergence in the molecular mechanisms of the RV compared with the LV under stress conditions (Reddy and Bernstein, 2015). The difference in the responses of the LV and RV to simulated microgravity as well as the signaling molecules involved in this process need to be explored further.

From the perspective of the cardiovascular system, rodent hindlimb unloading (HU) is a suitable model. There is extensive literature investigating the cardiovascular adaptation to simulated microgravity, predominantly using the HU rat or mouse model (Hasser and Moffitt, 2001). The mouse demonstrates a wide range of cardiovascular responses to HUsimulated microgravity, including alterations in heart function, heart rate, exercise capacity, peripheral arterial vasodilatory responsiveness, and the baroreflex response (Powers and Bernstein, 2004). Many of these responses are similar to those seen in humans. Following 28 days of HU-simulated microgravity, mice manifest many of the cardiovascular alterations that have been previously demonstrated in humans during space flight (Buckey et al., 1996; Fritsch-Yelle et al., 1996; Powers and Bernstein, 2004).

Here, we suppose that HU can lead to distinct remodeling of the LV and RV in mice, and that reloading after HU has a further effect on left and right ventricular remodeling. In this study, we detected the remodeling signals and structural changes of the LV and RV following HU and hindlimb reloading (HR). We determined that pathological remodeling signals are overactive in both the LV and RV following HU and/or HR, and are restored after 14 days of reloading. The physiological remodeling signal AMPK is downregulated in both the LV and RV, which leads to the functional decline of both ventricles. Finally, we found that recovery is more difficult in the RV than the LV. This study provides insight into the molecular mechanisms of cardiac remodeling and the decline of systolic function of both the LV and RV during simulated microgravity and recovery.

## MATERIALS AND METHODS

### Animals

All mice used in the experiments were bred and maintained at the SPF Animal Research Building of China Astronaut Research and Training Center (12-h light, 12-h dark cycles, temperature controlled for 23◦C and free access to food and water). The mice used on this study were 3 month old males and in a C57BL/6N background. The experimental procedures were approved by the Animal Care and Use Committee of China Astronaut Research and Training Center, and all animal studies were performed according to approved guidelines for the use and care of live animals.

### Hindlimb-Unloading Model

The hindlimb-unloading procedure was achieved by tail suspension, as described by Morey-Holton and Globus (2002). Briefly, the 3-month-old mice were individually caged and suspended by the tail using a strip of adhesive surgical tape attached to a chain hanging from a pulley. The mice were suspended at a 30◦ angle to the floor with only the forelimbs touching the floor, which allowed the mice to move and access to food and water freely. The mice were subjected to hindlimb unloading through tail suspension for 28 days, which we will identify as the "unloaded" state, for a total of 28 days, after which they were returned to the normal four-extremity weight bearing "reloaded" position (hindlimb reloading, HR). Similar numbers of control mice of the same strain background were instrumented and monitored in similar fashion under identical cage conditions but without tail suspension.

### Histological Analysis

Sections were generated from paraffin embedded hearts, and were stained with H&E for gross morphology, Masson's trichrome for detection of fibrosis, as described before Ling et al. (2012).

### RNA Extraction and Real-Time Polymerase Chain Reaction

Total RNA was extracted from heart tissues by using RNAiso Plus reagent (Takara) according to the manufacturer's protocol. The RNA (500 ng/sample) was reverse transcribed into cDNA and qPCR was performed using a SYBR Green PCR kit (Takara) in a Light Cycler (Eppendorf, Germany). PCR for each sample was carried out in duplicate for all cDNAs. The mRNA level of each gene was normalized to that of Gapdh, which served as an internal control. Primers (synthesized by Sunbiotech Co, China) for Col1a1 (Product size, 119 bp, Melting temperature, 61◦C), Col3a1 (Product size, 98 bp, Melting temperature, 60◦C), BNP (Product size, 185 bp, Melting temperature, 56◦C), and Gapdh (Product size, 122 bp, Melting temperature, 59◦C), were as follows:

Col1a1 sense primer: 5′ -CTGACTGGAAGAGCGGAG AGT-3′ ,

Col1a1 anti-sense primer: 5′ -AGACGGCTGAGTAGGGAA CAC-3′ ;

Col3a1 sense primer: 5′ -ACGTAAGCACTGGTGGACAG-3′ , Col3a1 anti-sense primer: 5′ - CAGGAGGGCCATAGCTGA AC-3′ ;

BNP sense primer: 5′ -TGTTTCTGCTTTTCCTTTATC TG-3′ ,

BNP anti-sense primer: 5′ -TCTTTTTGGGTGTTCTTTTGT GA-3′ ;

Gapdh sense primer: 5′ -ACTCCACTCACGGCAAATTCA-3′ , Gapdh anti-sense primer: 5′ -GGCCTCACCCCATTTGAT G-3′ .

### Transthoracic Echocardiography

Animals were lightly anesthetized with 2,2,2-tribromoethanol (0.2 ml/10 g body weight of a 1.2% solution) and set in a supine position. Two dimensional (2D) guided M-mode echocardiography was performed using a high resolution imaging system (Vevo 770, Visual-Sonics Inc., Toronto, ON, Canada). Two-dimensional images are recorded in parasternal long- and short-axis projections with guided M-mode recordings at the midventricular level in both views. Left ventricular (LV) cavity size and wall thickness are measured in at least three beats from each projection. Averaged LV wall thickness [interventricular septum (IVS) and posterior wall (PW) thickness] and internal dimensions at diastole and systole (LVIDd and LVIDs, respectively) are measured. LV fractional shortening ((LVIDd – LVIDs)/LVIDd), relative wall thickness [(IVS thickness + PW thickness)/LVIDd], and LV mass [LV Mass = 1.053 × [(LVIDd + LVPWd + IVSd)3 – LVIDd3]] are calculated from the M-mode measurements. LV ejection fraction (EF) was calculated from the LV cross-sectional area (2-D short-axis view) using the equation LV %EF = (LV Vold – LV Vols)/LV Vold × 100%. For RV, Two-dimensional images are recorded in right parasternal long- and short-axis projections with guided M-mode recordings at the maximum diameter level in both views. Right ventricular (RV) cavity size and wall thickness are measured in at least three beats from each projection. Averaged Right Ventricle Anterior Wall and internal dimensions at diastole and systole (RVIDd and RVIDs, respectively) are measured. Right Ventricle Percent Fractional Shortening (RVIDd – RVIDs)/RVIDd × 100%). Right Ventricle Percent Ejection Fraction (EF) was calculated from the RV cross-sectional area (2-D short-axis view) using the equation RV %EF = (RV Vold – RV Vols)/RV Vold × 100%.

For the RV mass weight calculation, firstly, we measured the RV endocardial borders in diastole (five measures) and systole (five measures) in five consecutive cardiac cycles in each flatimage, generating 10 RV endocardial areas (RVendo). Then, we got the epicardial borders and corresponding RV epicardial areas (RVepi) on the same frames. Ten RV free wall areas were then calculated by subtracting RVendo from RVepi (RVepi-RVendo). Finally, the total RV free wall volume of each plane was calculated by Simpson's method using the mean of the RV free wall area. RV free wall mass was obtained by multiplying this volume by the specific density of the myocardium (1.05 g/cc).

### Western Blot Analysis

Mouse hearts were crushed by homogenizer (Power Gen125, Fisher Scientific) and then lysed in lysis buffer (50 mM Tris, pH7.5, 250 mM NaCl, 0.1% SDS, 2 mM dithiothreitol, 0.5% NP-40, 1 mM PMSF, and protease inhibitor cocktail) on ice for 15 min. Protein fractions were collected by centrifugation at 12,000 g at 4◦C for 15 min and then applied to 8– 12% SDS-PAGE gels, electrophoresed at 80 V fof 30 min and 120 V for 90 min. After electrophoresis, protein was transfected to a polyvinylidene fluoride membrane using a Criterion blotter apparatus (Bio Rad). The membrane was then blocked in 5% non-fat dry milk (Becton, Dickinson, and Company) in TBST (10 mM Tris–Cl, 150 mM NaCl, 0.05% Tween-20, pH 7.5) for 1 h. After that, the membrane was incubated with primary antibody overnight at 4◦C followed by incubation with a secondary antibody conjugated to horseradish peroxidase (HRP), and visualized using an chemiluminescence kit (Thermo Pierce, No.32 109). Specific antibodies to p-HDAC4 (Cell Signaling Technology, #3443S), HDAC4 (Cell Signaling Technology, #5392S), p-AMPKα (Cell Signaling Technology, #2531S), AMPKα (Cell Signaling Technology, #2532), p-ERK1/2 (Cell Signaling Technology, #4370S), ERK1/2 (Cell Signaling Technology, #4696S), LC3- II/I (MBL, PM036-PN), mTOR (Cell Signaling Technology, #2983), p-mTOR (Cell Signaling Technology, #5536) Atrogin1 (ECM biosciences, # AP2041), MuRF1 (ECM biosciences, # MP3401), Gapdh (Santa Cruz Biotechnology, sc-25778) were used to detect protein levels. Gapdh was used as a loading control.

## Statistical Analysis

Data are presented as mean ± SEM per experimental condition. For the statistical differences among groups, considering the presence of unequal variance for the data, we firstly test the equality of variances across groups. If it shows the variances are unequal, we then use Welch's t-test for one-way analysis. Otherwise we use Student's t-test. Bonferroni adjustment was used for multiple comparisons. P < 0.05 is considered statistically significant. P < 0.01 is considered very significant. All the statistical tests are analyzed by Prism software (Graphpad prism for windows, version 5.01).

## RESULTS

## Heart Weight and Body Weight Changes Following HU and Recovery

Three groups of mice were subjected to HU by tail suspension for 28 days following which two groups were allowed to recover for 7 or 14 days (HU-28d, n = 10; HR-7d, n = 10; HR-14d, n = 10). The control group (n = 10) was treated equally, with the exception of tail suspension (**Figure 1A**). Body weight and heart weight were recorded before sacrifice (**Figures 1B,C**), and the mass of the LV and RV were calculated by echocardiography (**Figures 1E,F**). Compared with the control group, body weight showed an overall decrease, while heart weight increased following HU and HR. Thus, the ratio of heart weight to body weight increased

following 28 days of HU and 7 days of HR (**Figure 1D**). LV mass calculated by echocardiography remained unchanged following HU or HR (**Figure 1E**), but RV mass exhibited an overall decrease following HU, and recovered following HR (**Figure 1F**). All echocardiographic measurements were made while mice maintained heart rates of 450 ± 50 beats per minute (**Figure 1G**).

## Changes in Left and Right Ventricular Function Following HU and Recovery

To validate the effects of HU and HR on the LV and RV, transthoracic echocardiography was performed to determine ventricular function following HU and HR. After 28 days of HU, left ventricular fractional shortening (LV-FS), and left ventricular ejection fraction (LV-EF) decreased significantly in HU mice compared with the control, and did not recover during the first few days of HR. Full recovery was only apparent after reloading for 14 days (**Figure 2A**). However, after 28 days of HU, RV-FS and RV-EF decreased significantly in HU mice compared with the control, and did not recover even when reloaded for 14 days (**Figure 2B**). The results indicate that simulated microgravity can induce a decline in left and right ventricular function, and that recovery is slower in the RV after reloading.

## Changes in Left and Right Ventricular Structure Following HU and Recovery

To validate the influence of HU and HR in the LV, we performed transthoracic echocardiography to determine the structure of the LV following HU and HR (**Figure 3A**). Compared with control, the end-systolic LV internal diameter (LVIDs), and the enddiastolic LV internal diameter (LVIDd) did not change in the HU-28d mice, but increased following HR for 7 days, and recovered after 14 days of HR (**Figures 3B,C**). Furthermore, the change in end-systolic LV volume (LV-Vols) and the end-diastolic LV volume (LV-Vold) was the same as for LVIDs and LVIDd (**Figures 3D,E**). The LV posterior wall thickness (LVPW) and the LV anterior wall thickness (LVAW) didn't change following HU or HR (**Figures 3F–I**). Together, these data show that reloading after HU can induce enlargement of the left ventricular internal diameter and volume.

To validate the influence of HU and HR in the RV, we performed transthoracic echocardiography to assess the structure and function of the RV following HU and HR (**Figure 4A**). After 28 days of HU, the end-diastolic RV internal diameter (RVIDd) and the end-diastolic RV volume (RV-Vold) decreased in the HU mice compared with control, but recovered to its normal state in the HR-7d and HR-14d groups (**Figures 4B,D**). Furthermore, the end-systolic RV internal diameter (RVIDs) and the endsystolic RV volume (RV-Vols) did not change following HU, but increased in the HR-7d and HR-14d groups (**Figures 4C,E**). The RV anterior wall thickness (RVAWd and RVAWs) and the interventricular septal thickness (IVSd and IVSs) did not change following HU or HR (**Figures 4F–I**). Together, these data show that HU and HR induce different structural changes in the LV and RV (**Figure 4J**) and, during the recovery process, cardiac remodeling may be intensified because of reloading.

### HU and HR Lead to Cardiac Remodeling

To address the influence of HU and HR in the LV and RV, hearts from mice following HU and HR were assessed for changes in morphology and gene expression. Histological analysis showed heart remodeling occurred following HU and HR. In hematoxylin and eosin-stained (HE) sections, gross evidence of edema was easily observed by separation of the myofibers in the LV, the interventricular septum (IVS), and the RV following HU. Recovery to a normal state occurred following 14 days of HR (**Figure 5A**). Masson trichrome staining (MTT) showed a deeper staining of collagen in the LV, the interventricular septum, and the RV following HU. These changes also recovered after 14 days of reloading. In the HU-28d and HR-7d groups, relative Col1a1, Col3a1, and BNP mRNA levels increased in the LV, and recovered, although not significantly (**Figure 5B**). In the RV, the relative mRNA levels of Col1a1 and Col3a1 increased in the HU-28d group, and only the level of Col1a1 recovered to its normal state in the HR-14d group. The relative mRNA level of BNP increased following HU, and continued to increase

hindlimb unloading and recovery. (A) Echocardiographic assessment of fractional shortening (FS) and ejection factor (EF) of the left ventricle (LV) of mice following HU and recovery. (B) Echocardiographic assessment of fractional shortening (FS) and ejection factor (EF) of the RV of mice following HU and recovery. Values are means ± SEM. \**P* < 0.05, \*\**P* < 0.01.

after 7 days of HR (**Figure 5C**). These results demonstrate that HU leads to slight fibrosis and remodeling in both the left and right ventricle, and recovery occurs slowly after 14 days of reloading.

## Changes in HDAC4, AMPK, ERK1/2, and LC3-II Activity in the Left and Right Ventricles Induced By HU and HR

To gain more insight into the signaling pathways involved in the declining function of both the left and right ventricles, we examined several important signaling molecules involved in cardiac remodeling induced by external or intrinsic stimuli. As shown in **Figure 6A**, quantification of phosphorylation levels normalized to total protein in the LV revealed that HDAC4 phosphorylation at Ser246 did not change following HU but increased following 7 days of reloading, and was fully restored after 14 days. Compared with the control, Erk1/2 phosphorylation at Thr202/Tyr204 increased following HU, remained the higher level during the first 7 days of reloading, and was fully restored after 14 days. The phosphorylation level of AMPK at Thr172 decreased following HU, continued to decrease after 7 days of reloading, and was restored after 14 days. According to the research of Liu et al. (2015), autophagy is involved in HU-induced decline in heart function, so we assessed the change of LC3-II in

following hindlimb unloading and recovery. (J) Summary of Structure and function index of LV and RV. Values are means ± SEM. \**P* < 0.05, \*\**P* < 0.01.

our model. Quantification of LC3-II levels normalized to LC3-I revealed an increased ratio of LC3-II: LC3-I following HU, and this ratio did not return to its normal state until after 14 days of reloading. The changes of these signaling pathways in RV following HU and HR were also analyzed. As shown in **Figure 6B**, the level of HDAC4 phosphorylation at Ser246 in the RV increased following HU, however it was not restored to the normal level until 14 days of reloading, which is in contrast to the changes in LV. Phosphorylation of Erk1/2 at Thr202/Tyr204 was up-regulated increased following HU, and recovered after 7 days of reloading. The phosphorylation of AMPK at Thr172 was reduced following HU, and continued to decrease after 7 days of reloading, however, it was not restored to its normal state until 14 days after reloading, which is different from the changes in the LV. The changes of LC3-II: LC3-I was the same as that for the LV.

We also analyzed the changes of mTOR phosphorylation and MuRF1/Atrogin1 levels, which are involved in protein synthesis and degradation pathways, respectively. In LV, As is shown in **Figure S1A**, mTOR phosphorylation at S2448 was decreased following HU, and was restored following 7 days of reloading. In comparison with the control, the level of Atrogin1, an E3 ubiquitin ligase that mediates proteolysis events during muscle atrophy, obviously increased following HU, and remained much higher level during the first 7 days of reloading, then restored to the normal level after 14 days of reloading. The level of MuRF1, another ubiquitin ligases, was increased in HU group, however, it did not recover to the normal condition even after 14 days of reloading (**Figure S1A**). In RV, as shown in **Figure S1B**, the changes of phosphorylation level of mTOR was the same as that for the LV. The levels of Atrogin1 and MuRF1 were both substantially increased following HU, and restored till 14 days of reloading (**Figure S1B**). The changes of these signaling pathways were closely related to disorders of cardiac function observed both in the LV and RV and the differences between them.

### DISCUSSION

In this study, we report for the first time the differences between left and right ventricular remodeling induced by simulated microgravity and reloading. We also characterize the signaling molecules involved in this cardiac remodeling. Consistent with previous reports, our study indicates that left ventricular function declines following HU but recovers to its normal state after 14 days of reloading. Few studies have focused on the RV, however. We show here that the function of the RV also declines following HU, but does not recover even after 14 days of reloading. In other words, both the left and right ventricle exhibited a decline in function, but recovery of right ventricular function was much more difficult. We demonstrate that pathological remodeling signals, such as HDAC4, were activated following HU and recovered following HR in both the LV and RV. The physiological remodeling signal AMPK was inhibited in both the LV and RV following HU, but only restored in the LV following 14 days of HR.

Several studies have suggested that microgravity or simulated microgravity lead to cardiac remodeling, and result in cardiac deconditioning when reloaded. In humans exposed to 6 weeks of bed rest, LV mass decreased by 8.0 ± 2.2%, RV free wall mass decreased by 10 ± 2.7%, and RV end-diastolic volume decreased by 16 ± 7.9%. After 10 days of spaceflight, LV mass decreased by 12 ± 6.9%. Thus, cardiac atrophy occurs during prolonged horizontal bed rest, but may also occur after short-term spaceflight (Perhonen et al., 2001). Using an experiment with 60 days of sedentary head-down bed rest, one group demonstrated that the reduced LV mass in response to prolonged simulated weightlessness is not simply due to tissue dehydration but rather to true LV remodeling that persists well into recovery (Westby et al., 2016). Previous studies conducted on rats and mice have provided conflicting data. Bigard et al. (1994) demonstrated that LV mass decreased following HU for 21 days in rats. Ray et al. (2001), however, suggested that the mass of the rat heart was unchanged after 28 days of HU. Jennifer et al. (Powers and Bernstein, 2004) also reported that absolute heart weights were not altered significantly after 14 days of tail suspension in mice. Moreover, few studies have focused on right ventricular remodeling induced by space flight or simulated microgravity. In our research, LV mass and structure did not change following 28 days of HU, consistent with some of the previous studies, but RVIDd and RV Vold did decrease, and RV mass trended slightly lower following HU. So, we suggest that the RV is more sensitive than the LV following HU.

Many studies have demonstrated that cardiac remodeling is associated with a decline in heart function induced by microgravity and/or simulated microgravity. However, the changes in heart structure and function during reloading after simulated microgravity are not clear. In our study, heart weight increased significantly following HR for 7 days compared with the HU group, and recovered after 14 days of HR (**Figure 1C**). The masses of both the LV and RV increased following 7 days of HR, although the changes were not significant (**Figures 1E,F**). Echocardiography revealed that LVIDd, LVIDs, LV Vold, and LV Vols increased following 7 days of HR, and recovered after 14 days (**Figures 3B–E**). RVIDs and RV Vols also increased following HR for 7 and 14 days (**Figures 4C,E**). In summary, both the LV and RV were enlarged following 7 days of HR. These results indicate that simulated microgravity leads to cardiac remodeling, and in the early stages of recovery, reloading may intensify remodeling.

The mammalian heart is a muscle, the fundamental function of which is to pump blood throughout the circulatory system. In response to changed workload, typically caused by pathological or physiological stimulation, the heart undergoes remodeling in an attempt to maintain pump function in the new environment (Maillet et al., 2013). A variety of stimuli can induce the heart to grow or shrink. Exercise, pregnancy, and postnatal growth promote physiologic growth of the heart,

FIGURE 6 | Activity of signaling pathways in the mouse heart following hindlimb unloading and recovery. (A,F) Representative western blots of HDAC4 and its phosphorylation at Ser246, AMPKα and its phosphorylation at Thr172, ERK1/2, and its phosphorylation at Thr202/Tyr204, and LC3 of the left ventricle (LV) and right ventricle (RV). Gapdh levels served as a loading control. Quantification of phosphorylation levels normalized to total protein (LC3-II levels normalized to LC3-I levels) of the LV (B–E) and RV (G–J). (K) Summary of changed signaling molecules. HDAC4, Histone Deacetylase 4; ERK, Extracellular Regulated Protein Kinases; AMPK, AMP-activated Protein Kinase; LC3, Microtubule-associated Protein Light Chain 3. Gapdh, Glyceraldehyde phosphate dehydrogenase. Values are means ± SEM (*n* = 4). \**P* < 0.05, \*\**P* < 0.01, \*\*\**P* < 0.001.

while neurohumoral activation, hypertension, and myocardial injury can cause pathologic hypertrophic growth. As with other forms of cardiac remodeling, ventricular atrophy is induced by prolonged weightlessness during space travel, prolonged bed rest, and mechanical unloading with a ventricular assist device (Hill and Olson, 2008). Well-characterized signaling molecules that regulate cardiac remodeling include HDAC4, AMPK, ERK1/2, LC3-II, mTOR, Atrogin1, and MuRF1. HDAC4, a key member of class IIa HDACs (HDACs 4, 5, 7, and 9), is expressed in numerous tissues, and plays an important role in the modulation of biological responses and pathological disorders (Yang and Grégoire, 2005; Backs and Olson, 2006; Wang et al., 2014). Phosphorylated HDAC4 is exported to the cytoplasm from the nucleus, with consequent activation of MEF2 and its downstream target genes involved in pathological cardiac remodeling (Passier et al., 2000; Haberland et al., 2009; Ling et al., 2012). AMPK is a stress-activated kinase which functions as a cellular fuel gauge and master metabolic regulator, and is therefore crucial to cardiac homeostasis (Coughlan et al., 2014). The activation of heart AMPK is associated with the translocation of GLUT4 and phosphorylation of acetyl-CoA carboxylase (ACC), which promote ATP production by stimulating fatty acid oxidation, glucose uptake, and glycolysis (Coven et al., 2003; Maillet et al., 2013). AMPK is important for maintaining the physiological growth of the heart. ERK1/2 belongs to the mitogen-activated protein kinase (MAPK) family, and its activation has been reported to mediate both pathological and physiological cardiac remodeling (Tham et al., 2015). According to the research of Liu et al. (2015), autophagy is involved in HU-induced LV decline in function; LC3-II expression increased in the LV after HU. mTOR is an atypical serine/threonine protein kinase that belongs to the phosphoinositide 3-kinase (PI3K)-related kinase family and interacts with several proteins to form two distinct complexes named mTOR complex 1 (mTORC1) and 2 (mTORC2; Laplante and Sabatini, 2012). In muscle, activation of mTORC1 can stimulate protein synthesis to drive muscle hypertrophy (Philp et al., 2011). MuRF-1 and Atrogin 1 are two identified muscle specific ubiquitin ligases, which have been shown to be upregulated prior to the onset of atrophy in multiple models of muscle wasting (Bodine et al., 2001). In this study, we detected these molecular signals in the LV and RV after HU and HR, and we explored the molecular mechanism of LV and RV remodeling induced by simulated microgravity and recovery. We found that the phosphorylation of HDAC4 at Ser246 was upregulated following HU for 28 days. This phosphorylation remained high in the RV (**Figures 6F,G**), and increased in the LV following 7 days of HR (**Figures 6A,B**). Meanwhile, the phosphorylation levels of ERK1/2 increased in both the LV and RV following HU for 28 days (**Figures 6A,C,F,H**), and further increased in the LV following 7 days of HR. Our results also showed that the ratio of LC3-II:LC3-I increased in both the LV and RV following 28 days of HU and 7 days of HR (**Figures 6A,E,F,J**). Autophagy was activated in both the LV and RV, consistent with previous reports (Liu et al., 2015). These results indicate that both HU-simulated microgravity and reloading can activate pathological cardiac remodeling signaling pathways, which can initiate the expression of fetal genes in both the LV and RV, and ultimately lead to cardiac remodeling. Following HU, The phosphorylation of mTOR at S2448 was decreased both in LV and RV (**Figure S1**), protein synthesis pathway was inhibited. The levels of Atrogin1 and MuRF1 were increased in both LV and RV following 28 days of HU and 7 days of HR (**Figure S1**), which suggest that ubiquitin-proteasome system was activated both in LV and RV. The changes of these proteins contributed to the cardiac remodeling. The phosphorylation level of AMPK at Thr172 decreased following 28 days of HU and continued to decrease following 7 days of HR in both the LV and RV (**Figures 6A,D,F,I**). Moreover, the phosphorylation of AMPK returned to a normal level in the LV following 14 days of HR. This did not occur in the RV, however. Interestingly, levels of AMPK phosphorylation were consistent with the functional changes in both the LV and RV. The physiological remodeling signal AMPK decreased following HU in both the LV and RV, and did not return to its normal state in the RV following 14 days of HR. This may at least partially explain the different responses of the RV and LV following HU and HR.

This study provides evidence of the differences in the responses of the LV and RV under simulated microgravity and the signaling molecules involved in this process. We found that simulated microgravity leads to cardiac remodeling, and this remodeling could be reversed. In the early stages of recovery, reloading may intensify cardiac remodeling. Moreover, it is more difficult to restore the changes in the RV compared with the LV. Finally, we identified that following HU and HR, pathological remodeling signals, such as HDAC4, were activated, and physiological remodeling signals, such as AMPK, were inactivated in both the LV and RV, which led to cardiac remodeling and decline of heart function (**Figure 6K**).

### AUTHOR CONTRIBUTIONS

YXL and SL conceived the study, GZ performed the experiment with support from WS, DC, QX, HC, and HL; GZ, YHL, JL, and DZ analyzed and interpreted the results; JS, XW, GK, and QL provided intellectual contribution; GZ, SL wrote the manuscript with the assistance of XJ, HS, and XY; YXL, SL, and YHL revised the manuscript and gave final approval of the submitted manuscript. All authors have reviewed and approved the final manuscript. The English in this document has been checked by at least two professional editors, both native speakers of English. For a certificate, please see: http://www.textcheck.com/certificate/UQCwt5.

### FUNDING

This work was supported by the National Natural Science Foundation of China (No. 31300698, 31271225, and 31325012) and the Grant of State Key Lab of Space Medicine Fundamentals and Application (No. SYFD130051833, SYFD140041803).

### SUPPLEMENTARY MATERIAL

The Supplementary Material for this article can be found online at: http://journal.frontiersin.org/article/10.3389/fphys. 2016.00274

### REFERENCES


Figure S1 | The changes of protein levels involved in protein synthesis and

proteolysis. Representative western blots of mTOR and its phosphorylation at Ser2448, Atrogin1, and MuRF1 of the left ventricle (A) and right ventricle (B). Gapdh levels served as a loading control. mTOR, Mammalian Target of Rapamycin; MuRF-1, Muscle Ring Finger 1.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer JA and handling Editor declared their shared affiliation, and the handling Editor states that the process nevertheless met the standards of a fair and objective review.

Copyright © 2016 Zhong, Li, Li, Sun, Cao, Li, Zhao, Song, Jin, Song, Yuan, Wu, Li, Xu, Kan, Cao, Ling and Li. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

# A Disintegrin and Metalloprotease 17 in the Cardiovascular and Central Nervous Systems

Jiaxi Xu<sup>1</sup> , Snigdha Mukerjee<sup>1</sup> , Cristiane R. A. Silva-Alves <sup>2</sup> , Alynne Carvalho-Galvão<sup>2</sup> , Josiane C. Cruz <sup>2</sup> , Camille M. Balarini <sup>3</sup> , Valdir A. Braga<sup>2</sup> , Eric Lazartigues <sup>1</sup> \* and Maria S. França-Silva<sup>2</sup> \*

<sup>1</sup> Department of Pharmacology and Experimental Therapeutics and Cardiovascular Center of Excellence, Louisiana State University Health Sciences Center, New Orleans, LA, USA, <sup>2</sup> Centro de Biotecnologia, Universidade Federal da Paraíba, João Pessoa, Brazil, <sup>3</sup> Centro de Ciências da Saúde, Universidade Federal da Paraíba, João Pessoa, Brazil

### Edited by:

Yusuke Sata, Baker IDI Heart and Diabetes Institute, Australia

### Reviewed by:

Chikara Abe, Gifu University Graduate School of Medicine, Japan Yoshio Ijiri, Osaka University of Pharmaceutical Sciences, Japan

### \*Correspondence:

Eric Lazartigues elazar@lsuhsc.edu Maria S. França-Silva francasilva@cbiotec.ufpb.br

### Specialty section:

This article was submitted to Integrative Physiology, a section of the journal Frontiers in Physiology

Received: 26 August 2016 Accepted: 30 September 2016 Published: 18 October 2016

### Citation:

Xu J, Mukerjee S, Silva-Alves CRA, Carvalho-Galvão A, Cruz JC, Balarini CM, Braga VA, Lazartigues E and França-Silva MS (2016) A Disintegrin and Metalloprotease 17 in the Cardiovascular and Central Nervous Systems. Front. Physiol. 7:469. doi: 10.3389/fphys.2016.00469 ADAM17 is a metalloprotease and disintegrin that lodges in the plasmatic membrane of several cell types and is able to cleave a wide variety of cell surface proteins. It is somatically expressed in mammalian organisms and its proteolytic action influences several physiological and pathological processes. This review focuses on the structure of ADAM17, its signaling in the cardiovascular system and its participation in certain disorders involving the heart, blood vessels, and neural regulation of autonomic and cardiovascular modulation.

Keywords: hypertension, atherosclerosis, shedding, TACE, TNFα, EGFR, ACE2

## INTRODUCTION

A Disintegrin and Metalloproteases (ADAM), originally named metalloproteinases disintegrin cystein-rich (MDC), are membrane-anchored cell surface proteins containing both disintegrin and metalloproteinase domains. They belong to the adamalysin protein family in the zinc protease superfamily and combine features of both proteases and cell surface adhesion molecules (Wolfsberg and White, 1996; Seals and Courtneidge, 2003). By 2010, 40 ADAM homologs had been identified in the mammalian genome and 21 ADAM described, among which only 13 were proteolytically active (Edwards et al., 2008). The ADAMs responsible for protein cleavage are called sheddases. It is estimated that as much as 10% of the cell surface proteins undergo ectodomain shedding. Members of the ADAM family contribute to various physiological and pathophysiological processes by modulation of molecules like growth factors or cytokines. The prototype of ADAM with sheddase activity is exemplified in ADAM17.

ADAM17 (EC 3.4.24.86), also known as tumor necrosis factor-α converting enzyme (TACE), is a metallopeptidase which has extensive somatic distribution, being expressed significantly in the heart, vessels, brain, kidney, testicle, placenta, ovary, lung, spleen, and skeletal muscle, with levels varying from embryonic development to adulthood (Black et al., 1997; Patel et al., 1998; Sahin et al., 2004; Dreymueller et al., 2012). ADAM17 was described for the first time by Black et al., as a new member of the family of mammalian adamalysins, to specifically cleave the precursor of tumor necrosis factor α (pro-TNFα; Black et al., 1997; Moss et al., 1997). Until now, ADAM17 was revealed as the major sheddase, of which the substrates cover a diverse variety of membrane-anchored cytokines, cell adhesion molecules, receptors, ligands, and enzymes, such as TNFα, transforming growth factor α (TGFα), L-selectin, and angiotensin-converting enzyme type 2 (ACE2; Kishimoto et al., 1989; Sahin et al., 2004; Weskamp et al., 2004; Lambert et al., 2005; Scheller et al., 2011).

As predicted by the wide variety of ADAM17 substrates, global disruption of the ADAM17 gene in vivo results in the death of mice between embryonic day 17.5 and the first day after birth, due to a number of developmental defects from brain, heart, lung, skin, skeletal, and immune system (Peschon et al., 1998; Jackson et al., 2003). In human, a few cases have shown a rare syndrome, in which patients with a homozygous mutation in ADAM17 present severe diarrhea, skin rashes and recurrent sepsis, eventually leading to their early death (Bandsma et al., 2015). On the other hand, due to the structural and functional heterogeneity of ADAM17 substrates, the sheddase is also involved in various pathological processes such as cancer, inflammatory diseases, neurological diseases, cardiac failure, atherosclerosis, diabetes, cardiac hypertrophy, and hypertension (Sandgren et al., 1990; Black et al., 1997; Peschon et al., 1998; Li et al., 2006; Ohtsu et al., 2006b; Zhan et al., 2007; Wang et al., 2009; Scheller et al., 2011; Giricz et al., 2013; Menghini et al., 2013; Xia et al., 2013).

In this review, we summarize the seemingly paradoxical functions of ADAM17 with a particular emphasis on the cardiovascular and central nervous systems (CNS). The structure and structure-based modulation of ADAM17 are also described for better understanding of the various ADAM17 regulatory pathways in different cell types or tissues. Finally, we also discuss the contribution of ADAM17 as a potential therapeutic target in cardiovascular disorder and the neurogenic component of these cardiovascular diseases.

## ADAM17 STRUCTURE

ADAM17 has the structural characteristics of disintegrin and metalloproteases proteins that exist predominantly in two forms: as the full-length protein (∼100 kDa) and as a mature form lacking the pro-domain (∼80 kDa; Gilles et al., 2003). As shown in **Figure 1**, the full-length protein (or pro-ADAM17) is composed of 824 amino acids and consists of a series of conserved protein domains: an N-terminal signal sequence (aa 1–17), followed by a pro-domain (aa 18–216), in which there is a cysteine switch-like region (CysSL) PKVCGY<sup>186</sup> (aa 181–188), a catalytic domain (aa 217–474) with a Zn-binding domain region (Zn-BR) (aa 405–417), a disintegrin cysteinerich domain (aa 480–559), an EGF-like region (aa 571–602), a transmembrane domain (aa 672–694), and a cytoplasmic tail (aa 695–824). Tyr702, Thr735, Ser<sup>819</sup> have been shown as cytoplasmic phosphorylation (P) sites, Ser<sup>791</sup> has been shown as cytoplasmic desphosphorylation site (Patel et al., 1998; Ohtsu et al., 2006a; Gooz, 2010; Xu and Derynck, 2010; Niu et al., 2015).

The pro-domain of ADAM17 includes a cysteine switch box (Galazka et al., 1996), which is an unpaired cysteine residue and play a key role in the pro-domain release prior to ADAM17 activation (Van Wart and Birkedal-Hansen, 1990; Roghani et al., 1999). Normally, the pro-domain of ADAM17 acts as an inhibitor of the protease via linkage of a cysteine switch box (SH-group) to the zinc atom in the active catalytic site. Removal of the prodomain is a pre-requisite for ADAM17 activation (Reiss and Saftig, 2009). ADAM17 is synthesized and stored in the rough endoplasmic reticulum, the pro-domain removal occurs in a late Golgi compartment, most likely by furin or furin-like proprotein convertase or by autocatalysis, providing the mature form of ADAM17 (Adrain and Freeman, 2012; Dreymueller et al., 2012). Data from Srour et al., using in vitro and in vivo cleavage assays, highlight the pro-protein convertases PACE-4, PC5/PC6, PC1, and PC2, as some examples of furin-like enzymes that can directly cleave the ADAM17 protein (Srour et al., 2003). After maturation, ADAM17 translocates to the cell surface to perform proteolytic and non-proteolytic functions (Zhang et al., 2016). Since the pro-domain is highly sensitive to proteolysis, once detached from catalytic domain, it will be degraded rapidly, preventing its re-association with this domain. It has also been suggested that changes in conformation, mediated by the cysteine-rich domain, result in reducing the affinity between proand catalytic domains (Milla et al., 1999). Another function of the pro-domain is to chaperone the proper folding of ADAM17 and other ADAM (Loechel et al., 1999; Milla et al., 1999; Anders et al., 2001; Leonard et al., 2005; Li et al., 2009).

The catalytic or metalloproteinase domain starts downstream from the consensus furin cleavage site (RVKRR215) and contains a canonical zinc-binding site and a consensus (HEXXH) motif, crucial for the catalytic activity. It mediates shedding of membrane-bound proteins (Maskos et al., 1998; Ohtsu et al., 2006a).

The cysteine-rich domain of ADAM17 includes a disintegrinlike region that can bind to integrins, therefore ADAM17 exhibit both proteolytic and adhesive characteristics (Zhang et al., 2016). Furthermore, Reddy et al. (2000) suggested that the cysteinerich domain might be involved in recognition of substrates. Following the disintegrin domain, a single transmembrane domain defines the end of the catalytic domain of ADAM17 and then followed by a cytosolic tail that contains consensus sequences for binding to proteins containing Src homology 2 and Src homology 3 domains. Little has been known about the role of this domain in regulating catalytic activity and physiological substrates recognition (Qi et al., 1999).

The role of the cytoplasmic domain in ectodomain shedding still remains controversial. Fan et al. and Gechtman et al. reported that the cytoplasmic domain of ADAM17 is involved in the regulation of ectodomain cleavage in response to intracellular signaling events such as receptor protein tyrosine kinase and mitogen-activated protein kinases (MAPK) activation (Fan and Derynck, 1999; Gechtman et al., 1999). Arguably, soluble ADAM17 forms, lacking trans-membrane and cytoplasmic domain, can cleave trans-membrane proteins or synthetic peptide substrates but not very efficiently (Kahn et al., 1998; Reddy et al., 2000; Fan et al., 2003; Gonzales et al., 2004; Li and Fan, 2004). However, it was soon found that the ADAM17 transmembrane domain proved to play a critical role in appropriately regulating substrate recognition. When the ADAM17 trans-membrane domain was switched with the transmembrane domain of prolactin receptor and platelet-derived growth factor receptor (PDGFR) it strongly inhibit TGFα release, with no change in L-selectin shedding. However, switching the transmembrane domain

of TGFα with that of L-selectin actually restored cleavage activity of the transmembrane switched ADAM17 (Li et al., 2007).

### POST-TRANSLATIONAL REGULATION OF ADAM17

The fact that ADAM17 has a wide diversity of substrates raise the questions about where, when, and how it is activated. Shedding activity led by ADAM17 is sequestered in cholesterol microdomains within the cellular membrane. Cholesterol is often found to be distributed non-randomly in domains or pools in membranes, therefore these cholesterol-enriched microdomain, also known as "lipid raft," is thought to be an important microenvironment, in which the signaling pathway can be regulated specifically (Pucadyil and Chattopadhyay, 2006). Depletion of membrane cholesterol by methyl-β-cyclodextrin can induce ADAM17-dependent shedding (Tellier et al., 2006).

It was shown in an early report that ADAM17 cytoplasmic domain was not critical for the phorbol-ester (PMA)-induced TNFα shedding (Reddy et al., 2000). However, considering that PMA is a strong and pleiotropic activator, this cytoplasmic tail could still be necessary for stimulations working through physiological signaling pathways. The cytoplasmic domain of ADAM17, which contains putative phosphorylation sites, is thought to be required for regulation of the metalloprotease activity via several intracellular signals. Now the involvement of MAPK in ADAM17 activation is well-accepted. So far, two phosphorylation sites (Thr735, Ser819) and one dephosphorylation site (Ser791) located on the cytoplasmic tail have been recognized. Mutation or dephosphorylation on Ser<sup>791</sup> enhance ADAM17 phosphorylation at Thr735, which can be decreased by mutation on Ser<sup>819</sup> (Xu and Derynck, 2010). Replacing Thr<sup>735</sup> with alanine, which cannot be phosphorylated, resulted in accumulation of non-functional ADAM17 on the cell surface. Both extracellular signal-regulated kinase (ERK) and p38-MAPK pathways activate ADAM17 through Thr<sup>735</sup> phosphorylation. In mammalian cells, ERK-mediated phosphorylation of ADAM17 at Thr<sup>735</sup> highlights a key step in inducible ADAM17 protein trafficking and maturation (Soond et al., 2005). Protein kinase C (PKC) used to be considered as an upstream signal of ERK/MAPK pathway, however, it may also have direct role in ADAM17 phosphorylation. A recent study showed that the PKCδ inhibitor, rottlerin, significantly inhibited both constitutive, and high glucose-induced ACE2 shedding, which is mediated by ADAM17, and treatment with PKCδ-targeting siRNA reduced ACE2 shedding by ∼20% (Xiao et al., 2016). Other protein kinases, such as protein kinase G (PKG), phosphoinositide-dependent kinase 1 (PDK1), and Polo-like kinase 2 (PLK2), were also involved in ADAM17 phosphorylation in various cell types (Zhang et al., 2006; Chanthaphavong et al., 2012; Schwarz et al., 2014).

PMA can increase ADAM17 activity in multiple ways. An early report suggested that PMA might activate ADAM17 through ROS generation in monocytic cell line (Zhang et al., 2001). Actually, mutation of Thr<sup>735</sup> had little effect on PMAstimulated ADAM17 activation. Therefore, other than being an upstream signal of ERK/MAPK pathway, ROS may also have a role in translocating mature ADAM17 from Golgi to the cell surface. In p47phox KO mice, translocation of ADAM17 was partially blunted in cardiomyocytes when the ROS production pathway was blocked (Patel et al., 2014). In addition to ROS, ADAM17-mediated ectodomain shedding can be stimulated by calcium influx in a short term manner. Stimulation with ionomycin (a calcium ionophore) rapidly increased the level of interleukin-6 receptor (IL-6R) shedding in mice, however, it is still unclear whether other ADAM might contribute to this process (Garbers et al., 2011).

Though cytoplasmic phosphorylation is not required for PMA-induced ADAM17 activation, the presence of iRhom2 (inactive rhomboid type 2) was found to be indispensable. Actually, both iRhom1 and iRhom2 are involved in the maturation of ADAM17, and they can compensate each other's function in most of tissues except for the brain and immune system (Maretzky et al., 2013; Li et al., 2015). iRhom1/2 regulate forward trafficking of ADAM17 from endoplasmic reticulum (ER) to Golgi compartment where the pro-domain of ADAM17 can be cleaved (Adrain and Freeman, 2012). However, the underlying mechanism of this phenomenon remains unknown.

### ADAM17-MEDIATED SIGNALING IN DEVELOPMENTAL PROCESS

### Notch Signaling

ADAM-mediated shedding of Notch receptor is a key step in activating Notch signaling. The Notch receptor and its ligand Delta-l1 are required for neuroepithelial development during embryogenesis (Wakamatsu et al., 1999; De Bellard et al., 2002). The Notch signaling network is an evolutionarily conserved intercellular signaling pathway that regulates interactions between physically adjacent cells. In the brain, Notch1 promotes differentiation of progenitor cells into astroglia and knockingout the Notch1 gene is lethal at birth (Miller and Gauthier, 2007). Notch receptor is activated by one of five ligands: Jagged1, Jagged2, Delta-l (like) 1, Delta-l3, or Delta-l4, expressed on adjacent cells. Following ligand binding, Notch receptor is cleaved by ADAM17 at site 2 (S2) and then by γ-secretase at site 3 (S3) (Kopan and Ilagan, 2009), and so is the ectodomain release of Delta and Jagged (Murthy et al., 2012; Groot et al., 2013). Both murine Notch1 and Notch2 require ADAM17 instead of ADAM10 during ligand-independent activation, which is resisted by the human Notch2 (Habets et al., 2015). Since there are major cardiovascular defects in the ADAM17 global knockout mice, which failed to survive postnatal, ADAM17 would also play a role in vascular physiology. In the adult epidermis, ADAM17 permits tonic Notch activation to regulate epithelial cytokine production and maintain barrier immunity. During cerebral angiogenesis, Notch signaling is initiated by receptor-ligand recognition between adjacent cells (Serra et al., 2015). Overexpression of ADAM17 can promote angiogenesis by increasing blood vessel sprouting and pericytes number during brain micro vessel development (Lin et al., 2011).

### EGFR Signaling

Within the cardiovascular system, Shi et al. (2003) demonstrated that ADAM17 possibly regulates cardiomyocyte proliferation during the late fetal stage of cardiac development via an epithelial growth factor receptor (EGFR)-mediated pathway. It was showing in the ADAM17 global knockout mice, a remarkably enlarged heart with increased myocardial trabeculation and reduced cell compaction (Shi et al., 2003). It has also been reported that mice lacking functional ADAM17 suffer from several cardiac abnormalities in valvulogenesis, such as thickened, deformed semilunar, and atrioventricular valves (Jackson et al., 2003). Importantly, these lethal cardiac abnormalities resulted from insufficient EGFR activation, which were due to the lack of ADAM17 (Blobel, 2005). The upregulation of EGFR by ADAM17 stems from the critical ability of this metalloprotease to cleave multiple EGFR ligands, such as EGF itself, TGFα; epiregulin; heparin binding EGF-like growth factor (HB-EGF), and neuregulins β1 and β2 (Peschon et al., 1998; Merlos-Suárez et al., 2001; Hinkle et al., 2004; Sahin et al., 2004; Horiuchi et al., 2005; Saftig and Reiss, 2011). Neuregulins β1/2 are considered to be critical to cardiac development (Meyer and Birchmeier, 1995; Meyer et al., 1997; Britsch et al., 1998). Arguing against potential compensatory or redundant functions of other members from the ADAM family, quadruple knockout mice lacking ADAM 9, 12, 15, and 17 do not have a more severe phenotype than mice lacking only ADAM17 (Sahin et al., 2004).

In addition to the defects on cardiac development, removal of ADAM17 from subcortical white matter (SCWM) of postnatal mice (90 days) led to abnormalities in Schwann myelination causing impaired motor behaviors (Palazuelos et al., 2014). ADAM17-induced shedding of EGFR ligands, HB-EGF and TGFα, promotes the expansion of oligodendrocyte progenitor cells during the critical periods of Schwann myelination. However, neuregulin-1 cleavage mediated by ADAM17 can negatively regulate myelination in the peripheral nervous system. Down-regulation of ADAM17 expression or its inactivation in motor neurons can lead to abnormal myelination (La Marca et al., 2011).

## ADAM17-MEDIATED CYTOKINES AND ADHESION MOLECULES SIGNALING

### In the Heart

Soluble TNFα, a potent cytokine, is converted from pro-TNFα through ADAM17-mediated cleavage in various cell types. In heart, TNFα is crucially involved in the genesis and progression of several cardiovascular processes. Satoh et al. (1999) have shown that myocardial ADAM17 and TNFα expression in both mRNA and protein levels are increased in humans with dilated cardiomyopathy (Satoh et al., 1999), establishing a positive correlation between ADAM17 and TNFα in cardiovascular disorder. They later showed that increased expression of TNFα and ADAM17 had important implications in advanced cardiac dysfunction in myocarditis. In the myocardium, TNFα contributes to reversible and irreversible ischemia/reperfusion injury, post myocardial infarction remodeling, and heart failure development. Simultaneously, TNFα is also involved in cardioprotective processes in ischemic conditioning. The harmful and beneficial roles of TNFα appear to be dose- and time-dependent and in part related to the activation of specific receptor subtypes (Kleinbongard et al., 2010).

One possible pathway to activate TNFα in the myocardium is through EGFR and many studies have shown that ADAM17 is the main contributor to EGFR transactivation in cardiovascular system (Ohtsu et al., 2006b; Takayanagi et al., 2015). Furthermore, Küper et al. (2007) were the first to suggest a co-dependency between ADAM17-mediated up-regulation of EGFR and TNFα. They observed that within renal collecting duct cells LPS induces EGFR activation via TLR4/ADAM17 (Küper et al., 2007). Based on this study, Sun et al. (2015) have found that this promotes myocardial TNFα production and cardiac failure in endotoxemic mice (Sun et al., 2015). Another TNF-α activation pathway in cardiac cells by ADAM17 is through the non-receptor tyrosine kinase, Src. Src mediates ADAM17 activation in mechanically stretched rat cardiomyocytes by phosphorylating the Tyr<sup>702</sup> residue within the intracellular tail of ADAM17, leading to activation of p38 MAPK and thus TNF-α receptor activation (Niu et al., 2015).

The association between ADAM17 and cardiac function is not limited to local effects in the heart. Akatsu et al. (2003) demonstrated an increased expression of TNFα and ADAM17 in circulating leukocytes of patients with acute myocardial infarction (AMI) associated with higher plasma TNFα levels when compared with healthy control patients (Akatsu et al., 2003). Shimoda et al. (2005) confirmed the positive correlation between ADAM17 and TNFα levels in AMI (Shimoda et al., 2005). In addition, they observed that these levels were higher in peripheral blood mononuclear cells from AMI patients with in-hospital complications, such as pump failure, malignant ventricular arrhythmia, recurrent myocardial infarction, and cardiac death. Moreover, ADAM17 expression and TNFα levels were found to be increased in peripheral blood mononuclear cells, especially from individuals with advanced congestive heart failure (Satoh et al., 2004).

### In the Vasculature

ADAM17 expression is found throughout endothelial cells, smooth muscle cells, fibroblasts, and leukocytes (Dreymueller et al., 2012). The consequences of shedding events for vascular biology depend on the type of substrate shed. This can result in pro- or anti-inflammatory effects depending on the nature of the transmembrane protein cleaved, the generation of soluble receptor agonists or antagonists, modulation of cellular responsiveness, modulation of adhesive properties, and formation of transcription factors (Canault et al., 2006). Interestingly, ADAM17 has been described to be overexpressed in ruptured coronary plaques from infarcted patients and atherosclerotic plaques from apolipoprotein E knockout mice (apoE−/−, an important experimental model of atherosclerosis; Canault et al., 2006; Satoh et al., 2008), revealing its fundamental role in this vascular disease.

ADAM17 activity is correlated to adverse clinical outcomes in acute coronary atherosclerosis (Satoh et al., 2008; Gutiérrez-López et al., 2011; Rizza et al., 2015). An important group of substrates for ADAM17 includes adhesion molecules such as ICAM-1 (intercellular adhesion molecule-1), VCAM-1 (vascular cell adhesion molecule-1), L-selectin, and others. These adhesion molecules are involved in leukocyte migration through the vessel wall which is one of the first steps leading to atheroma formation (Gutiérrez-López et al., 2011; Freitas Lima et al., 2015). Shedding of cell adhesion molecules weakens cell–cell interactions and reduces adhesiveness of leukocytes (Arribas and Esselens, 2009; Dreymueller et al., 2012). It has been reported that ADAM17 null leukocytes present increase L-selectin levels (Tang et al., 2011). Interestingly, Zhang et al. demonstrated that nitric oxide (NO) is a critical factor involved in ADAM17 activation due to nitrosilation of thiol group in cysteine residues in the inhibitory pro-domain of ADAM17. This might explain the increased adhesion of leukocytes in dysfunctional endothelium (Zhang et al., 2000), which presents reduced NO bioavailability (Balarini et al., 2013).

### In the Central Nervous System

Kärkkäinen et al. (2000) was the first to demonstrate the expression of ADAM17 mRNAs in adult mouse and rat brains, using in situ hybridization (Kärkkäinen et al., 2000). ADAM17 showed a restricted pattern of distribution in the telencephalon and diencephalon. Within the mesencephalon, ADAM17 mRNA was detected in the hippocampus along with the inferior colliculus. In addition, ADAM17 mRNA was also detected in the cerebellar cortex, pontine nuclei, and cerebral cortex. According to these data, Hurtado et al. (2001) demonstrated constitutive expression of ADAM17 protein in rat brain (Hurtado et al., 2001). Aspects related to cell location and origin of ADAM17 have been clarified in later studies.

Goddard et al. (2001) in an immunohistochemical study showed that ADAM17 is expressed in astrocytes and endothelial cells from healthy adult human brain tissue and may have a role in normal brain function (Goddard et al., 2001). Hurtado et al. (2002) also found ADAM17 protein expression in astrocytes as well as rat microglial cells (Hurtado et al., 2002). Skovronsky et al. (2001), using Western Blots and immunolabeling approaches, found that ADAM17 was present in neurons of the hippocampus, cortex and cerebellum, and was slightly detectable in astrocytes, oligodendrocytes, and microglial cells (Skovronsky et al., 2001). Since ADAM17 is expressed in different cell types in the CNS and promotes ectodomain shedding of several molecules, this enzyme would participate in several cellular events.

Increased plasma ADAM17 activity was found in patients with mild cognitive impairment. Within the brain, ADAM17 can work as an α-secretase and cleave the amyloid precursor protein (APP) into soluble APPα fragment (sAPPα; Skovronsky et al., 2001; Asai et al., 2003). This shedding of APP is able to reduce the generation of neurotoxic amyloid β (Aβ) peptide in a competitive way. It is believed to be a possible neural protection and repair mechanism. On the contrary, ADAM17 is also a neuro-inflammatory target which responds to ischemia and other stress factors (Romera et al., 2004; Munhoz et al., 2008). Some of the subsequently increased inflammatory factors, like TNFα, can activate glial cells and be harmful to homeostasis within the brain (Suzumura, 2013). From this point of view, activation of ADAM17 can be neurotoxic. As described before, ADAM17 was initially reported to be responsible for the proteolytic activation of the membrane precursor of TNFα, which is critically involved in inflammation. The pro-inflammatory activity of TNFα is predominantly mediated by TNFα receptor type 1 (TNFR1) and to some extent by TNF receptor type 2 (TNFR2), both of which have been found to undergo shedding via ADAM17 (Chanthaphavong et al., 2012). Wang et al. (2003) had linked the shedding of TNF receptors to neutralization of soluble TNFα-induced actions. Either ADAM17 expression or activity have found to be altered in neuro-inflammatory conditions, in which, the level of TNFα is increased, such as stroke, multiple sclerosis, or traumatic brain injury (Romera et al., 2004; Plumb et al., 2006). Conditional deletion of the ADAM17 gene or inhibition of the protein, effectively blocks LPS-induced TNFα release and systemic inflammation in mice (Zhang et al., 2004). This demonstrates that ADAM17-mediated shedding is a critical trigger for pro-inflammatory signaling of TNFα. In addition, BB1101, a blocker of ADAM17, was shown to be effective in reducing TNFα levels in stressed animal models (Madrigal et al., 2002). Interestingly, the specific antagonist of NMDA receptor, MK-801, can, not only decrease stress-induced activity and expression of ADAM17, but also its constitutive expression, as well as TNFα levels (Madrigal et al., 2002). Therefore, in the CNS, excitation or stimulations targeting glutamatergic neurons may cause up-regulation of ADAM17 activity and will contribute to neuro-inflammation eventually. The mechanism of this neuro-excitation-induced ADAM17 activation is thought to be mediated via nuclear factor-κ B (NFκB). Stress-induced neuro-inflammation activates NFκB in the hippocampus as soon as 4 h after the onset of stress in rats (Madrigal et al., 2006). The implication of ADAM17 and TNFα in stressinduced activation of NFκB was confirmed after finding that decreased TNFα is mediated by hepatocyte growth factor-like

protein-induced decrease in NFκB activation and increased by the NFκB inhibitory protein, IκB (Nikolaidis et al., 2010).

Besides TNFα, inflammation is also characterized by elevated levels of interleukin-6 (IL-6). ADAM17 is the main sheddase for IL-6 receptor (IL-6R) to induce IL-6 trans-signaling (Pruessmeyer and Ludwig, 2009). Released IL-6R is capable of binding IL-6 and the formed receptor-ligand complex then stimulates gp130 on the cell surface, even in the absence of transmembrane IL-6R (Scheller et al., 2011). This pathway is thought to play an important role in chronic inflammatory processes. In immune cells, like neutrophils, ADAM17 and secondary shedding of IL-6R can be activated by apoptosis (Chalaris et al., 2007).

In the nervous system, expression of cell adhesion molecules (CAMs) can be up-regulated by glia-secreted cytokines. Shedding of the CAMs, such as CX3CL1 (fractalkine), L-selectin, VCAM-1, and JAM-A (junctional adhesion molecule-A) are induced by activation of ADAM10 or ADAM17 (Kalus et al., 2006; Pruessmeyer and Ludwig, 2009). In the periphery, CX3CL1 guides leukocytes to the site of inflammation. Within 4–6 h after onset of ischemia, circulating leukocytes adhere to vessel walls and migrate into the brain with subsequent release of additional proinflammatory mediators. The ADAM17-related adhesion procedure is critical during the process of inflammation (Lakhan et al., 2009). On the other hand, CX3CL1 can be secreted from damaged neurons and then activate microglia to rescue neurons by up-regulating phagocytosis of toxicants or damaged debris and production of anti-oxidant enzymes, like superoxide dismutase (Vernon and Tang, 2013; Limatola and Ransohoff, 2014). Furthermore, ADAM17-mediated shedding is important to neuronal outgrowth in the developmental process. ADAM17 regulates L1CAM (L1-cell adhesion molecule)-dependent neuronal cell adhesion, cell migration, as well as neurite outgrowth (Mechtersheimer et al., 2001; Maretzky et al., 2005). L1CAM has been observed within late embryonic/early postnatal cortical neurons and fibers in the corpus callosum and corticospinal tract (Fujimori et al., 2000; Jakovcevski et al., 2013). In mice embryonic cortex, knocking-down ADAM17 could affect L1-involed neuronal intermediate progenitor cells multipolar exit and migration (Li et al., 2013).

### ADAM17-MEDIATED GROWTH FACTORS SIGNALING

### In Vascular Remodeling and Formation of Atherosclerotic Plaques

ADAM17 is expressed in atherosclerosis-prone sites, however it is not clear whether this expression is due to an upregulation of enzyme expression or due to its expression by newly infiltrated vascular smooth muscle cells (VSMCs) and leukocytes (Canault et al., 2006). It is well-established that Ang-II can induce VSMC hypertrophy, proliferation and migration, which are important steps toward atheroma formation (Freitas Lima et al., 2015). Ang-II can induce ADAM17 protein expression in the vasculature and increase its activity by tyrosine phosphorylation (Elliott et al., 2013; Obama et al., 2015). ADAM17-induced EGFR transactivation is considered a major contributor in Ang-II-induced vascular remodeling. It was reported that Ang-II induces expression of fibronectin and transforming growth factor-β (TGFβ) through downstream signaling of EGFR transactivation and ER stress. This occurs via a signaling mechanism involving ADAM17-mediated shedding which is independent of hypertension (Moriguchi et al., 1999; Takayanagi et al., 2015).

In atherosclerotic plaques, neovascularization occurs as a physiological response to increased oxygen demand which causes adverse effects by facilitating inflammatory influx and favoring the conditions that destabilize the plaque. Thus, inhibition of neovascularization results in limited atherosclerotic lesion (van Hinsbergh et al., 2015). ADAM17 positively regulates angiogenesis by inhibiting the expression of the antiangiogenic factor thrombospondin 1, presenting a positive role in pathological angiogenesis (Caolo et al., 2015). ADAM17 is also critical for EGFR signaling due to the proteolytic release of several ligands of EGFR in the vessels, such as HB-EGF. Weskamp et al. demonstrated that inactivation of ADAM17 in endothelial cells reduced pathological neovascularization whereas its inactivation in smooth muscle cells revealed no evident effect (Weskamp et al., 2010). The underlying mechanism is probably EGFR signaling stimulated by EGFR ligands released by ADAM17 from endothelial cells. Selective inhibition of ADAM17 could be beneficial for the treatment of diseases implied in pathological neovascularization (Weskamp et al., 2010), such as atherosclerosis. Vascular endothelial growth factor (VEGF-A) and its receptor (VEGFR) are also critical for regulating angiogenesis in physiological and pathological conditions. VEGFR type 2 is a tyrosine kinase receptor which is considered the principal coordinator of adult angiogenesis (Swendeman et al., 2008). This receptor can be shed from cells by ADAM17. VEGF-A/VEGFR2 were shown to stimulate ADAM17, resulting in shedding of VEGFR2 and other ADAM17 substrates (Swendeman et al., 2008). This could limit angiogenesis process and reveal a potential novel target for treatment of pathological neovascularization associated to atherosclerosis. It is important to highlight that ADAM17 over-expression or inhibition could either increase or decrease neovascularization as it also presents pro- or anti-inflammatory functions.

Plaque rupture and subsequent thrombotic complications are adverse events associated with atherosclerosis (Freitas Lima et al., 2015). Local over-activity of ADAM17 may weaken atherosclerotic plaques, causing its rupture. In this context, Rizza et al. suggested that measuring ADAM17 activity may predict major cardiovascular events in subjects with established atherosclerosis (Rizza et al., 2015). Although ADAM17 activity was evaluated in an indirect manner through the measurement of its substrates in circulation, authors suggest that it is reasonable to assume that the increase in these molecules is related to increased ADAM17 activity in inflammatory sites. In addition, it was observed that unsaturated fatty acids in LDL particles are also involved in ADAM17 activation in the endothelial layer, due to increase in membrane fluidity, creating a link between endothelial dysfunction/atherosclerosis and increase in ADAM17 substrates in patients at risk (Reiss et al., 2011; Menghini et al., 2013).

## In Brain Injury and Tumor

Over the years, an important role for ADAM17 in neural injury and degeneration has gradually become clear. Via shedding of TGFα, ADAM17 can work as the constitutive sheddase of epiregulin and amphiregulin, both members of the epidermal growth factor (EGF) family playing important roles in the regulation of cell growth, proliferation, and survival (Freimann et al., 2004; West et al., 2008). Recombination of epiregulin and amphiregulin proteins can effectively inhibit endoplasmic reticulum stress and the subsequent induction of neuronal cell death. Therefore, up-regulation of epiregulin and amphiregulin in glial cells may have neuro-protective effects and provide a potential therapy for brain injury (Zhan et al., 2015). However, there is evidence suggesting an association between increased expression of ADAM17 and various types of cancer. EGFR binding with ligands after ADAM17-induced shedding can subsequently activate MEK/ERK and PI3K/Akt pathways, which contribute to the invasiveness and other malignant phenotypes of tumors. Glioma is the most common malignant intrinsic primary brain tumor. By up-regulating the ligands of EGFR, ADAM17 can promote glioma cells malignant phenotype by increased proliferation, invasion, and angiogenesis. In glioma cells and glioma-bearing nude mice, targeting ADAM17 with TAPI-2 (an ADAM17 inhibitor) or siRNA, can significantly attenuate tumor growth and invasiveness, compared to their untreated counterparts (Zheng et al., 2012). Further, it has been shown that beyond the EGFR pathway, TGFβ (transforming growth factor-β) also plays a key role in the regulation of glioma formation and progression (Lu et al., 2011). Interestingly, TGFβ, which can rapidly induce ADAM17 phosphorylation, is actually an upstream signal of ADAM17 (Wang et al., 2008). However, the mechanism of TGFβ-induced tumor formation and progression is not simply mediated by activation of the downstream EGFR pathway. Mu et al. found that TGFβ utilized TRAF6 (TNFR-associated factor 6), PKC ζ and ADAM17 to promote the formation of the TβRI (serine/threonine kinase receptor I) intracellular domain, which could be translocated to the nucleus, where it promotes tumor invasion by induction of Snail and MMP2 (matrix metalloproteinase-2; Mu et al., 2011).

## ADAM17-MEDIATED ACE2 SHEDDING

## In the Cardiac Renin Angiotensin-System (RAS)

Overactivity of the RAS contributes to the development and maintenance of cardiac hypertrophy in experimental models and humans. Whilst the primary physiological role of angiotensinconverting enzyme (ACE) in the RAS is to hydrolyze Ang-I into the potent vasoconstrictor Ang-II, ACE2 is able to cleave Ang-II to produce Ang-(1–7), a peptide which has opposing effects (Xia and Lazartigues, 2008). Therefore, ACE2 is an important regulator within the RAS. Cardiac hypertrophy and impaired contractility are associated with decreased ACE2, whereas ACE2 over-expression protects the heart from Ang II-mediated cardiac hypertrophy and myocardial fibrosis (Bodiga et al., 2011; Patel et al., 2012). Interestingly, neuronal overexpression of ACE2 has also therapeutic effects on Ang-II-induced cardiac hypertrophy (Feng et al., 2012), confirming the pivotal role of ACE2 within the brain RAS and in the central regulation of cardiovascular function. The compensatory role of ACE2 is compromised, as it is a subject of ADAM17-mediated shedding. The ACE2 ectodomain is cleaved from the cell membrane by ADAM17 and released into the extracellular milieu (Lambert et al., 2005). This further fuels the involvement of ADAM17 in cardiac hypertrophy, as has been shown in both animal and human studies. For example, systemic treatment of ADAM17-targeting siRNA for 30 days effectively stopped the progression of agonist-induced cardiac hypertrophy and fibrosis in adult spontaneously hypertensive rats and mice following Ang-II infusion (Wang et al., 2009).

Similarly, Patel et al. (2014) found that in mice infused with Ang-II, there was an angiotensin receptors type I (AT1R) mediated increase in myocardial ADAM17 expression and activity. Membrane translocation of ADAM17 was linked to a substantial reduction in myocardial ACE2 protein and activity with a corresponding increase in plasma ACE2 activity, suggesting that Ang-II-induced ACE2 shedding was mediated by ADAM17. Reactive oxygen species seem to play a key role since p47phox knockout mice were resistant to Ang-IIinduced ADAM17 activation with preservation of myocardial ACE2 and attenuated Ang-II-mediated cardiac dysfunction and hypertrophy (Patel et al., 2014).

## In the Brain RAS

Lautrette et al. (2005) demonstrated that peripheral inhibition of ADAM17 with TAPI-2 does not decrease systolic blood pressure in Ang II-infused mice, suggesting that cardiac ADAM17 participates in the development of cardiac hypertrophy, but does not play a primary role in the regulation of blood pressure in this model (Lautrette et al., 2005). However, our group has recently demonstrated that ADAM17 in the brain actually contributes to the development of neurogenic hypertension (Xia et al., 2013). Similar to the heart, overactivity of the brain RAS is a major contributor to the development and maintenance of neurogenic hypertension (Itaya et al., 1986; Davisson et al., 1998). Deoxycorticosterone acetate (DOCA) salt treatment, a well-characterized model for neurogenic hypertension (Yemane et al., 2010), led to enhanced ADAM17 expression and activity. This ultimately reduced the expression and activity of ACE2 in the mouse hypothalamus (Xia et al., 2013). In addition, the treatment with tempol or α-lipoic acid, blunted ADAM17 activity and preserved ACE2 compensatory effects, suggesting a possible role of oxidative stress in the up-regulation of ADAM17 during DOCA-salt hypertension (de Queiroz et al., 2015).

The activation of ADAM17 by DOCA-salt treatment could be dependent on AT1R. Following AT1R activation, downstream signaling pathways involving ROS, PKC, MAPK, and calcium (Ca2+) might not only up-regulate the activity of ADAM17 but also interfere with the protective effects of ACE2 by interacting with other molecules. In ACE2-expressed HEK cells, calmodulin (CaM) interacts with ACE2. Inhibiting CaM increased the release of the ACE2 ectodomain in a dose- and time-dependent manner (Lambert et al., 2008). This association between CaM and ACE2 has also been reported in the CNS. In mice with DOCA-salt hypertension, interaction between ACE2 and CaM was found to be decreased (Xia et al., 2013), covceivably because of an increase in intracellular [Ca2+]. The structure of CaM can be expanded by its specific Ca2<sup>+</sup> binding sites, which possibly contributes to the dissociation of CaM from the cytoplasmic site of ACE2. This might render ACE2 vulnerable to the shedding activity of ADAM17.

The balance between the classic RAS and the ACE2-related compensatory axis is important to the maintenance of a normal blood pressure as well as central regulation of autonomic function. In mouse basolateral amygdala, overexpression of ACE2 significantly increased the frequency of spontaneous inhibitory postsynaptic currents (indicative of presynaptic release of GABA) from pyramidal neurons. This effect was eliminated by central administration of a Mas receptor (MasR) antagonist, suggesting a possible role of ACE2 in GABA neurotransmission via MasR activation (Wang et al., 2016). Accordingly, ADAM17 can contribute to increased sympathetic outflow by downregulating ACE2 activity, which in turn exacerbates the contribution of classic RAS. In mice with ACE2 deletion, we observed an increase of baseline blood pressure, which could reach hypertensive levels with age (Xia et al., 2011).

Moreover, ADAM17 can control the transcription of matrix metalloproteinase-2 (MMP-2), which in turn mediated angiotensin-II (Ang-II)-induced hypertension in mice independently of cardiac hypertrophy or fibrosis, showing that the effects of ADAM17 in the cardiovascular system may be connected to other metalloproteinases (Odenbach et al., 2011).

The **Figure 2** summarizes the main targets of ADAM17 induced shedding described in the text, the downstream events of their proteolytic cleavage and pathological and physiological processes influenced by ADAM17-dependent shedding.

### ADAM17 AS A THERAPEUTIC TARGET

### A Brief Introduction of Endogenous ADAM17 Inhibitor

ADAM share characteristics with the wider family of matrix metalloproteinases (MMP), which are regulated by a group of endogenous inhibitors. Among them, tissue inhibitors of metalloproteinase 3 (TIMP3) not only blocks the activity of MMP but also inhibits the ectodomain shedding induced by ADAM17 (Amour et al., 1998). Both ADAM17 and TIMP3 have high abundance in the heart, kidney and brain. Mice with TIMP3 deficiency show significant increase in ADAM17 activity and soluble TNF-α abundance (Federici et al., 2005; Guinea-Viniegra et al., 2009). This is believed to be responsible for Ang-II-induced vascular inflammation and remodeling (Basu et al., 2012, 2013). Zheng et al. (2016) demonstrated that enhanced ADAM17 expression with decreased TIMP3 and increased TNFα expression, within 1 week after AMI, are associated with cardiac remodeling (Zheng et al., 2016). These data confirm that ADAM17 is an important regulator of TNF-α maturation and may be a potential target for the inhibition of cellular TNFα production in cardiovascular disorders. In addition, it was demonstrated that an imbalance between ADAM17 and TIMP3 is characteristic of unstable carotid plaques (Menghini et al., 2013; Rizza et al., 2015). On the other hand, treatment with soluble TIMP3 can impart a neuro-protective effect and enhance neurite out-growth both in vitro and in animals with braininjury, in vivo by activating the Akt-mTORC1 signaling pathway (Gibb et al., 2015).

In normal baseline conditions, the cytoplasmic tail of ADAM17 supports cell surface homo-dimerization. The ADAM17 homodimers associate with TIMP3, which silences the shedding activity of ADAM17. Phosphorylation on the cytoplasmic domain of ADAM17 disrupts the homodimers into monomers. This relieves TIMP3 association from ADAM17 enabling it to increase proteolysis of its substrates and hence shifting away from normal baseline condition (Xu et al., 2012).

## Existing Knockout Models

Because of its essential role in normal fetal development, disruption of the ADAM17 gene leads to perinatal lethality with opened eyes, defects in aortic, pulmonic, and tricuspid heart valves (Li et al., 2015). Mice with a targeted deletion of exon11 that encodes the catalytic active site of the metalloprotease domain (adam171Zn/1Zn), resulting in a lack of enzymatic activity, also display substantial perinatal lethality within 2 weeks. Though one study reported that some adam171Zn/1Zn null mice could survive to adulthood, they still have phenotypic abnormalities like dramatic weight loss and significant defects in the immune system (Gelling et al., 2008). Therefore, while the global knockout model may be useful for studies focusing on the physiological role of ADAM17, it might not be suitable for investigating its therapeutic role in pathological processes. Moreover, ADAM17-mediated shedding involves a variety of receptors and substrates, which are distributed in different cell types and various tissues. Accordingly, targeting of a specific cell population or tissue becomes a better solution for studies trying to identify ADAM17 and its potential role in certain disease or pathological processes. Since the therapeutic use of pharmacological inhibitors of ADAM17 also has its own limitation, several types of conditional ADAM17 knockout animals have been generated and already expanded our knowledge of ADAM17 roles (**Table 1**). Using Cre-LoxP technology, these models are amenable for conditional, preor post-developmental, deletion of neuronal ADAM17, making them attractive for in vivo studies in adult animals devoid of gross anatomical defects. These ADAM17 conditional knockout models should be quite useful for studies focusing on ADAM17 related neurological diseases and neurogenic cardiovascular disorders.

### CONCLUSION

ADAM17 has more than 70 different substrates, including cytokines, growth factors, receptors, adhesion molecules, and other proteins. There has been substantial progress in

FIGURE 2 | Summary of targets of the shedding induced by ADAM17, subsequent effects in the cell signaling and events influenced by proteolytical cleavage in the cardiovascular and nervous systems. ACE2, angiotensin-converting enzyme type 2; CX3CL1, fractalkine; EGFR, epithelial growth factor receptor; HB-EGF, EGF-like growth factor; ICAM-1, intercellular adhesion molecule-1; IL-6, interleukin-6; JAM-A, junctional adhesion molecule-A; L1CAM, L1-cell adhesion molecule; RAS, Renin Angiotensin-System; TGFα, transforming growth factor α; TNFα, tumor necrosis factor α; VCAM-1, vascular cell adhesion molecule-1; VEGF, vascular endothelial growth factor; VEGFR, vascular endothelial growth factor receptor.

### TABLE 1 | Summary of ADAM17 conditional knockout mice.


understanding how ADAM17 mediates a range of physiological functions and how ADAM17 mediates signaling pathway contributing to the pathological processes in a variety of neurological and cardiovascular disorders. Despite a plethora of data in various fields, the role of ADAM17 in the central regulation of cardiovascular modulation and related

cardiovascular diseases is far from being elucidated. Although there has been slow progress in translating the knowledge of ADAM17 into possible new treatments, because of the diversity of its substrates, it remains that ADAM17 could be used as potential novel target for the treatment of cardiovascular diseases. However, this will require treatment strategies with improved efficiency and specificity.

### AUTHOR CONTRIBUTIONS

All authors: JX, SM, CS, AC, JC, CB, VB, EL, and MF, drafted the work, contributed to work design, revised it critically, approved

### REFERENCES


the final version to be published and declare accountable for all aspects of the work.

### FUNDING

This study was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Brazil, Grant 88881.068184/2014-01 to MF, research grants from the American Heart Association (15POST25000010 to JX and 12EIA8030004 to EL) and the National Institutes of Health (HL093178 and GM106392) to EL.


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2016 Xu, Mukerjee, Silva-Alves, Carvalho-Galvão, Cruz, Balarini, Braga, Lazartigues and França-Silva. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.