# PATTERN RECOGNITION RECEPTORS AND CANCER

EDITED BY: Anton G. Kutikhin and Arseniy E. Yuzhalin PUBLISHED IN: Frontiers in Immunology

#### *Frontiers Copyright Statement*

*© Copyright 2007-2015 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 978-2-88919-674-6 DOI 10.3389/978-2-88919-674-6

### 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

## **PATTERN RECOGNITION RECEPTORS AND CANCER**

Topic Editors:

**Anton G. Kutikhin,** Research Institute for Complex Issues of Cardiovascular Diseases, Kemerovo, Russian Federation **Arseniy E. Yuzhalin,** University of Oxford, UK

The group of pattern recognition receptors (PRRs) includes families of Toll-like receptors (TLRs), NOD-like receptors (NLRs), C-type lectin receptors (CLRs), RIG-I-like receptors (RLRs), and AIM-2-like receptors (ALRs). Conceptually, receptors constituting these families are united by two general features. Firstly, they directly recognize common antigen determinants of virtually all classes of pathogens (so-called pathogen-associated molecular patterns, or simply PAMPs) and initiate immune response against them via specific intracellular signaling pathways. Secondly, they recognize endogenous ligands (since they are usually released during cell stress, they are called damage-associated molecular patterns, DAMPs), and, hence, PRR-mediated immune response can be activated without an influence of infectious agents. So, pattern recognition receptors play the key role performing the innate and adaptive immune response. In addition, many PRRs have a number of other vital functions apart from participation in immune response realization. The fundamental character and diversity of PRR functions have led to amazingly rapid research in this field. Such investigations are very promising for medicine as immune system plays a key role in vast majority if not all human diseases, and the process of discovering the new aspects of the immune system functioning is rapidly ongoing. The role of Toll-like receptors in cancer was analyzed in certain reviews but the data are still scattered. This collection of reviews systematizes the key information in the field.

**Citation:** Anton G. Kutikhin and Arseniy E. Yuzhalin, eds. (2015). Pattern recognition receptors and cancer. Lausanne: Frontiers Media. doi: 10.3389/978-2-88919-674-6

# Table of Contents



David L. Morse *173 Toll-like receptors and prostate cancer*

Shu Zhao, Yifan Zhang, Qingyuan Zhang, Fen Wang and Dekai Zhang


*197 Toll-like receptor-4 modulation for cancer immunotherapy* Shanjana Awasthi

## **Editorial: Pattern recognition receptors and cancer**

*Anton G. Kutikhin<sup>1</sup> \* and Arseniy E. Yuzhalin<sup>2</sup>*

*1 Laboratory for Genomic Medicine, Division of Experimental and Clinical Cardiology, Research Institute for Complex Issues of Cardiovascular Diseases under the Siberian Branch of the Russian Academy of Medical Sciences, Kemerovo, Russia, <sup>2</sup> Cancer Research UK/MRC Oxford Institute for Radiation Oncology, Department of Oncology, University of Oxford, Oxford, UK*

**Keywords: toll-like receptors, NOD-like receptors, C-type lectin receptors, RIG-I-like receptors, pattern recognition receptors, innate immunity, inflammation, cancer**

The problem of cancer remains one of the most immense challenges to current biomedical research. Affecting populations in all countries and all regions, this disease is responsible for millions of deaths annually (1). Evasion of the immune system is an ominous feature of cancers, which often leads to tumor outgrowth, epithelial–mesenchymal transition (EMT), and consequently, metastatic disease. The need to understand basic mechanisms governing immune response to tumors is increasingly acute, since contemporary cancer research gradually progresses toward highly specialized personalized medicine. In this respect, oncoimmunology of pattern recognition receptors (PRRs) is a promising area of research which requires more attention and broader interpretation.

The group of PRRs includes families of toll-like receptors (TLRs), NOD-like receptors (NLRs), C-type lectin receptors (CLRs), RIG-I-like receptors (RLRs), and AIM-2-like receptors (ALRs). United by two general features, these receptors are the key players in human immunity. First, they directly recognize antigen determinants of nearly all classes of pathogens [pathogen-associated molecular patterns (PAMPs)] and promote their elimination by triggering innate and adaptive immune response. Second, they recognize endogenous ligands released during cell stress [damageassociated molecular patterns (DAMPs)], and therefore can activate immune response in the absence of an infectious agent. In addition, PRRs are known to possess a number of other vital functions, regulating the processes of apoptosis, DNA repair, autophagy, and angiogenesis. Remarkable functional significance and diversity of biological functions are the reasons why PRRs today are an actively growing area of research.

During the last decade, much research has been done to investigate the role of PRRs in tumor immunity. Accumulating evidence demonstrate that anti-tumor immunity can be stimulated through the activation of PRRs (2, 3). It has been repeatedly shown that reinforced PRR activation may protect the host from infectious agents and prevent, inhibit, or block carcinogenesis whereas disrupted or deregulated functioning of PRRs may promote cancer through weakening the immune system (2, 3). At the same time, PRR activation may stimulate cancer by creating a proinflammatory microenvironment which is favorable for tumor progression and chemoresistance development (4). Furthermore, it may also result in immunosuppression caused by chronic inflammation (2), which is known to promote the development of breast carcinoma, colorectal cancer, pancreatic adenocarcinoma, and possibly several other cancer types (5, 6). In this case, on the contrary, lower PRR activity should minimize effects of chronic inflammation such as enhancement of cancer initiation and promotion/progression and, consequently, decrease probability of tumor development (4). Therefore, the situation resembles a double-edged sword, where both sides can cut unless golden mean is maintained. In this respect, it is clear that a subtle balance of low and high PRR activity is required for proper functioning of the immune system. This hypothesis, initially developed for PRRs (3), may also be successfully projected on PRR intracellular signaling pathways – if their elements are overexpressed/constantly activated, it may lead to consequences similar to that

*Edited and reviewed by: Anahid Jewett, UCLA School of Dentistry and Medicine, USA*

*\*Correspondence: Anton G. Kutikhin*

### *antonkutikhin@gmail.com*

*Specialty section: This article was submitted to Tumor*

*Immunity, a section of the journal Frontiers in Immunology Received: 02 August 2015*

*Accepted: 04 September 2015 Published: 16 September 2015*

#### *Citation:*

*Kutikhin AG and Yuzhalin AE (2015) Editorial: Pattern recognition receptors and cancer. Front. Immunol. 6:481. doi: 10.3389/fimmu.2015.00481* of enhanced PRR activation (7, 8). On the other hand, if downstream members of PRR pathways are underexpressed, inactivated, or unable to work properly, it may result in the same effects that of diminished PRR activity, and therefore a balance in functioning of all genes encoding proteins constituting PRR signaling pathways should be preserved for optimal immune system function (7, 8).

Three years ago, four milestone reviews on PRR biology were published in *Immunity* (9–12); we now think that *Frontiers in Immunology* can be an excellent platform for the constellation of review articles systematizing key information in the field with regard to the recent discoveries. With this aim in mind, we invited a number of recognized experts in the field to submit review papers on various aspects of PRR biology and their role in cancer. We sincerely thank all researchers who have agreed to contribute to our Research Topic.

This collection is divided into three sections. The first section describes basic functions of PRRs along with their signaling pathways, and was established with the participation of Taro Kawai and colleagues, Mansi Saxena and Garabet Yeretssian, Huimin Yan and colleagues, together with Stephanie Reikine, Jennifer Nguyen, and Yorgo Modis. We also sought to solicit a number of additional review articles on TLR and NLR biology, since we believe these topics deserve a particular attention. Regarding TLRs, Ajay Jain, Sabina Kaczanowska, and Eduardo Davila depict the newest schemes of the IL-1 receptor-associated kinase signaling, whereas Asif Amin Dar, Rushikesh Sudam Patil, and Shubhada Vivek Chiplunkar provide the insights into the relationship

### **References**


between TLRs and γδ T cell response. Readers interested in NLR structure and functioning will definitely appreciate elegant papers by Irving Coy Allen, Julie Magarian Blander, and Andrew Kent together with Silvia Lucena Lage and colleagues. In addition, a brilliant review by Nelson Di Paolo fills a substantial gap in the understanding of the recognition of human oncogenic viruses by PRRs. Finally, Raunaq Singh Nagi, Ashish Bhat, and Himanshu Kumar close up the first section with the description of the general conception on the role of PRRs in cancer development.

The second section is devoted to the role of PRRs in various vital cellular processes, including apoptosis, DNA repair, autophagy, and angiogenesis. It is contributed by Gustavo Amarante-Mendes and colleagues, Anton Kutikhin and colleagues, Ji Eun Oh, and Heung Kyu Lee along with Sheeba Murad.

Finally, the last piece of the collection consists of reviews that comprehensively analyze the impact of PRRs on the development of malignant tumors (esophageal cancer, gastric cancer, colorectal cancer, lung cancer, prostate cancer, breast cancer, ovarian cancer, and lymphoma). Furthermore, Simon Heidegger and colleagues discuss the role of PRRs in graft-versus-host disease and graftversus-leukemia following allogeneic stem cell transplantation. As a final point, Shanjana Awasthi underlines the importance of TLR agonists in cancer immunotherapy.

We created this Research Topic with the hope that it will be useful for a wide audience, particularly cancer researchers, immunologists, microbiologists, graduate, and undergraduate students of biomedical faculties as well as for their lecturers.


**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 © 2015 Kutikhin and Yuzhalin. 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.*

### Toll-like receptor signaling pathways

#### **Takumi Kawasaki and Taro Kawai \***

Laboratory of Molecular Immunobiology, Graduate School of Biological Sciences, Nara Institute of Science and Technology, Ikoma, Japan

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Uwe-Karsten Hanisch, University of Göttingen, Germany George Trendelenburg, Charité – Universitätsmedizin Berlin, Germany

#### **\*Correspondence:**

Taro Kawai, Laboratory of Molecular Immunobiology, Graduate School of Biological Sciences, Nara Institute of Science and Technology, 8916-5 Takayama, Ikoma, Nara 630-0192, Japan

e-mail: tarokawai@bs.naist.jp

#### **INTRODUCTION**

The innate immune system employs germline-encoded patternrecognition receptors (PRRs) for the initial detection of microbes. PRRs recognize microbe-specific molecular signatures known as pathogen-associated molecular patterns (PAMPs) and selfderived molecules derived from damaged cells, referred as damageassociated molecules patterns (DAMPs). PRRs activate downstream signaling pathways that lead to the induction of innate immune responses by producing inflammatory cytokines, type I interferon (IFN), and other mediators. These processes not only trigger immediate host defensive responses such as inflammation, but also prime and orchestrate antigen-specific adaptive immune responses (1). These responses are essential for the clearance of infecting microbes as well as crucial for the consequent instruction of antigen-specific adaptive immune responses.

Mammals have several distinct classes of PRRs including Toll-like receptors (TLRs), RIG-I-like receptors (RLRs), Nodlike receptors (NLRs), AIM2-like receptors (ALRs), C-type lectin receptors (CLRs), and intracellular DNA sensors such as cGAS (2, 3). Among these, TLRs were the first to be identified, and are the best characterized. The TLR family comprises 10 members (TLR1–TLR10) in human and 12 (TLR1–TLR9, TLR11–TLR13) in mouse. TLRs localize to the cell surface or to intracellular compartments such as the ER, endosome, lysosome, or endolysosome, and they recognize distinct or overlapping PAMPs such as lipid, lipoprotein, protein, and nucleic acid. Each TLR is composed of an ectodomain with leucine-rich repeats (LRRs) that mediate PAMPs recognition, a transmembrane domain, and a cytoplasmic Toll/IL-1 receptor (TIR) domain that initiates downstream signaling. The ectodomain displays a horseshoe-like structure, and TLRs interact with their respective PAMPs or DAMPs as a homo- or heterodimer along with a co-receptor or accessory molecule (4). Upon PAMPs and DAMPs recognition, TLRs

Toll-like receptors (TLRs) play crucial roles in the innate immune system by recognizing pathogen-associated molecular patterns derived from various microbes. TLRs signal through the recruitment of specific adaptor molecules, leading to activation of the transcription factors NF-κB and IRFs, which dictate the outcome of innate immune responses. During the past decade, the precise mechanisms underlying TLR signaling have been clarified by various approaches involving genetic, biochemical, structural, cell biological, and bioinformatics studies. TLR signaling appears to be divergent and to play important roles in many aspects of the innate immune responses to given pathogens. In this review, we describe recent progress in our understanding of TLR signaling regulation and its contributions to host defense.

**Keywords:TLRs, signal transduction, NF-**κ**B, IRFs, adaptors**

recruit TIR domain-containing adaptor proteins such as MyD88 and TRIF, which initiate signal transduction pathways that culminate in the activation of NF-κB, IRFs, or MAP kinases to regulate the expression of cytokines, chemokines, and type I IFNs that ultimately protect the host from microbial infection. Recent studies have revealed that proper cellular localization of TLRs is important in the regulation of the signaling, and that cell type-specific signaling downstream of TLRs determines particular innate immune responses. Here, we summarize recent progress on TLR signaling pathways and their contributions to host defense responses.

#### **PAMP RECOGNITION BY TLRs**

TLRs are expressed in innate immune cells such as dendritic cells (DCs) and macrophages as well as non-immune cells such as fibroblast cells and epithelial cells. TLRs are largely classified into two subfamilies based on their localization, cell surface TLRs and intracellular TLRs. Cell surface TLRs include TLR1, TLR2, TLR4, TLR5, TLR6, and TLR10, whereas intracellular TLRs are localized in the endosome and include TLR3, TLR7, TLR8, TLR9, TLR11, TLR12, and TLR13 (5, 6).

Cell surface TLRs mainly recognize microbial membrane components such as lipids, lipoproteins, and proteins. TLR4 recognizes bacterial lipopolysaccharide (LPS). TLR2 along with TLR1 or TLR6 recognizes a wide variety of PAMPs including lipoproteins, peptidoglycans, lipotechoic acids, zymosan, mannan, and tGPI-mucin (5). TLR5 recognizes bacterial flagellin (2). TLR10 is pseudogene in mouse due to an insertion of a stop codon, but human TLR10 collaborates with TLR2 to recognize ligands from listeria (7). TLR10 can also sense influenza A virus infection (8).

Intracellular TLRs recognize nucleic acids derived from bacteria and viruses, and also recognize self-nucleic acids in disease conditions such as autoimmunity (9). TLR3 recognizes viral double-stranded RNA (dsRNA), small interfering RNAs, and self-RNAs derived from damaged cells (10–12). TLR7 is predominantly expressed in plasmacytoid DCs (pDCs) and recognizes singlestranded (ss)RNA from viruses. It also recognizes RNA from streptococcus B bacteria in conventional DCs (cDCs) (13). Human TLR8 responds to viral and bacterial RNA (14). Structural analysis revealed that unstimulated human TLR8 exists as a preformed dimer, and although the Z-loop between LRR14 and LRR15 is cleaved, the N- and C-terminal halves remain associated with each other and participate in ligand recognition and dimerization. Ligand binding induces reorganization of the dimer to bring the two C termini into close proximity (15). TLR13 recognizes bacterial 23S rRNA (16–18) and unknown components of vesicular stomatitis virus (19). TLR9 recognizes bacterial and viral DNA that is rich in unmethylated CpG-DNA motifs; it also recognizes hemozoin, an insoluble crystalline byproduct generated by *Plasmodium falciparum* during the process of detoxification after host hemoglobin is digested (20). TLR11 is localized in the endolysosome and recognizes flagellin (21) or an unknown proteinaceous component of uropathogenic *Escherichia coli* (UPEC) as well as a profilin-like molecule derived from *Toxoplasma gondii* (22). TLR12 is predominantly expressed in myeloid cells and is highly similar to TLR11 and recognizes profilin from *T. gondii* (23). TLR12 functions either as a homodimer or a heterodimer with TLR11 (24, 25).

#### **TRAFFICKING OF TLRs**

All TLRs are synthesized in the ER, traffic to the Golgi, and are recruited to the cell surface or to intracellular compartments such as endosomes. Intracellular localization of TLRs is thought to be critical for ligand recognition as well as for preventing TLRs from coming into contact with self-nucleic acids, which could cause autoimmunity (26–29). The multi-pass transmembrane protein UNC93B1 controls the trafficking of intracellular TLRs from the ER to endosomes. Interestingly, UNC93B1 regulates excessive TLR7 activation by employing TLR9 to counteract TLR7. This was demonstrated by experiments in mice harboring an amino acid substitution (D34A) in UNC93B1, which exhibit a TLR7 hyperreactive and TLR9-hyporeactive phenotype associated with TLR7-dependent systemic lethal inflammation. Thus, a optimizing the balance between TLR7 and TLR9 is a potential mechanism for regulating autoimmunity (30). TLR trafficking is also controlled by the ER-resident protein PRAT4A, which regulates the exit of TLR1, TLR2, TLR4, TLR7, and TLR9 from the ER and their trafficking to the plasma membrane and endososmes (31). gp96, a member of the ER-resident heat-shock protein 90 family, functions as a general chaperone for most TLRs, including cell surface TLR1, TLR2, TLR4, and TLR5 and intracellular TLR7 and TLR9 (32).

In the endosome, nucleic acid-sensing TLRs undergo proteolytic cleavage by cathepsins B, S, L, H, and K and asparginyl endopeptidase to attain a functional form that mediates ligand recognition and initiates signaling (33–35). However, the Nterminal region of TLR9 is required for CpG-DNA recognition and binding (36). Interestingly, a recent study suggests that the N-terminal cleaved fragment (TLR9N) remains associated with truncated TLR9 (TLR9C) to form a complex, which acts as a functional DNA sensor (37).

#### **CONTRIBUTION OF TIR DOMAIN-CONTAINING ADAPTORS TO TLR SIGNALING**

Individual TLRs differentially recruit members of a set of TIR domain-containing adaptors such as MyD88, TRIF, TIRAP/MAL, or TRAM. MyD88 is utilized by all TLRs and activates NF-κB and MAPKs for the induction of inflammatory cytokine genes. TIRAP is a sorting adaptor that recruits MyD88 to cell surface TLRs such as TLR2 and TLR4 (**Figure 1**). However, a recent study demonstrated that TIRAP also participates in signaling through endosomal TLRs such as TLR9. The lipid-binding domain of TIRAP binds to PI(4,5)P<sup>2</sup> at the plasma membrane and to PI(3)P on endosomes, which mediates the formation of functional TLR4 and TLR9 signaling complexes at their respective sites. Thus, TIRAP associates with both cell surface and endosomal TLRs by binding to different lipids (38). However, a high concentration of TLR9 agonists activates cells in the absence of TIRAP, suggesting that TIRAP is required for TLR9 signaling in natural situations such as HSV-1 infection (39).

TRIF is recruited to TLR3 and TLR4 and promotes an alternative pathway that leads to the activation of IRF3, NF-κB, and MAPKs for induction of type I IFN and inflammatory cytokine genes. TRAM is selectively recruited to TLR4 but not TLR3 to link between TRIF and TLR4. TLR3 directly interacts with TRIF, and this interaction requires phosphorylation of the two tyrosine residues in the cytoplasmic domain of TLR3 by the epidermal growth factor ErbB1 and Btk (40, 41). Collectively, depending on the adaptor usage, TLR signaling is largely divided into two pathways: the MyD88-dependent and TRIF-dependent pathways.

#### **MyD88-DEPENDENT PATHWAY**

After TLR engagement, MyD88 forms a complex with IRAK kinase family members, referred to as the Myddosome (**Figure 1**) (42). During Myddosome formation, IRAK4 activates IRAK1, which is then autophosphorylated at several sites (43) and released from MyD88 (44). IRAK1 associates with the RING-domain E3 ubiquitin ligase TRAF6. TRAF6, along with ubiquitin-conjugating enzyme UBC13 and UEV1A, promotes K63-linked polyubiquitination of both TRAF6 itself and the TAK1 protein kinase complex. TAK1 is a member of the MAPKKK family and forms a complex with the regulatory subunits TAB1, TAB2, and TAB3, which interact with polyubiquitin chains generated by TRAF6 to drive TAK1 activation (45, 46). Although the mechanisms of TAK1 activation within this complex remain unclear, K63-linked ubiquitination or close proximity-dependent transphosphorylation may be responsible for TAK1 activation. TAK1 then activates two different pathways that lead to activation of the IKK complex-NF-κB pathway and -MAPK pathway. The IKK complex is composed of the catalytic subunits IKKα and IKKβ and the regulatory subunit NEMO (also called IKKγ). TAK1 binds to the IKK complex through ubiquitin chains, which allows it to phosphorylate and activate IKKβ. The IKK complex phosphorylates the NF-κB inhibitory protein IκBα, which undergoes proteasome degradation, allowing NF-κB to translocate into the nucleus to induce proinflammatory gene expression. TAK1 activation also results in activation of MAPK family members such as ERK1/2, p38 and JNK, which mediates activation of AP-1

family transcription factors or stabilization of mRNA to regulate inflammatory responses (2, 5).

TAK1 deficiency in mouse embryonic fibroblast cells (MEFs) reduces phosphorylation of IKKs, p38, and JNK after LPS stimulation. However, TLR4-mediated IKK, p38, and JNK activation and cytokine induction are increased in neutrophils derived from TAK1-deficient mice, suggesting a cell type-specific role for TAK1 in TLR signaling (47). Furthermore, the physiological roles of TAB proteins in TLR signaling also remain controversial: TAB1- or TAB2-deficient mice do not show any abnormality in TLR signaling pathways (48), and mice doubly deficient for TAB2 and TAB3 also exhibit normal cytokine production after TLR simulation in MEFs and macrophages (49). TAB family proteins may therefore compensate for each other in TLR signaling.

TLR2 and TLR4 ligations in macrophages increase the production of mitochondrial ROS for bactericidal action and recruit mitochondria to phagosomes (50). TRAF6 is translocated to mitochondria following bacterial infection, where it interacts with ECSIT. TRAF6 promotes ECSIT ubiquitination, resulting in increased mitochondrial and cellular ROS generation.

#### **TRIF-DEPENDENT PATHWAY**

TRIF interacts with TRAF6 and TRAF3. TRAF6 recruits the kinase RIP-1, which in turn interacts with and activates the TAK1 complex, leading to activation of NF-κB and MAPKs and induction of inflammatory cytokines (**Figure 1**). In contrast, TRAF3 recruits the IKK-related kinases TBK1 and IKKi along with NEMO for IRF3 phosphorylation. Subsequently, IRF3 forms a dimer and translocates into the nucleus from

the cytoplasm, where it induces the expression of type I IFN genes (2, 5).

The Pellino family E3 ubiquitin ligases are implicated in TLR signaling (51). Pellino-1-deficient mice display impaired TRIF-dependent NF-κB activation and cytokine production (52). Pellino-1 is phosphorylated by TBK1/IKKi and thereby facilitates ubiquitination of RIP-1, suggesting that Pellino-1 mediates TRIF-dependent NF-κB activation by recruiting RIP-1. Furthermore, Pellino-1 regulates IRF3 activation by binding to DEAF-1, a transcription factor that facilitates binding of IRF3 to the IFNβ promoter (51).

Recently, IRF3 activation was demonstrated to be regulated by an inositol lipid, PtdIns5P. PtdIns5P binds to both IRF3 and TBK1, and thus facilitates complex formation between TBK1 and IRF3. The accessibility of TBK1 to IRF3 mediated by PtdIns5P likely causes IRF3 phosphorylation in a closely proximal manner. Furthermore, PIKfyve was identified as a kinase responsible for production of PtdIns5P during virus infection (53).

#### **BALANCED ACTIVATION BETWEEN MyD88- AND TRIF-DEPENDENT PATHWAYS**

TLR4 activates both the MyD88-dependent and TRIF-dependent pathways. Activation of these pathways is controlled by several molecules to induce appropriate responses. Balanced production of inflammatory cytokines and type I IFN may be important for controlling tumor cell growth and autoimmune diseases.

TRAF3 was shown to be incorporated into the MyD88 complex as well as the TRIF complex in TLR4 signaling. TRAF3 within the MyD88 complex is then degraded, which causes TAK1 activation. Thus, in addition its role in promoting TRIF-dependent pathway activation, TRAF3 has a role in inhibiting the MyD88 dependent pathway. NRDP-1, an E3 ubiquitin ligase, binds and ubiquitinates MyD88 and TBK1, inducing the degradation of MyD88 and augmenting the activation of TBK1, which attenuates inflammatory cytokine production and induces preferential type I IFN production, respectively (54).

MHC class II molecules that are localized in endosomes in antigen-presenting cells interact with the tyrosine kinase Btk via the costimulatory molecule CD40 and maintain Btk activation. Activated Btk interacts with MyD88 and TRIF to promote the activation of the MyD88-dependent and TRIF-dependent pathways and thus to enhance production of inflammatory cytokines and type I IFNs, respectively (55).

#### **TLR7 AND TLR9 SIGNALING IN PLASMACYTOID DCs**

Plasmacytoid DCs are a subset of DCs with the capacity to secrete vast amounts of type I IFN in response to viral infection (**Figure 2**) (2, 5). In pDCs, TLR7 and TLR9 serve as primary sensors for

most TLRs.

IKKα and/or IRAK1. Localization of TLR7 and 9 is controlled by UNC93B1, PRAT4A, and AP3, which traffic TLRs from the ER to the endosome or the ER-resident heat-shock protein 90 family, functions as a general chaperone for

RNA and DNA viruses, respectively. Interestingly, the production of type I IFN by pDCs relies on a complex containing MyD88 and IRF7. This complex also contains TRAF3, TRAF6, IRAK4, IRAK1, IKKα, OPNi, and Dock2 (56, 57). Within this complex, IRF7 is phosphorylated by IRAK1 and/or IKKα and translocates into the nucleus to regulate the expression of type I IFN. Moreover, MyD88-IRAK4-TRAF6 complex drives NF-κB-dependent inflammatory cytokine induction. The signaling complex containing MyD88-IRAK1-TRAF6-IRF7 is formed within lipid bodies by the IFN-inducible Viperin, which activates IRAK1 by lysine 63 linked ubiquitination (58). It is notable that TLR9 signals through different cellular compartments that induce either MyD88-IRF7 dependent type I IFN or MyD88- NF-κB -dependent inflammatory cytokines (59). TLR9 initially traffics to VAMP3-positive early endosomes after CpG-DNA stimulation, where it triggers MyD88-IRAK4-TRAF6-dependent NF-κB activation. TLR9 then traffics to LAMP2-positive lysosome-related organelles (LROs), where it incorporates TRAF3 to activate IRF7 and induce type I IFN (**Figure 2**). AP3 has been shown to bind to TLR9 and control the trafficking of TLR9 to LROs, and is required for type I IFN induction (28). However, AP3 is not required for TLR9-dependent type I IFN induction triggered by DNA-antibody immune complexes (ICs) in pDCs. The intracellular compartment initiating type I IFN induction by DNA-antibody ICs is regulated by the autophagy pathway (60). Thus, pDCs have diverse cargoes for ligand recognition and triggering downstream signaling pathways.

#### **OTHER IRFs IN TLR SIGNALING**

In addition to IRF3 and IRF7, several other IRFs participate in TLR signaling. IRF1 interacts with MyD88 and contributes to TLR9 mediated cytokine production in the presence of IFNγ (61), while IRF5 is involved in the MyD88-dependent signaling pathway for inducing inflammatory cytokine production (62). IRF8 was proposed to be essential for TLR9-MyD88-dependent activation of NF-κB in pDCs (63). However, a subsequent analysis of IRF8 deficient mice demonstrated that IRF8 is involved in the second phase of feedback type I IFN production after treatment of DCs with TLR agonists (64).

#### **ACTIVATION OF TLR SIGNALING BY CO-RECEPTORS**

Recent studies have identified several transmembrane molecules that modulate TLR signaling pathways. CD14, a glycophosphatidylinositol-anchored protein, is a co-receptor with TLR4 and MD-2 for LPS recognition. It induces ITAM-mediated Syk- and PLCγ2-dependent endocytosis to promote TLR4 internalization into endosomes for activation of TRIF-dependent signaling (65). CD14 is also required for TLR7- and TLR9-dependent induction of proinflammatory cytokines (66).

CD36, a protein in the class B scavenger receptor family, acts as a co-receptor for oxidized low-density lipoprotein (LDL) and amyloid-β peptide. Ligand recognition induces the assembly of TLR4/TLR6 heterodimers through Src kinases and consequent sterile inflammation, by inducing inflammatory cytokines and ROS and priming NLRP3 inflammsome activation (67, 68).

#### **NEGATIVE REGULATORS**

TLR signaling is negatively regulated by a number of molecules through various mechanisms to prevent or terminate the excessive immune responses that lead to detrimental consequences associated with autoimmunity and inflammatory diseases. Negative regulators target each of the key molecules in TLR signaling (**Figure 1**). Activation of the MyD88-dependent pathway is suppressed by ST2825, SOCS1, and Cbl-b, and activation of the TRIF-dependent pathway is suppressed by SARM and TAG (69, 70). These molecules associate with MyD88 or TRIF to prevent them from binding to TLRs or downstream molecules. TRAF3 activation is negatively regulated by SOCS3 and DUBA (71). TRAF6 is targeted by a number of inhibitory molecules such as A20, USP4, CYLD, TANK, TRIM38, and SHP (72–74). TAK1 activation is inhibited by TRIM30α and A20 (75). In addition to these signaling molecules, the transcription factor NFκB is suppressed by Bcl-3, IκBNS, Nurr1, ATF3, and PDLIM2, while IRF3 activation is negatively regulated by Pin1 and RAUL (76). The stability of mRNAs encoding signaling molecules is regulated by miRNAs such as miR-146a, miR-199a, miR-155, miR-126, miR-21, miR-29, miR-148/152, and miR-466l (74). In addition to the stability of mRNAs for signaling molecules, stability of mRNA for cytokines is regulated by Regnase-1 and TTP (5, 74).

#### **CONCLUDING REMARKS**

During the past decade, tremendous progress has been made in our understanding of TLR signaling pathways. After genetic studies revealed the contribution of TIR domain-containing adaptor usage, cell biological and biochemical approaches have highlighted the importance of cellular localization of these adaptors in the regulation of downstream signaling. Moreover, numerous reports have demonstrated that TLR trafficking, TLR cleavage, and protein modification of signaling molecules such as ubiquitination and phosphorylation play important roles in the activation of TLR signaling. On the other hand, negative regulators of TLR signaling have been discovered, and their importance in preventing autoimmune and inflammatory diseases is recognized. More recently, much effort has been focused on identifying molecules that are involved in innate immunity through an integrated approach. Indeed, by combining transcriptomics, genetic/chemical perturbations and phosphoproteomics, Polo-like kinases (Plks) 2 and 4 have been found to regulate antiviral responses downstream of TRIF and MyD88 signaling (77). mRNA stability has also attracted attention because it is an important mechanism to regulate TLR-dependent inflammation. For example, the RNase Regnase-1 interacts with IL-6 and IL-12p40 mRNA and degrades them. Regnase-1-deficient macrophages produce large amounts of cytokines after treatment with various TLR ligands, and Regnase-1-deficient mice show elevated autoantibody production (78). Furthermore, it is notable that PAMP variants may activate distinct signaling pathways although they are recognized by the same PRRs. For example, LPS variant such as smooth or rough type activates either MyD88-dependent or TRIF-dependent pathway. These findings suggest that host makes a distinction between different types of LPS-containing bacteria by activating distinct signaling pathways (79).

Although PAMP recognition by TLRs is crucial for host defense responses to pathogen infection, aberrant activation of TLR signaling by PAMPs, mutations of TLR signaling molecules, and DAMPs-mediated TLRs signaling activation are responsible for the development of several diseases such as autoimmune, chronic inflammatory, and allergic diseases. Moreover, a link between cancer and TLRs has been proposed. The innate immune activation that caused after anti-cancer drug treatment is reportedly critical for cancer elimination through TLR-mediated recognition of endogenous molecules released from dying cancer cells (80). On the contrary, mutations in molecules involved in TLR signaling are associated with cancer development. Certain types of diffuse large B-cell lymphoma acquire oncogenic ability through MyD88 mutation and show aberrant activation of NF-κB, JAK and STAT3 (81). A mutation in A20, which is a negative regulator of TLR signaling, is also associated with B-cell lymphoma development (82, 83). Furthermore, it has been suggested that TBK1 functions as a negative regulator of cell growth in lung cancer (84). In summary, further elucidation of TLR signaling pathways should eventually allow us to manipulate them in strategies to treat various infectious and autoimmune diseases that are intimately associated with innate immune signaling, as well as cancer.

#### **REFERENCES**


type I interferon production in plasmacytoid dendritic cells. *Immunity* (2011) **34**:352–63. doi:10.1016/j.immuni.2011.03.010


**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.

*Received: 06 August 2014; paper pending published: 14 August 2014; accepted: 09 September 2014; published online: 25 September 2014.*

*Citation: Kawasaki T and Kawai T (2014) Toll-like receptor signaling pathways. Front. Immunol. 5:461. doi: 10.3389/fimmu.2014.00461*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Kawasaki and Kawai. 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.*

## Insights into the relationship between toll like receptors and gamma deltaT cell responses

#### **Asif Amin Dar, Rushikesh Sudam Patil and Shubhada Vivek Chiplunkar \***

Chiplunkar Laboratory, Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Tata Memorial Centre, Navi Mumbai, India

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Isaias Glezer, Universidade Federal de São Paulo, Brazil Rajesh Kumar Sharma, University of Louisville, USA

#### **\*Correspondence:**

Shubhada Vivek Chiplunkar, Chiplunkar Laboratory, Advanced Centre for Treatment, Research and Education in Cancer (ACTREC), Tata Memorial Centre, Sector 22 Kharghar, Navi Mumbai, Maharashtra 410210, India

e-mail: schiplunkar@actrec.gov.in

The tumor microenvironment is an important aspect of cancer biology that contributes to tumor initiation, tumor progression and responses to therapy.The composition and characteristics of the tumor microenvironment vary widely and are important in determining the anti-tumor immune response. Successful immunization requires activation of both innate and adaptive immunity. Generally, immune system is compromised in patients with cancer due to immune suppression, loss of tumor antigen expression and dysfunction of antigen presenting cells (APC). Thus, therapeutic immunization leading to cancer regression remains a significant challenge. Certain cells of the immune system, including dendritic cells (DCs) and gamma delta (γδ)T cells are capable of driving potent anti-tumor responses. The property of MHC-unrestricted cytotoxicity, high potential of cytokine release, tissue tropism and early activation in infections and malignant disease makes γδ T cells as an emerging candidate for immunotherapy.Various strategies are being developed to enhance anti-tumor immune responses of γδ T cells and DCs one of them is the use of novel adjuvants like toll like receptors (TLR) agonists, which enhance γδ T cell function directly or through DC activation, which has ability to prime γδ T cells. TLR agonists are being used clinically either alone or in combination with tumor antigens and has shown initial success in both enhancing immune responses and eliciting anti-tumor activity. TLR activated γδ T cells and DCs nurture each other's activation. This provides a potent base for first line of defense and manipulation of the adaptive response against pathogens and cancer. The available data provides a strong rationale for initiating combinatorial therapy for the treatment of diseases and this review will summarize the application of adjuvants (TLRs) for boosting immune response of γδ T cells to treat cancer and infectious diseases and their use in combinatorial therapy.

**Keywords: immunotherapy,** γδ**T cells, toll like receptors, tumors, dendritic cells**

#### **INTRODUCTION**

Innate and adaptive immune responses are sentinels of host against the diverse repertoire of infectious agents (viruses and bacteria) and cancer. Both components of immune system identify invading microorganisms or damaged tissues as non-self and activate immune responses to eliminate them. Efficient immune responses depend upon how close an interaction is between the innate and adaptive immune system. γδ T cells and toll like receptors (TLR) serve as an important link between the innate and adaptive immune responses (1–3). Extensive studies have suggested that γδ T cells play important roles in host defense against microbial infections, tumorigenesis, immunoregulation and development of autoimmunity. γδ T cells also have several innate cell-like characters that allow their early and rapid activation following recognition of cellular stress and infection (4, 5). However to accomplish these functions, γδ T cells use both the T cell receptor (TCR) and additional activating receptors (notably NKG2D, NOTCH, and TLR) to respond to stress-induced ligands and infection. γδ T cells express TLRs and modulate early immune responses against different pathogens (6). In this review, we summarize and discuss some of the recent advances of the γδ T cell biology and how direct control of γδ T lymphocyte function

and activation is monitored by TLR receptors and ligands. The review highlights involvement of TLR signaling in γδ T cell functions and their implications in harnessing γδ T cells for cancer immunotherapy.

#### γδ **T CELLS, ANATOMICAL DISTRIBUTION AND ANTIGENIC DIVERSITY**

Based on the type of TCR they express, T lymphocytes can be divided into two major subsets, αβ and γδ T cells. γδ T cell represents a small subset of T lymphocytes (1–10%) in peripheral blood. While in anatomical locations like small intestine, γδ T cells comprise a major bulk of T cells (25–60% in human gut) (7). γδ T cells are the first T cells to appear in thymus during T cell ontogeny in every vertebrate (8), which suggests that their primary contribution could be neonatal protection because at this point conventional αβ T cell responses are severely functionally impaired and DCs are immature (9). In neonates, the Vδ2 <sup>+</sup> cells derived from human cord blood showed early signs of activation. These cells secrete IFN-γ and express perforin after short-term *in vitro* stimulation (10). In comparison to the neonate derived αβ T cells of peripheral blood, γδ T cell subset produces copious amount of IFN-γ and are precociously active (11). Hence, γδ T cells are well engaged in newborns to contribute to immuneprotection, immune-regulation and compensate for impaired αβ T cell compartment.

γδ T cells are unconventional CD3<sup>+</sup> T cells and differ from the conventional αβ T cells in their biology and function (**Table 1**). Although a sizeable fraction of γδ T cells in the intraepithelial lymphocyte compartments of human and mice are CD8αα<sup>+</sup> but the peripheral blood γδ T cells are predominantly double negative (CD4−CD8−) T cells. The absence of CD4 or CD8 expression on majority of the circulating γδ T cells is well in line with the fact that antigen recognition is not MHC restricted (12, 13). Crystal structure analysis of the γδ TCR revealed that γδ TCR is highly variable in length resembling immuno-globulins (Ig) more than the αβ TCR. The antigen recognition property of γδ T cells is fundamentally different from αβ T cells but similar to antigen– antibody binding, which is more likely to occur independent of MHC cross presentation (14). However, recently butyrophilin BTN3A1, a non-polymorphic ubiquitously expressed molecule was identified as an antigen presenting molecule of Vγ9Vδ2 T cells. Soluble BTN3A1 binds (Isopentenyl diphosphate) IPP and (E)-4-hydroxy-3-methyl-but-2-enyl diphosphate (HMBPP) with different affinities in 1:1 ratio to stimulate γδ T cells (15).

The important feature of γδ T cells is their tropism to epithelial tissues. With respect to anatomical localization, γδ T cell population can be divided into two groups: lymphoid-homing γδ T cells that can be primed in the circulation and clonally expand in a conventional "adaptive" manner; and innate-like cells that respond rapidly and at a relatively high frequency in many tissue sites. Migration and anatomical localization of T lymphocytes is crucial for their antigen specificity and maintaining homeostasis in the mammalian immune system. Although γδ T cells are well represented among peripheral blood mononuclear cells (PBMC) and in afferent and efferent lymph, they are rarely found in lymph node parenchyma, spleen, Peyer's patches and thymus. Moreover, unlike αβ T cells, splenic γδ T cells, if present, are not confined to the lymphoid areas (the white pulp) but are also found throughout the red pulp of spleen and marginal zones of cell trafficking (16). γδ T cells are abundantly present in the epithelia of skin, genital and intestinal tract (17). In the small intestines of humans, mice, chickens and cattle, γδ T cells comprise a substantial fraction of intestinal intraepithelial lymphocytes (IELs); in mice γδ<sup>+</sup>

IELs constitute 50–60% of the IEL pool (18–20). The epidermal γδ<sup>+</sup> IELs of mice and cattle (but not humans) have a marked dendritic morphology and are hence known as dendritic epidermal T cells (DETCs) (21). DETCs are maintained at steady state in normal adult murine skin but on activation execute specialized functions like tissue repair (22). DETCs also maintain keratinocyte homeostasis, which along with Langerhan cells forms its neighborhood (23). Under pathological conditions, γδ T cells quickly expand and infiltrate into lymphoid compartments and other tissues.

Another striking difference between αβ and γδ T cells is the range of antigens or ligands that are recognized by the respective TCRs. Unlike αβ T cells, which recognize protein antigen processed inside the cell and presented by MHC molecules, γδ T cells recognize antigens like B cells as revealed by structural and functional studies (24).γδ T cells can respond to a variety of stimuli irrespective of their molecular or genetic nature. In mice, the non-classical MHC class I molecules T10 and T22 are recognized by γδ T cells (25–28). Similar to T10 and T20, murine class II MHC (IA) antigens IE and IA are identified to act as ligands for γδ T cell clones (29, 30). In addition, herpes glycoprotein GI-reactive γδ T cell clones protect mice from herpes simplex virus (HSV) induced lethal encephalitis (31, 32). γδ TCRs can also bind to an algal molecule, phycoerythrin inducing upregulation of CD44 and downregulation of CD62L in γδ T cells (33). B6 murine splenic and hepatic γδ T cells respond to cardiolipin (bacterial cell-wall phospholipid and endogenous component of mitochondria) presented by CD1d molecules (34). Insulin derived peptide B:9–23 is also recognized by the γδ T cell clones derived from non-obese diabetic mice (NOD mice) (35). SKINT1, a mouse immunoglobulin superfamily member, bears structural similarity to human CD277 (butyrophilin 3A1) and is expressed by medullary thymic epithelial cells (mTECs) and keratinocytes that is crucial for the development of Vγ5Vδ1 <sup>+</sup> DETCs (36).

In humans, majority of γδ T cells express a rearranged T cell receptor (TCR) composed of Vγ9 andVδ2 domains; thus, this population is referred to as Vγ9Vδ2. The Vγ9Vδ2 T cells recognize self and microbial phosphorylated metabolites generated in eukaryotic mevalonate pathway and in the microbial 2-C-methyl-derythritol 4-phosphate (MEP) pathway (37). Initially, it was reported that the non-peptidic ligands isolated from mycobacterial cell lysates were


#### **Table 1 | Comparison between** αβ **and** γδ**T cells**.

stimulatory for Vγ9Vδ2 T cell clones. Later, IPP, an intermediate metabolite of the mevalonate pathway, was isolated and identified as a stimulatory molecule. Characterization of the microbial antigens recognized by human γδ T cells predicted that these are non-proteinaceous in nature and have critical phosphate residues (37, 38). Subsequent studies, conducted with *M. tuberculosis*, identified HMBPP, an intermediate metabolite of the MEP pathway, as a strong agonist of γδ TCR. The measured potencies of IPP and HMBPP show an enormous difference. The ED50 of IPP is ~20µM, whereas that of HMBPP is ~70 pM, i.e., more than 105 times lower (38).

Another stimulatory molecule is *Staphylococcus aureus* enterotoxin A (SEA) that directly interacts with the TCR Vγ9 chain independently of the pairedVδ chain. The mechanism of recognition of this superantigen is different from that of phosphorylated metabolites and requires the interaction with MHC class II molecules. γδ T cells kill target cells and release cytokines upon interaction with SEA but do not proliferate (39).

Recently, the TCR from a γδ T cell clone derived from a cytomegalovirus (CMV)-infected transplant patient was shown to directly bind to endothelial protein C receptor (EPCR), which is a lipid carrier with a similar structure to CD1, showing again that γδ TCR engagement is cargo independent (40). ATP F1 synthase has been identified as stimulatory ligand of the TCR Vγ9Vδ2. ATP F1 synthase is an intracellular protein complex involved in ATP generation. However, optimal responses of Vγ9Vδ2 T cells by tumor target cell lines expressing F1-ATPase requires apolipoprotein A1. A monoclonal antibody interacting with apolipoprotein A1 was shown to inhibit TCR γδ activation as it disrupted the trimolecular complex of ApoA1, ATP F1 synthase, and γδ TCR required for optimal response (41).

The second major population of human γδ T cells utilizes the Vδ1 chain, which pairs with a variety of Vγ chains. This subset of Vδ1 <sup>+</sup> T cells is mainly found in tissues and is activated by CD1c and CD1d-expressing cells. The group 1 CD1 molecules have ability to present lipid A to human γδ T cells. The human γδ T cells also recognize the related group 2 CD1 molecule as CD1d/lipid complex. Phosphatidyl ethanol amine (PE), a phospholipid, activates γδ T cells in a CD1d manner dependent suggesting its CD1d restricted recognition (42). In addition, some populations of γδ T cells in normal human PBMCs also recognize lipid molecules such as cardiolipin (a marker of damaged mitochondria), sulfatide (a myelin glycosphingolipid), or α-galactosylceramide (α-GalCer) in association with CD1d, which are noted ligands of natural killer T (NKT) cells (34, 43–45). Human γδ T cells also recognize the stress-induced MHC class I-related MICA/MICB molecules and the UL16-binding proteins that are upregulated on malignant or stressed cells (46–48). Heat shock proteins (HSPs) expressed on the cell membrane play an important role in cancer immunity. Hsp60 expressed on oral tumors act as ligand for Vγ9Vδ2 T cells (49, 50). Hsp60 and Hsp70 expressing human oral and esophageal tumors are lysed by Vγ9Vδ2 T cells (49–51). Hsp72 expressing neutrophils were rapidly killed by γδ T cells through direct cell to cell contact, indicating that hsp72 expression on cell surface pre-disposes inflamed neutrophils to killing by γδ T cells (52). In Another study, hsp90 expression on EBV infected B cells rapidly promoted γδ T cell proliferation (53). This confirms that γδ T cells recognize

qualitatively distinct antigens, which are profoundly regulated by their anatomical localization.

#### **CO-RECEPTORS AND** γδ **T CELL ACTIVATION**

Most γδ T cells respond to non-peptidic antigens even in the absence of antigen presenting cells (APCs). However, the presence of APCs can greatly enhance the γδ T cell response (54). This suggests that accessory molecules/receptors may be involved in effector functions of these cells. Some of important co-receptors used by γδ T cells include NOTCH, NKG2D, and TLR (55).

Our study has identified Notch as an additional signal contributing to antigen specific effector functions of γδ T cells. We have shown that γδ T cells express Notch1 and Notch2 at both mRNA and protein level. Inhibition of Notch signaling in anti-CD3 MAb stimulated γδ T cells resulted in marked decrease in proliferation, cytotoxic potential, and cytokine production by γδ T cells confirming the involvement of Notch signaling in regulating antigen specific responses of γδ T cells (55).

γδ T cells express NKG2D on their cell surface resulting in their activation. Treatment of PBMC with immobilized NKG2Dspecific mAb or NKG2D ligand MHC class I related protein A (MICA) resulted in the up-regulation of CD69 and CD25 on Vγ9Vδ2. Furthermore, NKG2D increased the production of TNFalpha and release of cytolytic granules by Vγ9Vδ2 T cells (56). Later, it was shown that the protein kinase C transduction pathway as a main regulator of the NKG2D-mediated costimulation of anti-tumor Vγ9Vδ2 T cell cytolytic response (57).

TLR agonists are also known to trigger the early activation and the IFN-γ secretion by Vγ9Vδ2T cells (58). TLR ligands indirectly increase the anti-tumoricidal activity of Vγ9Vδ2T cells (59). In this review, we will focus on TLR as an additional co-receptor modulating the function of immune cells with special focus on γδ T cells.

#### **TOLL LIKE RECEPTOR AND IMMUNE CELLS**

The immune system functions in anti-microbial defense by recognizing groups of molecules unique to microorganisms (60). These unique microbial molecules are called pathogen-associated molecular patterns (PAMPs) and are recognized by a family of cellular receptors called pattern recognition receptors (PRRs) (61). TLRs along with retinoic acid-inducible gene (RIG)-I-like receptors (RLRs) and nucleotide-binding oligomerization domain (NOD)-like receptor (NLRs) are prototype PPRs, which recognize pathogen-associated molecular patterns (PAMPs) from microorganisms or danger-associated molecular patterns (DAMPs) from damaged tissues (62). Recognition of PAMPs by TLRs trigger release of inflammatory cytokines and type 1 interferon's (IFN) for host defense (60, 63–65). The adaptive immune system, on the other hand, is responsible for elimination of pathogens in the late phase of infection and in the generation of immunological memory mediated by B and T cells (66).

TLRs derived their name from *Drosophila melanogaster* Toll protein based on their homology (67). In mammals, till date 13 members of TLR family has been identified (63, 68–71). TLR1-9 is conserved in humans and mice while TLR10 is non-functional in mice because of a retroviral insertion while TLR11-13 is lost from the human genome. The first TLR identified was TLR4 and recognizes bacterial lipopolysaccharide (LPS) from Gramnegative bacteria (67, 72, 73). TLRs are classified into several groups based on the types of PAMPs they recognize. TLR1, 2, 4 and 6 recognize lipids whereas the highly related TLR7, TLR8 and TLR9 recognize nucleic acids. Murine TLR11 recognizes a protozoan derived profilin-like protein while TLR13 recognizes *Vesicular stomatitis virus* (63). TLRs are localized in the distinct cellular compartments, for example; TLR1, TLR2, TLR4, TLR5, TLR6, and TLR11 are expressed on the cell surface whereas TLR3, TLR7, TLR8 TLR9, TLR11, TLR12 and TLR13 are expressed in intracellular vesicles such as the endosome and ER. The intracellular TLRs are transported to the intracellular vesicles via UNC93B1, a trans-membrane protein, which is localized in the ER of the cell (70, 71, 74–77). TLR family receptors have a common structural architecture. TLRs are type I integral membrane glycoproteins characterized by multiple extracellular leucine-rich repeats (LRRs) and a single intracellular Toll/interleukin-1 (IL-1) receptor (TIR). TLRs mostly form homo-dimers with a few exceptions, which form heterodimers to trigger a signal. For example, TLR2 forms heterodimers with TLR1 or TLR6 enabling differential recognition of lipopeptides. The TIR domain of TLRs is required for the interaction and recruitment of various adaptor molecules to activate downstream signaling pathway. After recognizing PAMPs, TLRs activate intracellular signaling pathways that lead to the induction of inflammatory cytokine genes such as TNF-α, IL-6, IL-1β and IL-12 through the recruitment of adaptors such as MyD88, TRIF, TRAM, TIRAP and SARM1 (78). MyD88 is a universal adaptor used by all TLRs, except TLR3, to induce inflammatory pathways through activation of MAP Kinases (ERK, JNK, p38) and transcriptional factor NF-κB (63, 79). TLR3 and TLR4 use TRIF to bring activation of alternative pathway (TRIF-dependent pathway) through transcription factors IRF3 and NF-κB to induce type 1 IFN and inflammatory cytokines (80–82). TRAM selectively participates in the activation of the TRIF-dependent pathway downstream of TLR4, but not TLR3 (83, 84). TIRAP functions to recruit MyD88 leading to activation of MyD88-dependent pathway downstream of TLR2 and TLR4 (85, 86).Sterile-α- and armadillo-motif-containing protein 1 (SARM1), was shown to inhibit TRIF and is also critical for TLR-independent innate immunity (87). Thus, signaling pathways can be broadly classified as either MyD88-dependent pathway or TRIF-dependent pathway.

Hornung et al. have showed differential expression of TLR1- 10 on human APCs and lymphocytes including T cells and their functional discrepancy in recognition of specific TLR ligands (88). CD4<sup>+</sup> T cells express almost all TLRs at mRNA levels but may not express all as functional protein (89, 90). Moreover, they do not respond to all TLR ligands. Stimulation with TLR5, 7, or 8 agonists combined with TCR activation of CD4+T cells resulted in increased proliferation and production of IL-2, IL-8, IL-10, IFN-γ and TNFα (91). There are other reports as well suggesting the functional modulation of subtypes of CD4<sup>+</sup> T cells by TLR ligands. The mouse Th1 but not Th2 cells responded to TLR2 agonist and resulted in enhanced proliferation and IFN-γ production independent of TCR stimulation (92). This work validated that the TLR can regulate function of CD4<sup>+</sup> T cells even in absence of TCR engagement. CD4+CD25<sup>+</sup> regulatory T cells (Tregs) express majority of TLRs with selectively higher expression of TLR2, 4, 5, 7/8, and 10 compared to CD4+CD25<sup>−</sup> conventional T cells (93). Liu et al. showed that CD4+CD25<sup>+</sup> regulatory T cells and CD4+CD25<sup>−</sup> conventional T cells express TLR2 and proliferated upon stimulation with its agonist. TLR2 stimulation also led to transient loss of Treg suppressive potential through suppression of FOXP3 (94, 95). However, Tregs also express TLR5 but upon stimulation with flagellin (ligand of TLR5), do not proliferate rather showed increased suppressive capacity and enhanced expression of FOXP3 (96). These reports suggest that the suppressive function of Treg can be either enhanced or dampened by the type of TLR ligand engaged. TLR2 stimulation not only abrogates suppressive functions of CD4<sup>+</sup> Tregs but also drives naïve as well as effector Treg population toward IL17 producing Th17 phenotype (97). Th17 cells express TLR2 along with TLR6 compared to Th1 and Th2 subsets and promote Th17 differentiation upon Pam3Cys stimulation and accelerates experimental autoimmune encephalomyelitis (98). Like TLR2, TLR4 also regulate the functions of CD4<sup>+</sup> T cells. In a mouse model of arthritis, mice lacking TLR2 showed enhanced histopathological scores of arthritis by a shift in T cell balance from Th2 and T regulatory cells toward pathogenic Th1 cells. TLR4, in contrast, contributes to more severe disease by modulating the Th17 cell population and IL-17 production (99, 100). Recently, Li et al. showed that high-mobility group box 1 (HMGB1) proteins decrease Treg/Th17 ratio by inhibiting FOXP3 and enhancing RORγt in CD4<sup>+</sup> T cells via TLR4–IL6 axis in patients with chronic hepatitis B infections (101). This shows that HMGB1 (TLR4 ligand) act as a modulator of CD4<sup>+</sup> T cells responses in chronic viral inflammation. CD4<sup>+</sup> T cells also express intracellular TLRs such as TLR9 and TLR3. Both these TLRs promote T cell survival via activation of NF-κB and MAPK signaling (102). Although the effector functions of CD4<sup>+</sup> T cells are regulated by TLRs but the molecular pathway involved in skewing of CD4<sup>+</sup> T cell function is poorly understood.

Like CD4<sup>+</sup> T cells, CD8<sup>+</sup> T cells also show differential expression of TLRs with high expression of TLR3 but lower expression of TRL1,2,5,9,10 compared to CD4<sup>+</sup> T cells at mRNA level. It is important to note that the expression of TLR2, TLR3 and TLR5 increases on CD8 T cells in infected tonsils compared to controls (89) indicating immune activating role of TLRs in infections. Stimulation of CD8<sup>+</sup> T cells through TLR2 agonists enhances their proliferation and IFN-γ production (103, 104). It also promotes cytolytic activity of CD8<sup>+</sup> T cells and enhances anti-tumor response mediated through MyD88-dependent TLR1/2 pathway (105). Recently, Mercier et al. showed that TLR2 cooperate with NOD-containing protein 1 (NOD1) to enhance TCR mediated activation and can serve as alternative co-stimulatory receptor in CD8<sup>+</sup> T cells (106). CD8<sup>+</sup> T cells also express intracellular TLRs such as TLR3, TLR9 which are more potent in inducing CD8<sup>+</sup> T cell activation *in vivo* (107).

Natural killer (NK) cell is a vital player in innate immune system. They recognize infected and transformed cells with downregulated major histocompatibility complex (MHC) class 1 molecules. They are the primary producers of IFN-γ and are protective against infections. Unlike CD4 and CD8 T cells NK cells as well as CD56+CD3<sup>+</sup> NKT cells constitutively express TLR 1–8 with high expression of TLR2 and 3 at mRNA level. They recognize bacterial PAMPs and respond by producing α-defensins (108– 111). Human NK cells can also directly recognize *Mycobacterium bovis* via TLR2 and enhance their cytolytic activity against tumor cells (112). Tumor-associated macrophages induce NK cell IFN-γ production and cytolytic activity upon TLR engagement (113). TLRs modulate NK cell function directly or indirectly to promote antibody dependent cell mediated cytotoxicity and cross presentation of viral antigens to T lymphocytes (114, 115). This highlights that the cells of adaptive immune system do express TLRs and their function can be directly or indirectly modulated by TLR ligands.

#### **ACTIVATION OF** γδ **T CELLS BY TLR LIGANDS**

In 1997, the first human homolog of *Drosophila* Toll protein was cloned and characterized. It was also established that γδ T cells also express hToll mRNA (67). Purified γδ T cells were found to respond to the *E. coli* native lipid A in a TCR-independent fashion and the LPS/lipid A-reactive γδ T cells strongly expressed TLR2 mRNA. TLR2 antisense oligonucleotide inhibited the proliferation of γδ T cells in response to the native lipid A as well as the TLR2-deficient mice showed an impaired response of the γδ T cells following injection of native lipid A. These results suggest that TLR2 is involved in the activation of canonical Vγ6/Vδ1 T cells by native lipid A (116). Again, functional presence of TLR2 on Vγ2Vδ2 T cells (also known as Vγ9Vδ2 T cells) was reported when the dual stimulation of Vγ2Vδ2 T cells with anti-TCR antibody and Pam3Cys increased synthesis and secretion of IFN-γ and elevated the levels of CD107a expression. IFN-γ secretion and cell surface CD107a levels are markers of increased effector function in Vγ2Vδ2 T cells (117). Similarly, Bruno et al. reported that IL-23 and TLR2 co-stimulation induces IL17 expression in γδ T cells. However, TLR1 and TLR2 expression was found only on CCR6<sup>+</sup> IL-17 producing murine peritoneal γδ T cells but not others. Thus, γδ T cells with innate receptor expression coupled with IL-17 production establishes them as first line of defense that can orchestrate an inflammatory response to pathogen-derived and environmental signals long before Th17 can sense the bacterial invasion (118). Pam3CSK4, TLR2 agonist was able to stimulate only splenic γδ T cell proliferation but not the dermal γδ T cells demonstrating that TLR2 signaling shows tissue tropism. (19). Furthermore, a profound change in the circulating γδ T-cell population was observed in early burn injury (24 h). These γδ T-cells showed TLR2 and TLR4 expression, priming them for TLR reactivity, However TLR expression was specific to circulatory γδ T cell subset and was transient, since it was not observed after postinjury (7 days). Transient nature of the post-burn increase in γδ T-cell TLR expression is likely to be protective to the host, most likely via regulation of inflammation and initiation of healing processes (119).Mitochondrial danger-associated molecular patterns (MTDs) induce TLR2 and TLR4 expression on γδ T cells in dose dependent manner. MTDs also induced the production of IL-1β, IL-6, IL-10, RANTES, and vascular endothelial growth factor by γδ T-cells thereby resulting in initiation of sterile inflammation leading to tissue/cellular repair (120).

Different studies have reported that γδ T cells express TLR3 (121, 122). TLR3 recognizes viral dsRNA, synthetic analogs of dsRNA, polyinosinic–polycytidylic acid [poly (I:C)] and small interfering (si) RNA. The direct stimulation of freshly isolated γδ T cells via TCR and surrogate TLR3 ligand poly (I:C) dramatically increased IFN-γ production. Addition of neutralizing anti-TLR3 mAb inhibited the co-stimulatory effect of poly (I:C), presumably by antagonizing the TLR3 signaling (122). Thus, the integrated signals of TLR3 and TCR induce a strong antiviral effector function in γδ T cells supporting the decisive role of γδ T cells in early defense against viral infection. In other study, it has been reported that γδ cells of term babies and of adults express TLR3 and TLR7 while the preterm babies have reduced levels. The greater levels of IFN-γ protein was observed in adult and cord blood cells costimulated with anti-CD3 and poly(I:C) whereas this was not seen in γδ T cell clones of preterm babies. Thus, reduced level of TLR3 expression by preterm-derived clones had an overt functional consequence on IFN-γ levels (11). Interestingly, a primary role of TLR3 in humans appears to mediate resistance to HSV-induced encephalitis (123). Hence, premature babies are particularly susceptible to HSV infection because of reduced levels of TLR3 on γδ T cells.

TLR4 was reported to be absent in the γδ T cells but can become functional in γδ T cells depending on localization, environmental signals, or γδ TCR usage (19, 118, 124). However, our own data has shown that TLR4 is expressed on human γδ T cells. Stimulation of γδ T cells with LPS (TLR4 ligand) increased their proliferation, IFN-γ release, and cytotoxic potential (125). DETCs lack cell surface expression of TLR4–MD2. MD-2 physically associates with TLR4 on the cell surface and is required for LPS signaling. However, TLR4–MD2 expression was upregulated when DETCs emigrated from the epidermis during cutaneous inflammation. The migration signals of DETCs may promote the TLR4–MD2 expression (126). Cairns et al. showed that late post-burn injury increased expression of TLR-4 on splenic T-cells (127). However, Martin et al. reported transient TLR-4 expression post-burn in the circulation or spleen but were specific for the γδ T-cell subset (119). Several evidences suggest that murine γδ T cells recognize LPS/LA through TLR2 or TLR4 (128, 129). Importantly activated γδ T cells, especially Vδ2 T cells, in peripheral blood cells recognize LA, a major component of LPS, via TLR4 resulting in extensive proliferation and production of IFN-γ and TNF-α *in vitro* (130). The data suggest that γδ T cells play an important role in the control of infection induced by gram negative bacteria. Reynolds et al. showed that a heterogeneous population of γδ T cells responds to LPS via TLR4 dependent manner and demonstrate the crucial and innate role of TLR4 in promoting the activation of γδ T cells, which contributes to the initiation of autoimmune inflammation (100). Another study showed the indirect role of TLR4 in HMGB–TLR4– IL-23–IL17A axis between macrophages and γδ T cells, which contribute to the accumulation of neutrophils and liver inflammation. Necrotic hepatocytes release HMGB1, a damage-associated molecule or TLR4 ligand, which increased IL-23 production of macrophages in a TLR4 dependent manner. IL-23 aids γδ T cells in liver in the generation of IL-17A, which then recruits hepatic neutrophils (131).

Human γδ T cells were found to express appreciable levels of TLR7. Costimulation with poly I:C upregulated the TLR7 expression in TCR-cross linked freshly isolated γδ T cells (124). In addition, tumor-infiltrating γδ T cells also express TLR7 (132). In case of mouse dermal γδ T cells, both TLR7 and TLR9 signaling promoted IL-17 production, which could be synergistically enhanced with the addition of IL-23 (19).

The identification of dominant γδ T cells in the total population of tumor-infiltrating lymphocytes (TILs) in renal, breast, and prostate cancer suggested that these cells might have the potent negative immune regulatory function (132,133). The breast tumor-derived bulk γδT cell lines and clones efficiently suppressed the proliferation and IL-2 secretion of naïve/effector T cells and inhibited DC maturation and function. Hence, their depletion or the reversal of their suppressive function could enhance antitumor immune responses against breast cancer. Indeed as in CD4<sup>+</sup> regulatory T cells (Tregs), the immunosuppressive activity of γδ T cells could be reversed by human TLR8 ligands both *in vitro* and *in vivo*. Study revealed that MyD88, TRAF6, IKKα, IKKβ and p38α molecules in γδ1 cells were required for these cells to respond to TLR8 ligands (132, 134, 135). **Table 2** shows expression and co-stimulatory effects mediated by TLR activation of γδ T cells

#### **TLRs MODULATE CROSSTALK BETWEEN** γδ **T AND DENDRITIC CELLS**

The functional fate of effector T cells is governed by antigen presentation and the cytokine milieu in the local environment. Dendritic cells (DCs) being professional APCs, recognize the danger signal, process it, and present it to the T lymphocytes thereby modulate adaptive immune response. γδ T cells influence the antigen presenting property of DCs. DCs pre-incubated with activated γδ T cells enhance the production of IFN-γ by alloreactive T cells in mixed lymphocyte reaction (136). Moreover, γδ T cells not only upregulated CD86 and MHC I expression on DC but themselves get activated, leading to up-regulation of CD25, CD69, and cytokine production (137). These studies showed how γδ T cell and DCs regulate each other's function. There are reports, which have shown how γδ T cells interact with DC or *vice versa* via TLR ligands. Leslie et al. reported that stimulation with TLR ligands in γδ/DC cocultures enhanced the maturation and production of IL12p70 by DCs (138). TLR also regulate the γδ T cells and DC crosstalk in microbial context. TLR2-stimulated DCs enhanced IFN-γ production by Vδ2 T cells; conversely, phosphoantigen activated Vδ2 T cells enhanced TLR2-induced DC maturation via IFN-γ, which co-stimulated interleukin-12 (IL-12) p70 secretion by DCs (139). Further, γδ T cells stimulated with TLR7 (CL097) or TLR3 (poly I: C) agonists produce IFN-γ, TNFα and/or IL-6 thereby inducing DC maturation, which prime effector T cells against West Nile Virus (WNV) infection (140). This study

confirmed that the antiviral effector immunity may be regulated by interplay of DCs, γδ T cells and TLRs. Similarly, in human's γδ T cells and DCs regulate each other's immunostimulatory functions. TLR3 and TLR4 ligands stimulation of human PBMCs induced a rapid and exclusive IFN-γ production by Vγ9Vδ2 subset dependent on type 1 IFN secreted by monocytic DC. TLR-induced IFN-γ response of Vγ9Vδ2 T cells led to efficient DC polarization into IL-12p70-producing cells (58). In another study, it was reported that Vδ2 cells are indirectly activated by BCG and IL-12p70 secreted by DCs. IL-12p70 production by DC is modulated by Toll like receptor 2/4 ligands from BCG and IFN-γ secreted by memory CD4 T cells (141). This study portrayed the complex interplay between cells of the innate and adaptive immune response in contributing to immunosurveillance against pathogenic infections.

#### **TLRs COMPLEMENT CYTOTOXIC POTENTIAL OF** γδ **T CELLS AGAINST TUMOR CELLS**

γδ T cells have capability to lyse different types of tumors and tumor-derived cell lines (49, 50, 142–145). Circulating as well as tumor-infiltrating γδ T cells have the ability to produce abundant proinflammatory cytokines like IFN-γ and TNF-α, cytotoxic mediators and MHC-independent recognition of antigens, render them as important players in cancer immunotherapy (143, 145). In addition to TCR, γδ T cells use additional stimulatory co-receptors or ligands including TLRs to execute effector functions and TLR agonists are considered as adjuvants in clinical trial of cancer immunotherapy (146). Kalyan et al. even quoted that "TLR signaling may perfectly complement the anti-tumor synergy of aminobisphosponates and activated γδ T cells and this combined innate artillery could provide the necessary ammunition to topple malignancy's stronghold on the immune system" (147). Paradoxically, TLR agonists execute dual role of enhancing immune response (148) as well as increasing invasiveness of tumor cells (149–152). Hence, the tripartite cooperation of tumor cell, TLRs, and γδ T cells should be carefully analyzed. In concordance to this, Shojaei et al. reported that Toll like receptor 3 and 7 agonists enhanced the tumor cell lysis by human γδ T cells. The enhanced capability of γδ T cells to lyse tumor cells was attributed to increased expression of CD54 and downregulation of MHC class 1 on tumor cells. Poly(I:C) treatment of pancreatic adenocarcinomas resulted in overexpression of CD54 and concomitant coculture of tumor cells with γδ T cells led to interaction between CD54 and its ligand CD11a/CD18 triggering effector function in γδ T cells. However, TLR7 surrogate ligand induced


#### **Table 2 | Expression and functions mediated by TLRs on** γδ**T cells**.

downregulation of MHC class 1 molecule on tumor cells resulting in a reduced affinity for inhibitory receptor NKG2A on γδ T cells (59). Manipulation of TLR signaling by using TLR8 agonists reversed the suppressive potential of γδ Tregs found elevated in breast cancer (132). Polysaccharide K (PSK) known for its antitumor and immuno-modulatory function can also activate TLR2 leading to increased secretion of IFN-γ by γδ T cells on stimulation. The cell–cell contact between γδ T cells and DC was required for optimal activation of γδ T cells. However, PSK along with anti-TCR could co-activate γδ T cells even in the absence of DC. The study confirmed that the anti-tumor effect of PSK was through activation of γδ T cells (153).

Studies from our lab have shown that the TLR signaling in γδ T cells derived from the oral cancer (OC) patients may be dysfunctional. We reported that γδ T cells from healthy individuals (HI) and OC patients express higher levels of TLR2, TLR3, TLR4, and TLR9 than in αβT cells. Higher TLR expression was observed in HI compared to OC patients. Stimulation with IL2 and TLR agonists (Pam3CSK, Poly I:C, LPS, and CpG ODN) resulted in higher proliferative response of peripheral blood lymphocytes from HI compared to OC patients. However, the role of other immune cells that may influence the TLR ligand stimulation induced activation status of lymphocytes cannot be ignored (125). Impairment in TLR expression/signaling can be viewed as a strategy employed by tumor cells to avoid immune recognition.

#### **TLRs AND** γδ **T CELLS IN DISEASES**

Studies have demonstrated the protective role of γδ T cells in infection and inflammation (154–157). Inoue et al. showed that during mycobacterial infection, γδ T cells precedes the αβ T cells, indicating role of γδ T cells as first line of defense against infections (158). The conserved molecular patterns associated with pathogens are directly recognized by γδ T cells leading to rapid protective response against the danger signal. Unlike αβ TCR, γδ TCR acts as pattern recognition receptor providing advantage in anti-infection immunity by directly initiating cytotoxicity against infected cells or through production of cytokine to involve multiple immune system components to combat infection (159, 160). Activated γδ T cells through TLR3 and TLR4 ligands rescue the repressed maturation of virus-infected DCs and mount a potent antiviral response (58, 140). Malarial infection in MyD88 deficient mice resulted in impairment in CD27−IL-17A-producing γδ T cell without affecting the IFN-γ producing γδ T cells (161). This study specifies the role of TLR in promoting proliferation

**FIGURE 1 | Improving** γδ**T cell functions by TLRs in combinatorial therapy**. **(A)** TLR agonists induce effector function of γδ T cells through IFN-γ, TNF-α, IL-6 secretion, and increased expression of CD107a. **(B)** IFN-γ, TNF-α, and IL-6 secreted by γδ T cells and TLR agonists promote the maturation of dendritic cell. **(C)** γδ T cells upregulate CD86 and MHC I expression on DCs and are themselves activated through up-regulation of CD25, CD69, and cytokine production thereby modulating each other's function. **(D)** Co-stimulation of γδ T cells with TLR agonists and IL-1β secreted by dendritic cells promote their polarization toward IL17 producing cells. **(E)** γδ TCR also recognizes the specific molecular patterns such as IPP, which are induced upon inhibition of mevalonate pathway by bisphosphonates. Moreover, NKG2D receptor on γδ T cells recognizes MICA/B or ULBP expressed on tumor cells. This binding enhances release of perforins and granzymes by the γδ T cells leading to tumor cell lysis. **(F)** TLR agonists act as adjuvants and can induce CD54 expression and downregulation of MHC class 1 on tumor cells. Interaction between CD54 and its ligand CD11a/CD18 trigger effector functions in γδ T cells. Downregulation of MHC class 1 molecule on tumor cells result in reduced signaling through the inhibitory receptor NKG2A on γδ T cells, which enhances the cytotoxic potential of γδ T cell.

of proinflammatory γδ T cells. Another study by Martin et al. showed that IL17 producing γδ T cells express TLR1 and TLR2 and expand in response to their ligands and mount an adequate response against heat-killed *M. tuberculosis* or *C. albicans* infection (118). However, γδ T cell are also known to directly recognize the pathogen-derived molecules and mediate cytotoxic effector function either through secretion of perforin and granzyme B or by secretion of proinflammatory cytokine IL17 (162–164). The involvement of TLRs in regulating anti-microbial γδ T cell function should be investigated in depth to exploit it as a cell based therapy for infectious diseases.

#### **CONCLUDING REMARKS**

The characteristic copious IFN-γ or IL17 secretion, MHCindependent antigen recognition, tissue tropism, and potent cytotoxicity make γδ T cells promising targets for immunotherapy. Similar to αβ T cells, γδ T cells exhibit functional and phenotypic plasticity, which influences the nature of the downstream adaptive immune response. The adoptive transfer of *ex vivo* expanded Vγ9Vδ2 T cells or *in vivo* activation of Vγ9Vδ2 T cells (phosphoantigens or amino-bisphosphonates) can be utilized as adjuvant to conventional therapies. Clinical trials of Vγ9Vδ2 T cells as immunotherapeutic agents have shown encouraging results that could be attributed to its low toxicity grade. Combinations of cellular immune-based therapies with chemotherapy and other anti-tumor agents may be of clinical benefit in the treatment of malignancies. Combinatorial treatment using, chemotherapeutic agents or bisphosphonate zoledronate (ZOL) sensitizes tumorderived cell lines to rapid γδT cells killing.Vγ9Vδ2 T cell triggering may be also enhanced by combining TCR stimulation with engagement of TLRs. Various TLR agonists are currently under investigation in clinical trials for their ability to orchestrate anti-tumor immunity. In one study, simultaneous use of both Imiquimod (TLR7 agonist) and CpG–ODN (TLR9 agonist) loaded onto virus like nanoparticles was found to be effective in triggering effector and memory CD8<sup>+</sup> T cell response (165). Similarly, combination of γδ T cells and DCs along with nanoparticle loaded TLR agonists can be employed for developing effective immunotherapeutic strategies. The direct or indirect stimulation of γδ T cells by TLR agonists could be a strategy to optimize Th1-mediated immune responses as adjuvant in vaccines against infectious or malignant diseases.

Administration of an "immunogenic chemotherapy" (such as oxaliplatin or anthracycline or an X-ray-based regimen) or local delivery of TLR surrogates in the tumor microenvironment (which stimulate local DCs and provides a source of IL-1β) may be also instrumental in polarization of γδ TILs into IL17 producing cells. Tγδ17 cells play a crucial role in anti-microbial immunity but their role in tumor immunity remains controversial. Tγδ17 have both pro and anti-tumor properties. TLR use in combinatorial therapy, therefore, could be a double edged sword. Careful use of TLR agonists in combinatorial γδ T cell based therapy is needed to strike the balance between pro and anti-tumor effects (**Figure 1**).

#### **REFERENCES**

1. Tipping PG. Toll-like receptors: the interface between innate and adaptive immunity. *J Am Soc Nephrol* (2006) **17**(7):1769–71. doi:10.1681/ASN. 2006050489


ex vivo from patients with poor-prognosis neuroblastoma lyse autologous primary tumor cells. *J Immunother* (2010) **33**(6):591–8. doi:10.1097/CJI. 0b013e3181dda207


**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.

*Received: 27 May 2014; accepted: 15 July 2014; published online: 31 July 2014.*

*Citation: Dar AA, Patil RS and Chiplunkar SV (2014) Insights into the relationship between toll like receptors and gamma delta T cell responses. Front. Immunol. 5:366. doi: 10.3389/fimmu.2014.00366*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Dar, Patil and Chiplunkar. 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.*

## Toll-like receptors and cancer: MYD88 mutation and inflammation

#### **James Q.Wang\*† ,Yogesh S. Jeelall \*† , Laura L. Ferguson and Keisuke Horikawa**

Department of Immunology, John Curtin School of Medical Research, Australian National University, Canberra, ACT, Australia

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Ronald B. Corley, Boston University School of Medicine, USA Muriel Moser, Université Libre de Bruxelles, Belgium

#### **\*Correspondence:**

James Q. Wang and Yogesh S. Jeelall, Immunogenomics Laboratory, Department of Immunology, The John Curtin School of Medical Research, Australian National University, Building 131, Garran Road, Canberra, ACT 0200, Australia e-mail: james.wang@anu.edu.au; yogesh.jeelall@anu.edu.au

†James Q. Wang and Yogesh S. Jeelall have contributed equally to this work.

Pattern recognition receptors (PRRs) expressed on immune cells are crucial for the early detection of invading pathogens, in initiating early innate immune response and in orchestrating the adaptive immune response. PRRs are activated by specific pathogen-associated molecular patterns that are present in pathogenic microbes or nucleic acids of viruses or bacteria. However, inappropriate activation of these PRRs, such as the Toll-like receptors (TLRs), due to genetic lesions or chronic inflammation has been demonstrated to be a major cause of many hematological malignancies. Gain-of-function mutations in the TLR adaptor protein MYD88 found in 39% of the activated B cell type of diffuse large B cell lymphomas and almost 100% of Waldenström's macroglobulinemia further highlight the involvement of TLRs in these malignancies. MYD88 mutations result in the chronic activation of TLR signaling pathways, thus the constitutive activation of the transcription factor NFκB to promote cell survival and proliferation. These recent insights into TLR pathway driven malignancies warrant the need for a better understanding of TLRs in cancers and the development of novel anti-cancer therapies targetingTLRs.This review focuses onTLR function and signaling in normal or inflammatory conditions, and how mutations can hijack the TLR signaling pathways to give rise to cancer. Finally, we discuss how potential therapeutic agents could be used to restore normal responses to TLRs and have long lasting anti-tumor effects.

**Keywords: cancer, drug targets, inflammation, lymphoma, MYD88 L265P, pattern recognition receptors, self-nucleic acid,Toll-like receptors**

#### **INTRODUCTION**

Pattern recognition receptors (PRRs) are germline-encoded receptors with the ability to relay "danger signals" to the host in order to mediate an early innate immune response. The term "pattern recognition receptors" comes from their ability to recognize specific pathogen-associated molecular patterns (PAMPs) and danger-associated molecular patterns (DAMPs) (1, 2). PRRs can be broadly divided into five distinct subfamilies: Toll-like receptors (TLRs), C-type lectin receptors (CLRs), NOD-like receptors (NLRs), RIG-1-like receptors (RLRs), and AIM2-like receptors (ALRs). These PRR subfamilies differ in their structures, localization patterns, the distinct types of ligands they recognize, and the activation of specific intracellular signaling cascades to mediate a range of responses such as the regulation of gene transcription, cell activation, and proliferation, and the production of proinflammatory cytokines,chemokines,and anti-viral molecules (3).

One of the most well characterized PRR is the TLR (4). TLRs are type I transmembrane proteins with an extracellular domain consisting of leucine-rich repeats and a cytoplasmic domain homologous to that of the interleukin (IL)-1 receptor (5, 6). These evolutionarily conserved receptors are absolutely critical for the host innate immune response against many pathogens (7). Activation of TLRs depends on the number of different ligands they may encounter, which is by large, governed by their subcellular localization. Much insight has been gained in recent years on the localization and trafficking of TLRs and the important roles their localization play in the way they recognize their ligands. TLRs can be divided into two groups based on their subcellular localization, either on the cell surface or within intracellular compartments (8). Given the ability of TLRs to recognize a large number of pathogen-associated ligands such as glycoproteins, lipopolysaccharides, flagellin, and viral double-strand or single-strand RNAs or DNAs, TLRs have emerged as an important family of PRRs in shaping both the innate and adaptive immunity (7). However, inappropriate activation of these pathways can often lead to chronic inflammatory diseases or cancer.

This review will focus on TLRs and malignancies associated with the dysregulation of TLR signaling pathways. TLR activation by somatic MYD88 mutations and chronic inflammations has been implicated in a number of hematological malignancies. Targeting the TLR signaling network has gained increasing attention from researchers and clinicians seeking strategies to achieve long lasting anti-tumor outcomes. Here, we discuss the signal transduction and immune regulation by TLRs and the immunological malignancies that manifest from dysregulation of TLR pathways, including how targeting these pathways could be an attractive therapeutic regime.

#### **TOLL-LIKE RECEPTORS**

Toll-like receptors are probably the best studied PRRs that participate in the first line of host defense against pathogens. TLRs belong to an evolutionarily conserved family of adaptors sharing homology with the *Drosophila* protein Toll, which is best known for its essential role in establishing dorsoventral polarity during embryogenesis in insects (9). Amino acid sequencing and hydropathy profiling identified Toll as a type I transmembrane protein with a membrane-spanning segment and multiple tandem leucine-rich repeats directed at the extracellular surface (9). Further biochemical and functional studies conducted on the receptor Toll and its leucine repeats established it as a critical pathogen sensing receptor for recognizing bacteria and fungus in *Drosophila* (10). This study later became critical for the discovery of Toll-like homologs (TLRs) in mammals as mediators of the innate immunity (4, 10, 11).

A total of 10 TLRs have been identified in humans and 12 in mice (7). Due to the small repertoire of TLRs available to recognize a virtually unlimited combination of pathogen-associated patterns, each individual TLR must be able to detect and respond to a large number of pathogens ranging from bacteria, fungi, protozoa, and viruses (12, 13). For instance, TLRs 1, 2, and 6 recognize lipo-, glycol-, and acyl-peptides expressed on the surfaces of many Gram-positive and Gram-negative bacteria and mycobacteria (7). Additional cooperation between TLRs 1, 2, and 6 enables them to further discriminate different microbial components (14). TLR4 recognizes lipopolysaccharide components of the cell wall of Gram-negative bacteria through its co-receptor MD-2 (15, 16). In addition, TLR4 can also recognize endogenous ligands such as heat-shock proteins, extracellular matrix components including fibronectin, hyaluronic acid, and heparin sulfate in response to tissue injury (7). The nucleic acid sensing subfamily of TLRs consists of TLRs 3, 7, and 9 and exhibit unique endosomal localization in contrast to the surface expression of the other TLRs (17). These TLRs have the ability to detect nuclear material such as ssRNAs, dsRNAs, and dsDNAs and are vital for anti-viral responses (18–21). Importantly, these nucleic acid sensing TLRs must discriminate between foreign and selfnuclear material to prevent autoimmunity. Due to the relative lack of specificity of TLRs compared to the B cell receptors (BCRs), restriction of self-TLR activation must be achieved through other means. TLRs are protected from engaging self-nuclear material by Unc93b mediated restriction to the endosome (22). In such way, self-nucleic acids are prevented from entering the endosome, but foreign material can enter via endocytosis and be processed in the acidified endosomes in order to activate the endosomal TLRs (23, 24).

Together, the 10 human TLRs can recognize a virtually unlimited combination of pathogens, however, the downstream signaling pathways they share are striking. All TLRs except for TLR3 signal through the adaptor protein MYD88 (25). Upon ligand binding, TLRs induce the dimerization of their ectodomains, bringing the cytoplasmic TIR domains together, and initiating a signaling cascade via signal adaptor molecules. The four main TLR adaptor molecules are the myeloid differentiation response protein 88 (MYD88), Toll-interleukin 1 receptor (TIR) domain containing adaptor protein (TIRAP; also known as MAL), TIRAP inducing IFN-β (TRIF), and TRIF-related adaptor molecule (TRAM) (**Figure 1**). These adaptors are used in various combinations by the different TLRs, but these signaling pathways can be broadly classified into either MYD88 dependent or MYD88 independent.

#### **MYD88 DEPENDENT TLR SIGNALING**

With the exception of TLR3, all TLRs initiate a MYD88-dependent signaling pathway (26). The signal adaptor protein MYD88 contains two main conserved protein domains; a C-terminal TIR and a N-terminal death domain (DD) (27, 28). Upon TLR activation, MYD88 is recruited to the TIR domain of the activated TLR via TIR–TIR interaction (29). The serine–threonine kinase, IL1-receptor associated kinase 4 (IRAK4), is then recruited to MYD88 through the interaction of their DD domains. IRAK4 then recruits and phosphorylates IRAK1 and IRAK2 to form a structure known as the "Myddosome" (30). Phosphorylation of IRAKs 1 and 2 allows them to interact with the E3 ubiquitin ligase, TRAF6, via their TRAF binding domain (31). TRAF6 then ubiquitylates and activates TAK1 (32), which has the dual ability to activate both the NFκB pathway and the mitogen-activated protein kinase (MAPK) pathway (26). In resting cells, NFκB dimers are sequestered in an inactive form in the cytoplasm by the IκB protein (33). During NFκB activation, TAK1 phosphorylates and activates IκB kinase β (IKKβ), which in turn phosphorylates IκB, targeting it for proteosomal degradation (34). The degradation of IκB releases NFκB, enabling it to enter the nucleus and bind to sequences known as κB sites to activate transcription of genes (35). TAK1 also activates the MAPK pathway, leading to the activation of c-Jun N-terminal kinase (JNK), which activates the Jun family of transcription factors (36) (**Figure 1**).

The MYD88-dependent pathway can be initiated by TLR5 and TLR7-9 using the adaptor MYD88 alone, while the adaptor protein TIRAP is required with MYD88 to initiate signaling downstream of TLR2 and TLR4 (37, 38). In this subset of TLRs, TIRAP acts as a sorting molecule that is necessary for efficient recruitment of MYD88 to the activated TLRs to initiate signal transduction to activate NFκB and produce pro-inflammatory cytokines (39). During TLR 7 and 9 activation, MYD88 also recruits TRAF3 to activate TBK1 and IKKε, which phosphorylates the transcription factor interferon-regulatory factor 7 (IRF7) and leads to IFN-α production (40, 41). IFN-α production, as with production of other IFNs, is particularly important for anti-viral responses (42) (**Figure 1**).

#### **MYD88 INDEPENDENT TLR SIGNALING**

MYD88-independent TLR3 signaling requires the adaptor molecule TRIF to activate downstream signaling pathways, including the activation of IRF3 and the production of IFN-β (43). TRIF has also been known to participate in signaling downstream of TLR4 for type 1 interferon responses (44). Upon ligand binding, TRIF recruits TRAF3, which acts as a scaffold for the activation of the IKKs, TBK1, and IKKε, leading to the phosphorylation and activation of the transcription factor IRF3 and IFN-β transcription (45, 46). While TLR3 can activate this pathway using TRIF alone, the adaptor TRAM is required for TLR4, where TRAM facilitates the recruitment of TRIF to TLR4 (47). Upon the activation of TRIF, TRAF6 is recruited, which then activates TAK1 through ubiquitination and leading to the subsequent activation of NFκB (48)(**Figure 1**). Interestingly, TRAF3 has been shown to play important roles in regulating both the MYD88 dependent and independent response through its differential ubiquitination (49). MYD88-independent signaling triggers the non-degradative

self-ubiquitination of TRAF3, promoting IRF3 activation. On the other hand, the MYD88-dependent pathway results in the degradative ubiquitination of TRAF3 and the activation of TAK1 (49). In this manner, TRAF3 acts to balance pro-inflammatory and IFN response by the MYD88 dependent and independent pathways.

#### **HEMATOLOGICAL MALIGNANCY AND MYD88 MUTATION**

Inappropriate activation of TLRs due to the somatic acquisition of gain-of-function mutations in the TLR adaptor protein MYD88 has been implicated in many hematological malignancies. Activated B cell type diffuse large B cell lymphoma (ABC-DLBCL), a particularly aggressive subtype of DLBCL whose pathogenesis relies on constitutively active NFκB, frequently accumulates MYD88 mutations. 39% of tumor samples contain mutations in MYD88, and strikingly, 29% of those mutations result in a single nucleotide change from leucine into proline at position 265 (L265P) (50). shRNA knockdown of MYD88 in lymphoma cell lines demonstrated that MYD88 mutations are critical for their survival and high NFκB transcription factor activity (50). A hyperphosphorylated isoform of IRAK1 was strongly associated with the

L265P mutant form of MYD88, suggesting that this mutation is a gain-of-function mutation that leads to the constitutive activation of downstream IRAKs (50). The effects of the L265P mutation include increased NFκB activity as well as increased JAK-STAT3 signaling and the production of pro-inflammatory cytokines such as IL6, IL10, and IFN-β (50). The production of these cytokines further activates JAK-STAT3 signaling as part of an autocrine loop that enhances the survival of the lymphoma cells (51, 52) (**Figure 2**).

MYD88 mutations have since emerged in a number of other human malignancies, with the L265P mutation found in including almost 100% of Waldenström's macroglobulinemia (WM), 2–10% of chronic lymphocytic leukemia (CLL), 69% of cutaneous diffuse large B cell lymphoma (CBCL), and 38% of primary central nervous system lymphoma (PCNSL) (previously reviewed in Ref. (53)). However, the effect of single MYD88 L265P mutation on tumor growth is confounded by the accumulation of other potential damaging mutations in the same malignant clones. Recently, a retroviral gene transfer strategy to study the effects of single MYD88 mutation in other wise normal mature B cells found that the MYD88 L265P mutation alone was able to drive limited

rounds of mitogen independent B cell proliferation both *in vitro* and *in vivo* (54). Nevertheless, the drive for B cell proliferation was dependent on intact nucleic acid sensing TLR activity since *Unc93b13d* mutation or *Tlr9* deficiency inhibited the proliferation of MYD88 L265P B cells *in vitro* (54). Other studies have also shown that oncogenic MYD88 depends on TLRs by using the depletion of UNC91B1, PRAT4A, and CD14 in ABC-DLBCL lines as well as by using pharmacological inhibitors to TLR7 and TLR9 (55). Given that intact TLR activity is critical for lymphoma cells carrying MYD88 mutations, targeting this pathway appears to be attractive for treating these malignancies. Indeed, blocking endosome acidification using chloroquine selectively inhibits MYD88 L265P mutation driven B cell proliferation *in vitro* (54). The use of chloroquine to treat hematological malignancies should be further explored, as evidence suggests that there is a strong involvement of the activation of nucleic acid sensing TLRs that depends on normal endosome acidification in promoting proliferative abnormality in these tumors.

#### **HEMATOLOGICAL MALIGNANCY AND INFLAMMATION**

Remarkably, inflammation enables most of the key cellular and molecular capabilities that are required for carcinogenesis, such as genomic instability, proliferative abnormality, and reprograming of the stromal environment (56). Although, the mechanisms by which inflammation promotes neoplastic transformation are not fully understood, it is apparent that, in many cases, tumor development is linked to chronic inflammation (57, 58).

The link between inflammation and tumor formation was first speculated by Virchow in the 1800s as he observed that tissue injury and inflammation induced by irritants could promote cell proliferation (59). Infection has been accepted as a major driver of inflammation-induced tumorigenesis, with up to one-fifth of all cases of cancer associated with infection (60, 61). For instance, persistent *Helicobacter pylori* infection is associated with gastric cancer and mucosa-associated lymphoid tissue (MALT) lymphoma, infections with hepatitis B and C viruses are associated with hepatocellular carcinoma, and infections with *Bacteroides* species are linked to colon cancer (62, 63). The inflammatory response triggered by infection is a part of normal host defense to eliminate the pathogen. However, some tumorigenic pathogens subvert host immunity and establish persisting infections, leading to chronic inflammation (64, 65).

Persistent inflammation establishes a microenvironment, which contains macrophages, dendritic cells, natural killer cells, and T and B lymphocytes in addition to the surrounding stroma (66) (**Figure 3A**). These diverse cells communicate with each other by means of direct contact or cytokine and chemokines, which influence tumor formation and growth (67, 68). This network of inflammatory cells promotes the formation of cancerous cells, which further complicates the initial chronic inflammation induced by infection. The neoplastic cells trigger anti-tumor immunity, which further adds to the established inflammation. Early during tumor formation, whether tumor-promoting inflammation or anti-tumor immunity follows seems to be stochastic and is influenced by a combination of cell-intrinsic and cell-extrinsic processes (69,70). In established cancers,pro-tumor inflammation seems to be favored, as without therapeutic intervention advanced tumors rarely regress.

Continuous stimulation of TLRs by microbial products constitutively engages the activation of the NFκB and STAT3 transcription factors, which exert pro-cancerous activity through multiple effectors (62, 71, 72). Additionally, the production of cytokines by the host inflammatory cells activates these transcription factors (62) (**Figure 2**). These cytokines facilitate the establishment of feed-forward signal amplification loops, which ultimately promote cell proliferation and resistance to cell death. For instance, the expression of the anti-apoptotic proteins Bcl-2 and Bcl-X<sup>L</sup> is promoted by both NFκB and STAT3, as is the expression of c-IAP1, c-IAP2, Mcl-1, c-FLIP, and survivin (62, 72). Moreover, both transcription factors interfere with p53 expression and function, representing another potential tumor-promoting mechanism (73).

An additional mechanism linking inflammation to tumor formation is the expression of activation-induced cytidine deaminase (AID), an enzyme that promotes immunoglobulin gene class switching by catalyzing deamination of cytosines in DNA (74). In addition to B lymphocytes, where it was originally discovered, AID is overexpressed in many cancers of diverse origin, and its expression is induced by inflammatory cytokines in a NFκB-dependent manner (74). AID induces genomic instability and increases mutation probability during the error-prone joining of double-stranded DNA breaks. This mutagenic process causes

mutations in critical cancer-associated genes such as Tp53 and c-Myc (75, 76).

#### **HELICOBACTER PYLORI, INFLAMMATION AND MALT LYMPHOMA**

MALT lymphomas, which occur in the context of chronic inflammation caused by infectious agents, such as *H. pylori* (gastric lymphoma), *Chlamydia psittacii* (ocular adnexal lymphoma), and *Borrelia burgdorferi* (cutaneous lymphoma) are a prime example of lymphoid malignancies associated with chronic inflammation (77, 78). Interestingly, in some patients, gastric MALT lymphoma and diffuse large B cell lymphoma (DLBCL) co-occur, indicating that MALT lymphomas can develop into more aggressive DLBCL (79). The pathogenesis of MALT lymphoma involves several steps, which result in the transformation of a single B cell clone, initially part of the polyclonal B lymphocyte response against *H. Pylori* into a monoclonal tumor (78) (**Figure 3B**). Under physiological conditions, the stomach lacks MALT because the low pH prevents the survival of lymphocytes in the gastric wall. However, *H. pylori* infection results in buffering of the gastric pH owing to the secretion of bacterial urease. The decreased acidity of stomach

environment, along with the presence of the infection, triggers lymphoid infiltration and the establishment of MALT (78).

Subsequently, the continuous presence of *H. pylori* induces an upregulation of TLR4 and MD-2 expression in gastric epithelial cells, which contributes to establishing a persistent inflammatory environment (80–82). Although, the role of TLRs in the pathogenesis of MALT lymphoma has been poorly investigated, the immune response to chronic stimulation by *H. pylori* infection is thought to induce NFκB activation in B cells, which plays a crucial part in the development of MALT lymphoma (83, 84). In addition, the presentation of *H. pylori* by dendritic cells recruits and activates T cell responses, which enhance B cell activation through CD40– CD40L interactions (85) (**Figure 3A**). Thus, both direct activation of TLR signaling by *H. pylori* and T cell-mediated B cell activation could be involved in the pathogenesis of MALT lymphoma (86).

Interestingly, several lines of evidence indicate that chronic antigen stimulation precedes MALT lymphoma pathogenesis. The rearranged *IGVH* genes from MALT lymphomas have a high frequency of variants, which have been implicated in autoantibody production (87). In addition, approximately half of the MALT lymphoma cases display evidence of intraclonal variation in the *IGVH* locus, indicating that continued antigenic stimulation is a key driver of clonal B cell expansion (87). As both somatic hypermutation and intraclonal variations are antigen-driven processes, their occurrence in gastric MALT lymphoma strongly indicates a role for antigens during both initiation and progression of this neoplasm.

Remarkably, tumor-derived immunoglobulins from MALT lymphomas bind to various autoantigens as well as *H. pylori* with varying affinities (87). The autoantigens include DNA and stomach-associated antigens, which could be abundant in the MALT-microenvironment under a situation of continuous inflammation. Given that *H. pylori* eradication with antibiotics is the preferred therapy for patients with *H. pylori*-positive gastric MALT lymphoma (88, 89), and the evidence that MALT lymphoma cells proliferate when stimulated with *H. pylori* in tissue culture, one possible hypothesis is that neoplastic B cells receive proliferative signals from both the B cell receptor and TLRs, which are continuously and simultaneously engaged by self-antigens and LPS from *H. pylori* respectively. Thus, the eradication of *H. pylori* by antibiotics disrupts a critical 'weak' link in the inflammatory process, which gradually resolves and shuts off the supply of autoantigens available to lymphoma cells.

#### **ROLE OF INFLAMMATION AND CYTOKINES IN CLL AND MULTIPLE MYELOMA**

It is apparent that antigenic stimulation, autoimmunity, and inflammation contribute to the development of CLL (90). One mechanism through which these stimuli promote CLL development is induction of B cell activating factor (BAFF), a member of the TNF family, recently shown to accelerate development of CLL-like disease in mice (91). In addition, cytokines such as IL6 and interactions with bone marrow stromal cells support CLL expansion and suppress apoptosis through the expression of Bcl-2, Survivin, and Mcl-1 (92, 93). Increased IL6 production activates the JAK-STAT, MAPK, and PI3K pathways to promote cell survival, proliferation, and resistance to apoptosis (94–96), with the constitutive activation of STAT3 being a hallmark for CLLs (97, 98). Similarly, through the secretion of IL6, TNF-α, and BAFF, bone marrow stromal cells promote the survival of neoplastic plasma cells and also confer drug resistance in multiple myeloma (99). Interestingly, IL6-deficient mice are resistant to induction of multiple myeloma (100, 101). Thus, despite cell-intrinsic constitutive NFκB activation, multiple myeloma cells depend on an extrinsic source of IL6 for their development and survival. High levels of plasma IL6 have been associated with increased disease progression and decreased survival, thus providing the rationale for the evaluation of combination therapies including drugs targeting IL6 for the treatment of this malignancy (102–104).

#### **TARGETING INFLAMMATION AND TLRs IN CANCER**

Constitutively, active NFκB signaling due to the aberrant activation of TLRs during chronic inflammation or by MYD88 mutation determines the poor clinical outcome of many hematological malignancies. Desirable outcomes in treating these diseases can be achieved by using a combination of inhibition of signal transducers and transcription factors, sequestration of chemokines and cytokines that sustain inflammatory cells, and the depletion of immune or inflammatory cells that promote tumor development.

Gain-of-function MYD88 mutations have emerged as a potent driver of constitutively active NFκB signaling in many tumors. Targeting this pathway is likely going to be useful as part of a multi-component therapy for many hematological malignancies that are addicted to NFκB activity for their survival. MYD88 signaling is critically dependent on its homo-dimerization through conserved residues within the BB-loop structure of the TIR domain (29, 105). Interfering with this interaction by heptapeptides mimicking the BB-loop has achieved significant reduction in NFκB activity (106). Another novel synthetic compound, ST2825, developed by the same group of researchers is currently under pre-clinical evaluation for the treatment of chronic inflammatory diseases (107). Other peptide-based synthetic small molecule inhibitors such as hydrocinnamoyl-l-valyl pyrrolidine (compound 4a) and Pephinh-MYD88 have also been developed to target MYD88 dimerization in the treatment of lymphoma patients with MYD88 mutations (108). However, these potential MYD88 specific therapeutic options are yet to be trialled in large clinical cohorts.

Constitutive NFκB activity in certain lymphoid tumors suggests that the activation of this pathway is crucial for their survival and thus making them attractive drug targets for anti-cancer therapy (62, 72, 109, 110). However in most cases, such therapy is likely to be effective only in combination with more conventional approaches. Furthermore, as genotoxic therapies often lead to NFκB activation in remaining malignant cells, it makes sense to combine genotoxic dugs with NFκB inhibitors to overcome drug resistance. However, prolonged NFκB inhibition can result in severe immune deficiency and may lead to neutrophilia and greatly enhanced acute inflammation due to enhanced IL1β secretion. Such complications as well as increase propensity for liver damage have hindered the clinical development of NFκB and IKKβ inhibitors (57, 111, 112). An attractive alternative target is the STAT3 transcription factor and the signaling pathway that leads to its activation (113, 114). Several STAT3 and JAK2 inhibitors have been described and shown to inhibit the growth of various cancers that exhibit STAT3 activation (115, 116). So far, none of the complications associated with NFκB inhibitor have been reported for STAT3 or JAK2 inhibitors.

It is unlikely that inhibition of NFκB or STAT signaling alone will be sufficient for tumor regression, yet the combination of an NFκB inhibitor and an apoptosis inducing drug or cytokine could be highly effective. Selective inhibition of NFκB in cancer cells blocks the stimulatory effect of TNF and markedly increases susceptibility to TRAIL-induced cell death, resulting in tumor regression (117, 118). NFκB inhibition and anti-TNF therapy, together with the administration of IFN or TRAIL might offer an attractive combined strategy for immunomodulatory cancer therapy. A recent study has found such synergy between lenalidomide and the BTK inhibitor Ibrutinib in killing ABC-DLBCL by the induction of IRF7 and IFN-β production to cause cell cycle arrest and apoptosis (119, 120). Combinatorial strategies provide a distinct advantage where by certain IFN induced side-effects might be diminished after NFκB inhibitor treatment, shifting the

balance of cytokines in the tumor microenvironment to promote tumor regression.

Although it is widely accepted that dampening inflammation and diminishing TLR activity are beneficial for tumor regression, several new lines of evidence have emerged to suggest that TLR agonists could be used as potent anti-tumor agents. When treated with a TLR9 agonist, type B CpG oligodeoxynucleotides (CpG-B ODNs), and CLL B cells that selectively express high levels of TLR9 undergo profound apoptosis by the activation of STAT1, reduction of Bcl-xl pro-survival protein, and elevation of Fas and Fas ligand (121). TLR9 triggered apoptosis seems to be dependent on the altered NFκB status of lymphoma cells compared to normal cells. Moreover, the use of TLR agonists has been known to activate the cognate immune system against cancer cells (122–124). TLRs in lymphoid malignancies appear to be a "double-edged sword" in actively driving disease progression in some but exhibit tumor regressive roles in others. Activation of TLRs by MYD88 mutations has often been associated with poor clinical outcome in lymphoma patients. However, a recent study has reported improved patient survival in a subset of young CLL patients with the identical mutation (125). Interestingly, patients with MYD88 mutations were much younger and had lower expression of CD38 and ZAP-70 than patients with unmutated MYD88. CD38 expression on CLL cells is important for their proliferation and chemotaxis through a signaling pathway involving ZAP-70 (90). Elevated CD38 expression often marks CLL patients with poor clinical outcome and responsiveness to therapy (90). Complex interactions between MYD88 mutation, IGHV mutation status, and CD38 and ZAP-70 levels confound the explanation behind why patients with MYD88 mutations had reduced CD38 expression and show better survival (125).

#### **CONCLUSION**

Pattern recognition receptors protect us from danger and damage associated signals, however, inappropriate activation of these pathways can cause cancer. TLRs can also use ubiquitously available self-ligands such as our own DNA to drive aberrant cell growth when the adaptor protein MYD88 is mutated. This recent finding is one of the many pieces of supportive evidence for Virchow's hypothesis that chronic inflammation is linked with cancer development. Studies into mutations in the TLR signaling pathways have significantly advanced our understanding on the involvement of TLRs in cancer. However, the potential for targeting TLRs as anti-cancer therapy remains an area that is not yet fully understood. Often TLRs act as a "double-edged sword" in cancer, over active TLR signal provides a microenvironment that is necessary for malignant cell proliferation; on the other hand, TLR agonists can also be used to inhibit cancer cell growth. A better understanding of the involvement of TLRs in cancer would help in tipping the balance between tumor stimulatory and inhibitory effects and the development of novel anti-cancer agents.

#### **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.

*Received: 28 May 2014; accepted: 16 July 2014; published online: 31 July 2014.*

*Citation: Wang JQ, Jeelall YS, Ferguson LL and Horikawa K (2014) Toll-like receptors and cancer: MYD88 mutation and inflammation. Front. Immunol. 5:367. doi: 10.3389/fimmu.2014.00367*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Wang , Jeelall, Ferguson and Horikawa. 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.*

## IL-1 receptor-associated kinase signaling and its role in inflammation, cancer progression, and therapy resistance

#### **Ajay Jain<sup>1</sup>\*, Sabina Kaczanowska<sup>2</sup> and Eduardo Davila2,3\***

<sup>1</sup> Division of Hepatobiliary and Pancreatic Surgery, Department of Surgery, State University of New York Upstate Medical University, Albany, NY, USA

<sup>2</sup> Department of Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, MD, USA

<sup>3</sup> Greenebaum Cancer Center, Baltimore, MD, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Bernhard H. Rauch, University Medicine Greifswald, Germany Haiqi He, US Department of Agriculture, USA

#### **\*Correspondence:**

Ajay Jain, State University of New York, Suite M 550 Harrison Street, Albany, Syracuse, NY 13202, USA e-mail: jaina@upstate.edu; Eduardo Davila, Bressler Research Building Room 10-041, 655 West Baltimore Street, Baltimore, MD 21201, USA e-mail: edavila@som.umaryland.edu Chronic inflammation has long been associated with the development of cancer. Among the various signaling pathways within cancer cells that can incite the expression of inflammatory molecules are those that activate IL-1 receptor-associated kinases (IRAK).The IRAK family is comprised of four family members, IRAK-1, IRAK-2, IRAK-3 (also known as IRAK-M), and IRAK-4, which play important roles in both positively and negatively regulating the expression of inflammatory molecules.The wide array of inflammatory molecules that are expressed in response to IRAK signaling within the tumor microenvironment regulate the production of factors which promote tumor growth, metastasis, immune suppression, and chemotherapy resistance. Based on published reports we propose that dysregulated activation of the IRAK signaling pathway in cancer cells contributes to disease progression by creating a highly inflammatory tumor environment. In this article, we present both theoretical arguments and reference experimental data in support of this hypothesis.

**Keywords: IRAK-4, cancer, toll-like receptors, therapeutics, inflammation**

#### **INTRODUCTION**

Interleukin-1 receptor-associated kinases (IRAK) play a central role in inflammatory responses by regulating the expression of various inflammatory genes in immune cells. These signals are critical for elimination of viruses, bacteria, and cancer cells, as well as for wound healing. Inflammation plays contradictory roles in tumor development, exhibiting both the potential to promote anti-tumor immune responses and also paradoxically to support tumor growth and metastases. What role the expression of IRAK family members in cancer cells plays in tumorigenesis and cancer progression remains relatively unknown and is the focus of this review. We also describe how these proteins may be novel therapeutic targets that can be inhibited in order to sensitize cancer cells to cytotoxic therapies.

The IRAK family is composed of IRAK-1, -2, and -4, which are expressed in a variety of human immune cell types and IRAK-M whose expression is largely limited to monocytes and macrophages (1), **Figure 1**. Greater details regarding the structures of the IRAK family proteins were extensively described in a recent review by Flannery and Bowie (1). All four IRAK family proteins contain an N-terminal death domain (DD), a ProST domain, and a centrally located kinase domain (1). With the exception of IRAK-4, all IRAK family members also contain a C-terminal domain. The DD serves as a platform that allows protein–protein interaction with other DD-containing proteins, the most important of which is the adaptor protein myeloid differentiation factor 88 (MyD88) (1, 2).

The proST domain, which contains serine, proline, and threonine residues, is important for regulating some of the IRAK family proteins. For example, in IRAK-1, auto-phosphorylation occurs several times in the ProST domain, which is located between the N-terminal DD and the kinase domain. Phosphorylation at multiple sites allows IRAK to dissociate from MyD88 while maintaining interactions with downstream proteins such as TNF receptor-associated factor 6 (TRAF-6) to initiate signaling (1, 3). Furthermore, all IRAK proteins contain an invariant lysine in subdomain II of the kinase domain. This invariant lysine is essential for ATP binding and catalytic function, and disruption of this lysine abrogates kinase activity (1, 4). IRAKs also contain a tyrosine "gatekeeper" residue (Tyr262) that alters the conformation of the IRAK protein, allowing it to maintain an active orientation. The term"gatekeeper"arises from its role in blocking a hydrophilic pocket located behind the ATP-binding site where small-molecule ATP competitive inhibitors bind and impair function (5). In a database search of over 400 kinases, this Tyr<sup>262</sup> residue was seen exclusively on IRAK family members (5). Finally, IRAK proteins can initiate downstream activation of NF-κB and JNK through engagement and activation of TRAF-6 (1, 6). Interaction with TRAF-6 occurs through Pro-X-Glu-X-X-(Ar/Ac) motifs located in the C-terminal region of IRAK1-3 (1, 6).

#### **IRAK ACTIVATION**

IL-1 receptor-associated kinase signaling can be initiated from Toll-like receptors (TLRs) or from the interleukin-1 family receptors (IL-1R), **Figure 2** (7, 8). Thirteen TLRs have been identified in human beings. TLRs recognize conserved pathogenassociated molecular patterns (PAMPs) expressed on a variety of microbes including bacteria, fungus, yeast, and viruses. Some

TLRs can also be stimulated by endogenous danger signals released from stressed or dying cells such as HMBG-1 and A100 (9, 10). A wide variety of cancers have been shown to express functional TLRs. A detailed review regarding the expression of TLRs and the consequence of ligating these receptors on tumor cells was recently published by Kaczanowska et al. (11). The IL-1Rs bind pro-inflammatory cytokines in the IL-1 family, the most wellknown of which are IL-1α, IL-1β, and IL-18. The signaling cascade is initiated by the adaptor MyD88 binding to the toll/interleukin-1 receptor (TIR) domain, which is shared by these receptors. MyD88 oligomerizes and recruits IRAK-4 via the DD. IRAK multimerization is dependent on DD interactions, which in turn result in kinase activation and propagation of the downstream signal.

Of the four IRAK proteins, IRAK-1 and IRAK-4 are active serine/threonine kinases (12). IRAK-4, the most recent IRAK family protein to be discovered, is the most proximal IRAK family protein in the TIR-mediated signaling pathway and directly downstream of MyD88 (8, 13, 14). IRAK-4 and IRAK-1 are able to associate with each other upon engaging MyD88 through their DD. IRAK-4 is thought to phosphorylate IRAK-1, which allows IRAK-1 to initiate an auto-phosphorylation cascade occurring in three sequential steps (15). IRAK-1 is first phosphorylated at Thr209, which causes a conformational change in the protein (14, 15). The second step is phosphorylation at Thr387. IRAK-1 does not become fully active until this residue is phosphorylated. There are data suggesting that either Thr<sup>209</sup> or Thr<sup>387</sup> may be sites for initial IRAK-1 phosphorylation by IRAK-4. However, this question remains unresolved as both of these residues are also sites of auto-phosphorylation. The third step is auto-phosphorylation at several residues in the proST region; this allows IRAK-1 to be released from the active receptor complex. IRAK-1 and TRAF-6 dissociate from the complex, bind TAB-1 (TAK-1 binding protein-1) followed by binding of TAK-1 (transforming growth factor-β-activated kinase) and TAB-2. IRAK-1 ubiquitination and degradation are rapidly induced. The remaining complex translocates into the cytoplasm, associates with ubiquitin ligase such as ubiquitin conjugating enzyme-13 (UBC-13) and ubiquitin conjugating enzyme E2 variant-1 (UEV-1a), leading to ubiquitination and degradation of TRAF-6. This activates TAK-1 and phosphorylation of the inhibitor of κB kinase (IKK) complex (IKKα, IKKβ, and IKKγ), as well as mitogen activated protein kinases (MAPKs). The resulting NF-κB activation regulates the transcription of pro-inflammatory genes. IRAK-1 activity and induction of NF-κB is also regulated by ubiquitination at Lys<sup>134</sup> and Lys180. It is worth noting that mutant forms of IRAK containing arginine at these sites have an impaired capacity to induce NF-κB (16).

While the IRAK-1 kinase activity is also not essential for IL-1R-mediated NF-κB activation, its role as an adaptor protein that brings together MyD88, IRAK-4, and Tollip is essential for IL-1R-mediated NF-κB activation (17–19). IRAK-1 expression and activation is, of course, subjected to regulation. In addition to inducing activation, auto-phosphorylation renders IRAK-1 susceptible to proteasome-mediated degradation (17, 19). Regulation may also occur at a transcriptional level (19). For example, a

family proteins to the receptor complex. Upon activation, IRAK members associate with TRAF6, which leads to the activation of a variety of transcription factors, including NF-κB, IRF5, AP-1, and CREB. The activation of these transcription factors results in the expression of a broad array of inflammatory molecules and apoptosis-related proteins. Moreover, TRAF6 can alter protein stability though its ability to polyubiquitinated various proteins including anti-apoptotic proteins.

human IRAK-1b splice variant that lacks kinase activity is resistant to proteasome-mediated degradation, and an IRAK-1c splice variant with a truncated sequence at the C-terminal end of the kinase domain functions as a *negative* regulator of TLR and IL-1R signaling (17, 20, 21).

IRAK-2 was initially thought to be a "pseudokinase" because a critical aspartate residue in the catalytic domain is replaced with asparagine and unlike IRAK-1 and IRAK-4, IRAK-2 cannot autophosphorylate (22–25). However, IRAK-2 possesses catalytic activity and has been implicated in maintenance of proinflammatory cytokine release induced by TLR4 and TLR9 engagement (24). Wesche et al. demonstrated that wild-type IRAK-2 can be phosphorylated when co-cultured with IRAK-1. Although it is not as good a substrate as wild-type IRAK-3, it can replace IRAK-1 when IRAK-1 is knocked down (25). However, a mutant IRAK-2 containing a substitution (K237A) in its ATP-binding pocket is not able to be phosphorylated (23, 25). Kawagoe et al. confirmed that IRAK-4, and not IRAK-1, phosphorylates IRAK-2, resulting in activation which essential for IRAK-2 kinase and effector function.

Similar to the other IRAK proteins, IRAK-3 (a.k.a. IRAK-M) can form complexes with MyD88 and TRAF-6, Like IRAK-2, it is considered to be a pseudokinase with very limited capacity for auto-phosphorylation, but with the potential to become activated by other IRAK proteins and serve as a functional kinase. In contrast to other IRAK proteins, IRAK-M is thought to function as a negative regulator that prevents the dissociation of IRAK-1 and IRAK-2 from the receptor complex, inhibiting their interaction with TRAF-6 and interrupting the downstream inflammatory cascade (26, 27).

More recent data show that IRAK-M may promote antiinflammatory effects through a paradoxical "second wave" of NF-κB activation. In this model, IRAK-M interacts with the MyD88/IRAK-4 complex to form an IRAK-M Myddosome. Upon ligation of the IL-1R, the IRAK-M Myddosome can induce a second wave of NF-κB activation and is dependent on MEKK3 signaling (26). However, this secondary NF-κB activation is believed to decrease overall inflammation by inducing the expression of several inhibitory molecules such as SOCS1, SHIP1, A20, and IκBα (20). IRAK-M can also interact with IRAK-2 in order to inhibit mRNA transcription of inflammatory cytokines and chemokines.

#### **ROLES OF THE DIFFERENT IRAK FAMILY PROTEINS IN CANCER IRAK-1**

There is an increasing body of data to suggest that IRAK-1 signaling may be important to the development and progression of cancer. *Helicobacter pylori*, bacteria strongly associated with gastric inflammation and the development of gastric cancer has been shown to cause upregulation of TLR2 and TLR5 expression in various cell types and subsequent engagement of these receptors increases IRAK-1 phosphorylation and NF-κB activation (1). Importantly, gastric carcinogenesis was recently reported to be associated with increased TLR expression and reduced expression of the TLR inhibitors Tollip and PPAR (2). As another example, an evaluation of over 300 tumor samples from non-squamous cell lung cancer (NSCLC) patients showed that tumor tissue had significantly increased cytosolic IRAK-1 expression and decreased nuclear expression relative to adjacent normal tissue (3). Our group has also found IRAK-1 and/or IRAK-4 to localize to the nucleus of melanoma cells, but not melanocytes (Geng, unpublished data). IRAK's role in the nucleus and how this contributes to tumor progression has not been defined. In order to gain a better sense of the expression levels of each IRAK family member in various cancer types, we analyzed immunohistochemistry data using the online data base ProteinAtlas (http://www.proteinatlas.org/), **Figure 3**. These data highlight the heterogeneity of different IRAK family members in different cancer types. Of all the IRAK family members, IRAK-4 was the most frequently expressed (at the medium to high range) and found on the highest percentage of tumor samples. IRAK-1 was the next most frequently expressed with appreciable levels (medium to high) in all tumor samples analyzed. IRAK-2 and IRAK-3 were the least detected IRAK family members, respectively. Despite the high-expression levels of IRAK-1 and IRAK-4,

it is important to note that the level of activation (phosphorylation) was not examined but plays an important role in IRAK signaling.

Additional evidence indicating the importance if IRAK-1 in cancer came from studies of microRNAs (miRNAs) (4). miRNAs are small non-coding RNA sequences that play critical roles in regulating cellular mRNA stability, protein expression, proliferation, apoptosis, and cancer metastasis (5, 6). It has been shown that expression of a specific miRNA (miR-146a) is frequently diminished in metastatic prostate cancers. Intriguingly, upregulation of miR-146a and miR-146b in metastatic breast cancer cell lines has been shown to downregulate TRAF-6 and IRAK-1 expression and subsequently reduce NF-κB expression (5, 28, 29). Moreover, inhibiting miR-146a expression also reduced cancer cell invasiveness of pancreatic and colon cancer cell lines. Panc-1 and Colo-1 pancreas and colon cancer cell lines, respectively, also have lower miR-146 expression in comparison to non-malignant pancreas cells, and induction of miRNA in these cancers lines decreases their invasiveness. This phenotypic change is also accompanied by down-regulation of EGFR and metastasis-associated protein 2 (MTA-2) (5).

IRAK-1 may be particularly relevant to the pathogenesis of melanoma. The use of rapid subtraction hybridization analysis was used to identify IRAK-1 as one of eight genes that are differentially expressed in metastatic cells compared to parental human melanoma cell lines, with IRAK-1 expression being upregulated in the metastatic variants (5, 30). Srivastava et al. reported that a large percentage of established human melanoma cell lines exhibit constitutive expression of phosphorylated forms of IRAK-1 and IRAK-4 (31). Patient-derived melanoma tumor samples also exhibited increased expression of phosphorylated IRAK-4 although there did not appear to be a correlation between p-IRAK levels and melanoma stage. Inhibition of IRAK-1 and IRAK-4, using pharmacological inhibitor or siRNA, sensitized melanoma tumors expressing phosphorylated forms of these IRAKs to cytotoxic chemotherapies*in vivo*, raising the possibility that IRAK family proteins may be potential therapeutic targets in cancer. In agreement with these studies, recent data indicate that inhibiting IRAK-1,-4 signaling in a variety of leukemias including Waldenstrom macroglobulinemia, diffuse large B-cell lymphoma, myelodysplasia, and acute myeloid leukemia substantially impaired proliferation *in vitro* and *in vivo*, and treatment with IRAK inhibitors prolonged mouse survival (32, 33). We recently found that IRAK-4 signaling in T cell acute lymphoblastic leukemia (T-ALL) is critical for their ability to proliferate but did not induce cell death (Li, unpublished data). In order to determine whether IRAK inhibitors could enhance the cytotoxic effects of chemotherapeutic agents, we screened nearly 500 FDA-approved drugs for their ability to

kill T-ALL cells when combined with IRAK inhibitors. We identified three classes of drugs that worked synergistically with IRAK inhibitors and,in some cases,restored sensitivity of chemoresistant samples. Whether a similar effect will be observed in other cancer types merits further investigation. This is especially true given that many cancers exhibit increased protein levels of IRAK-1 and IRAK-4 and are resistant to chemotherapy (**Figure 3**).

Finally, IRAK-1 activation may also be important for cross talk between cancer cells and other cell populations present in the tumor microenvironment. IL-1β release by lingual squamous cell carcinomas causes upregulation of the IL-1R and increased levels of p-IRAK-1 in cancer associated fibroblasts. This results in nuclear translocation of NF-κB and induction of genes important for tumor progression including IL-6, Cox-2, BDNF, and IRF-1 (34).

#### **IRAK-2**

In terms of signaling and function, there is some redundancy between IRAK-2 and IRAK-1. Using single and double IRAK knockout mice, Kawagoe and colleagues confirmed that both IRAK1 and IRAK2 have common functionality in the early phase of TLR signaling (23). IRAK2 kinase activity, however, was longer sustained than that of IRAK-1, and IRAK-2 was critical in latephase TLR responses. This raises the possibility that IRAK-2 may be relevant to chronic inflammatory responses often associated with cancer. Whether downstream signaling differs between IRAK-1 and IRAK-2 remain to be determined. Recent studies by Cui and colleagues suggest that a stress-induced NF-κB-activated, miRNA-146a-mediated down-regulation of IRAK-1 coupled to an NFκB-driven upregulation of IRAK-2 supports a self-perpetuating inflammatory signaling loop (35).

The role of IRAK-2 as a regulator of TLR signaling may be more complex than originally thought. IRAK-2 is known to induce NF-κB activation through TLR3, TLR4, and TLR8 (14). Of note, IRAK-2 is the only member of the family thought to mediate signaling through TLR3. Interestingly, IRAK-2 has recently been shown to have a dual function (immunosuppressive and immunostimulatory) in TLR9 related signaling and inflammatory responses. Wan and colleagues demonstrated that IRAK-2 suppresses TLR9 signaling in the early post-stimulation phase, raising the activation threshold for TLR9-induced inflammatory response and potentially preventing autoimmunity (36). However, if the higher activation threshold is successfully triggered through a strong stimulus, IRAK-2 mediates a positive feedback loop allowing for sustained release of pro-inflammatory cytokines. It is conceivable that loss of negative regulatory function could allow sustained IRAK-2 activation and inflammation, thus, promoting carcinogenesis. Importantly, whereas TLR9 was previously thought to be expressed only on immune cells, it has been shown that it also expressed on a number of different cancers (oral, prostate, breast, lung, Burkitt lymphoma), and signaling through TLR9 promotes proliferation and/or cell survival (37–44).

#### **IRAK-3 (a.k.a. IRAK-M)**

Unlike other IRAK family members that are widely expressed on a variety of cell types, IRAK-M is thought to chiefly reside in monocyte and macrophage populations. As mentioned previously, IRAK-M activation generally acts as a negative regulator of NF-κB activation in TLR and IL-1R signaling (45). Also, even though IRAK-M induces a paradoxical "second wave" of MEKK3 dependent NF-κB activation, the overall effect of IRAK-M favors immunosuppression (26).

IRAK-M is a negative regulator of IRAK-4/IRAK-1 and IRAK-4/IRAK-2 and thus serves to inhibit the expression of a variety of inflammatory molecules induced by IRAK-4. Our working hypothesis is that in cancers with reduced levels of IRAK-M but elevated levels of IRAK-1, -2, and/or -4 will show increased IRAK-4 signaling and consequently elevated levels of inflammatory molecules. In addition to augmenting the amounts of inflammatory factors, the lack of IRAK-M might further sustain IRAK-4 signaling and perpetuate a chronically inflamed tumor environment; chronic inflammation is a hallmark of tumorigenesis and tumor progression (46). That IRAK-3 expression levels are reduced in some cancer types is further highlighted in **Figure 3** and supports our hypothesis.

Even though it is an anti-inflammatory mediator, IRAK-M may still play an important role in tumorigenesis through modulation of the activity of tumor-associated macrophages (TAMs). It is generally thought that there are two types of macrophages associated with cancer (47). These include classically activated (M1) macrophages that secrete pro-inflammatory cytokines and present antigens to cytotoxic immune effector cells, and alternatively activated (M2) macrophages with impaired Th1-like cytokine release (and one favoring Th2 cytokines) and decreased capacity to activate T cells. The M1 type is thought to play a more prominent role in the early stages of carcinogenesis through NF-κB activation and chronic inflammation to initiate carcinogenesis. As cancers become more established, M1 macrophages may become "re-educated" to take on a M2 phenotype. M2 macrophages can secrete tumor growth factors, promote angiogenesis and invasiveness through remodeling of the tumor matrix, and induce immune tolerance. The term "tumor-associated macrophage" or TAM is typically associated with the M2 phenotype. Indeed, macrophage re-education may be a critical aspect of cancer pathogenesis, and IRAK-M may play a significant role in this process.

IRAK-M may promote cancer progression through modulation of macrophage activity. IRAK-M is known to be an important negative regulator in macrophages in models of inflammation. For example, in mouse models of myocardial infarction, upregulation of IRAK-M in cardiac macrophages reduces myocardial inflammation and prevents adverse cardiac remodeling (45). Naïve monocytes and macrophages exposed to tumor cell lines exhibited decreased expression of TNFα, IL-12p40, and IRAK-1 (48, 49). Moreover, these characteristics, as well as the ability to present antigens, were diminished with prolonged exposure to tumor cells as the macrophages take on an M2 phenotype. A hallmark feature of this transition is the rapid upregulation of IRAK-M in macrophages upon exposure to tumor cells (48, 49). *In vivo* mouse studies using Lewis lung cancer (LLC) cell lines have shown that tumor infiltrating macrophages have higher IRAK-M expression and impaired ability to secrete IL-12, TNFα, and IFN-γ compared to peritoneal macrophages isolated from the same mouse (50). Interestingly, the ability of TAMs to secrete TNFα could be restored by knocking down IRAK-M expression using siRNA (48). These data indicate that IRAK-M upregulation can be induced by surface-associated or soluble factors from tumor cells to promote tumor growth and immune evasion. Proposed mechanisms include the engagement of hylauronan (a tumor cell surface glycosaminoglycan) to monocyte-expressed CD44 or secretion of TGF-β. Furthermore, monocytes isolated from patients with chronic myleogeneous show upregulation of IRAK-M mRNA, monocytes from chronic lymphocytic leukemia patients (in whom IRAK-M expression was not evaluated) showed impaired ability to secrete cytokines and present antigen. Analysis of a cohort of 439 lung cancer patients showed that the level of IRAK-M expression on tumor cells was a significant and independent predictor of mortality. In contrast, these data suggest that IRAK-M is a critical mediator of cross talk that occurs between tumor cells and macrophages to allow a more favorable tumor microenvironment and facilitate cancer progression (48, 49).

#### **IRAK-4**

IRAK-4, the most recently identified member of the family, is considered the"master IRAK"because it is required for all MyD88 dependent NF-κB activation and for inducing IFNα expression through TLR 7, 8, and 9 (51). Loss of IRAK-4 renders mice completely resistant to LPS-induced shock, and deficiencies in human beings have been associated with increased susceptibility to encapsulated bacterial infections (especially pneumococcal) (52, 53). Data regarding the specific role of IRAK-4 in cancer have not been fully investigated, and its potential role in cancer progression is just now beginning to emerge. As previously discussed (in the Section IRAK-1) some melanomas constitutively express active, phosphorylated forms of IRAK-1 and IRAK-4. Inhibiting IRAK-4 rather than IRAK-1 using shRNA was more effective at sensitizing melanoma tumors and T-ALL cells to chemotherapies. It is still unclear, however, whether this is a direct phenomenon or whether upstream signaling events drive phosphorylation. As IRAK-4 is a lynchpin for MyD88-mediated pro-inflammatory signaling, it can promote carcinogenesis regardless of whether it is directly mutated or not. For example, a subset (29%) of activated B-cell type diffuse large B-cell lymphomas (ABC DLBCL) with a very aggressive phenotype were recently found to carry an oncogenic MyD88 mutation (L265P) that promotes survival. This mutation allowed spontaneous formation of a stable complex between MyD88, IRAK-4, and a phosphorylated form of IRAK-1. However, knockdown of IRAK-1 kinase activity was not required for survival of ABC DLBCLs, while IRAK-4 kinase activity was essential (54). To date, no group has reported any mutations in any of the IRAK family members specifically in cancer but this subject merits further investigation considering recent data uncovering an important role for dysregulated IRAK signaling via MyD88 mutations.

#### **IRAK FAMILY PROTEIN INHIBITORS AS NOVEL CANCER THERAPEUTICS**

#### **SMALL-MOLECULE INHIBITORS**

Given the strong data indicating that IRAK family proteins are critical mediators of inflammation, there has been considerable interest in developing targeted agents to treat autoimmune and inflammatory diseases. As we previously addressed, IRAK inhibitors (especially IRAK-1 and -4) may also have therapeutic applications in cancer. Several classes of IRAK-4 inhibitors have been developed, including amino-benzimidazole, thiazole, or pyridine amides, imidazo[1,2-*a*] pyridines, imidazo[1,2 *b*]pyridazines, and benzimidazole–indazoles (47–50, 52, 54). IRAK inhibitors may have particular utility in the treatment of Waldenstrom's macroglobulinemia, a B-cell lymphoproliferative disorder that is critically dependent upon NF-κB activation. *However, compounds that target molecules downstream of IRAK-1 are also potential candidates*. One such compound is 5Z-7-oxozeaenol, which selectively inhibits TAK-1 and has been shown to reduce inflammation and enhance the sensitivity of breast and pancreatic cancer cells to various chemotherapeutic agents, further highlighting the central role that IRAK signaling plays in chemotherapy resistance (54–56).

#### **BOTANICAL DERIVATIVES**

It is possible that plant-derived compounds may also induce antiinflammatory and anti-cancer therapeutic effects through inhibition of IRAK family members. For example, ginseng (*Panax ginseng* ), which is anecdotally described to have a many health benefits including anti-inflammatory and anti-cancer properties, contains protopanaxatriol ginsenoside. This agent has been shown to inhibit IRAK-1 and IKK-β phosphorylation in LPS stimulated macrophages, as well as alleviate inflammation induced by 2,4,6 trinitrobenzene sulfonic acid-induced colitis in mice (54, 56–59). The xanthone derivative 1,3,5-trihydroxy-4-prenylxanthone (TH-4-PX) isolated from *Cudrania cochinchinensis*, a plant used as a traditional remedy for diseases in Asia, inhibits LPS/TLR-mediated release of nitrous oxide through inhibition of IRAK-1 (60). A second agent from this plant (isoalvaxanthone) has anti-neoplastic properties, as it can inhibit matrix metalloproteinase-2 expression (a factor associated with tumor invasiveness) *in vitro* in SW620 colon cancer cells. Admittedly, it is unclear if the isoalvaxanthone effects are the result of IRAK family member inhibition, as this agent did not inhibit expression of NF-κB.

#### **NITROGEN BISPHOSPHONATES**

There has been increasing evidence that nitrogen bisphosphonates (NPBs), a class of drugs used to treat osteoporosis, may also have potential for treating cancer. Paradoxically, NPBs are associated with inhibition of IRAK-M expression. The NBP xoledronate reduces IRAK-M levels when cultured with PBMCs from a subset of human blood donors (50%). In these individuals, the reduction in IRAK-M is associated with enhanced cytokine release after TLR stimulation or administration of IL-1 (61). Depletion of IRAK-M in dendritic cells (DCs) using siRNA has been shown to enhance DC migration to lymph nodes, augment cytokine release, and enhance antigen presentation, proliferation, and activation of antigen-specific T cells. Thus, pharmacologic inhibition of IRAK-M using NBPs may likewise improve the induction of cell-based anti-tumor immune responses. A summary of the various IRAK inhibitors is shown in **Table 1**.

#### **SUMMARY**

Dysregulated IRAK signaling in tumors is beginning to emerge as an important factor in cancer initiation, tumor progression, and therapy resistance. Studies from several groups highlight the

#### **Table 1 | A summary of small molecules that can inhibit IRAK family members**.


potential of IRAK family members as therapeutic targets for cancer treatment alone or when combined with other therapies. A better understanding of how IRAK signaling drives inflammation through interaction with TLR and IL-1 family members will be critical for developing targeted therapies that work synergistically with systemic chemotherapies. Furthermore, such an understanding may allow manipulation of these proteins to favor anti-tumor cytotoxicity rather than carcinogenic downstream effects.

#### **REFERENCES**


in human melanoma cells by rapid subtraction hybridization. *Gene* (2004) **343**(1):191–201. doi:10.1016/j.gene.2004.09.002


**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.

*Received: 13 June 2014; paper pending published: 31 July 2014; accepted: 17 October 2014; published online: 17 November 2014.*

*Citation: Jain A, Kaczanowska S and Davila E (2014) IL-1 receptor-associated kinase signaling and its role in inflammation, cancer progression, and therapy resistance. Front. Immunol. 5:553. doi: 10.3389/fimmu.2014.00553*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Jain, Kaczanowska and Davila. 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.*

## NOD-like receptors: master regulators of inflammation and cancer

#### **Mansi Saxena<sup>1</sup> and GarabetYeretssian1,2\***

<sup>1</sup> Department of Medicine, Immunology Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA <sup>2</sup> Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Dario S. Zamboni, Universidade de São Paulo, Brazil Maya Saleh, McGill University, Canada Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Garabet Yeretssian, Department of Medicine, Clinical Immunology, Immunology Institute, Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, 1425 Madison Avenue, 12-20E, New York, NY 10029, USA e-mail: garabet.yeretssian@ mssm.edu

Cytosolic NOD-like receptors (NLRs) have been associated with human diseases including infections, cancer, and autoimmune and inflammatory disorders.These innate immune pattern recognition molecules are essential for controlling inflammatory mechanisms through induction of cytokines, chemokines, and anti-microbial genes. Upon activation, some NLRs form multi-protein complexes called inflammasomes, while others orchestrate caspaseindependent nuclear factor kappa B (NF-κB) and mitogen activated protein kinase (MAPK) signaling. Moreover, NLRs and their downstream signaling components engage in an intricate crosstalk with cell death and autophagy pathways, both critical processes for cancer development. Recently, increasing evidence has extended the concept that chronic inflammation caused by abberant NLR signaling is a powerful driver of carcinogenesis, where it abets genetic mutations, tumor growth, and progression. In this review, we explore the rapidly expanding area of research regarding the expression and functions of NLRs in different types of cancers. Furthermore, we particularly focus on how maintaining tissue homeostasis and regulating tissue repair may provide a logical platform for understanding the liaisons between the NLR-driven inflammatory responses and cancer. Finally, we outline novel therapeutic approaches that target NLR signaling and speculate how these could be developed as potential pharmaceutical alternatives for cancer treatment.

**Keywords: apoptosis, autophagy, colorectal cancer, innate immunity, intestinal inflammation, inflammasome, nod-like receptors, nodosome**

#### **INTRODUCTION**

Over the past two decades, immunologists have begun to appreciate the complexity of the innate immune system, its importance as the first wave of defensive action against perceived harmful microbes or foreign particles and its functions in triggering antigen-specific responses by engaging the adaptive immune system. Innate immune responses are orchestrated by germlineencoded pattern recognition receptors (PRRs) (1). PRRs recognize conserved pathogen-derived and damaged self-derived molecular components, commonly referred to as pathogen associated molecular patterns (PAMPs) and danger associated molecular patterns (DAMPs), respectively (2, 3). PRR superfamilies are broadly classified based upon structural homology and the requirement of different adaptor proteins that ensure their function and downstream signal transduction (4). The PRRs include members of the Toll-like receptors (TLRs) (3), nucleotide-binding, and oligomerization domain containing receptors [NOD-like receptors (NLRs)] (5, 6), retinoic acid-inducible gene (RIG) I-like RNA helicases (7), C-type lectins (8), and AIM2 like receptors (ALRs) (9). Evidence in the field points to a paramount importance of NLRs in human diseases with increasing interest in translating this knowledge toward clinical benefits. Due to the active role of NLRs in regulating pro-inflammatory signals and recruiting the adaptive arm of the immune system, dysregulation of microbial sensing has been reported to influence disease outcomes and tumorigenesis (10). In this review, we will describe the crucial roles of NLRs in cancer development and progression, and discuss the possibility of NLRs as targets for tumor therapy.

#### **FACTORS THAT INFLUENCE TUMORIGENESIS**

Observations by Rudolf Virchow in the nineteenth century indicated a link between inflammation and cancer, and suggested that immune and inflammatory cells are frequently present within tumors. Indeed, chronic inflammation plays critical roles in various stages of cancer development and progression (11–13). Many cancer risk factors are associated with a source of inflammation or act through inflammatory mechanisms such as those evoked by bacterial and viral infections (14), tobacco smoke (15), obesity (16, 17), and aging or cell senescence (18, 19). While some cancers arise from chronic inflammation or after immune deregulation and autoimmunity, solid malignancies elicit intrinsic immune mechanisms that guide the construction of a tumorigenic microenvironment (12, 13, 20). Although the exact mechanism of how inflammation leads to neoplastic transformation is not fully known, it is suggested that inflammatory immune cells like macrophages and T cells are the main orchestrators of inflammation-mediated tumor progression. These cells secrete cytokines and chemokines that cause DNA damage, generate mutagenic reactive oxygen species (ROS), and supply cancer cells with growth factors (13). In addition, inflammatory mechanisms

were shown to promote genetic instability by impairing DNA repair mechanisms, altering cell cycle checkpoints, and often facilitating epigenetic silencing of anti-tumor genes, thus contributing to the high degree of genetic heterogeneity in tumors (21). Oncogenic mutations prompted by an inflammatory microenvironmentfrequently cause neoplastic transformation by promoting excessive proliferation and resistance to cell death (22). Indeed, impaired expression and activity of proteins that control cell survival, such as the inhibitor of apoptosis proteins (IAPs) and the BCL2 family of proteins, is a common occurrence in many cancers (23, 24). Typically known to exert strong anti-apoptotic functions, IAPs neutralize pro-apoptotic second mitochondrial activator of caspases (SMAC) and inhibit activation of apoptotic caspases, thereby promoting cell survival during both physiological stresses and pathogenic stimulations (25–29). Owing to their strong prosurvival potency, enhanced expression of IAPs has been correlated with several human cancers (22). Unlike IAPs, the BCL2 family of proteins consists of both pro- and anti-apoptotic proteins that control critical checkpoints of intrinsic apoptosis by regulating mitochondrial integrity and release of cytochrome *c* into the cytosol (30). Deregulation of the functions of BCL2 proteins, i.e., down-regulation of pro-apoptotic members and over expression of pro-survival members, has been strongly correlated with tumorigenesis and resistance to chemotherapy (31). Interestingly, the pro-apoptotic BID, PUMA, and NOXA are transcriptional targets of the tumor suppressor gene p53 and loss of their expression enhances tumorigenesis and morbidity of MYC overexpressing transgenic mice (32, 33). It was described that the transcription factor p53 senses physiological stresses and is critical for restraining tumor growth. Indeed, loss of p53 expression or function in both humans and mice has been proven to promote sporadic tumorigenesis (34, 35). Induction of target genes that inhibit cancer progression is generally considered to be the canonical mechanism of p53-mediated tumor-suppression. These target genes directly modulate cellular programs involving induction of apoptosis, cell cycle arrest, and promotion of cellular senescence and DNA repair (36). Recently, non-canonical functions of p53 have come to light, like the regulation of cellular metabolism, cell-to-cell communication, autophagy, tumor invasion, and metastasis, making p53 an attractive pharmaceutical target for treating cancers [reviewed in Ref. (37)]. Early detection of rogue tumor cells by the innate immune cells and their rapid removal is a key host defense strategy for evading tumorigenesis. In particular, natural killer (NK) cells are primary sentinels that guarantee such immune surveillance by differentiating normal cells from stressed or tumor cells via the expression of specific NK receptors (38). Indeed, increased presence of NK cells at tumor sites has been reported to improve remission, whereas decreased NK cell antitumor activity has been correlated with a greater likelihood for developing cancer (39).

#### **NOD-LIKE RECEPTORS IN CANCER**

#### **OVERVIEW OF NLRs**

NOD-like receptors are a relatively recent addition to the PRR superfamily (40–42). All NLRs contain a central NACHT domain that facilitates oligomerization, and bear multiple leucine-rich repeats (LRRs) on their C-terminal for ligand sensing (5, 43). The 22 human NLRs can be distinguished into five subfamilies by their N-terminal effector domains that bestow unique functional characteristics to each NLR (43) (**Figure 1**). NLRs with an N-terminal acidic transactivation domain are termed NLRA (CIITA) and serve as transcriptional regulators of MHC class II antigen presentation (44). NLRB (NAIP) proteins have an N-terminal baculoviral inhibition of apoptosis repeat (BIR) domain and are largely recognized for their roles in host defense and cell survival. For instance, NAIP5 is known to induce host defense against bacterial infections by curtailing macrophage permissiveness to *Legionella pneumophila*, the causative agent of the Legionnaires' disease (45–47). N-terminal caspase activation and recruitment domain (CARD) distinguishes the NLRC subfamily (NLRC 1–5) and allows direct interaction between members of this family and other CARD carrying adaptor proteins. NOD1 (NLRC1) and NOD2 (NLRC2), the founding members of the NLRs, are key sensors of bacterial peptidoglycan (PGN) and are crucial for tissue homeostasis and host defense against bacterial pathogens (48). Notably, single-nucleotide polymorphisms (SNPs) in the *NOD2* (*CARD15*) gene are among the most significant genetic riskfactors associated with Crohn's disease (CD) susceptibility (49, 50), hence the rising interest in unraveling the functions of NOD1 and NOD2 receptors in microbial sensing, intestinal homeostasis, and disease. Members of the pyrin domain (PYD) containing NLRP subfamily (NLRP 1–14) are best known for their role in inducing the formation of the oligomeric inflammatory complex "Inflammasome" (51). NLRX1, the only described member of the NLRX subfamily contains an N-terminal mitochondria-targeting sequence required for its trafficking to the mitochondrial membrane (**Figure 1**). Mechanistically, NLRX1 was shown to down-regulate mitochondrial anti-viral signaling protein (MAVS)-mediated type I interferon (IFN) production (52), interfere with the TLR-TRAF6-NF-κB pathways (53, 54), and enhance virus induced-autophagy (55, 56). On the other hand, NLRX1 was implicated in the generation of ROS induced by TNFα and Shigella infection magnifying the JNK and NF-κB signaling (57). Interestingly,NLRX1-mediated ROS generation was involved in promoting *Chlamydia trachomatis* replication in epithelial cells (58). However, recent data from Soares et al. revealed that bone marrow macrophages (BMMs) and mouse embryonic fibroblasts (MEFs) fromWild type (WT) or *Nlrx1*−/<sup>−</sup> mice respond equally to *in vitro* infection with Sendai virus or following *in vivo* challenge with influenza A virus and TLR3 ligand Poly I:C (59). Additionally, Rebsamen et al. reported no significant contribution of NLRX1 in RLR–MAVS signaling both *in vitro* and *in vivo* (60). Overall, the precise role of NLRX1 remains controversial and further research is required to validate its pro or anti-inflammatory properties.

Dysregulated apoptosis and autophagy pathways, as well as excessive chronic inflammation are major drivers of carcinogenesis. NLRs are innate immune sensors that actively communicate with a myriad of cell death regulators. Hence, these PRRs are well-positioned to influence tumor development and progression particularly at sites with high host-microbiome interactions like the gut. One of the mysteries of the innate immune system is how do NLRs sense molecular patterns from both commensal and pathogenic microorganisms and manage to tolerate one while help eradicate the other (5, 61). This disparity in NLR functions is particularly useful in the intestinal epithelia where host cells are in

Domain architecture of human NLRs is depicted here. Human NLRs are sub-classified into five categories: NLRA, NLRB, NLRC, NLRP, and NLRX. All 22 human NLRs contain a central NACHT domain and a C-terminal ligand sensing LRR domain, with the exception of NLRP10. The N-terminal domains ascribe functional properties to the NLRs; however, the function of some of the domains is still unclear like for the N-terminal domain of NLRC3 and NLRC5, as well as the C-terminal FIIND in NLRP1. CARD; caspase association and recruitment domain, ATD; acidic transactivation domain, FIIND; function to find domain, PYD; pyrin domain, BIR; Baculoviral inhibition of apoptosis protein repeat domain, LRR; leucine-rich repeats, MT; targets NLRX1 to the mitochondria but no sequence homology with traditional mitochondrial targeting sequence has been reported.

constant contact with millions of microbes. Consequently, it came as little surprise when common variants in the NLR genes were correlated with the incidence of CD and susceptibility to cancers (50, 62–64). Due to these correlations, most of the studies have been focused on understanding the mechanisms by which NODs and inflammasome NLRs regulate intestinal inflammation and tumorigenesis.

#### **NOD1 AND NOD2 IN CANCER**

#### **NOD-DEPENDENT SIGNALING CASCADES**

NOD1 and NOD2 are cytosolic proteins that sense intracellular bacterial PGN and trigger signal transduction via NF-κB and MAPK activation. NOD1 is expressed in both hematopoietic and non-hematopoietic cells and responds to intracellular gamma-d-glutamyl-meso-diaminopimelic acid (iE-DAP) mostly present on Gram-negative bacteria and only on some select Grampositive bacteria, like *Listeria* and *Bacillus* species (65–67). Unlike NOD1, NOD2 expression is largely restricted to hematopoietic cells and certain specialized epithelial cells such as the small intestinal Paneth cells (68). NOD2 recognizes cytosolic muramyl dipeptide (MDP) found in the PGN of all bacteria (69). Besides

providing immunity against intracellular bacteria, NODs were revealed to be critical for host defense against non-invasive Gramnegative bacteria like *Helicobacter pylori,* following delivery of its PGN into the host cells through the bacterial type IV secretion system (70). Moreover, NOD1 and NOD2 ligands were also described to gain access to the cytosol by endocytosis with the help of transporter proteins like SLC15A3 and SLC15A4 (71– 73). Notably, NOD1 and NOD2 have been reported to localize to the plasma membrane at the sites of infection; however, the biological relevance of this translocation remains elusive (74, 75). Interestingly, a recent report accentuated the importance of NOD proteins in monitoring the activation state of small Rho GTPases (e.g., RAC1, CDC42, and RHOA) and inducing unusual immune responses in the host in response to bacterial infection (76). Upon activation by their cognate ligands both NOD1 and NOD2 self-oligomerize, undergo a conformational change, and through homotypic CARD–CARD interactions allow the recruitment of the CARD containing adaptor Receptor-interacting protein kinase 2 (RIP2 or RIPK2) (41, 42, 77, 78) (**Figure 2**). This event facilitates the formation of a multi-protein signaling complex termed "Nodosome," which leads to downstream NF-κB and MAPKmediated inflammatory and anti-microbial output. Indeed, cells or mice lacking RIP2 do not respond to NOD agonists and fail to produce pro-inflammatory and anti-microbial molecules (78–80). Initially, it was thought that NOD oligomerization initiated RIP2 aggregation and activation by"induced proximity" (81).While this model still stands true, over the years new body of research has contributed a wealth of data regarding specific sequence of events that leads to RIP2 activation. In contrast to the earlier studies (82– 85), recent *in vitro* data using pharmacological inhibitors as well as *in vivo* evidence using a knock-in mouse with kinase-dead RIP2 (K47A) have highlighted the key role of the kinase activity of RIP2 in NOD-mediated immune responses (86, 87).

Lately, it was described that the pathways activated downstream of NOD proteins are closely related to those activated by death receptors, notably TNF receptor 1 (TNFR1). For instance, hierarchical recruitment of selective TNFR-associated factors (TRAF2, TRAF5, or TRAF6) facilitates Lys63 poly-ubiquitination and activation of RIP2 (88–90). Activated RIP2 facilitates ubiquitination of NEMO (also called IKKγ) leading to the recruitment of tumor growth factor β-activated kinase 1 (TAK1) and TAK1 binding proteins (TAB) 1, TAB2, or TAB3 (91, 92). Following this complex formation, IKKs (IKKα and IKKβ) get phosphorylated eventually driving the phosphorylation and degradation of IκBα and subsequent transcription of NF-κB target genes (5, 89, 92) (**Figure 2**). RIP2 activation also constitutes a key event that links the NOD– RIP2 cascade with the p38, extracellular signal-regulated kinase (ERK), and c-Jun N-terminal kinase (JNK) MAPK pathways (93).

In addition to TRAFs, members of the IAP family including X-linked IAP (XIAP) and cellular IAP1 (cIAP1) and cIAP2 were described to physically interact with RIP2 and facilitate NODmediated immunity (94–98). Both *in vitro* and *in vivo* studies suggest a strong role for cIAP1 and cIAP2 in promoting NOD signaling (**Figure 2**); however, the mechanism for such positive regulation is still not fully understood (94, 99–101). Similarly, XIAP was reported to recruit a linear ubiquitin chain assembly complex (LUBAC) for RIP2 ubiquitination and this step was proven

critical for downstream NF-κB regulation (96, 97). Upon microbial sensing another E3 ubiquitin ligase, ITCH, also ubiquitinates RIP2, and it is speculated that ITCH-mediated ubiquitination acts like a molecular switch dictating the fate of the signaling circuit to NF-κB or p38 and JNK activation (102). Pathogen-mediated NOD1 activation has also been shown to elicit protective immune responses via RIP2-TRAF3-IRF7-mediated transcription of IFNβ (79). Overall,it is tempting to speculate that similar to pro-survival association of RIP1 with cIAP1 and cIAP2 (103), interactions between RIP2 and the IAPs may also lead to modulation of cellular apoptosis. However, neither NODs nor RIP2 has been demonstrated to exploit these associations to affect cell survival. Similarly, several studies have alluded to NODs as being regulators of caspase-mediated apoptosis (82, 104, 105); yet, no direct link has so far been reported. Recently, the pro-apoptotic BH3 only BCL2 family protein BID (BH3 interacting-domain death agonist) was identified in a genome wide siRNA screen as a positive regulator of NOD signaling (101). BID was demonstrated to bind to RIP2 bridging both NOD and IKK complexes to specifically transduce NF-κB and ERK signaling events (101). Notably, BID was phosphorylated upon activation with NOD agonists and these innate immune functions of BID were found to be independent of its pro-apoptotic processing by caspase-8 (101). The discovery involving a classical pro-apoptotic protein, such as BID, in NOD– RIP2 signaling strengthens the concept that inflammatory and cell death pathways do not function as discrete mechanisms but share common adaptors. Such adaptors can exert multiple functions depending upon the nature of the stimuli (5, 106–108) (**Figure 2**). One recent study have reported that BID-deficient mice exhibit a normal NOD-mediated immunity (109), suggesting that further investigations are still needed to clearly decipher the implication of BID in NOD signaling.

Similar to *NOD2*, a SNP encoding a missense variant in the autophagy gene *ATG16L1* was strongly associated with the incidence of CD, raising a possible common role of both genes in host defense mechanisms (110, 111). Intriguingly, it has been described that NOD1 and NOD2 stimulation enhances autophagy, either directly by interacting with ATG16L1 (112) or indirectly (112– 115). Conversely, pharmacological inhibition of both early and late autophagy has been proven to down-regulate MDP-mediated NF-κB and MAPK activation, suggesting that autophagocytic trafficking of MDP may be required for efficient NOD2 signaling (114). Surprisingly, ATG16L1 was recently shown to negatively regulate NOD1- and NOD2-mediated inflammatory signaling by interfering with RIP2 ubiquitination and recruitment to the Nodosome (116) (**Figure 2**). Taken together, this information suggests that different NLRs can have opposing regulatory effects on autophagy and cell death, yet the molecular triggers that dictate these actions are not fully understood.

#### **NOD PROTEINS AND CANCER**

Three mutations within the LRR region of the *NOD2* gene have been associated with increased CD susceptibility. Interestingly, these same mutations have also been found to directly interfere with NOD2's bacterial sensing faculties and downstream NF-κB activation (49, 50). Notably, such inactivation of NOD2 immunity has been indicated to enhance the risk of bacterial infections following chemotherapy in patients with acute myeloid leukemia (117). In addition, *NOD2* polymorphisms have been correlated with modifications in gastric mucosa and increased risk for *H. pylori* induced gastric cancer (118). Apart from intestinal disorders, mutations in NOD2 have been linked with increased prevalence of early onset breast (119) and lung cancers (120, 121). However, how NOD2 contributes to the initiation and the progression of cancer remains ill defined. Although no mutations in the *NOD1* gene have been so far associated with the incidence of intestinal inflammation or even colorectal cancer (CRC), murine models clearly designate a central antitumorigenic function for NOD1 in the pathophysiology of disease. For instance, *Nod1*−/<sup>−</sup> mice have been described to be susceptible to dextran sulfate sodium (DSS), a sulfated polysaccharide highly toxic to enterocytes (122). Upon combination of a single hit of the carcinogen, azoxymethane (AOM), with DSS (123), NOD1-deficient mice were found to develop significantly more and larger colonic tumors as compared to WT mice (122). This experimental CRC model is particularly applicable when the focus is on understanding colitis-driven tumor initiation and progression. The *ApcMin/*<sup>+</sup> mouse is a *N*-Ethyl-*N*-Nitrosourea (ENU) mutant model carrying the multiple intestinal neoplasia (Min/+) mutation and recapitulates many aspects of human hereditary or sporadic CRCs with mutations in the adenomatous polyposis coli (Apc) gene (124–127). Intriguingly, it has been reported that treatment with low doses of DSS leads to increased colonic tumors in

*ApcMin/*+*Nod1*−/<sup>−</sup> mice suggesting that NOD1 serves as a negative regulator of the tumor-promoting Wnt/β-catenin cascade (128, 129). Further analysis revealed that absence of NOD1 exacerbated NF-κB-mediated inflammation early during colitis causing gut barrier damage and prompted a second wave of microbiota driven inflammation and intestinal epithelial cell (IEC) proliferation, thus initiating tumor development. These conclusions are supported by the observation that antibiotic treatment of *Nod1*−/<sup>−</sup> mice ameliorated DSS-induced CRC (122). While most investigations have been focused on the role of NOD1 in models of intestinal tumorigenesis, one report provided experimental evidence for the protective role of NOD1 in breast cancer (104). Herein, it was shown that NOD1-deficient MCF-7 breast cancer cells were resistant to iE-DAP and cycloheximide mediated cell death. Interestingly, SCID mice grafted with NOD1 overexpressing cells exhibited rapid tumor regression. In sharp contrast, mice grafted with NOD1-deficient MCF-7 cells displayed increased and continued tumor growth (104).

Like *Nod1*−/<sup>−</sup> mice,NOD2-deficient mice have been revealed to be highly susceptible to DSS-induced colitis by inheritance of dysbiotic microbiota that markedly sensitizes mice to injury (130). Furthermore, *Nod2*−/<sup>−</sup> mice have been found to display worse disease outcome with increased epithelial dysplasia, heightened tumor burden, and elevated expression of the pro-inflammatory cytokine IL-6 when subjected to AOM–DSS treatment. This transmissible phenotype was significantly ameliorated upon treatment with broad-spectrum antibiotics or using the neutralizing IL-6 receptor antibody (130). Altogether, these findings reinforce the idea that aberrant NOD signaling gives rise to dysbiosis that in an inflammatory setting ultimately causes mucosal injury and drives CRC. So far, the translational value of this knowledge is limited but with the recent technological advances in the microbiome research it is predicted that modulation of dysbiosis could be used as a therapeutic strategy for patients with CD as well as CRC.

Contrary to the protective role for NODs in intestinal tumorigenesis, increased expression of both NOD1 and NOD2 has been reported in the head and neck squamous cell carcinoma biopsies as compared to the healthy nasal biopsies. These findings implicate NODs in enhancing head and neck cancers; however, thus far no corroborating experimental evidence has been reported (131). Furthermore, iE-DAP stimulation of human pharyngeal squamous carcinoma cell lines (Detroit 562 and Fadu) has been determined to augment the production of β-defensins, which can serve as chemoattractants, thus fostering an inflammatory and pro-tumorigenic environment (131).

#### **INFLAMMASOME NLRs IN CANCER**

#### **INFLAMMASOME NLRs: NLRP3-MEDIATED SIGNALING CASCADES**

While NOD1 and NOD2 form the Nodosome, other NLRs assemble macromolecular inflammasome complexes. To date, various inflammasome platforms have been described (132), but the NLRP3 inflammasome is the most commonly studied. The reason behind this could be the initial discovery of mutations in the *NLRP3* (*CIAS1*) gene implicating this PYD containing protein in both familial cold auto-inflammatory syndrome (FCAS) and Muckle–Wells Syndrome (MWS) (133). Thus, the NLRP3 inflammasome will be described here as a prototype for these NLRs (**Figure 3**). Classically, the inflammasome has been described to consist of an NLRP, the inflammatory protease caspase-1, and the apoptosis-associated speck like protein (ASC) (51). ASC contains an N-terminal PYD and a C-terminal CARD making it uniquely suited for bringing into close proximity the two key components, caspase-1 and NLRPs (134, 135). Upon activation, NLRP3 recruits ASC and caspase-1, which is a prerequisite for the cleavage and maturation of the inflammatory cytokines IL-1β and IL-18 and consequent inflammatory cell death named pyroptosis (136–141). Lately, a more complex model for NLRP3 inflammasome activation has been proposed where two adaptors, ASC and mitochondrial MAVS, are required for optimal inflammasome triggering (142).

Owing to its widespread expression in numerous cell types such as neutrophils, monocytes, DCs, epithelial cells, and T cells (140, 143, 144), NLRP3 is exposed to a wide array of PAMPs and DAMPs that instigate the assembly and activation of the inflammasome [reviewed in Ref. (5, 132, 145–148)]. The NLRP3-inflammasome formation requires a two-step process (149). The priming step (or signal 1) involves TLR-NF-κB-driven induction of inflammasome components, as basal expression of NLRP3 in resting cells is insufficient for effective inflammasome activation (149, 150). However, certain cells like the human blood monocytes and murine macrophages appear to activate the NLRP3 inflammasome in response to LPS stimulation alone (151, 152). It is noteworthy that a transcriptionally silent mechanism for TLR4-mediated inflammasome priming has been lately discovered (153, 154). This mechanism involves mitochondrial ROS (mtROS)-driven deubiquitination of NLRP3, suggesting that constitutive ubiquitination of NLRs may be a homeostatic mechanism to prevent overt inflammasome activity (154). The second activation step (or signal 2) promotes the NLRs to undergo homotypic oligomerization and assemble the inflammasome.

While several models have been proposed to define the signals behind NLRP3 activation, the precise mechanisms remain hitherto unresolved. Various bacterial pathogens induce potassium efflux and activate the NLRP3 inflammasome via the action of secreted pore-forming toxins (e.g., nigericin from *Streptomyces hygroscopicus*, listeriolysin O from *Listeria monocytogenes*, pneumolysin from *Streptococcus pneumoniiae*, alpha-hemolysin, etc.) (138, 155, 156) (**Figure 3**). In addition, NLRP3 inflammasomes have been known to assemble in response to cytosolic bacterial and viral RNA both *in vivo* and *in vitro* (137, 157–160). Extracellular adenosine tri-phosphate (ATP) released from dying or damaged cells also causes NLRP3-inflammasome activation through either paracrine or autocrine sensing of ATP by the purinergic receptor P2X7 (138, 161–163). Besides, it has been defined that ATP released from phagocytosed dying cells acts similarly on P2X7 and prompts pannexin-1 (PANX1) channels to open, thus resulting in potassium (K+) efflux and allowing other agonists to further engage and activate NLRP3 (164) (**Figure 3**).

Monosodium urate (MSU) and calcium pyrophosphate dehydrate crystals, alum, amyloid-β fibrils, as well as environmental pollutants like asbestos and silica strongly activate the NLRP3 inflammasome (139, 165–170). According to one model for this mode of activation, uptake of the crystalline and particulate matters into the cell causes lysosomal destabilization and release of cathepsin B, which is sensed by NLRP3 (168, 169). Interestingly, however, opposing results were obtained when cathepsin B-deficient BMMs were used to test this hypothesis, as no differences in IL-1β or caspase-1 cleavage were observed in response to several inflammasome activators such as hemozoin, MSU, or alum (171). Another model suggests that these activators prompt generation of mtROS and mitochondrial DNA, both of which are responsible for NLRP3-inflammasome activation (172–174). Evidently, pharmacological inhibition of mtROS production has been shown to prevent NLRP3-inflammasome formation indicating that ROS generation is an upstream event for NLRP3 activation (165, 166) (**Figure 3**). Liposomes have been found to induce mtROS and NLRP3-inflammasome activation by triggering calcium (Ca2+) influx via the transient receptor potential melastatin 2 (TRPM2), although the exact mechanism linking ROS production to TRPM2 channel opening is still not wellcharacterized (175). On the other hand, the mitochondrial protein cardiolipin has been shown to directly bind and activate NLRP3 in a ROS-independent manner suggesting that ROS may not be the common denominator engaging the NLRP3 inflammasome (176). Recent advances have put forward additional mechanisms underlying NLRP3-inflammasome activation. In BMMs stimulated with PAMPs, extracellular calcium has been shown to activate the calcium sensing receptor (CASR) mediating signal transduction pathways that culminate in the release of calcium stores from the endoplasmic reticulum (ER), eventually activating the NLRP3 inflammasome (177–179). The diverse nature of the NLRP3-inflammasome agonists allude to the likelihood that, instead of directly sensing PAMPs and DAMPs, NLRP3 may be activated by converging pathways with a final common ligand for NLRP3. Guanylate binding protein 5 (GBP5) has been recently proposed as one such component that directly participates in NLRP3-inflammasome activation; however, further investigation is needed to decipher how the GBP5 is activated and why it is required for select inflammasome assembly (180). Finally, studies by Munoz-Planillo et al. suggest that potassium efflux may perhaps be the sole intracellular event necessary for NLRP3 activation in response to a wide array of stimuli arguing for a unifying model for the NLRP3-inflammasome activation (181) (**Figure 3**).

Production of mtROS often culminates in mitophagy, an autophagic clearance of dysfunctional mitochondria. It has been demonstrated that inhibition of mitophagy enhances NLRP3 caspase-1-mediated secretion of IL-1β and IL-18 in response to LPS and ATP (172). In addition, deletion of ATG16L1 was found to promote IL-1β release in response to ATP, MSU, or LPS alone (182). Moreover, it has been recently suggested that autophagy may restrict NLRP3 activity by directly sequestering and targeting inflammasome components for degradation (183, 184). Overall, it is reasonable to speculate that autophagy could serve as a mechanism for preventing excessive NLRP3-inflammasome activation (172, 173, 183–185).

Mitochondrial dysfunction plays a central role in regulating the mechanisms involved in both inflammasome and apoptosis pathways. Loss of mitochondrial membrane potential is a pivotal event in intrinsic apoptosis and is tightly regulated by the BCL2 family of proteins through a system of checks and balances (30). Interestingly, anti-apoptotic BCL2 and BCL-XL proteins have

#### **FIGURE 3 | Simplified mechanisms for the canonical NLRP3-**

**inflammasome activation**. Various PAMPs and DAMPs provide the signal 2 required to assemble and activate the NLRP3 inflammasome comprised of NLRP3, ASC, and caspase-1. Although the precise mechanism leading to NLRP3 activation is still controversial, it is speculated that K<sup>+</sup> efflux may be the common cellular response that triggers inflammasome activation. However, this notion has not been fully verified and it is possible that an unidentified or intermediate adaptor may be required for transmitting signals between K<sup>+</sup> efflux and the NLRP3 inflammasome. Crystals and particulate DAMPs enter the cell via endocytosis directly inducing K<sup>+</sup> efflux and NLRP3-inflammasome formation. In addition, the endolysosomes carrying these DAMPs undergo lysosomal rupture and release cathepsin B, which acts as an intracellular DAMP and can induce K<sup>+</sup> efflux. However, contradicting studies indicate that lysosomal rupture may cause K <sup>+</sup> efflux and inflammasome activation even in the absence of cathepsin

B. ATP binds to the P2X7 receptor on the cell membrane and causes opening of the PANX1 channels allowing K<sup>+</sup> efflux and influx of any PAMPs and DAMPs present in the extracellular space. PAMPs such as pore-forming toxins activate the NLRP3 inflammasome and facilitate K<sup>+</sup> efflux. Liposomes instigate Ca<sup>2</sup><sup>+</sup> influx through opening of the TRPM2 channels. Accumulation of excessive Ca<sup>2</sup><sup>+</sup> in the cytosol causes mitochondrial dysfunction and release of mtROS and oxidized mtDNA, which may activate the NLRP3 inflammasome either directly or by inducing K<sup>+</sup> efflux. Clearance of distressed mitochondria by mitophagy serves to evade such inflammasome activation. Mitochondrial Cardiolipin binds to NLRP3 and is required for the NLRP3-inflammasome activation. Following NLRP3-inflammasome assembly, caspase-1 undergoes proximity driven proteolytic cleavage and further processes pro-IL-18 and pro-IL-1β into their mature active forms. Activation of the NLRP3-caspase-1 axis results in inflammation and pyroptotic cell death.

been reported to directly interact with NLRP1 (CARD and PYD domain containing NLRP) to negatively regulate caspase-1 activation (186, 187). Similarly, BCL2 overexpression was shown to limit NLRP3-inflammasome activation (173, 174). In addition to BCL2 proteins, cIAP1, cIAP2, and XIAP have also been linked with inflammasome activation. Unlike their role in NOD signaling, initial studies have proposed that expression of these proteins might prevent caspase-1-dependent cell death (188). However, more recently cIAP1 and cIAP2 along with TRAF2 were found to enhance inflammasome activation seemingly by ubiquitinating and stabilizing caspase-1 and consequently prompting IL-1β release (189). In another report, genetic ablation of cIAP1 or cIAP2 had no effect on NLRP3-inflammasome activation, but concurrent pharmacological degradation of XIAP, cIAP1, and cIAP2 using SMAC mimetics was shown to limit caspase-1 activation (190). Interestingly, further inquiries revealed that in the absence of XIAP, cIAP1, and cIAP2, cell death in response to LPS was primarily incited by RIP3 activation causing NLRP3-caspase-1 as well as caspase-8-dependent IL-1β secretion (190). Lately, the concept of non-canonical inflammasome has been defined, which requires activation of caspase-11 in response to Gram-negative bacteria to facilitate either caspase-1-mediated IL-1β secretion or caspase-1-independent pyroptosis (191–194). Interestingly, apoptosis mediators FADD and caspase-8 have been involved in canonical and non-canonical NLRP3-inflammasome signaling. Indeed, FADD and caspase-8 facilitate the priming in "signal 1" by instigating both, LPS-TLR-MyD88-triggered induction of pro-IL-1β and NLRP3, as well as TLR-TRIF-mediated upregulation of procaspase-11 (195). Upon infection with *Citrobacter rodentium* or *Escherichia coli*, FADD and caspase-8 have been found to promote the "signal 2" by interacting with the NLRP3-inflammasome complex, thus influencing both canonical (caspase-1-dependent IL-1β maturation) and non-canonical (caspase-11-dependent pyropotosis) inflammasomes (194, 195). Conversely, it has been exhibited that caspase-8-deficient murine DCs are hyper-responsive to LPSinduced NLRP3-inflammasome assembly and activation (196). Overall, these studies place caspase-11 and caspase-8 at the center of inflammasome activation; however, a general lack of consensus in the field makes it hard to aptly judge their contribution in inflammasome-induced inflammation.

#### **INFLAMMASOME NLRs AND CANCER**

*NLRP3*, previously associated with rare and severe autoinflammatory disorders, has been lately implicated in CD susceptibility and correlated with decreased NLRP3 expression and IL-1β production (62). Indeed, mice lacking NLRP3 have been shown to display exacerbated colonic inflammation upon DSS-induced colitis characterized by greater gut barrier damage, inflammatory immune cell infiltration, and cytokine production (197, 198). In accord, a central role has been ascribed for caspase-1 and ASC in intestinal epithelial repair after DSS-injury (199). Specifically, caspase-1, ASC, or NLRP3 deficiency in mice has been shown to be detrimental in DSS-induced intestinal inflammation, a mechanism attributed to the lack of IL-18 production by IECs (198, 199). Concomitantly, the increased colitogenic phenotype was completely reversed when mice were exogenously administered with the recombinant IL-18 cytokine (198, 199). The same lack

of inflammatory regulation was found to render *Nlrp3*−/<sup>−</sup> and *Casp1*−/<sup>−</sup> mice more susceptible to AOM–DSS carcinogenesis (197, 200). The heightened tumor growth in the caspase-1 deficient mice was accompanied with drastically low levels of colonic IL-18. Overall, NLRP3 was shown to be important for IL-18 secretion, which in turn through IFNγ production induces STAT1 (Signal transducers and activators of transcription 1) phosphorylation and thus promotes an anti-tumorigenic environment (200). Moreover, it has been shown that *Il18*−/<sup>−</sup> or *Il18r*−/<sup>−</sup> mice are more susceptible to DSS-induced colitis and CRC, mimicking the increased tumor burdens observed in NLRP3 and caspase-1 deficient mice (201). Recent findings have put forward a novel concept for the dual function of IL-18 in intestinal inflammation and colitis-driven CRC (202, 203). For instance, during acute injury IEC-derived IL-18 triggers repair and restitution of the ulcerated epithelial barrier, whereas under chronic inflammatory settings the excessive release of IL-18 both from IECs and lamina propria macrophages and DCs is deleterious (203, 204). A protective role for NLRP3 has also been described in hepatocellular carcinoma (HCC) (205). This correlation is primarily based on mRNA and protein expression data showing reduced levels of NLRP3 and other related inflammasome components seen in hepatic parenchymal cells derived from HCC tissue specimens as compared to non-cancerous liver sections (205). On the other hand, a gain of function SNP (Q705K) within the *NLRP3* gene has been associated with increased mortality in CRC patients (206). Significantly, the same SNP was also found to be more prevalent in patients with malignant melanoma (207). Human monocytic THP-1 cells overexpressing a mutant variant of NLRP3 bearing the Q705K SNP have been reported to greatly respond to the inflammasome agonist alum and to trigger the production of IL-1β and IL-18, implying that overt NLRP3 activation could be detrimental for certain types of cancer (208). Similarly, another group implicated constitutive NLRP3-inflammasome signaling in the development and progression of melanomas (209).

Loss of function in the tumor suppressor gene p53 has been associated with a large number of sporadic cancers (36). One of the mechanisms for p53-induced clearance of potentially carcinogenic cells has been found to be via transcriptional up regulation of cell death activators (210). In light of this knowledge, the discovery of NLRC4 as a downstream transcriptional target of p53 was a promising evidence for the anti-tumorigenic functions of this NLR (211). Moreover, lack of NLRC4 inflammasome has been associated with the attenuation of p53-mediated cell death, indicative of a protective role of NLRC4 during tumor development (211). Several groups have investigated the role of NLRC4 in colitis and CRC. However, lack of consensus in the susceptibility of *Nlrc4*−/<sup>−</sup> mice to DSS as well as AOM–DSS treatment makes it difficult to gage the protective effect of NLRC4 in these models (197, 212). It has been demonstrated that mice deficient in NLRC4 develop higher tumor burdens than WT mice when subjected to DSS-induced CRC (212). In addition, bone marrow chimera experiments verified that NLRC4 expression within the radioresistant compartment was the major driver of CRC protection (212). Surprisingly, similar colitic phenotypes have been observed between WT and *Nlrc4*−/<sup>−</sup> mice following DSS administration, suggesting that tumor regulation by NLRC4 is mostly

cell intrinsic and not through down-regulation of inflammation (213). Given the unique capacity of NLRC4 to sense and differentiate between commensal and pathogenic microbes in the gut (214), it is surprising that the tumor restraining roles of NLRC4 have been ruled to be independent of its immune regulatory functions. One unifying theory addressing these discrepancies could be that anti-tumor functions of NLRC4 are attributed to the cells of non-hematopoietic origin, whereas intestinal mononuclear phagocytes are the primary source of NLRC4 for microbial sensing and pathogen clearance (213, 214). Overall, these assumptions warrant deeper inquiries to clearly elucidate the mechanisms by which NLRC4 exerts protective functions during CRC and to decipher the relevance of p53-mediated role of NLRC4 in tumorigenesis.

Akin to NLRP3, both NLRP6 and NLRP12 have been recently described to use ASC-caspase-1 molecular platforms and assemble inflammasomes. A first hint of NLRP6 being an inflammasome NLR was gleaned from *in vitro* experiments showing increased caspase-1 cleavage when ASC and NLRP6 were co-expressed (215). Further *in vivo* evidence emphasized a protective role for NLRP6 in intestinal inflammation and tumorigenesis as *Nlrp6*−/<sup>−</sup> mice showed high susceptibility to DSS-induced colitis and AOM–DSSinduced CRC (216–218). Unlike NLRC4, dampening of inflammation is purported to be one of the primary mechanismsfor NLRP6 mediated protection and tissue homeostasis. NLRP6 has been shown to promote a gut microbiome that limits chronic inflammation. In fact, it has been evidenced that *Nlrp6*−/<sup>−</sup> mice display a distinct transmissible pro-colitogenic microbiome with increased prevalence of the bacterial genus *Prevotellaceae* (217). These mice presented a steady state colitic phenotype and an enhanced sensitivity to DSS colitis (217). Overall, a mechanism has been suggested wherein dysbiosis in the gut, caused by aberrant NLRP6 inflammasome signaling, drives excessive CCL5-mediated IL-6 production, barrier damage, and inflammation (217). In agreement with the findings in *Casp1*−/<sup>−</sup> mice (199), NLRP6-deficient mice had impaired IL-18 production mainly from the intestinal epithelial compartment further diminishing the capacity of these mice to recover from colitis. Likewise, overt inflammation and lack of IL-18 in the *Nlrp6*−/<sup>−</sup> mice has been associated with increased colonic tumor development (216), however, as seen for *Nlrp3*−/<sup>−</sup> mice it is still unknown whether administration of IL-18 is capable of rescuing the susceptibility phenotype. Interestingly, gene expression profiling of colorectal tumors derived from WT and *Nlrp6*−/<sup>−</sup> mice revealed an increased expression of paracrine factors of the Wnt and NOTCH signaling cascades, underscoring a novel function of NLRP6 in controlling intestinal proliferation (218). Sensing of damaged or dying cells by NLRP6 and NLRP3 inflammasomes has lately been hypothesized to prevent CRC through maintaining the balance between IL-22 and IL-22 binding protein (IL22-BP) (219). It has been speculated that sensing of DAMPs by both NLRs instigates IL-18-dependent down-regulation of the inhibitory molecule IL-22BP, thus allowing IL-22 to repair the injured tissue. However, dysregulated NLRP6 or NLRP3 signaling could potentially lead to inappropriate IL-22BP expression, thus creating a pro-tumorigenic environment caused by either excessive cell proliferation or lack of tissue repair (219). Although the dual function of IL-22 in CRC has been well-described, further experimental validation is needed to pinpoint the exact mode by which NLRP3 or NLRP6 regulate IL-22/IL-22BP ratio during colon tumorigenesis.

NLRP12 was originally defined as an inflammasome NLR due to its co-localization with ASC and caspase-1, induction of IL-1β and IL-18 secretion as well as NF-κB activation (220, 221). SNPs within the *NLRP12* gene have been associated with increased susceptibility to atopic dermatitis and periodic fever syndromes accompanied mostly with caspase-1 activation and IL-1β release (222–225). It has been observed that NLRP12 can negatively regulate both canonical and non-canonical NF-κB pathways by targeting the IL-1R-associated kinase 1 (IRAK1) and NF-κB inducing kinase (NIK) for proteasomal degradation (226–228). Two independent studies proposed that NLRP12 acts as a tumor suppressive molecule *ex vivo* and in *in vivo* animal models of colitis and colitisinduced CRC (229, 230). Mice lacking NLRP12 have been found to be more susceptible to DSS-injury with increased body weight loss, enhanced pathology scores coupled with massive infiltration of inflammatory cells and high inflammatory cytokine production (229, 230). Furthermore, AOM–DSS treatment of *Nlrp12*−/<sup>−</sup> mice has been shown to further provoke colonic tumor development and progression (229, 230). In the first study, it was clearly demonstrated that lack of NLRP12 increases NIK-dependent noncanonical NF-κB signaling and drives the regulation of cancer promoting genes like CXCL12 and CXCL13 (230). In the second report, the enhanced tumorigenicity in knockout mice was traced to excessive canonical NF-κB activation due to lack of NLRP12 in hematopoietic cells. Indeed, enhanced LPS-induced canonical NF-κB activation was exhibited in *Nlrp12*−/<sup>−</sup> macrophages *ex vivo*, suggesting that microbial sensing and negative regulation of inflammation may account for NLRP12-mediated tumor suppression (229). Altogether, these results underscore the importance of anti-inflammatory signals provided by NLRP12 in maintaining colonic homeostasis and protecting from colitis and colon tumorigenesis.

#### **THERAPEUTIC STRATEGIES AND CONCLUSION**

It has been suggested that the strong immunomodulatory properties of NLRs could be exploited for mounting potent antitumorigenic responses. In fact, mice injected with B16 melanoma cells or EL4 thymoma cells expressing flagellin from *Salmonella typhimurium* were shown to display dramatic resistance to tumor establishment in NLRC4 dependent manner (231). In addition, immunization with flagellin expressing cancer cells lead to impressive antigen-specific CD4 and CD8 T cell responses via NLRC4 and NAIP5 signaling and bestowed anti-tumor immunity against a secondary inoculation with tumor cells (231). Similarly, activation of NODs, in particular NOD2, to elicit robust cell-based anti-tumor immunity has been under scrutiny for several years. Indeed, instillation of MDP in patients with lung cancer has been reported to enhance expression of inflammatory cytokines and neutrophils in the pleural fluid (232). Relatedly, it has been suggested that the local immune-modulatory activity of MDP helps improve prognosis in hamsters suffering from osteosarcoma (233) and significantly reduces tumor metastasis in several murine cancer models, such as B16–BL6 melanoma, colon 26-M#1 carcinoma, and L5178Y-ML25T T lymphoma (234, 235).

Overt activation of the NLRP3 inflammasome has been demonstrated to elicit cancer progression. For instance, in mouse models of methylcholanthrene (MCA, a highly potent carcinogen) induced fibrosarcoma, NLRP3 was demonstrated to promote cancer progression. Moreover, NLRP3 expression in myeloid cells was shown to interfere with the suppression of cancer metastasis by inhibiting recruitment of anti-tumor NK cells to the site of carcinogenesis (236). Besides interfering with natural tumor control, NLRP3 inflammasome-mediated IL-1β has been described to attenuate anti-tumor effects of chemotherapeutic agents, gemcitabine (Gem), and 5-fluorouracil (5FU) (237). Mice lacking NLRP3 were far more receptive to thymoma regression upon treatment with Gem or 5FU as compared to WT mice. Furthermore, enhanced NLRP3-driven IL-1β release was linked with the induction of T helper 17 (Th17) cells that promoted chemoresistance in WT mice (237). Keeping these observations in view, several studies support the use of specific inhibitors, antagonists, and monoclonal antibodies against components of the inflammasome, e.g., caspase-1, IL-1β, and IL-18, as therapeutic approaches beneficial for controlling inflammation and improving cancer prognosis (238).

An early phase clinical study suggests that administration of the IL-1R antagonist,Anakinra, alone or in combination with dexamethasone could potentially impede human multiple myeloma progression (239). Furthermore, it was demonstrated that IL-18 derived from tumor cells had the ability to subvert the NK cell-mediated tumor immunosurveillance and to promote tumor progression in a programed death receptor 1 (PD1)-dependent manner (240, 241). These findings suggest the potential of using IL-18 as well as PD1 neutralization for cancer immunotherapy. Overall, selective attenuation of the activities of certain NLRs could potentially boost regression and improve responsiveness to chemotherapy. The variability in NLRP3- and IL-18-mediated effects in different cancers highlights the complexity in NLR circuits and suggests that any broad implications regarding NLR intervention in tumorigenesis should be carefully investigated.

Microbial environment, diet, mouse strain, tumor ontogeny, etc. are all part of the complex network that dictates how an NLR influences inflammation and tumorigenesis. Sensitivity to these factors has lead to conflicting disease phenotypes in genetically modified mice lacking specific NLRs. Furthermore, NLR expression in hematopoietic or non-hematopoietic cellular compartments appears to have distinct influence on inflammatory regulation and tumorigenesis. Due to such discrepancies, it is still uncertain how dysregulation of these innate immune sensors incites inflammation that leads to carcinogenic transformation of cells. Although several mechanisms have been suggested like control of NF-κB signaling, regulation of tissue repair factors, and IL-18 secretion, no unifying hypothesis exists. In addition, interaction of NLRs with different members of the TNFR pathway, BCL2 family of proteins, IAPs, apoptotic caspases, and autophagy regulators point toward more intricate mechanisms for NLR regulation than currently acknowledged. Future studies focusing on the biochemistry of interactions between cell death regulators and NLRs are required to delineate the co-integration of NLRcell death mechanisms so as to facilitate implementation of NLR

modifying therapeutic strategies for inflammatory diseases and cancer.

#### **ACKNOWLEDGMENTS**

We thank all the Yeretssian lab members for the insightful and critical reading of the manuscript. This work was supported by the Helmsley Foundation.

#### **REFERENCES**


that confers potent transactivation activity. *J Biol Chem* (2001) **276**:19089–93. doi:10.1074/jbc.M101295200


containing diaminopimelic acid. *Nat Immunol* (2003) **4**:702–7. doi:10.1038/ ni945


NFkappaB/TNFAIP3- and NLRP3-inflammasome signaling affect susceptibility and outcome of colorectal cancer. *Carcinogenesis* (2012) **33**:2126–34. doi:10.1093/carcin/bgs256


with recurrence of an NLRP12 mutation. *Arthritis Rheum* (2011) **63**:830–9. doi:10.1002/art.30170


**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.

*Received: 27 April 2014; paper pending published: 02 May 2014; accepted: 27 June 2014; published online: 14 July 2014.*

*Citation: Saxena M and Yeretssian G (2014) NOD-like receptors: master regulators of inflammation and cancer. Front. Immunol. 5:327. doi: 10.3389/fimmu.2014.00327 This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Saxena and Yeretssian. 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.*

## Nod-like receptors: key molecular switches in the conundrum of cancer

#### **Andrew Kent <sup>1</sup> and J. Magarian Blander 1,2\***

1 Immunology Institute, Department of Medicine, Graduate School of Biological Sciences, New York, NY, USA <sup>2</sup> Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

J. Magarian Blander, Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA

e-mail: julie.blander@mssm.edu

It is believed the immune system can contribute to oncogenic transformation especially in settings of chronic inflammation, be activated during immunosurveillance to destroy early neoplastic cells before they undergo malignant outgrowth, and finally, can assist growth of established tumors by preventing clearance, remodeling surrounding tissue, and promoting metastatic events. These seemingly opposing roles of the immune system at the different stages of cancer development must all be mediated by innate signaling mechanisms that regulate the overall state of immune activation. Recently, the cytosolic nod-like receptor (NLR) pathway of innate immunity has gained a lot of attention in the tumor immunology field due to its known involvement in promoting inflammation and immunity, and conversely, in regulating tissue repair processes. In this review, we present all the current evidence for NLR involvement in the different stages of neoplasia to understand how a single molecular pathway can contribute to conflicting immunological interactions with cancer.

**Keywords: nod-like receptors, cancer, immunoediting, immunosurveillance, innate immunity, transformation**

#### **INTRODUCTION**

The pervading conception of the immune system today depicts it simply as the body's means of warding off infection. In her *Anthropology of Immunology*, Martin eloquently describes "the body as nation state at war over its borders, containing internal surveillance systems (encompassed in the immune system) to monitor foreign intruders" (1). However, this "infection-centric" view does not consider profound facets of the immune system, now well established in the literature, and largely forgotten since the earliest immunologists predicted their existence. As early as the 1890s, Ilya Metchnikoff conceived of the theory of "physiological inflammation," in which the immune system, especially phagocytic cells, were essential for maintaining homeostasis within all tissues of the body (2). He postulated that phagocytic cells uphold the balance between competing cell types and organs as they arise within a multicellular organism, establishing a unified "organismal identity" (2). This did not ignore the role of phagocytes in fighting infection, but suggested a "wide functional spectrum, of which host defense against pathogens was only one aspect" (2). Included were roles in regulating tissue development, clearance of damaged tissue, promotion of wound repair after any insult, be it infectious or sterile, and resolution of unwarranted inflammatory processes.

There is no better example of a question of organismal identity, of the need for a restoration of homeostasis, or of cell types or tissues in competition with one another, than that of cancer. Because they are initially derived from self-tissue, transformed cells pose a dilemma – to destroy or repair? It seems the immune system is responsible for answering this question, and is now known to be intimately involved in the oncogenic process from the very emergence of the first transformed cells through to malignant disease (3–5). Due to the nature of the predicament at hand, the immune

system has been described to have conflicting roles depending on which stage of cancer progression is being studied (6). How the opposing immunological phenotypes in cancer are controlled is not well known, but nod-like receptors (NLRs) have been implicated in various stages of the disease process and have the required capacity to act as key regulators of physiological and pathological inflammation (7–9). NLRs are initiators of the inflammasome pathway, a cytosolic signaling apparatus that canonically activates caspase-1, and IL-1β and IL-18 thereafter (10). NLRs can respond to both pathogen- and danger-associated molecular patterns (PAMPs and DAMPs, respectively), and the pathway has been shown to have important roles in mounting immune responses to both microbial pathogens and damaged self, as well as regulating tissue repair after damage (11, 12). Here, we will review the evidence for NLR involvement in the initial emergence of neoplastic lesions, in the control and destruction of transformed cells during a phase of immunosurveillance, and finally the immune shift to supporting growth of established disease. We will argue that the conflicting roles of the immune system during oncogenesis can be reconciled within the framework of Metchnikoff's theory of immune control of tissue homeostasis, and that NLRs and their downstream signaling elements serve as key molecular switches in this process.

#### **EMERGENCE OF TRANSFORMATION**

Schreiber and colleagues categorized immune interaction with cancer into three stages of immunoediting: elimination by immunosurveillance mechanisms; equilibrium, when cancer attains a latent balance between aberrant growth and destruction; and escape, when the tumor overcomes suppression as an edited malignancy (13). Although overlooked in the "Three E's model" of immunoediting, the involvement of inflammatory processes in the initial emergence of cancer is well established within the literature. Chronic inflammation is a major risk factor for neoplasia in the clinic, working to both disrupt the microenvironment to favor neoplastic outgrowth, and contribute to genetic instability and altered turnover rates of stromal cells, promoting accelerated emergence of malignant clones (14). Many studies have now implicated the inflammasome pathway and the NLRs in this context, but with contrasting influences depending on the context and specifics under scrutiny.

A predominant model used to study NLR and inflammasome contributions to carcinogenesis is the AOM/DSS model (15). DSS causes damage to the colonic epithelium, while AOM causes G-to-A mutations in DNA of cells undergoing DNA replication. Deficiency in NLRP6, an NLR primarily expressed in colonic myofibroblasts, resulted in decreased repair of the intestinal epithelium following DSS treatment, but conversely, was associated with increased epithelial colonocyte proliferation and transcript expression of molecules involved in cell cycle progression (16). Another study showed prolonged colitis and epithelial destruction in *Nlrp6*−/<sup>−</sup> mice after DSS treatment was related to alterations in commensal microbiota, and was phenocopied when mice were deficient in any of the NLRP6 inflammasome components ASC (a common adapter to many inflammasomes), and caspase-1 (17). The IL-18 cytokine, cleaved into its biologically active form by activated caspase-1, has emerged as a key cytokine downstream of inflammasome activation that enables epithelial repair after damage, but also prevents cancer progression through its induction of the tumor suppressors STAT1 and IFN-γ (18). When treated with AOM/DSS, the resulting increased epithelial proliferation and exacerbated inflammation in *Nlrp6*−/<sup>−</sup> mice led to accelerated outgrowth of colonic cancer (16). In addition to NLRP6, loss of NLR family members NOD1, NOD2, NLRP3, NLRC4, and NLRP12 has resulted in similar exacerbated colitis and accelerated rates of cancer (19–24). Together, results from these gut studies suggest NLRs and their associated inflammasome components are essential for controlling wound repair responses and preventing transformative events and unwarranted epithelial proliferation early in potentially neoplastic settings (20). Much work needs to be done to clarify the mechanisms of NLR regulation in these processes, especially their connection to regulation of epithelial regrowth.

Paradoxically, over-expression of NLR pathway components also drives cancer rather than suppresses its emergence. As might be predicted from the above evidence, the derepression of caspase-1 that occurs in *Casp12*−/<sup>−</sup> mice results in accelerated recovery from colitis after DSS. However, after AOM/DSS, these mice have accelerated rather than decreased colorectal cancer development, a pathology linked to increased levels of inflammatory cytokine gene expression including *Il1b* (25). In a model of HCV infection, IL-1β production downstream of NLRP3 by hepatic macrophages was linked to chronic hepatitis (26). Similarly, CCl<sup>4</sup> treated *Nlrp3*−/<sup>−</sup> and*Asc*−/<sup>−</sup> mice exhibited reduced levels of liver fibrosis,and wildtype hepatic stellate cells treated with monosodium urate crystals upregulated the *Tgfb* and *Col1a* genes in an inflammasomedependent manner (27). Thus in the liver, NLRs contribute to chronic inflammatory processes, both infectious and sterile, that result in the hepatitis and fibrosis commonly found prior to hepatocellular carcinoma.

IL-1β has many pleiotropic effects involved in inflammation, immunosuppression, cell proliferation and differentiation, tissue regeneration, tumor-promotion, and chemoresistance (28). In addition to its roles in hepatic carcinoma, the cytokine has been implicated in accelerating tumor development in mammary epithelial (29), gastric (30), and skin (31) cancer models, further establishing its role as an inflammatory instigator of oncogenesis. Drexler et al. were able to show both anti- and pro-tumorigenic effects of ASC in a single model of chemically induced skin carcinogenesis (31). ASC expression in infiltrating myeloid cells helped drive carcinogenesis, while ASC expression in keratinocytes suppressed epithelial cell proliferation and carcinogenesis (although in a caspase-1-independent manner). While the specific NLR implicated in these opposing roles of ASC was not identified, involvement of the inflammasome pathway was strongly implicated.

These studies all demonstrate opposing roles of the inflammasome in the early initiation of neoplastic disease. NLR activation can inhibit malignant transformation by controlling epithelial cell regeneration, but can also contribute to chronic inflammation that eventually results in carcinogenesis. The NLRs mediate a fine balance between inflammation and repair to maintain homeostasis in each tissue. If tipped in either direction, malignancy can result.

#### **ELIMINATION OF TRANSFORMED CELLS**

Once a transformed cell appears, it immediately presents a unique challenge to the immune system. Its uncontrolled proliferation threatens the evolutionarily defined healthy function of the tissue of its origin. Although derived from self,it no longer obeys the rules of organismal identity. From observations of homograft rejection, and increased cancer incidence in immunocompromised individuals, Lewis Thomas and Sir MacFarlane Burnett postulated the theory of immunosurveillance – the ability of the immune system to recognize and destroy abnormal self despite its ontogenic origins (32). Schreiber and others have built a strong case for the existence of adaptive immunosurveillance, and now evidence is emerging in spontaneous models of neoplasia (33–36).

Every adaptive response requires innate priming, thus innate immunity must be involved. Some studies have shown innate cell involvement (34, 37, 38), but thorough examinations of the molecular pathways that enable immune activation against tumor antigens are scarce. However, there are a few studies directly demonstrating NLRs can be involved in immunosurveillance. In an allograft model, Ghiringelli et al. show that chemotherapeutic killing of tumor cells causes a release of ATP that binds the P2RX7 purinergic receptor on dendritic cells (DCs), eventually leading to the activation of the NLRP3 inflammasome in these cells (37). By synergizing with HMGB1, released from dying tumor cells and signaling through toll-like receptor (TLR) 4, activated DC are licensed to prime an anti-tumor immune response in a caspase-1 and IL-1β-dependent manner. Another study found that extracts from an anti-tumorigenic mushroom functioned by activating the same P2RX7/NLRP3 pathway in macrophages, but did not draw a direct link to altered tumor kinetics (39). Although these conclusions derive from experimental models, anthracycline-treated breast cancer patients with mutations in the *P2rx7* gene were found to develop metastatic disease faster than those with normal *P2rx7* genes, suggesting the NLRP3-dependent pathway may be activated in humans with spontaneous disease (37). In addition to NLRP3, in 2012 we published on the ability of flagellin to synergistically activate TLR5 and the NLRC4 inflammasome, resulting in effective priming of CD4 and CD8 immunity against subcutaneously implanted allografts in mice (40). Besides priming of adaptive immunosurveillance, NLRs have been implicated in antitumor immunity through the link between IL-18 and increased NK cell activity against tumors (41–44). However, these latter findings were made in the presence of exogenous administration or expression of IL-18 above normal levels.

All these studies involve some artificial intervention that enhances NLR activity, but present a strong case for the ability of the pathway to influence immunosurveillance. It remains to be shown if the inflammasome pathway is involved in intrinsic immunosurveillance mechanisms, or is activated at this early stage of disease in any capacity. It is difficult to capture the elimination phase due to its transience and lack of overt disease phenotypes. Spontaneous models with a definable pre-malignant stage must be employed to further analyze which innate signaling pathways, and in which cell types, are naturally engaged to clear transformed cells before they cause disease. Selectively enhancing this engagement could greatly benefit therapeutic intervention. Additionally, these studies suggest a critical function of the inflammasome in priming adaptive immunity against transformed self-cells. It remains to be shown if this ability is mediated entirely through cytokine production, or if the inflammasome can influence T cell priming in a more direct manner. Conversely, it is possible there are strictly innatemediated immunosurveillance or tumor-suppressing mechanisms engaged that help inhibit malignancy without priming T or NK cells (45). NLR involvement in these processes is unknown.

#### **MAINTENANCE OF ESTABLISHED DISEASE**

Malignant disease is the result of failed immunosurveillance mechanisms. The editing process selects for clones of the rapidly dividing and mutating transformed cell that are progressively less immunostimulatory (13). Eventually, the developing tumor attains a phenotype that no longer incites immune destruction and can grow uncontrolled. Furthermore, established tumors are known to usurp immune mechanisms to not only prevent destruction, but facilitate growth (46). Tumors have been described as wounds that will not heal due to their self origin, the stress they undergo as they rapidly expand, and their elicitation of reparative and protective immune functions (47, 48).

In light of this analogy, it is not surprising to find NLRs activated in malignant disease, in this context attempting to repair the "wound" to restore homeostasis and protect it from further immune destruction. A host of evidence supports various roles for NLR-activated IL-1β in malignancy, notably in humanized models (49, 50). Okamoto et al. found that malignant human melanoma cells spontaneously activated their intrinsic NLRP3 inflammasome, resulting in caspase-1 cleavage and spontaneous secretion of IL-1β (51). This secreted IL-1β became increasingly autonomous with later stage disease, implicating it as an evolutionarily advantageous trait for the developing tumor. *In vitro*, the inflammasome pathway and IL-1β were shown to increase macrophage chemotaxis and angiogenesis, both features linked to worse prognosis in various cancers (52). Another study found that IL-1β and caspase-1-deficient mice were much less susceptible to melanoma liver metastases by an injected allograft, improving their overall survival (53). *In vitro*, secreted factors from the melanoma cell line induced IL-18-dependent upregulation of VCAM-1 on hepatic sinusoidal endothelial cells, as well as IL-1β secretion. In opposition to the results in the previous section, endogenous IL-18 from melanoma cells was also found to inhibit NK cell-mediated killing of melanoma cells by upregulating Fas ligand expression (54). Additionally, IL-18 was found to enhance immunosuppression of NK cells by inducing upregulation of the inhibitory molecule PD-1 (55).

Nod-like receptors are also implicated in the ability of myeloidderived suppressor cells (MDSCs) to inhibit anti-tumor immunosurveillance. Related to the gut studies in the first section, IL-1β over-expression in the stomach was shown to induce inflammation and cancer (30). This was associated with an increase in MDSC numbers homing to the stomach in an IL-1R and NF-κBdependent fashion. In a model of DC-based vaccination against melanoma, van Deventer et al. demonstrated that *Nlrp3*−/<sup>−</sup> mice had improved outcomes due to decreased numbers of MDSCs homing to the tumor site (56). However, they did not observe a change in MDSC function, such as the ability to suppress T cell responses. Finally, chemotherapy was found to trigger cathepsin B release within MDSCs, triggering NLRP3 within the same cells (57). The resultant IL-1β production induced IL-17 secretion by CD4 T cells. Allograft tumor growth was slower in *Il17a*−/−, *Il1r1*−/−, *Nlrp3*−/−, and *Casp1*−/<sup>−</sup> mice after chemotherapy treatment, demonstrating all elements in this pathway play a part in tumor protection although the exact mechanism is unclear.

This evidence clearly implicates the NLRs and inflammasome pathway in tumor-promotion and defense. They directly facilitate tumor cell growth and metastasis, and help prevent any anti-tumor immune responses. It is curious to speculate how accurate the analogy of tumor to "unhealing wounds" is with regards to NLR involvement. Are NLRs engaged in the same way by malignant disease as they are by damaged tissues prior to malignant transformation, in both cases inducing repair and protective properties? Fitting with the tumor editing hypothesis, any pro-inflammatory DAMPs or other signals resulting from initial transformation that would trigger tumor clearance have in theory been selected away, leaving only those characteristic of damaged self in need of repair. Inflammasome involvement in such diverse functions as tissue repair, immune suppression, and inflammation warrants a search for more inflammasome-activated targets besides IL-1β and IL-18 that could fine-tune downstream effector mechanisms. Are these two cytokines alone able to control such diverse effects, or are they working in collaboration with many other pathways, the overall milieu defining the result? Concerted efforts to consolidate information across tumor models and treatments, being mindful of cell-type specificity, will help clarify these points.

#### **CONCLUSION**

We have now seen how NLRs switch roles in every stage of cancer progression (**Figure 1**). In each, the NLRs can be conceptualized

as attempting to restore homeostasis. First, in situations where damage to self has occurred, the NLRs contribute both to fighting off infection and repairing the damaged epithelial layers. The latter implicates an ability of the NLR pathway to regulate growth of surrounding tissues, with a strong link to IL-18. These processes require perfect coordination to maintain equilibrium in the tissue. The fact that too much or too little NLR signaling in this type of setting can result in neoplasia betrays how essential this pathway is to maintaining balance and organismal integrity. Second, when the very idea of self is challenged by oncogenic mutations, again NLR signaling is observed. Presumably here in early pre-neoplastic situations, NLR activation functions as an innate defense against localized transformation events. When clinical pathology is observed, these endogenous protective functions of the NLR have failed. Therapeutic enhancement of this activation has been shown to be beneficial in mouse models, especially in concert with activation of other inflammatory pathways such as TLRs. Thus, development of therapies that employ NLRs could have great impact in the clinic, especially if used very early in neoplasia. Finally, after tumors become established and are immunologically indistinguishable from other self-tissues, NLR activation reverts to helping protect and maintain this neoself, establishing a new, pathological state of homeostasis. Malignant disease is extremely hard to treat in part because of this unique pseudo-self phenotype and consequent immunoprotective state, reiterating the need for early intervention for successful treatment. Metchnikoff's prescient description of physiological inflammation is thus embodied within the recently discovered NLR pathway. Theories from this founding father of immunology can still help us conceptualize the perplexing and, in the case of NLRs and cancer, diametrically opposed functions of the immune system.

#### **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.

*Received: 04 April 2014; accepted: 08 April 2014; published online: 23 April 2014. Citation: Kent A and Blander JM (2014) Nod-like receptors: key molecular switches in the conundrum of cancer. Front. Immunol. 5:185. doi: 10.3389/fimmu.2014.00185 This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Kent and Blander. 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.*

#### **Irving Coy Allen\***

Department of Biomedical Sciences and Pathobiology, Virginia-Maryland Regional College of Veterinary Medicine, Virginia Polytechnic Institute and State University, Blacksburg, VA, USA

#### **Edited by:**

Anton G. Kutikhin, Research Institute for Complex Issues of Cardiovascular Diseases, Russia

#### **Reviewed by:**

Anton G. Kutikhin, Research Institute for Complex Issues of Cardiovascular Diseases, Russia Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Irving Coy Allen, Department of Biomedical Sciences and Pathobiology, Virginia-Maryland Regional College of Veterinary Medicine, Virginia Polytechnic Institute and State University, IDRF 140, 295 Duck Pond Drive, Blacksburg, VA 24061, USA e-mail: icallen@vt.edu

Aberrant inflammation is an enabling characteristic of tumorigenesis. Thus, signaling cascades that alter inflammatory activation and resolution are of specific relevance to disease pathogenesis. Pattern recognition receptors (PRRs) are essential mediators of the host immune response and have emerged as critical elements affecting multiple facets of tumor pathobiology. The nucleotide-binding domain and leucine-rich repeat containing (NLR) proteins are intracellular PRRs that sense microbial and non-microbial products. Members of the NLR family can be divided into functional sub-groups based on their ability to either positively or negatively regulate the host immune response. Recent studies have identified a novel sub-group of non-inflammasome forming NLRs that negatively regulate diverse biological pathways associated with both inflammation and tumorigenesis. Understanding the mechanisms underlying the function of these unique NLRs will assist in the rationale design of future therapeutic strategies targeting a wide spectrum of inflammatory diseases and cancer. Here, we will discuss recent findings associated with this novel NLR sub-group and mechanisms by which these PRRs may function to alter cancer pathogenesis.

**Keywords: Nod-like receptors, NLRP12, NLRX1, NLRC3, NF-**κ**B,TRAF, cancer, pattern recognition receptors**

#### **INTRODUCTION**

The intimate association between inflammation and cancer was first noted over 150 years ago by Rudolf Vierchow (1, 2). Indeed today, aberrant inflammation is considered both an emerging hallmark of tumorigenesis and an enabling characteristic of cancer (3). Tumorigenesis is a multistep process and inflammation functions at multiple levels to both antagonize and enhance tumor initiation and progression (3). During the early stages of tumorigenesis, an inflammatory microenvironment serves as an enabling characteristic to activate diverse signaling pathways and drive the progression of pre-malignant and malignant lesions toward cancer (3–5). In later stages, cancer cells typically acquire a diverse repertoire of defense mechanisms that allow the cells to both passively and actively evade immune surveillance and elimination (3, 6, 7). This immune system subversion is an emerging hallmark of cancer and serves to remove the most effective barriers employed by the host to defend against neoplasia, late-stage tumor, and micro-metastasis progression (3).

Pattern recognition receptors (PRRs) are an essential component of the host immune system and significantly contribute to cancer pathobiology. There are 4 major families of PRRs that have been implicated in tumorigenesis, including the toll-like receptors (TLRs), the nucleotide-binding domain and leucine-rich repeat containing (NLR) family of sensors, C-type lectin receptors (CLRs), and RIG-I-like receptors (RLRs) (8). These receptor families function to initiate inflammatory signaling cascades following the direct or indirect recognition of pathogens, damage and stress through sensing highly conserved pathogen-associated molecular patterns (PAMPs), and damage-associated molecular patterns (DAMPs). In addition to their roles in facilitating the immune response, PRRs also play fundamental roles in the regulation of proliferation, cell survival and death, reactive oxygen species generation, angiogenesis, and tissue remodeling and repair (8). In the context of cancer, PRRs drive the immune response following exposure to potentially carcinogenic pathogens, environmental exposures to mutagenic agents and insults, and cancer-associated cellular damage and stress (9–16). In general, increased PRR signaling creates an enriched, pro-inflammatory microenvironment that is favorable for tumor initiation and progression (17). Thus, we find that PRRs are stuck in a "Goldilocks Conundrum." Robust PRR activation is critical in driving the host immune response following PAMP and DAMP exposure; whereas, an overzealous and persistent immune response driven by PRR activation can cause significant collateral damage to the host tissue that ultimately results in chronic inflammation and cancer.

To date, the majority of studies evaluating PRR signaling in cancer have focused on members of the TLR family. However, new and emerging findings have revealed a significant role for members of the NLR family in contributing either directly or indirectly to a variety of hallmarks associated with cancer, including inflammation, cell death, tumor growth, angiogenesis, invasion, and metastasis (18–26). There are at least 23 distinct NLR and NLR-like proteins that have been identified in humans and 34 family members identified in mice (23, 27–29). The NLR proteins function as cytosolic receptors and sensors to detect intracellular PAMPs and DAMPs. Since their discovery, a variety of names have been used to describe the members of this gene family and their respective proteins. For example, these PRRs

have been previously referred to as CATERPILLERs, NOD-like receptors, NACHT-leucine-rich repeats (LRR), and NBD-LRR proteins (28). This resulted in a lack of consistency in the field and resulted in the currently accepted and standardized nomenclature defining the NLRs as the NLR gene family (28). These proteins contain a highly conserved tripartite domain structure (28). The N-terminal domain of the protein is comprised of a variable, but limited number of effector domains that can include combinations of acidic transactivation domains (NLRA proteins), baculoviral inhibitory repeat (BIR)-like domains (NLRB proteins), caspase recruitment domains (NLRC proteins), and pyrin domains (NLRP proteins) (28). These N-terminal domains function to recruit adaptor, intermediary, or effector molecules that drive downstream signaling. The core of the protein is comprised of a conserved NACHT nucleotide-binding domain, which facilitates oligomerization (28). The C-terminal domain of the protein contains multiple LRR elements, which are essential for ligand sensing (28). Each LRR element is typically 28–29 residues in length and each NLR may contain up to 33 individual LRR elements (30, 31).

#### **INFLAMMASOME FORMING NLRs IN CANCER**

One of the most fundamental roles of the NLR family is to regulate pro-inflammatory cytokines and chemokines that drive the host innate immune response to pathogens and environmental insults. Key to this response is the proper regulation of IL-1β and IL-18, which are both potent pro-inflammatory cytokines that affect diverse aspects of health and disease (32–37). Both of these cytokines are generated in an immature pro-form that requires post-translational cleavage for activation. A functional sub-group of NLRs has been identified as driving this process through the formation of a multi-protein complex termed the inflammasome (32, 35, 36). Upon activation, the NLR is thought to undergo a conformational change that allows the recruitment and binding of adaptor and effector proteins and inflammasome formation (35). The inflammasome is composed of an NLR that recognizes a specific repertoire of PAMPs and DAMPs, the adaptor protein ASC, and pro-Caspase-1 (32). These subunits continue to multiplex, ultimately resulting in the maturation and activation of Caspase-1, which subsequently drives the cleavage and activation of IL-1β and IL-18. These inflammasome forming NLRs are by far the best characterized and most highly studied members of the NLR family. To date, at least 6 NLR and NLR-like proteins have been strongly implicated in inflammasome formation, including NLRP1, NLRP3, NLRP6, NLRC4, NLRC5, and the PYHIN family member AIM2 (NLRlike) (32–37). Inflammasome forming NLRs significantly regulate the tumor microenvironment by modulating cytokine production. For example, many of the inflammasome forming NLRs have been shown to significantly attenuate inflammation and tumorigenesis in mouse models of colitis-associated colorectal cancer (CAC) by regulating IL-18 production (18, 19, 21, 22, 38–40). In addition to being a potent pro-inflammatory cytokine, IL-18 is also secreted by epithelial cells to stimulate regeneration and repair and improve barrier function in the colon, thus loss of this cytokine in NLR inflammasome deficient mice enhances tumorigenesis (41). Beyond colon cancer, NLR inflammasome activation may

also play important roles in many other types of cancer, including breast cancer, skin cancers, and virus-associated hepatocellular carcinoma (25, 26, 42–47).

#### **NON-INFLAMMASOME FORMING NLRs THAT NEGATIVELY REGULATE INFLAMMATION**

While the inflammasome forming NLRs are the best characterized members of this PRR family, recent studies have identified a functional sub-group of NLRs that negatively regulate inflammation (48–54). This sub-group is currently composed of three NLR family members, NLRP12, NLRX1, and NLRC3 (**Figure 1**). NLRP12 was one of the first NLR proteins to be described and is the best characterized member of this functional NLR subgroup. NLRP12 was previously known as monarch-1 and PYPAF7 and was originally suggested to form an inflammasome with ASC in overexpression systems (55, 56). In these overexpression studies, transient transfection of NLRP12 and ASC was also shown to induce the transcription of an NF-κB reporter construct (56). Thus, these early *in vitro* studies initially suggested that NLRP12 was an inflammasome forming NLR and a positive regulator of NF-κB signaling. These findings are also consistent with human data that has identified mutations in NLRP12 linked to a spectrum of hereditary periodic fever syndromes. The disorders associated with *NLRP12* mutations are characterized by redox alterations and enhanced secretion of IL-1β, which are similar to the characteristics associated with the family of diseases linked to gainof-function mutations in the *NLRP3* gene (57–59). Interestingly, these diseases are associated with increased caspase-1 activity, are sensitive to therapeutics targeting IL-1β (anakinra), and appear to be independent of NF-κB activation (57–59). However, the ability of NLRP12 to form a functional inflammasome under physiological situations and in the context of human disease appears to occur only under highly specific conditions and is an area of current investigation (60, 61). Indeed, several studies have evaluated NLRP12 inflammasome formation *ex vivo* and using *Nlrp12*−/<sup>−</sup> mice under a variety of conditions and have directly shown that this NLR does not regulate IL-1β/IL-18 maturation (62–69). The prevailing literature associated with NLRP12 indicates that this protein functions as a negative regulator of inflammation by modulating canonical and non-canonical NF-κB signaling (48, 49, 62, 65, 66, 68, 70–73). NLRP12 negatively regulates non-canonical NF-κB signaling through its association with TRAF3 and NF-κB inducing kinase (NIK) (49, 68). This interaction leads to the degradation of NIK and subsequent attenuation of p100 cleavage to p52 (**Figure 1**). Similarly, NLRP12 attenuates canonical NF-κB signaling through the inhibition of IRAK-1 phosphorylation (48, 66, 71) (**Figure 1**). In addition to directly mediating the NF-κB cascade, NLRP12 has also been shown to attenuate ERK signaling, though the exact mechanism has yet to be fully resolved (66, 68). Thus, while some conflicting data has been reported, most issues can be resolved by considering the technical limitations of the assays used to define the respective mechanisms and the specific models being evaluated.

NLRX1 was originally characterized in 2008, and was shown to negatively regulate the host anti-viral immune response (51). NLRX1 is unique among the NLRs due to its mitochondrial localization and its relatively undefined N-terminal domain. Similar

to NLRP12, NLRX1 negatively regulates canonical NF-κB signaling (50, 52) (**Figure 1**). NLRX1 associates with TRAF6 and IκB kinase (IKK) through an activation signal-dependent mechanism (50). Following stimulation, NLRX1 is rapidly ubiquitinated and disassociates from TRAF6 to bind the IKK complex and inhibit subsequent canonical NF-κB activation (50). In addition to attenuating NF-κB signaling, NLRX1 also negatively regulates type-I interferon (IFN-I) signaling through inhibiting the interaction between the PRR Rig-I and the mitochondrial anti-viral signaling (MAVS) protein following virus exposure (50–52, 74, 75) (**Figure 2**). NLRX1 also functions as a positive regulator of autophagy following virus exposure through interacting with the protein TUFM and the mitochondrial immune signaling complex (MISC), which also includes ATG5, ATG12, and ATG16L1 (74, 75) (**Figure 2**). Interestingly, autophagy also functions as a negative regulator of IFN-I signaling and provides an additional route for the negative regulatory properties of NLRX1. In addition to regulating NF-κB and IFN-I signaling, subsequent studies have also shown that NLRX1 functions as a positive regulator of ROS production in epithelial cells following *Chlamydia trachomatis* infection, likely through interactions with the UQCRC2 protein (76, 77) (**Figure 2**). Thus, it is clear that NLRX1 regulation is quite

complex and appears to occur through cell type, temporal and signal-dependent mechanisms.

NLRC3 is the most recently characterized member of this functional sub-group and has been shown to negatively regulate NF-κB and IFN-I signaling (54, 78). NLRC3 was originally identified as a negative regulator of T cell function, in part through delaying the degradation of IκBα (78). Subsequent studies have since revealed that NLRC3 attenuates TLR signaling through interacting with and modulating TRAF6 activity and inhibiting canonical NF-κB signaling (54). NLRC3 has also been recently shown to fine tune the host innate immune response to intracellular DNA, DNA viruses, and c-di-GMP (53). NLRC3 impedes STING-TANK-binding kinase 1 (TBK1) interactions and inhibits STING trafficking, which results in an attenuation of subsequent downstream activation of IFN-I genes (53).

While NLRP12, NLRX1, and NLRC3 each influence a variety of signaling pathways, the convergence on NF-κB signaling appears to be a common strategy among the NLRs in this functional sub-group to attenuate inflammation (**Figure 1**). Additional mechanistic studies have revealed prevalent NLR–TRAF interactions in these models and support the emerging hypothesis that these NLRs function to inhibit NF-κB signaling through the

formation of a multi-protein"TRAFasome"complex (54). Dysregulated NF-κB signaling and the additional pathways modulated by these NLRs are critical features in cancer initiation and progression. Thus, the NLRs that modulate these signaling cascades are highly relevant to cancer pathobiology and additional mechanistic insight will be critical for developing future therapeutic strategies.

#### **NEGATIVE REGULATORY NLRs IN CANCER PATHOBIOLOGY**

While several studies have characterized the contribution of the NLRP3, NLRC4, and NLRP6 inflammasomes in tumorigenesis, significantly less is known regarding the role of NLRs that negatively regulate inflammation. Initial studies have focused on NLRP12. In the context of cancer, somatic mutations in human *NLRP12* have been detected in several large scale screening studies evaluating a variety of cancer sub-types, including glioblastoma, breast cancer, lung squamous cell carcinoma, melanoma, prostate adenocarcinoma, and colon adenocarcinoma (http://cancergenome.nih.gov/). However, broader linkage with specific populations, causation, and mechanism for each mutation has not yet been established. In mice, NLRP12 has been shown to attenuate colorectal cancer. Using the AOM/DSS model of CAC, *Nlrp12*−/<sup>−</sup> mice were shown to develop increased inflammation and tumorigenesis (66, 68). Colon histopathology revealed significant epithelial cell damage and loss of barrier integrity in these

animals, which resulted in increased pro-inflammatory cytokine and chemokine production (66, 68). These animals eventually develop extensive pre-cancerous lesions, which result in significantly increased areas of hyperplasia, dysplasia, and adenocarcinoma (66, 68). These studies revealed that NLRP12 attenuates inflammation and tumorigenesis through negatively regulating NF-κB and ERK signaling (66, 68).

While the overall results of each study are quite complementary, it should be noted that a few mechanistic differences were proposed. In one study, the increased tumorigenesis was attributed to an increase in canonical NF-κB signaling (66). NF-κB signaling was evaluated *in vivo* and in macrophages isolated from wild type and *Nlrp12*−/<sup>−</sup> mice following PAMP stimulation and a significant increase in the levels of p-p105, Rel-A, and p65 activity was observed (66). Furthermore, loss of NLRP12 was shown to significantly increase the transcription of a variety of pro-inflammatory mediators associated with canonical NF-κB signaling and colon tumorigenesis,including *Il-6*,*Tnf-*α, and*Cox2* (66). These findings are consistent with earlier*in vitro* studies,which demonstrated that NLRP12 functions as an antagonist of TLR and TNFR-induced pro-inflammatory signals, in part through inhibiting IRAK-1 hyper-phosphorylation (48). In the second study, NLRP12 was shown to attenuate colon tumorigenesis through negatively regulating non-canonical NF-κB signaling. While some markers of canonical NF-κB signaling were found to be transiently increased in the absence of NLRP12, this study revealed a significant increase in NIK activation and p100 to p52 cleavage in primary cells and in colon tissues isolated from *Nlrp12*−/<sup>−</sup> mice during disease progression (68). These data are highly consistent with previous *in vitro* studies associating NLRP12 activity with NIK suppression and attenuation of non-canonical NF-κB signaling (49, 79). Loss of NLRP12 resulted in a significant increase in *Cxcl12* and *Cxcl13* expression in the colons from *Nlrp12*−/<sup>−</sup> mice (68). These chemokines are highly associated with non-canonical NF-κB activation and cancer (49, 68, 80–82). CXCL12 (SDF-1) and CXCL13 (BLC) and their respective receptors CXCR4 and CXCR5 have been implicated in tumor growth, metastasis, and are critical for the regulation of the tumor microenvironment in multiple cancer sub-types as a component of the tumor "Immunome" (3, 83–85). Regulation of the NF-κB signaling pathway is highly complex. The apparent discrepancies between these two studies can be reconciled by previous findings, which show that non-canonical NF-κB signaling can influence both the canonical pathway and MAPK signaling (86, 87). It is also highly likely that NLRP12 regulates canonical and non-canonical NF-κB signaling through currently undefined cell type, temporal and/or stimuli-specific mechanisms.

To date, neither NLRX1 nor NLRC3 have been directly evaluated in the context of cancer. As previously stated, both of these NLRs negatively regulate NF-κB signaling and would be expected to attenuate tumorigenesis through mechanisms similar to those described for NLRP12. However, each also regulates pathways other than NF-κB that could dramatically influence cancer pathobiology. For example, NLRX1 has been shown to additionally regulate ROS production and autophagy. The dysregulation of oxidative stress signaling is a well-established and important element of tumor development (88). Similarly, autophagy is thought to have a dual function in cancer, where it can attenuate tumor initiation by suppressing tissue damage and inflammation signaling or it can function as a tumor promoter to sustain metabolism, growth, and survival through metabolite recycling (89, 90). Thus, it is highly likely that NLRX1 will contribute to tumorigenesis; however, it is difficult to speculate which of its many biologic functions will have a greater influence on disease pathogenesis.

#### **CONCLUSION**

The recent characterization of this unique sub-group of NLRs that function to attenuate inflammation emphasizes the point that a significant number of the identified NLR proteins in humans have yet to be adequately characterized. Identifying the unique regulatory and signaling pathways modulated by these NLRs is an essential step toward ultimately developing effective therapeutics targeting these proteins and the pathways they modulate. Characterizing unidentified ligands, cell type and temporal regulatory mechanisms, and redundant functions of these NLR family members will significantly improve our understanding of the contribution of these proteins in maintaining immune system homeostasis. It is also clear that NLRs significantly impact cancer pathobiology, beyond colorectal cancer. Additional studies are necessary to better define the contribution of both inflammasome forming NLRs and non-inflammasome forming NLRs in modulating the hallmarks of cancer.

#### **ACKNOWLEDGMENTS**

This work was supported by NIH grant K01 DK092355 and a Virginia-Maryland Regional College of Veterinary Medicine Internal Research Grant. The software application ScienceSlides Suite 2011 (VisiScience) was used to generate **Figures 1** and **2**.

#### **REFERENCES**


resistance to anti-interleukin-1 therapy. *Arthritis Rheum* (2011) **63**:2142–8. doi:10.1002/art.30378


**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.

*Received: 22 March 2014; paper pending published: 24 March 2014; accepted: 29 March 2014; published online: 22 April 2014.*

*Citation: Allen IC (2014) Non-inflammasome forming NLRs in inflammation and tumorigenesis. Front. Immunol. 5:169. doi: 10.3389/fimmu.2014.00169*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Allen. 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.*

## Emerging concepts about NAIP/NLRC4 inflammasomes

#### **Silvia Lucena Lage<sup>1</sup> , Carla Longo1,2, Laura Migliari Branco<sup>1</sup> ,Thaís Boccia da Costa<sup>1</sup> , Carina de Lima Buzzo<sup>1</sup> and Karina Ramalho Bortoluci 1,2\***

<sup>1</sup> Centro de Terapia Celular e Molecular (CTC-Mol), Universidade Federal de São Paulo, São Paulo, Brazil <sup>2</sup> Departamento de Ciências Biológicas, Universidade Federal de São Paulo, São Paulo, Brazil

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Dario S. Zamboni, Universidade de São Paulo, Brazil Amedeo Amedei, University of Florence, Italy Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Karina Ramalho Bortoluci, Departamento de Ciências Biológicas e Centro de Terapia Celular e Molecular (CTC-Mol), Universidade Federal de São Paulo, R. Mirassol, 207 Vila Clementino, São Paulo, São Paulo 04044-010, Brasil e-mail: kbortoluci@gmail.com

Neuronal apoptosis inhibitory protein (NAIP)/NOD-like receptor (NLR) containing a caspase activating and recruitment domain (CARD) 4 (NLRC4) inflammasome complexes are activated in response to proteins from virulent bacteria that reach the cell cytosol. Specific NAIP proteins bind to the agonists and then physically associate with NLRC4 to form an inflammasome complex able to recruit and activate pro-caspase-1. NAIP5 and NAIP6 sense flagellin, component of flagella from motile bacteria, whereas NAIP1 and NAIP2 detect needle and rod components from bacterial type III secretion systems, respectively. Active caspase-1 mediates the maturation and secretion of the pro-inflammatory cytokines, IL-1β and IL-18, and is responsible for the induction of pyroptosis, a pro-inflammatory form of cell death. In addition to these well-known effector mechanisms, novel roles have been described for NAIP/NLRC4 inflammasomes, such as phagosomal maturation, activation of inducible nitric oxide synthase, regulation of autophagy, secretion of inflammatory mediators, antibody production, activation of T cells, among others. These effector mechanisms mediated by NAIP/NLRC4 inflammasomes have been extensively studied in the context of resistance of infections and the potential of their agonists has been exploited in therapeutic strategies to non-infectious pathologies, such as tumor protection. Thus, this review will discuss current knowledge about the activation of NAIP/NLRC4 inflammasomes and their effector mechanisms.

**Keywords: NAIP, NLRC4, flagellin, caspase-1, inflammasomes, lysosomes, cell death**

#### **INTRODUCTION**

Inflammasomes are multiprotein platforms containing specialized cytosolic sensors for a wide range of pathogen-associated molecular patterns (PAMPs) or damage-associated molecular patterns (DAMPs) that are able to activate the inflammatory caspase-1 and caspase-11 (caspase-4 in humans) in a manner dependent or independent of adaptor molecules (1–4). Inflammasomes are composed of a cytosolic receptor from the nucleotide-binding domain-leucine-rich repeat (NBD-LRR) [also named NOD-like receptors (NLR)] or the pyrin and HIN domain-containing protein (PYHIN) families; the adaptor molecule ASC [apoptosisassociated speck-like protein containing a caspase activating and recruitment domain (CARD)]; and pro-caspase-1 or pro-caspase-11. AIM2 is the only member of the PYHIN family described to form inflammasomes. AIM2 is composed of two domains: a C-terminal HIN200 domain and an N-terminal pyrin (PYD) domain. The members of the NLR family contain three domains: a central NBD that is responsible for protein oligomerization and common to all members; a C-terminal region composed of LRR sequences that are supposed to sense PAMPs or DAMPs; and an N-terminal portion that is responsible for the specificity of their molecular interactions and, therefore, their effector functions. The NLR proteins can be classified into NLRBs [NLR containing the baculovirus inhibitory (BIR) domain], NLRCs (NLRs containing the CARD domain), and NLRPs (NLRs containing the PYD domain) (5).

NOD-like receptor proteins are maintained in an autoinhibited state under physiological conditions. After agonist recognition, they undergo a conformational rearrangement, triggering the NBD domains. Then, these proteins expose the effector domain to allow the assembly of oligomeric complexes. The NLRs that lack the CARD domain to recruit and activate pro-caspases-1 and 11 require the assistance of the adapter molecule ASC, which contains the PYD and CARD domains for binding caspases (6, 7). The NLRC members can directly recruit pro-caspase-1 through homotypical interactions between CARD domains, or they can recruit the adaptor ASC to activate caspase-1 (2). The canonical effector mechanisms mediated by caspase-1 are the maturation and secretion of IL-1β and IL-18 and the induction of pyroptosis, a pro-inflammatory form of cell death. Furthermore, caspase-11 seems to be able to induce pyroptosis (8).

After a decade of inflammasome discovery (9), little is known about the molecular complex formed by most members of the NLR family. AIM2, NLRP3, and NLRC4 are the best-characterized inflammasome complexes. The importance of these complexes to control bacterial, viral, fungal, and protozoan infections and their influence in inflammatory processes are gaining prominence in the literature, although their precise activation mechanisms remain to be elucidated. Here, we focus on NLRC4 inflammasomes, the recent advances in the understanding of their assembly and the consequences of their activation to the immune response.

#### **ASSEMBLY AND ACTIVATION OF NAIP/NLRC4 INFLAMMASOMES**

The first reports about the recognition of cytosolic flagellin, the monomeric subunit from flagella present in motile bacteria, demonstrated that the neuronal apoptosis inhibitory protein (NAIP)-5 was responsible for the detection of cytosolic flagellin from *L. pneumophila* and for the restriction of infection (10, 11). In the same year, studies with *S. typhimurium* revealed that another member of the NLR family, NLRC4, was also able to detect cytosolic flagellin (12, 13). NLRC4 was first described in 2001 as a mammalian protein homologous to CED4 of *C. elegans*, whose function is to recruit and activate caspases through its CARD domain (14, 15). Because of the ability to activate caspase-1, previously known as interleukin-1-converting enzyme (ICE), NLRC4 was first named IPAF (ICE-protease-activating factor). Although the involvement of NLRC4 in the control of infections was previously reported, their agonists remained a mystery until 2006.

Flagellin is one of the best-characterized agonists of the innate immune system. Extracellular flagellin is recognized by TLR5 (16) but it can be delivered to the cell cytosol though the secretion systems present in virulent bacteria strains, such as the *S. typhimurium* type III secretion system (T3SS SPI-1) and *L. pneumophila* type IV (T4SS). In the cell cytosol, flagellin induces the formation of the NAIP5/NLRC4 inflammasome, leading to the subsequent activation of caspase-1 (17, 18, 23). Notably, the activation NAIP5/NLRC4 inflammasomes by cytosolic flagellin occurs independently of TLR5 (20),and these two receptors recognize distinct regions of flagellin (16). TLR5 senses a region present in the D1 domain of the protein, whereas the amino acid sequences recognized by NAIP5/NLRC4 inflammasomes are in the D0 domain of the molecule (18, 23, 19, 21, 22).

Previous studies have pointed to the involvement of NAIP5 in controlling *L. pneumophila* flagellated bacteria (24, 25) and to the involvement of NLRC4 in caspase-1 activation and the induction of macrophage death (14, 15), although the role of flagellin in these processes was unidentified at that time. The simultaneous demonstration of cytosolic flagellin recognition by NAIP5 and NLRC4 prompted a model that proposed the existence of two distinct inflammasomes that recognize slight differences in the structure of flagellin (10–13). In 2008, with the advent of NAIP5-deficient mice, Lightfield and collaborators confirmed that NAIP5 is required for NLRC4-containing inflammasome activation in response to *L. pneumophila* infection in a flagellin-dependent manner; however, the NLRC4-mediated macrophage responses against *S. typhimurium* were only partially dependent on NAIP5 (21). A subsequent work from the same group demonstrated that the differential requirement for NAIP5 in response to *S. typhimurium* and *L. pneumophila* infection is not due to intrinsic differences between distinct flagellins, as a genetically engineered *L. pneumophila* developed to express the *S. typhimurium* flagellin also activated the NLRC4 inflammasome in a manner strictly dependent on NAIP5 (17). These data indicated that another agonist from *S. typhimurium* could activate NLRC4 independent of the presence of NAIP5. In fact, these studies confirm that NLRC4 responds to the *S. typhimurium* PrgJ protein independently of NAIP5, thus explaining why NLRC4-mediated responses to *S. typhimurium* are only partially dependent on NAIP5.

The inflammasome structure formed by these proteins was unveiled only recently when two independent groups proposed a model for NAIP5/NLRC4 inflammasome assembly (18, 23). Using the transfection of inflammasome components and microbial molecules in HEK 293T cells or followed by biochemical assays, the authors demonstrated the ability of flagellin from different bacterial species to bind NAIP5. This interaction was dependent on the three leucine residues of the C-terminal portion of flagellin, confirming prior data (17). Furthermore, after the recognition of flagellin, a physical association between NAIP5 and NLRC4 was demonstrated, resulting in the formation of an oligomeric complex. Reconstitution experiments using truncated receptor variants showed that NAIPs are upstream of NLRC4 and suggest that they interact via the NBD domain. Notably, NAIP6 worked similarly to NAIP5, as it induced the oligomerization of NLRC4 in response to flagellin, and this could explain the response of NAIP5−/<sup>−</sup> cells to high concentrations of flagellin. NAIP1 and NAIP2 also recruit NLRC4 in response to the bacterial needle and inner rod proteins of T3SS, respectively (18, 23). Therefore, NAIP proteins seem to be the universal sensors of cytosolic flagellin and secretory complex proteins, whereas NLRC4 acts as an adapter molecule and is responsible for the recruitment and activation of caspase-1. It is noteworthy that there is only one functional NAIP found in humans, which is not activated by flagellin but is able to detect needle proteins of T3SS, similar to NAIP1 (18).

Despite these recent contributions to the understanding of NAIP/NLRC4 assembly, the molecular requirements of bacterial proteins for the formation of the inflammasome complex still requires further clarification. Lightfield et al. (21) originally demonstrated that the final 35 amino acids of the C-terminal portion of the flagellin molecule are essential for the activation of NAIP5. Moreover, the replacement of three leucine residues by alanine in this region abrogated the potential of flagellin to activate NAIP5. However, these studies were based on constructs containing only the C-terminal portion of the flagellin structure. A recent study using whole flagellin with or without these regions have shown that although the three leucine residues were essential for the detection of the C-terminus, their involvement seems to be less important for full-length flagellin recognition, as whole flagellin containing three alanines instead of three leucines still induces cell death and inflammasome complex formation, although fewer complexes are formed (22). Surprisingly, although the absence of the N-terminal domain does not affect the ability of whole flagellin to interact with NAIP5, constructs containing only N-terminus also retain the ability to activate NAIP5/NLRC4. Thus, the molecular interaction between flagellin and NAIP5/6 still requires clarification. Moreover, although flagellin was found inside the NAIP5/NLRC4 complex, as demonstrated by immunoprecipitation (19, 26) and yeast two-hybrid (18) assays, providing a basis for the model of direct interaction between flagellin and NAIP5, our group recently demonstrated the ability of cytosolic flagellin to activate a lysosomal pathway and the requirement of cathepsin B for NLRC4-dependent IL-1β secretion and pyroptosis (27). These observations raise the possibility that NAIP5/NLRC4 can also be activated by cytosolic alterations induced by the

presence of flagellin, as proposed for the activation of the NLRP3 inflammasome (28).

Challenging prior models that hypothesized that LRR domains are responsible for the detection of NLR agonists, a recent study found that these domains are dispensable for the ligand specificity of NAIPs (26). By using a series of chimeric proteins in which the N-terminal domains of NAIP5 or NAIP6 were fused to the C-terminal domains of NAIP2 or vice-versa, the authors demonstrated that NAIP proteins lost the ability to oligomerize with NLRC4 only when NOD domain-associated α-helical domains were absent, suggesting that ligand specificity maps to this region. Interestingly, a similar region in NLRC4 was recently associated with its autoinhibition (29), whereas LRR domain from NAIPs was shown to be required for the maintenance of this protein in an autoinhibited conformation (19). Despite, these unsolved pieces of the puzzle, it has been demonstrated that the interaction of NAIPs with their ligands and the association of NLRC4 with NAIPs induce conformational changes in these molecules that enable their oligomerization and activation (22, 30). Predicted models for the NAIP/NLRC4 inflammasome suggest that these complexes contain an excess of NLRC4 for each NAIP protein (22, 26) and that NLRC4 molecules are able to recruit and activate caspase-1 either directly or through an ASC adapter. The association of pro-caspase-1 with an inflammasomes-containing ASC allows its autoproteolytic cleavage to become an enzymatically active heterodimer capable of processing pro-IL-1β and pro-IL-18 into mature cytokines (2). In contrast, an ASC-independent complex activates caspase-1 without autoproteolysis, which is sufficient for caspase-1 to target a distinct subset of substrates critical for the induction of pyroptosis.

#### **CANONICAL EFFECTOR MECHANISMS INDUCED BY NAIP/NLRC4 INFLAMMASOMES PYROPTOSIS**

The NAIP5/NLRC4 inflammasome is perhaps the best-studied inflammasome complex with regard to resistance to infections. Their involvement has been reported against infections such as *S. typhimurium* (31, 32), *L. pneumophila* (25), *P. aeruginosa* (33, 34), *Y. pestis* (35), *S. flexneri* (36), and *A. veronii* (37). NAIP/NLRC4 mediated responses are related to the restriction of bacterial growth due to the active caspase-1-mediated canonical and noncanonical effector mechanisms, highlighting the importance of this inflammasome as a host defense mechanism against a large number of bacterial infections. The best elucidated effector mechanisms involved in the control of infections mediated by caspase-1 are the secretion of inflammatory cytokines IL-1β and IL-18 and the induction of pyroptosis (38).

The term pyroptosis (from the Greek "pyro" meaning fire or fever, and "ptosis" to a fault) was coined in 2001 to describe a pro-inflammatory programed cell death during *S. typhimurium* infection (39). Morphological and biochemical changes displayed by *S. typhimurium*-infected dying cells were more closely related to those found in classic necrosis compared with those observed during apoptosis, including the following: (1) diffuse DNA fragmentation with no chromatin condensation; (2) early loss of membrane integrity observed by the simultaneous uptake of annexin V with an impermeable membrane dye; (3) lactate dehydrogenase (LDH) release, suggesting a loss of intracellular content; and (4) independence of any apoptotic caspase. Although cells dying by pyroptosis displayed features of necrosis with an inflammatory outcome, the authors found that this process was highly regulated by active caspase-1, as the addition of inhibitors of caspase-1 (z-YVAD-fmk) abolished *S. typhimurium*-induced cell death.

The induction of pyroptosis by pathogenic bacteria depends on an active secretion system that translocates bacterial proteins into the cell cytosol, such as the T3SS (SPI-1) of *S. typhimurium* and type IV (T4SS) of *L. pneumophila* (12, 13, 40–42). Mutant *L. pneumophila* (43) or *P. aeruginosa* (34, 44) lacking flagellin fail to activate caspase-1 and, therefore, are not able to induce pyroptosis and IL-1β secretion in infected macrophages. Accordingly, the transfection of purified flagellin from *L. pneumophila* and *S. typhimurium* directly into the cell cytosol is sufficient to trigger caspase-1-dependent pore formation, pyroptosis, and IL-1β secretion (45, 46). Importantly, infection with the nonflagellated bacteria *S. flexneri* also induces NLRC4-mediated pyroptosis, most likely in response to the inner rod component of T3SS (36).

Although the molecular mechanisms that regulate pyroptosis remain to be elucidated, the model of *S. typhimurium* infection has given us important knowledge about this form of cell death. The cell lysis observed during pyroptosis seems to result from a highly regulated process of pore formation in the plasma membrane (45, 46). Pores dissipate cellular ionic gradients but allow the retention of larger cytoplasmic constituents, leading to increased liquid osmotic pressure and water influx. These events are followed by cell swelling and subsequent osmotic lysis with the release of intracellular contents, which are potentially inflammatory (45, 46). Caspase-1-dependent DNA cleavage also occurs during pyroptosis (45, 47). However, the DNA cleavage observed during *S. typhimurium*-induced pyroptosis is independent of caspase-activated DNase (CAD) (45, 47), unlike what is observed during apoptosis, in which the proteolysis of inhibitor of CAD (ICAD) by apoptotic caspases mediates the release of CAD to the nucleus, where it cleaves DNA between nucleosomes. Therefore, pyroptotic cells do not display the typical pattern of oligonucleosomal fragmentation observed during apoptosis, a fact that can be used to distinguish between these two processes of cell death (48).

There is good evidence implicating pyroptosis as an important host defense mechanism mediated by NAIP/NLRC4 that clears intracellular pathogens*in vitro*. The death of infected macrophages by pyroptosis seems to correlate with a rapid loss of the replicative niche and high bacterial loads are recovered from macrophages deficient in components of inflammasomes or infected with mutant bacterial strains that fail to trigger their activation [reviewed by Bortoluci and Medzhitov (1) and Bergsbaken et al. (49)]. Moreover, a study conducted *in vivo* demonstrated that the NLRC4-dependent flagellin-mediated lysis of bacteria-containing macrophages not only results in the early loss of the intracellular replication niche but also creates an inflammatory milieu with the recruitment of effector cells to the infection site, which are involved in pathogen clearance (32). Although the possible targets of caspase-1 and caspase-11 mobilized during pyroptosis remain unidentified, the studies involving NAIP/NLRC4 hugely contribute to the idea that this inflammatory form of cell death is an important effector mechanism against infections.

#### **IL-1**β **AND IL-18 SECRETION**

IL-1 was the first identified cytokine and has been related to several inflammatory processes. IL-1 plays a role in virtually all cells and organs, ranging from fever and resistance to microorganisms to the activation of the hypothalamus–pituitary–adrenal axis (HPA) (50–56). IL-18 was first described in 1989 as a potent IFN-γ-inducing factor and an important component of polarized type-1 T helper cells (Th1) and type-1 macrophages (M1) responses, cells with a pro-inflammatory profile (57–59). Macrophages, monocytes, lymphocytes, keratinocytes, microglia, neutrophils, dendritic cells, and other cells are described as important sources of IL-1β and IL-18 (60–64). IL-18 and IL-1β have similar processing; they are both synthesized in an inactive form that requires processing by active caspase-1 to become biologically active (61, 65, 66). Although extensively studied, the mechanism responsible for IL-1β and IL-18 release has not been fully elucidated. These cytokines can be passively released during cell lysis; however, there is recent evidence supporting the existence of active mechanisms involved in the secretion of IL-1β and IL-18, such as caspase-1-induced membrane pores, vesicle shedding and lysosomal exocytosis (45, 49).

Although the precise effector mechanisms of IL-1β and IL-18 remain to be elucidated, these cytokines have been reported to be important mediators induced by NAIP/NLRC4 to host resistance to bacterial infections (67). In addition to the effects of IL-1β and IL-18 in the activation and recruitment of innate immune cells, these cytokines have important roles in the activation and differentiation of T lymphocytes (52). IL-1β and IL-18 have been shown to drive the establishment of T CD4<sup>+</sup> adaptive responses in mice and in humans and are responsible for the differentiation of Th17 and Th1, respectively (68–70). However, little is known about the involvement of IL-1β and IL-18 in NAIP/NLRC4-induced adaptive immune responses. Kupz et al. demonstrated that IL-18, when produced by the activation of NLRC4 during infection by *S. typhimurium*, is required for the activation of non-cognate CD8<sup>+</sup> T cells and the production of IFN-γ (71), supporting a role for this cytokine in the induction of cellular responses.

Additional evidence of the role of NAIP/NLRC4 in the activation of T cells came from an experimental vaccination with irradiated flagellin-expressing tumor cells. Authors demonstrated that the immunization of mice with flagellin-fused tumor cells induced tumor-specific CD4<sup>+</sup> and CD8<sup>+</sup> T cell responses and prevented parental tumor growth. Despite the well-known role of TLR5, the recognition of flagellin by the NAIP5/NLRC4 inflammasome was also required for the induction of a protective CD8<sup>+</sup> T cell response and tumor suppression. Although the NAIP5/NLRC4 inflammasome-mediated IL-1β secretion in response to the injection of flagellin-modified tumor cells, it is unclear whether the involvement of this cytokine was necessary for the success of this immunotherapy. The role of IL-1β and IL-18 in tumorigenesis remains controversial. There is strong evidence supporting pro-tumorigenic properties of these cytokines via the induction of chronic inflammation. Although the induction of Tregs and Th17 could impair the immune response against tumor cells, it is reasonable to consider that the activation of Th1 and cytotoxic CD8 T cells by IL-1β and IL-18 may be beneficial to the host (72, 73).

#### **EMERGING EFFECTOR MECHANISMS MEDIATED BY THE NAIP/NLRC4 INFLAMMASOME HUMORAL EFFECTOR MECHANISMS**

In addition to the well-characterized functions of NAIP/NLRC4 inflammasomes described above, non-canonical effector mechanisms have emerged. Recent data describe a range of effector functions mediated by NAIP/NLRC4 inflammasomes that operate independently of IL-1β, IL-18 and pyroptosis. The NAIP5/NLRC4 inflammasome has been implicated in the activation of phospholipase A2 (cPLA2) with a consequent production of lipid mediators, such as prostaglandins and leukotrienes (74). Authors demonstrated that systemic cytosolic flagellin stimulation leads to an "eicosanoid storm" that initiates inflammation and the loss of vascular fluids, resulting in a very fast death in mice. Of note, these effects are mediated by NAIP5/NLRC4 and occur independently of IL-1β/IL-18 or pyroptosis.

Inflammasomes have also been implicated in the active secretion of endogenous molecules known as DAMPs, challenging the idea that these molecules are only passively released during the process of cell lysis (75). IL-1α is an alarmin, whose release has been recently linked to inflammasomes. Both IL-1β and IL-1α present some common features, such as belonging to the same family, synthesis in the cytoplasm and secretion by an unconventional pathway independent of the endoplasmic reticulum and Golgi complex (55); additionally, they are released simultaneously by various stimuli, and they act on the same receptor, IL-1R1, thus sharing some biological functions (52). However, despite these similarities, there are some important differences in the production, secretion, and function of these cytokines. Unprocessed forms of both IL-1α and IL-1β are thought to be produced in response to TLR ligands, but they have distinct activities. Unlike IL-1β, which needs to be processed by caspase-1 to become biologically active (65), the uncleaved form of IL-1α is able to engage IL-1R1 (60, 76), although it's full activity seems to require cleavage by calpain (77). Although IL-1α is not a substrate for caspase-1, there are some reports that have demonstrated that macrophages from caspase-1-deficient mice release less IL-1α (27, 78–80), suggesting the involvement of inflammasomes.

The mechanism by which caspase-1 mediates IL-1α secretion is still a matter of debate. A recent study demonstrated that the requirement of inflammasomes for IL-1α secretion depends on the nature of agonists (81). Caspase-1 has been described as a shuttle that facilitates the secretion of leaderless proteins, such as IL-1α (80). However, it is not clear whether active caspase-1 is the shuttle itself or whether it activates another enginery that is dependent on its activity, e.g., IL-1β (82) or IL-1R2 (77), as has been proposed for the secretion of IL-1α in response to NLRP3 agonists. The involvement of NLRC4 inflammasomes in IL-1α secretion is poorly understood. In one previous study, infection by *S. typhimurium* resulted in NLRC4- and caspase-1-dependent secretion of IL-1α (81). Interestingly, in contrast with most of the NLRP3 agonists, the secretion of IL-1α in response to *S. typhimurium* was completely independent of ASC, indicating a differential requirement for this adaptor molecule in cytokine secretion in response to NLRC4 agonists,as IL-1β is entirely dependent on ASC (2). However, Barry et al. showed that IL-1α initiates the inflammatory response driven by *L. pneumophila* independent of caspase-1 and NLRC4 (83). We recently reported that the activation of macrophages with purified flagellin inserted into lipidic vesicles induced IL-1α secretion in a manner partially dependent on caspase-1 and cathepsin B (27). Therefore, the reasons for the discrepancies in the literature and the precise mechanisms involved in the cross talk between IL-1α and NLRC4/caspase-1 axis remain to be addressed.

Another factor whose secretion has been linked to inflammasomes is the"High Mobility group box-1" (HMGB-1). HMGB-1 is a nuclear protein involved in the regulation of nucleosome function and DNA transcription that functions as an inflammatory mediator when released to the extracellular milieu (84). Lamkanfi et al. reported a critical role for HMGB-1 secreted through the NLRP3/ASC/caspase-1 axis in LPS-induced endotoxic shock (85). Interestingly, macrophages infected with *S. typhimurium* released significant amounts of HMGB-1 in a NLRC4 and caspase-1 dependent manner but independently of ASC, which is similar to previous reports of IL-1α secretion (81). During pyroptosis induced by a variety of stimuli, including *S. typhimurium* infection, HMGB-1 did not undergo caspase-1-mediated processing before its secretion, but extracellular HMGB-1 was hyperacetylated at the nuclear localization sequences (NLSs) (86). Because this translational modification is essential for HMGB-1 translocation from the nucleus to the cytoplasm (87, 88), HMGB-1 release upon inflammasome activation seems to be a coordinated process. More recently, Nystrom et al. (89) reported that NLRC4-mediated pyroptosis is the prevalent factor in the regulation of HMGB-1 secretion, leading to the release of the chemotactic acetylated HMGB-1 isoform without requiring TLR-derived priming. Although the mechanisms by which inflammasome components can regulate DAMPs secretion still need to be better understood, DAMPs are already considered important therapeutic targets because of their role in host resistance against infection and their involvement in inflammatory disorders.

With respect to antibodies production NLRC4, NAIP5, and caspase-1 have been reported to have a redundant role with TLR5 in the induction of total IgG (90) or IgG1 (91) against flagellin or co-administered OVA and an additive effect to TLR5 in the induction of IgG2a (91). In the absence of MyD88, in which TLR5, IL-1β, IL-1α, and IL-18 signaling is compromised, the production of antibodies induced by flagellin was reduced but not abolished, and a large amount of antibodies was still produced (91). The same results were obtained with TLR5/caspase-1 double-knockout mice (91), supporting previous data that demonstrated that no significant difference was observed in specific anti-flagellin IgG titers in mice deficient for IL-18 (92) or IL-1R (93). These reports suggest that some yet-undiscovered mechanism that acts in addition to TLR5 and inflammasome-mediated cytokines could be involved in the adjuvant properties of flagellin, requiring new investigations into this agonist.

#### **CELLULAR EFFECTOR MECHANISMS**

In addition to inflammatory mediators and cell death processes, some cellular effector mechanisms mediated by NLRC4 have emerged. Previous studies from our group described a requirement of NAIP5, NLRC4, and caspase-1 for the activation of inducible nitric oxide synthase (iNOS) and nitric oxide (NO) secretion in response to cytosolic flagellin (94). Interestingly, cytosolic flagellin-induced iNOS activation is preserved in the absence of MYD88, ruling out the participation of TLR5, IL-1β, and IL-18. Moreover, NO secretion through the NAIP5/NLRC4 caspase-1 axis in response to flagellin is involved in the control of *L. pneumophila* (94) and *S. typhimurium* (unpublished data from our group) by macrophages, pointing to this pathway as an additional effector mechanism mediated by NAIP5/NLRC4.

Autophagy is another effector mechanism used by NAIP5/NLRC4 to control *L. pneumophila*. In the presence of NAIP5, NLRC4 macrophages present a rapid turnover of LC3<sup>+</sup> *L. pneumophila*containing vesicles, preventing the establishment of secondary infections (95). This response is mediated by the detection of flagellin, and the inhibition of autophagy in macrophages infected with flagellin-sufficient *L. pneumophila* increased the rate of pyroptosis in these cells. These data confirm a previous study that demonstrated that NLRC4 plays a role in the regulation of autophagy by binding Beclin-1 in steady-state conditions (96). Because the initiation of autophagy seems to precede the induction of pyroptosis, autophagy can be considered a pathway through which macrophages raise the threshold of contaminants necessary to result in the loss of cell by inflammatory cell death. NAIP5/NLRC4 can also restrict flagellin-competent *L. pneumophila* replication by promoting the delivery of *L. pneumophila*containing phagosomes (LCP) to lysosomes for degradation (43, 97). In the absence of NAIP5/NLRC4/caspase-1, LCP avoids fusion with lysosomes, which allows the pathogen to exponentially replicate inside macrophages. This effect is dependent on caspase-1 mediated caspase-7 processing and does not require IL-1β/IL-18 and the classical apoptosis pathway involving caspase-8 and -9 (98). These data corroborate a previous report that demonstrated a requirement of NLRC4, caspase-1, and ASC for caspase-7 processing during infection with flagellin-competent *S. typhimurium* (99). NLRC4 and ASC-dependent caspase-8 proteolysis was also reported during *S. typhimurium* infection (100). Interestingly, caspase-8 contributes to *Salmonella*-induced IL-1β production, but it is dispensable for inducing pyroptosis, whereas caspase-1 processes pro-IL-1β and coordinates pyroptosis. These data highlight the fact that inflammasomes are dynamic complexes that are able to recruit distinct members of the caspase family to induce diverse effector functions in response to *Salmonella* infection.

Similar to what has been demonstrated during apoptosis (101, 102) and necrosis (103), the cleavage of PARP1 (also called ARTD1) was also observed during pyroptosis induced by *S. typhimurium* (104). PARP1 processing in *S. typhimurium*-infected macrophages was abrogated in *Nlrc4*−/<sup>−</sup> but not in *Nlrp3*−/<sup>−</sup> cells, consistent with the role of the NAIP5/NLRC4 inflammasome in the induction of pyroptosis during *S. typhimurium* infection (12, 31, 105). PARP1 is a nuclear chromatin-associated multifunctional enzyme that catalyzes the polymerization of ADP-ribose units from donor NAD<sup>+</sup> molecules (106, 107). Although it has been

historically studied in the context of genotoxic stress signaling and consequent apoptosis, PARP1 has been related to chromatin structure regulation, transcription, and chromosomal organization (108, 109). Previous reports showed that inflammasomes are able to use PARP1 to induce the transcription of NF-κB-dependent target genes independently of any type of programed cell death (110). Upon LPS stimulation, caspase-7 is activated by caspase-1, which is translocated to the nucleus to induce PARP1 cleavage at the promoters of a subset of NF-κB-dependent target genes that are negatively regulated by PARP1. Mutating the PARP1 cleavage site D214 renders PARP1 uncleavable and inhibits PARP1 release from chromatin and, therefore, chromatin decondensation, thereby restraining the expression of cleavage-dependent NF-κB target genes, such as *il-6*, *cfs2*, and *lif*, but not *ip-10* (110). Preliminary and unpublished data from our group suggest the involvement of caspase-1-dependent PARP1 cleavage in iNOS gene expression upon cytosolic flagellin stimulation, as iNOS expression is significantly reduced in macrophages that harbor non-cleavable PARP1 (D214N). This is important evidence of the involvement of inflammasomes in epigenetic regulation and gene expression, although many of these outputs require further evaluation.

An important process of lysosomal exocytosis occurs during pyroptosis. Bergsbaken and Cookson (111) demonstrated that caspase-1-mediated pore formation induced during *S. typhimurium* infection promotes an influx of extracellular calcium, which is critical for lysosomal exocytosis. The release of lysosomal proteases with known antimicrobial activity contributes to the control of extracellular bacteria. In addition to the effect of lysosomal contents in the extracellular compartment, recent data from our group demonstrated that cytosolic flagellin is also able to activate a lysosomal pathway that culminates in an inflammasome-independent inflammatory form of cell death. This inflammasome-independent cell death induced by cytosolic flagellin is regulated by cathepsins B and D and is temporally correlated with the restriction of *S. typhimurium* infection by macrophages (27). Together, these data indicate a cross talk between lysosomes and NAIP/NLRC4 inflammasomes that impact the control of bacterial infections and opens new avenues for the development of inflammasome-based therapeutic strategies for non-infectious pathologies such as tumors. In fact, lysosomes have been considered important targets for the development of antitumor drugs (112). Lysosomes from cancer cells appear to be less stable than normal cells, which has given rise to the development of therapies based on lysosomotropic detergents. In this sense, flagellin could be an alternative that in addition to the induction of lysosomal cell death, is able to mediate several effector mechanisms as described throughout this review (**Figure 1**).

#### **CONCLUSION AND FUTURE DIRECTIONS**

More than 10 years after their discovery (14, 15), the molecular mechanisms involved in the activation of NAIP/NLRC4 began to be elucidated (18, 19, 26). From two distinct inflammasome complexes, NAIPs emerged as universal sensors for cytosolic bacterial proteins, whereas NLRC4 became an adaptor molecule responsible for the recruitment and activation of caspase-1. At the same time, in addition to NAIP5, novel NAIPs members were described,

inflammasome-independent cell death that contributes to macrophage control of infection and regulation of NAIP5/NLRC4-dependent

responses.

Beclin-1, thus inhibiting autophagy. When flagellin is detected by NAIP5/6, NLRC4 is recruited to assembly inflammasome complex and release Beclin-1 to initiate autophagy. As a host protection response,

amplifying the potential of these proteins to detect bacterial infections (18, 19, 113, 114). Despite this important information, the molecular signatures of agonists recognized by NAIP/NLRC4 inflammasomes still require further study. Moreover, NLRC4 has been associated with host resistance against a mucosal *Candida albicans* infection (115) and in a colitis-associated colorectal cancer (CAC) model (116, 117). Interestingly, in both cases, NLRC4 seems to exert a protective role in non-hematopoietic compartments. However, the precise mechanism of NLRC4 activation in these models is unknown, raising the possibility that NLRC4 functions as an adaptor molecule for other NLR members in addition to NAIP and providing new insights into inflammasome signaling.

NAIP/NLRC4 are most likely the best-studied inflammasomes in the context of host resistance against infections. In addition to the extensively described IL-1β and IL-18 secretion and pyroptosis, other important effector mechanisms mediated by these inflammasomes have recently emerged (**Figure 1**). Moreover, flagellin, the best studied NAIP/NLRC4 ligand, has been reported to activate distinct pathways, such as autophagy (95) and a lysosome pathway (27) (**Figure 1**). Although the precise mechanism involved in the lysosome disruption by flagellin is still under investigation, it culminates in an inflammatory process of cell death that is accompanied by IL-1α secretion and contributes to the control of *S. typhimurium* by macrophages. This peculiar process of cell death occurs in the absence of inflammasome components. Additionally, the inhibition of cathepsin B disrupted IL-1β secretion and pyroptosis in response to cytosolic flagellin, indicating a role for lysosomal proteases in the regulation of NAIP/NLRC4 dependent responses. Because human cells do not express NAIP5 or NAIP6 (18), the activation of the lysosomal pathway by flagellin might be an alternative pathway used when human cells interact with flagellated bacteria that reach cell cytosol. In the context of therapeutic strategies, this knowledge could be an important gain, as the immune properties of flagellin have been extensively exploited in different models. At least in the context of antitumor vaccination (118) and antibody production (90, 91), the protective and adjuvancy roles of flagellin require its cytosolic detection. Together, these reports open up new avenues to explore the immune potential of NAIP/NLRC4 agonists as therapeutic targets.

#### **ACKNOWLEDGMENTS**

This work was supported by Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP – Brazil) Proc 2013/16010-5, Brazilian Research Council (CNPq-Brazil), CAPES, and INCTV.

#### **REFERENCES**


implicated in cytokine and chemokine responses following stressor exposure. *Brain Behav Immun* (2013) **28**:54–62. doi:10.1016/j.bbi.2012.10.014


**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.

*Received: 24 April 2014; accepted: 17 June 2014; published online: 02 July 2014.*

*Citation: Lage SL, Longo C, Branco LM, da Costa TB, Buzzo CdL and Bortoluci KR (2014) Emerging concepts about NAIP/NLRC4 inflammasomes. Front. Immunol. 5:309. doi: 10.3389/fimmu.2014.00309*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Lage, Longo, Branco, da Costa, Buzzo and Bortoluci. 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.*

## **Targeting C-type lectin receptors for cancer immunity**

#### *Huimin Yan1,2, Tomomori Kamiya1,3, Papawee Suabjakyong1,4 and Noriko M. Tsuji <sup>1</sup> \**

*1 Immune Homeostasis Laboratory, Biomedical Research Institute, National Institute for Advanced Industrial Science and Technology (AIST), Tsukuba, Japan, <sup>2</sup> Institute for Liver Disease, Fifth Hospital of Shijiazhuang, Shijiazhuang, China, <sup>3</sup> Research Institute for Biomedical Sciences, Tokyo University of Science, Noda-shi, Japan, <sup>4</sup> Department of Clinical and Analytical Biochemistry, Graduate School of Pharmaceutical Sciences, Chiba University, Chiba-shi, Japan*

C-type lectin receptors (CLRs) are a large family of soluble and trans-membrane pattern recognition receptors that are widely and primarily expressed on myeloid cells. CLRs are important for cell–cell communication and host defense against pathogens through the recognition of specific carbohydrate structures. Similar to a family of Toll-like receptors, CLRs signaling are involved in the various steps for initiation of innate immune responses and promote secretion of soluble factors such as cytokines and interferons. Moreover, CLRs contribute to endocytosis and antigen presentation, thereby fine-tune adaptive immune responses. In addition, there may also be a direct activation of acquired immunity. On the other hand, glycans, such as mannose structures, Lewis-type antigens, or GalNAc are components of tumor antigens and ligate CLRs, leading to immunoregulation. Therefore, agonists or antagonists of CLRs signaling are potential therapeutic reagents for cancer immunotherapy. We aim to overview the current knowledge of CLRs signaling and the application of their ligands on tumor-associating immune response.

#### *Edited by:*

*Anton G. Kutikhin, Research Institute for Complex Issues of Cardiovascular Diseases, Russia*

#### *Reviewed by:*

*Jian Zhang, The Ohio State University, USA Dirk Werling, Royal Veterinary College, UK*

#### *\*Correspondence:*

*Noriko M. Tsuji, National Institute of Advanced Science and Technology, Biomedical Research Institute, Tsukuba, 305–8566, Ibaraki, Japan nm-tsuji@aist.go.jp*

#### *Specialty section:*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology*

> *Received: 01 April 2015 Accepted: 26 July 2015 Published: 24 August 2015*

#### *Citation:*

*Yan H, Kamiya T, Suabjakyong P and Tsuji NM (2015) Targeting C-type lectin receptors for cancer immunity. Front. Immunol. 6:408. doi: 10.3389/fimmu.2015.00408* **Keywords: C-type lectin receptors, innate immunity, cancer immunity, immunoregulation**

### **Introduction**

Interaction between tumors and the immune system is a complex and dynamic process. The immune system consists of innate and adaptive immunity whose cooperative interactions are required for eliminating pathogens efficiently. Similar protective mechanisms are effective against cancer cells; the endogenous non-self which potentially grow into harmful cell mass. To prevent and suppress such tumor progression, the immune system utilize host defense mechanisms (1, 2).

Protecting self from harmful pathogens, and facilitating the symbiosis with harmless environmental microorganisms are the original mission of immune system. Above all, the innate immune system provides the first line of host defense against invading pathogens, with use of soluble factors, anti-microbial peptides, compliments, and natural antibodies. Initial activation of innate immune cells are mediated via pattern recognition receptors (PRRs) by recognizing characteristic structures of microorganisms (3, 4). Known PRRs are categorized into Toll-like receptors (TLRs), Nod-like receptors (NLRs), RIG-I-like receptors (RLRs), C-type lectin receptors (CLRs), and cyclic GMP–AMP synthase (cGAS) that has been recently identified.

Toll-like receptors and CLRs are involved in antigen capture, presentation, and activation of immune responses by enhancing cytokine/chemokine production and up-regulation of MHC class II molecules (5–7). NLRs predominantly recognize microbial products and endogenous danger signals, and enhance caspase activity to produce activated IL-1β (8). RLRs and cGAS are involved in cytosolic recognition of nucleic acids and other microbial components, i.e., RLRs are sensors of

cytosolic dsRNA and cGAS are sensors of DNA, respectively, and both induce type I IFN production (9, 10).

C-type lectin receptors are a large family of receptors that encompass upwards of 1000 members with diverse functions including cell adhesion, complement activation, tissue remodeling, platelet activation, endocytosis, phagocytosis, and activation of innate immunity (11, 12). CLRs contain one or more C-type lectin-like domains, which are important for the recognition of specific carbohydrate structures of pathogens and self-antigens (13). Because of their specificity for glycans, such as mannose structures, Lewis-type antigens, or GalNAc (14, 15), CLRs may also mediate specific interactions with tumor antigens and facilitate tumor rejection. On the other hand, tumor cells devise multiple strategies to inhibit effector anti-tumor immune responses through modulating CLRs signaling (16, 17). It is therefore important to identify CLRs signaling toward immune evasion and regulate them in a specific way, while making the best application of beneficial side of CLRs signaling to mount anti-tumor immunity (**Figure 1**).

### **The Immune Regulation by CLRs and Signaling Pathways**

C-type lectin receptors are widely expressed on myeloid cells, such as macrophages, neutrophils, and dendritic cells (DCs). They contain one or more C-type lectin-like domains, which are important for recognition and internalization of glycosylated antigens. Ligand activation of CLRs initiates intracellular signaling pathways that regulate the immune response. Mounting evidence has been shown that CLRs play roles in sharping innate immune response. Many CLRs such as dectin-1, dectin-2, dectin-3, Mincle, and DEC-205 have been demonstrated to trigger cellular immune responses, including DC maturation, chemotaxis, reactive oxygen species production, and inflammasome activation (18, 19). The innate immune cells stimulated through CLRs acquire the capacity to secrete pro-inflammatory and anti-inflammatory cytokines such as TNF-α, IL-12, IL-6, IL-1β, and IL-10 (20–22). On the other hand, ligand engagement of some CLRs, such as MICL and DCIR, has inhibitory effects on host immunity through controlling DC maturation, activation, and proliferation (23–25).

The ability of CLRs to exhibit activation or inhibition of immune response is regulated by the specific motifs in their cytoplasmic tails. Intracellular signaling through CLRs with immunereceptor tyrosine-based activation motif (ITAM) domains result in cell activation, whereas CLRs which possess immune-receptor tyrosine-based inhibition motif (ITIM) domains usually mediate inhibitory functions (18, 26). The tyrosine residues are phosphorylated by Src family kinases and a tri-molecular complex composed of CARD9, Bcl10, and MALT1 is involved in the subsequent activation of NF-κB and expression of inflammatory

Frontiers in Immunology | www.frontiersin.org 84 August 2015 | Volume 6 | Article 408

contribute to enhance anti-tumor immunity via two independent mechanisms.

cancer.

cytokines (6, 27, 28). Syk/CARD9 pathway is utilized by dectin-1, dectin-2, dectin-3, or Mincle and plays important roles in bridging the innate immunity and adaptive immunity. Dectin-1 directly signals through Syk using cytoplasmic ITAM and activates NFκB, whereas dectin-2, dectin-2/dectin-3 heterodimer, and Mincle couple to Syk via the FcRγ and mediate NF-κB activation (29– 32) (summarized and depicted in **Figure 2**). Signaling through Syk/IRF5 is crucial for the production of dectin-1-mediated IFNβ (33). Furthermore, it is reported that dectin-1 activates inflammasomes and caspase-1, leading to production of IL-1β (34).

Stimulation of these CLRs has been shown to drive the development of Th1, Th17, and CD8<sup>+</sup> cytotoxic T lymphocytes (CTLs) cells immune responses through triggering the production of multiple cytokines (26, 35–37). In particular, dectin-1 has been found to activate NFAT also and enhance IL-2 and IL-10 production in DCs (38). A further study found that Src-homology phosphatase (SHP)-2 is an essential component, which facilitates the recruitment of Syk to the dectin-1 or the ITAM-containing adaptor FcRγ of dectin-2/3 and Mincle, and mediates the induction of Th17 responses (39). Given that T-cell immunity is essential for anti-tumor immunity, activation of ITAM-based CLRs signaling should support the development of protective immunity.

Recently, the important role of CLRs in inducing immunological tolerance has also been demonstrated. In the case of inhibitory CLRs containing ITIMs, such as DCIR (on dendritic cells) or MICL (on granulocytes and monocytes), SHP is an essential element. Ligation of these CLRs results in phosphorylation of ITIM domain, leading to SHP-1 and SHP-2 activation and inhibits cellular activation (25). Ligation of DCIR increases the number and function of Foxp3<sup>+</sup> Treg cells, thus attenuates airway hyper responsiveness and inflammation (40). BDCA-2 and DC-SIGN do not contain a cytoplasmic ITIM motif but signaling through these CLRs has been shown to modulate TLR signaling through alternative pathways (41) and be critical for the maintenance of Foxp3<sup>+</sup> Treg cells (42, 43). Moreover, several CLRs such as DC-ASGPR, SIGNR1, and dectin-1 are shown to play an important role in triggering IL-10-producing suppressive CD4<sup>+</sup> T cells (44– 47). Recently, it is highlighted that inflammation-induced cancers are prevented by anti-inflammatory mechanisms including Tregs (48). Therefore, the anti-inflammatory pathway lead by CLRs activation may also become a therapeutic strategy for reducing the risk of such diseases (**Figure 1**).

### **Recognition of Tumor-Associated Antigen by CLRs**

Tumors are recognized by the immune system through tumor antigens, including membrane proteins and altered carbohydrate molecules of glycoproteins or glycolipids on the cell surface (49). Tumor-associated carbohydrate antigens (TACAs) can be specifically recognized by CLRs. It has been shown that DC-SIGN recognizes carcinoembryonic antigen (CEA), a well-known tumor-associated antigen overexpressed on almost all human colorectal, gastric, and pancreatic adenocarcinomas, 70% of nonsmall cell lung carcinomas, and 50% of breast carcinomas A (50). DC-SIGN also exhibits high affinity for Mac-2-binding protein (Mac-2BP), which increases in patients with pancreatic, breast, and lung cancers (51).

Macrophage galactose type C-type lectin (MGL) is involved in the recognition and binding of tumor-associated Neu5Ac-Tn and Neu5Gc-Tn antigens (52). It has also been demonstrated that DCs are able to recognize cancer-specific glycosylation changes of the mucin 1 (MUC1), in particular, the carbohydrate sialyl Lewis X, and the sialyl TN epitope through MGL and DC-SIGN (53, 54). In addition, MUC1, CA-125, and TAG-72 show strong binding activity to mannose receptor (MR) and induce its internalization (55–57). Further, mannose-binding lectin (MBL) has been shown to recognize glycoproteins from a human colorectal carcinoma cell line in a fucose-dependent manner (58–60).

A critical role of dectin-1, a receptor for β-glucans (61, 62), has recently been shown in recognition of N-glycan structures on tumor cells. N-glycosidase treatment markedly reduced the binding of dectin-1 to tumor cells. Importantly, tumoricidal activity of splenocytes was reduced when tumor cells were pretreated with N-glycosidase (63).

Plasmacytoid dendritic cells (pDCs) are responsible for production of type I interferons (IFN-α and β), type III IFNs (IFNλ/IL-28/29), and pro-inflammatory cytokines. Antigen presentation by CpG-activated pDC influenced anti-tumor immune responses by promoting efficient Th17 differentiation (64). A study showed that BDCA-2 exclusively expressed on pDCs binds tumor cells via asialo-oligosaccharides containing terminal residues of galactose (65) and potently suppresses the ability of pDCs to produce type I IFNs. Such direct regulation and/or cross-regulation of TLRs signaling by BDCA-2, an inhibitory CLR, may also suppress beneficial adaptive immune response *in vivo* (**Figure 3**).

### **CLRs in Induction of Anti-Tumor Immune Response**

Effective immunological eradication of tumors requires NK cells and tumor-specific CD8<sup>+</sup> and CD4<sup>+</sup> T cells. The potential role of CLRs improving anti-tumor activity of immune cells has been investigated. A study showed that MGL interacts with tumorassociated Tn antigens and efficiently internalized with antigens for presentation to CD4<sup>+</sup> T cells (5). Furthermore, engagement of MGL using α-*N*-acetylgalactosamine-carrying tumor-associated antigens promotes the up-regulation of maturation markers of DCs, decrease phagocytosis, enhance motility, and most importantly increase antigen-specific CD8<sup>+</sup> T-cell activation (54).

DC-SIGN is another important CLR in inducing anti-tumor immune responses. It is reported that Lewis X oligosaccharides–heparanase complex activate and enhance the maturation of DCs, leading to enhancement of antigen-specific IFN-γ production and cytotoxic T-cell response. Furthermore, the modified DCs also significantly suppress the established tumor growth and prolong the life span of tumor-bearing mice (66). In addition, glycan-modified liposomes lead to efficient antigen presentation of DCs in the presence of LPS and augment CD4<sup>+</sup> and CD8<sup>+</sup> effector T-cell activation via DC-SIGN-dependent pathway (67). The potency of MR to improve anti-tumor immune responses has also been conducted. Cross-presentation of antigen and strong antigen-specific immune response were induced by conjugation of glycan ligands to MR (68), which resulted in an efficient antitumor response and tumor clearance (69).

Dectin-1 is one of the most important CLRs and its contribution to anti-tumor immunity has been intensively studied.

Dectin-1 engagement is apparent to up-regulate costimulatory molecules such as CD80, produce TNF-α, IL-6, IL-2, IL-10, IL-12, and IL-23, and elicit potent CTL responses that protect mice from tumor challenge (35). Targeting of dectin-1 with its ligands β-glucan has been shown to increase the infiltration of activated T cells into the tumor. On the other hand, the number of tumorcaused immunosuppressive regulatory T cells and myeloidderived suppressor cells are decreased (70, 71). More recently, the critical role of dectin-1 on enhancement of NK-mediated killing of tumor cells has been demonstrated. Dectin-1 recognize Nglycan structures on the surface of some tumor cells, and cause the activation of IRF5 transcription factor and downstream gene induction, for the full-blown tumoricidal activity of NK cells (63).

As described above, MR and DC-SIGN are major players for both immune evasion and eradication of tumor cells. Further information is necessary to clarify how these CLRs signaling affect the direction of the immunological outcome. Whether cell types or expression level is important, or ligands and microenvironment is the key, or maybe both are closely related. It is known the nature of ligands (i.e., size, form, or chemical side chains of ligands) directly modulate CLRs signaling (62). Further investigation on such regulation of CLRs signaling should lead to make the best application of beneficial side of CLRs signaling to mount antitumor immunity.

### **CLRs and Tumor Immune Evasion**

C-type lectin receptors mediate beneficial effect on anti-tumor immunity via enhancement of type I and type II interferon production. On the other hand, CLRs signaling also play roles on induction of anti-inflammatory factors and molecules (23), and suppress TLRs-mediated protective immunity, thereby tolerating cancer cells escape from immune surveillance. Some examples of such process are induction of specific tolerance to tumor antigens, TGF-β and/or IL-10 production, down-regulation of MHC molecules, or up-regulation of FasL expression (72). Several studies have shown the involvement of CLRs on dysfunction of anti-tumor immune responses. The interaction between DC-SIGN and tumor-associated Le glycans results in enhanced IL-10 production, and impairs production of pro-inflammatory cytokines in tumor-associated macrophages (TAMs) from breast adenocarcinoma and melanoma patients, which leads to decrease capacity to elicit anti-tumor T-cell responses (73). Ligation of DC-SIGN and tumor-associated Le glycans also strongly enhance LPS-induced anti-inflammatory cytokine secretions of IL-6 and IL-10 by monocyte-derived DCs (50). Therefore, ligation of DC-SIGN might cause tumor progression by contributing to the maintenance of an immunosuppressive environment.

Other CLR associated with tumor immune evasion is MR. The research study showed that tumor-activated liver sinusoidal endothelial cells (LSECs) affect liver sinusoidal lymphocytes (LSLs) anti-tumor cytotoxicity and IFN-γ/IL-10 secretion through MR-dependent mechanisms. Further, immunosuppressive effects of tumor-activated LSECs on LSLs were abrogated by way of anti-mouse MR antibodies or MR*−*/*<sup>−</sup>* mice (74).

Recently, the important role of CLRs on modulating the function of tumor-associated cells in tumor microenvironment has been demonstrated. TAMs are a major component of the tumor stroma, which contribute to the evasion of tumors from immune control by producing immune-suppressive cytokines such as IL-10 and TGF-β (75). It has been found that TAMs from human ovarian carcinoma abundantly express MR and dectin-1, MDL-1, MGL, DCIR. MR engagement by tumoral mucins and an agonist anti-MR antibody modulates cytokine production by TAMs toward an immune-suppressive profile: increase of IL-10, absence of IL-12, and decrease of the Th1-attracting chemokine CCL3, indicating that tumoral mucin-mediated activation of the MR on TAMs is important for their immune-suppressive phenotype (57).

In addition to expressing in immune cells, some CLRs have been shown to express on tumor cells, and involved in suppressing human immune system function. LSECtin, a cell-surface member of the C-type lectin DC-SIGN, has been found to express in B16 melanoma cells and inhibit tumor-specific T-cell responses (76). It is therefore important to identify such self-recognition toward immune evasion and regulate them in a specific way.

### **Genetic Variation of CLRs and Cancers**

Host genetic background is one of important factors influencing susceptibility to cancer. Recently, study on single nucleotide polymorphisms (SNP) has been widely used to explore genetic susceptibility. SNPs in CLRs loci have been investigated to clarify its relationship to inflammatory responses. Because chronic inflammation is highly associated with the onset and progression of a multiplicity of human cancer, it is possible SNPs in CLRs associate with cancer susceptibility. Lu et al. (77) evaluated the correlation between colorectal cancer (CRC) risk and SNPs in three C-type lectin genes, i.e., DC-SIGN, MBL, and REG4. They found that polymorphisms in DC-SIGN gene promoter were associated with increased risk in CRC patients, while a SNP in REG4 might be a useful marker for CRC progression. The association of polymorphisms of genes encoding DC-SIGN with nasopharyngeal carcinoma risk has also been investigated. Three SNPs in the GG genotype of the rs2287886, AA genotype of the *−*939 promoter polymorphism, and the G allele of the rs735239 are connected with increased risk of nasopharyngeal carcinoma (78).

Mannose-binding lectin, soluble CLRs, is a plasma collectin and one of the key molecules involved in modulating innate immune system. Low level of serum MBL is associated with increased risk of colon cancer. Polymorphisms in the 3*′* -untranslated region of MBL2 at rs10082466, rs2120132, rs2099902, and rs10450310 reduce MBL plasma levels and activity (79). Odds ratio for homozygous variants versus wild-type ranged from 3.17 to 4.51, whereas the 3*′* -UTR region haplotype consisting of these four variants had an OR of 2.10.

### **Ligand Treatment or Blockade of CLRs and Cancer**

Based on the immune-regulatory effects of CLRs on cellular immunity, application of their ligands to cancer therapy is a scheme of promising scope. Several CLR agonists or antagonists are candidates for anti-cancer drugs. β-glucan as dectin-1 agonists has been extensively investigated for their anti-tumor activity. In murine lung carcinoma models, orally administered particulate β-glucans significantly inhibited tumor growth (71, 80). Both oral and intraperitoneal injection of highly purified soluble β-glucan derived from *Grifola frondosa* were reported to exert anti-tumor effects in experimental murine mammary and colon adenocarcinoma tumor models (70, 81). In addition to their direct effects on specific immunity, β-glucans significantly augment the therapeutic efficacy mediated by anti-tumor monoclonal antibodies (mAbs) in murine breast, liver metastasis, lung, and lymphoma tumor models as well as in human neuroblastoma, lymphoma, and melanoma xenograft models (82). In human, the combination therapy of β-glucan and conventional chemotherapy was reported to improve the long-term survival of patients with ovarian cancer (83). A meta-analysis shows that the addition of lentinan (a purified β-glucans isolated from shiitake mushroom) to chemotherapy prolonged the survival of patients with advanced gastric cancer as compared to chemotherapy alone (84).

Some mechanisms have been proposed to explain the therapeutic response of β-glucan on anti-tumor activity. First, β-glucans are capable of eliciting anti-tumor innate and adaptive immune response via dectin-1-dependent pathway. As discussed above, βglucans play an essential role in activating DCs and macrophages both *in vitro* and *in vivo*, leading to enhanced antigen-specific CD4<sup>+</sup> and CD8<sup>+</sup> T-cell responses. Moreover, β-glucans modulate the suppressive tumor microenvironment and facilitate antitumoral cellular immunity.

The other important role of CLRs is to serve as sensors that transduce tumor antigen into DCs. Some CLRs, including MGL, MR, DNGR-1, and DEC-205, have been found to deliver exogenous antigens on MHC-I for inducing efficient CTL immune response and MHC-II for stimulation of CD4<sup>+</sup> T cells (68, 85, 86). Moreover, targeted delivery of tumor antigens via DC-SIGN, DNGR-1, and DEC-205 with an appropriate adjuvant is capable to prevent development or mediate eradication of tumor in grafted mouse models (87–90).

### **References**


Along with the rapid and thorough innate immune systems, targeting CLRs has emerged as a translational approach to treat a wide variety of cancers. However, there still are some problems yet resolved and further research is required for improving the anti-tumor strategies via CLRs. Some CLRs signaling results in immunosuppressive responses, for instance, and lead to tumor immune escape. Drugs targeting immune checkpoint molecules such as PD-1, PD-L1, and CTLA-4 have recently been demonstrated beneficial and safe (91, 92). The combination of strategy targeting CLRs and immune checkpoints may improve anti-tumor effectiveness.

### **Concluding Remarks**

C-type lectin receptors are multifunctional receptors that have a key role in the recognition of pathogens and regulating innate and adaptive immune responses. In fact, abundant evidence supports that CLRs, especially on DCs, contribute to the recognition of TACA. CLRs also play important roles in inducing anti-tumor immune response and regulate tumor-promoting inflammation. On the other hand, the function of CLRs in tumor remains unknown, therefore CLRs may act as double-edged swords in tumor-associated immune response. Specific regulation of CLRs signaling by modulating tumor microenvironment such as glycoligands and immune cells should lead to the best application of CLRs biology.

### **Acknowledgments**

This work was supported by a Grant-in-Aid from Strategic International Collaborative Research Program (SICORP), by the fund from Japan society for the promotion science 15H04504 (JSPS), by a Grant-in-Aid from Cross-ministerial Strategic Innovation Promotion Program (SIP), by Yakult Bio-Science Foundation, and by the Canon Foundation (to NMT).

mycobacterial cord factor. *Immunity* (2013) **38**:1050–62. doi:10.1016/j.immuni. 2013.03.010


IL-6 and IL-10. *J Immunol* (2011) **186**(4):2192–200. doi:10.4049/jimmunol. 1000475


92. Postow MA, Chesney J, Pavlick AC, Robert C, Grossmann K, McDermott D, et al. Nivolumab and ipilimumab versus ipilimumab in untreated melanoma. *N Engl J Med* (2015) **372**(21):2006–17. doi:10.1056/NEJMoa1414428

**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 © 2015 Yan, Kamiya, Suabjakyong and Tsuji. 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.*

## Pattern recognition and signaling mechanisms of RIG-I and MDA5

### **Stephanie Reikine, Jennifer B. Nguyen andYorgo Modis\***

Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, CT, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Christian Muenz, University of Zurich, Switzerland Gaya Amarasinghe, Washington University School of Medicine, USA Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Yorgo Modis, Department of Molecular Biophysics and Biochemistry, Yale University, 266 Whitney Avenue, Bass 430, New Haven, CT 06520, USA e-mail: yorgo.modis@yale.edu

Most organisms rely on innate immune receptors to recognize conserved molecular structures from invading microbes. Two essential innate immune receptors, RIG-I and MDA5, detect viral double-stranded RNA in the cytoplasm. The inflammatory response triggered by these RIG-I-like receptors (RLRs) is one of the first and most important lines of defense against infection. RIG-I recognizes short RNA ligands with 5<sup>0</sup> -triphosphate caps. MDA5 recognizes long kilobase-scale genomic RNA and replication intermediates. Ligand binding induces conformational changes and oligomerization of RLRs that activate the signaling partner MAVS on the mitochondrial and peroxisomal membranes.This signaling process is under tight regulation, dependent on post-translational modifications of RIG-I and MDA5, and on regulatory proteins including unanchored ubiquitin chains and a third RLR, LGP2. Here, we review recent advances that have shifted the paradigm of RLR signaling away from the conventional linear signaling cascade. In the emerging RLR signaling model, large multimeric signaling platforms generate a highly cooperative, self-propagating, and context-dependent signal, which varies with the subcellular localization of the signaling platform.

**Keywords: pathogen-associated molecular pattern, nucleic-acid sensor, RecA-like DEAD-box (DExD/H-box) RNA helicase, caspase recruitment domain, signal transduction, signalosome, prion-like switch, amyloid-like aggregation**

#### **INTRODUCTION**

Eukaryotic organisms rely on their innate immune system to detect viruses and other microbes. Innate immune receptors detect chemical patterns or structures that are broadly conserved in microbes, including bacterial cell wall components, microbial nucleic acids, and certain highly conserved proteins. These pathogen-associated molecular patterns (PAMPs) are recognized by pattern recognition receptors that fall into several families, including Toll-like receptors (TLRs), NOD like receptors (NLRs), C-type lectin receptors (CLRs), and RIG-I-like receptors (RLRs). At the cell surface and in endocytic compartments, TLRs are the most important family of molecular sentries for the innate immune recognition of a wide range of microbial patterns outside the cytosol (1). CLRs, such as Dectin1, are localized on the cell surface and principally recognize fungal pathogens (2). In the cytosol, NODs and other NLRs recognize cell wall fragments and other bacterial components (3). This review will focus on the RLRs, which are found in the cytosol and recognize viral double-stranded RNA (dsRNA). Innate immune receptors from all families have in common that they nucleate the assembly of large multimeric protein complexes with their signaling adaptors, which include most notably MyD88, MAVS, ASC, and RIP2 (4). These oligomeric assemblies rapidly activate and amplify potent inflammatory antimicrobial responses, principally through the activation of NF-κB, type I interferons, or caspase 1.

Nucleic acids are the largest, and arguably the most important class of ligands for innate immune receptors. To avoid signaling in response to endogenous nucleic acids, which are ubiquitous in the cytoplasm and nucleus, innate immune sensors must recognize specific patterns in specific subcellular locations. (1) A subfamily of TLRs (TLRs 3, 7, 8, and 9) recognizes microbial DNA and RNA ligands exclusively in endolysosomal compartments (5–9). In the cytosol, two essential immune sensors, RIG-I and MDA5, detect viral dsRNA (10–12). Several different sensors recognize double-stranded DNA (dsDNA) in the cytoplasm, including proteins from the AIM2 family, the DDX family, RNA polymerase III, and cyclic GMP–AMP synthase (13, 14). Ligand binding by each of these sensors induces a conformational change that directs the cooperative assembly of large oligomeric signaling platforms, leading to the recruitment and activation of signaling adaptors (4). The rapidly ensuing inflammatory response culminates in activation of the NF-κB and type I interferon signaling pathways (**Figure 1**). This response is one of the first and most important lines of defense against infection and is responsible for the activation of the adaptive immune system (1). Innate immune receptors therefore play pivotal roles as master-regulators of inflammation.

Many viruses deliver an RNA genome into the cytoplasm or rely on a replication or transcription step that generates viral RNA in the cytoplasm. Infection by these viruses is primarily detected by RIG-I and MDA5, also referred to as the RLRs. RIG-I and MDA5 sense complementary sets of viral RNA ligands (10–12, 15). RIG-I recognizes 5<sup>0</sup> -phosphorylated blunt ends of viral genomic dsRNA, whereas MDA5 binds internally to long dsRNA with no end specificity (10–12). RIG-I and MDA5 both

have tandem N-terminal caspase recruitment domains (CARDs) with death domain folds, a DExD/H-box helicase (consisting of two RecA-like helicase domains, Hel1 and Hel2, and an insert domain, Hel2i), and a C-terminal domain (CTD) (**Figure 2A**). In the absence of dsRNA, RIG-I has a closed inactive conformation (16). RNA binding through the helicase and CTD domains (17, 18) releases the CARDs, which then recruit and activate the signaling adaptor MAVS (IPS-1) (19). In contrast, MDA5 does not sequester its CARDs (20) and cooperatively assembles into ATPsensitive filaments on dsRNA (20–22). Moreover, the MDA5 CTD is required for cooperative filament assembly but not for RNA binding (20, 23, 24). The MDA5 CARDs have been proposed to nucleate the assembly of MAVS into its active polymeric form (20, 25) in a process that can be promoted by K63-linked polyubiquitin chains (26). The self-propagating amyloid-like properties of MAVS polymers amplify signaling (25). RLR signaling is regulated by numerous host and viral factors through various mechanisms, including ubiquitin-dependent proteolytic degradation and cleavage of MAVS by virally encoded proteases (27–29). A third RLR, LGP2, lacks CARDs and exerts co-stimulatory and inhibitory functions on MDA5 and RIG-I, respectively (30–33).

Recent biochemical, biophysical, and cellular studies have greatly advanced our understanding at the molecular level of the mechanisms of pattern recognition and signaling by RIG-I and MDA5. Here, we review these studies and their implications on the current models of microbe-induced inflammation, auto-inflammation, and inflammation-induced cancer.

#### **RECOGNITION OF dsRNA IN THE CYTOSOL BY RIG-I AND MDA5**

**THE MOLECULAR DETERMINANTS OF LIGAND RECOGNITION BY RLRs** RIG-I preferentially binds to short (<300 bp) dsRNAs that have blunt ends and a 5<sup>0</sup> triphosphate (5<sup>0</sup> -ppp) moiety, facilitating discrimination between host and viral dsRNA (10–12). Crystal structures of RIG-I bound to a 12-bp dsRNA ligand and of unliganded RIG-I have provided detailed insights into the mechanism of activation of this receptor. In the absence of dsRNA ligand, RIG-I is in an auto-repressed state: the domains in the helicase domain are in an open conformation and the tandem CARDs form contacts with the Hel2i domain. This conformation sterically prevents the CARDs from binding to polyubiquitin or to CARDs from other binding partners, thereby preventing signaling to MAVS (16).

Upon the presentation of a viral dsRNA, RIG-I undergoes significant conformational rearrangement. The CTD binds tightly to the 5<sup>0</sup> -ppp and the helicase domains wrap around dsRNA, adopting a more compact configuration (16–18) (**Figure 2B**). RIG-I recognizes RNA primarily through non-specific interactions with the phosphate sugar backbone, predominantly by the Hel2i domain. This conformational change allows ATP to bind RIG-I, a necessary step for the activation of RIG-I (16–18). Although the CARDs were absent from the RNA-bound RIG-I crystal structures, biochemical studies and small angle X-ray scattering data indicate that the tandem CARDs are released from the Hel2i domain in the active form of RIG-I (17, 18).

In contrast to RIG-I, MDA5 preferentially binds internally to long dsRNA (>1,000 bp) with no end specificity (10–12) and cooperatively assembles into a filament on the dsRNA (20, 21). Unlike RIG-I, the CARDs of MDA5 are not sequestered in the absence of ligand (20). The forced proximity of the CARDs upon MDA5 filament formation induces oligomerization of MDA5 CARDs, forming a scaffold for binding and oligomerization of MAVS CARD (see Activation of MAVS and Downstream Signaling). Notably, the atomic structures of the MDA5 CARDs have not yet been determined.

A crystal structure of the MDA5 helicase domains and CTD bound to dsRNA revealed how MDA5, despite having a similar domain architecture as RIG-I, recognizes dsRNA in a different manner (**Figure 2B**). The helicase domains of MDA5 wrap around dsRNA similarly to the helicase domains of RIG-I (34, 37). However, consistent with the observation that MDA5 is not preferentially activated by 5<sup>0</sup> -ppp dsRNA (10–12), the MDA5 CTD is rotated by 20°, bringing it closer to the dsRNA, as compared to the RIG-I structure. The CTD also forms contact with Hel1 in MDA5, such that MDA5 forms a closed ring around the dsRNA (37). This orientation of the CTD promotes cooperative filament formation along dsRNA, initiated from internal sites in the dsRNA rather than from one of the ends (20, 21, 34).

The RLRs are part of the DExD/H-box helicase family based on their domain architecture (33), but they do not appear to have dsRNA helicase activity. Instead, ATP binding and hydrolysis have been implicated in filament formation. ATP binding strengthens the interaction between MDA5 and the dsRNA (34). ATP hydrolysis, however, causes MDA5 to dissociate from the dsRNA (20, 38). At the ends of the MDA5-RNA filaments, ATP hydrolysis causes

depolymerization, providing a mechanism for shutting down the signal and for recycling of MDA5. MDA5 filament assembly and disassembly dynamics provide the specificity for long dsRNA (20, 38). RIG-I was also shown recently to form ATP-dependent filaments, although the RIG-I filaments are shorter and less stable than MDA5 filaments (34, 39).

LGP2, the third RLR, has similar helicase and CTD domains as RIG-I and MDA5, but it lacks the tandem CARDs (33). LGP2 recognizes the termini of dsRNA through similar types of protein-RNA contacts as RIG-I and MDA5 (23, 33, 40, 41). ATP hydrolysis enhances RNA recognition by LGP2 (42). Because it does not have CARDs, LGP2 does not recruit MAVS or induce MAVS signaling. LGP2 affects signaling in response to viral stimuli, however, by modulating the RIG-I and MDA5 signals (see Regulation of RLR Signaling) (30–33).

#### **ROLE OF UNANCHORED LYSINE 63-LINKED UBIQUITIN CHAINS IN RLR ACTIVATION**

The oligomerization of the RNA sensors RIG-I and MDA5 that activates the antiviral innate immune response depends on unanchored lysine 63-linked polyubiquitin chains (19, 26). In 2010, Chen and colleagues reconstituted the RIG-I pathway *in vitro* and demonstrated that unanchored K63-linked polyubiquitin chains are required for a full signaling response as measured by IRF3 dimerization (19). Polyubiquitin chains containing as few as four ubiquitin molecules bind non-covalently to the RIG-I CARDs and can be covalently attached to RIG-I by the E3 ligase TRIM25 (19, 43). Furthermore, RIG-I interacted with K63-linked polyubiquitin chains from HEK293T cells in co-immunoprecipitation experiments (19). Similar studies generalized these findings to MDA5 and showed that K63-ubiquitin chains promoted oligomerization of the MDA5 CARDs (26).

A recent crystal structure of the tandem CARDs of RIG-I bound to K63-diubiquitin revealed the molecular basis of the CARDubiquitin interaction (**Figure 2C**) (35). K63-ubiquitin chains promote the assembly of RIG-I CARDs into a tetrameric "lockwasher" structure by stabilizing intermolecular CARD–CARD interactions. This RIG-I tetramer recruits and activates MAVS (see next section) (35). Monoubiquitin is not sufficient to promote RIG-I signaling because a single ubiquitin domain does not make enough contacts to significantly stabilize RIG-I oligomerization through CARD–CARD interactions (19, 35).

Although ubiquitin chains promote RIG-I tetramerization, RIG-I and MDA5 can both assemble into oligomeric filaments and induce MAVS filament formation and signaling in the absence of polyubiquitin chains. Indeed, under certain experimental conditions, namely in the absence of polyubiquitin and as a result of ATP hydrolysis, RIG-I has been observed to form filaments along dsRNA (34, 39, 44). Similarly, MDA5 signaling is thought to be triggered by the formation of MDA5 filaments along dsRNA, which is a ubiquitin-independent process (20, 21). The forced juxtaposition of RLR CARDs upon RLR filament formation is thought to be sufficient to activate MAVS signaling (34). Both RIG-I CARDs and MDA5 CARDs have, however, been shown to

bind K63 polyubiquitin chains (26). Hence the question arises of whether K63-linked ubiquitin chains always participate in RLR signaling, or whether they are only required under specific physiological conditions that do not favor RLR filament formation. Because RIG-I has much higher affinity for the 5<sup>0</sup> -ppp end of viral ligands than it does for the phosphate backbone alone, it has been proposed that RIG-I is more likely to bind to the 5<sup>0</sup> -ppp end of the dsRNA (34). If sufficient polyubiquitin is available, RIG-I does not form a filament and instead remains at the end of the dsRNA, and the tetrameric CARD lock-washer scaffold is formed (34, 35). K63-linked polyubiquitin chains stabilize the CARDs oligomer through non-covalent interactions. Covalent linkage of the ubiquitin chains to RIG-I by TRIM25 can provide further stabilization of the RIG-I oligomer, thereby increasing interferon signaling capacity (19, 35, 43). If the local concentration of polyubiquitin is insufficient to induce RIG-I CARDs tetramer formation, ATP hydrolysis may enable RIG-I to translocate along dsRNA and assemble into filaments (39), bringing the CARDs together by cooperative stacking of the helicase domains and leading to ubiquitin-independent signal activation. Unlike RIG-I, MDA5 has no known RNA end-preference and MDA5 has a higher propensity to form filaments than RIG-I (26,34). Hence, the physiological role of unanchored polyubiquitin chains in MDA5 signaling remains less well understood than in RIG-I.

#### **ACTIVATION OF MAVS AND DOWNSTREAM SIGNALING**

In the textbook view of RLR signaling, the signal is propagated sequentially from the ligand-bound RLR to MAVS to the cytosolic protein kinases IKK and TBK1, which in turn activate the transcription factors NF-κB and IRF3, respectively (45). Activated NFκB and IRF3 are translocated into the nucleus, where they induce expression of type I interferons and other inflammatory antimicrobial molecules. The discovery that ligand binding induces RIG-I and MDA5 to assemble into large oligomeric platforms with MAVS on the mitochondrial and peroxisomal membranes has, however, shifted the paradigm for RLR signaling away from the model of a linear signaling cascade. As reviewed in the previous section, both RIG-I and MDA5 form filaments along dsRNA ligands. For RIG-I the forced juxtaposition of its CARDs, along with binding of K63-linked polyubiquitin chains, promotes the formation of a tetrameric lock-washer structure (**Figure 2C**), which serves as a platform to recruit MAVS (35). Structural and biochemical data suggest that the minimal signaling unit for MDA5 is much larger than for RIG-I and contains at least 11 MDA5 molecules (34). These oligomeric RLR CARD assemblies have been proposed to nucleate the formation of MAVS polymers (**Figure 2D**) (20, 25). Notably, the polymeric form of MAVS, but not its monomeric form, activates downstream RLR signaling (25). Moreover, once MAVS polymers have been nucleated they are self-propagating, drawing soluble-form MAVS monomers into the polymer. The MAVS CARD, even when isolated from the C-terminal and transmembrane domains, recapitulates this behavior *in vitro* (25). MAVS CARD polymers were recently found to consist of helical filaments (36), similar to those formed by the death domains of MyD88 (4, 46). The switch from a soluble form to a selfpropagating helical fiber is reminiscent of amyloids and prions, and indeed MAVS CARD functions like a *bona fide* prion in yeast (47). Thus, MAVS has a prion-like mechanism of signal activation and amplification. ASC, the adaptor of the NLRP3 inflammasome, was recently shown to have a similar prion-like mechanism of signal transduction (47).

A transmembrane domain tethers MAVS to the mitochondrial or peroxisomal membrane. MAVS polymerization may therefore cause some remodeling of the membrane in these organelles (**Figure 2D**) (36). In support of this notion, MAVS facilitates cell death by disrupting the mitochondrial membrane potential and by activating caspases (48). Notably, the signaling output from MAVS is different depending on whether it occurs at the peroxisomal or mitochondrial membrane. Peroxisomal MAVS induces the rapid interferon-independent expression of defense factors, which precedes the activation of the principal interferon-dependent pathway by mitochondrial MAVS that amplifies and stabilizes the antiviral response (49). Thus, MAVS signaling is dependent on cellular localization, and peroxisomes are an important site of antiviral signal transduction (49).

#### **REGULATION OF RLR SIGNALING**

The inflammatory response resulting from RLR signaling unavoidably occurs at a cost to normal tissue function. Multiple regulatory mechanisms have evolved to allow rapid activation, amplification, and inactivation of RLR signaling, and to achieve the optimal trade-off between the cost and benefit of the inflammatory response (50). Polyubiquitination has been one of the most extensively studied modifications of RIG-I and MDA5, so it is not surprising that E3 ligases and deubiquitinases have been implicated in modulating the RLR response. TRIM25, the most exhaustively studied E3 ligase, covalently attaches K63-linked polyubiquitin to RIG-I CARDs to initiate or promote signaling (26, 43). The E3 ligase Riplet has recently been identified as a necessary component of RIG-I signaling (51). USP21 negatively regulates RIG-I signaling by deubiquitinating RIG-I (52).

In addition to ubiquitination, phosphorylation is slowly emerging as an important regulatory mechanism for RLR signaling. Phosphorylation of Ser8 and Thr170 in the CARDs of RIG-I antagonizes RIG-I signaling (53, 54). Based on the crystal structure of RIG-I in complex with K63-linked diubiquitin (35), we expect phosphorylation of Ser8 but not Thr170 to interfere with ubiquitin binding. Phosphorylation of RIG-I CARD has also been proposed to inhibit recruitment of TRIM25 and MAVS (53, 54). The RIG-I phosphorylation sites are not conserved in MDA5, but MDA5 does have a suppressing phosphorylation site in its first CARD, at Ser88 (55). Conventional protein kinases Cα and β (PKCα/β) have been identified to be responsible for RIG-I phosphorylation (56). RIG-I and MDA5 are thought to be constitutively phosphorylated until presentation of viral RNA, at which time the RLRs must be dephosphorylated by phosphoprotein phosphatase 1 α and γ (PP1α/γ) (55).

Besides post-translational modification of the RLRs, RLR signaling is also modulated by several different proteins, derived both from the host and from pathogens. One such protein is the third RLR, LGP2. Because it lacks CARDs, LGP2 cannot activate MAVS; however, its ability to recognize dsRNA allows it to modulate the signaling capacities of RIG-I and MDA5. LGP2 downregulates signaling by RIG-I (32, 33). This activity was attributed to LGP2 competitively recognizing the same viral ligand as RIG-I. In contrast, LGP2 enhances MDA5 signaling (30, 33, 42). The molecular mechanism of this enhancement remains unclear, but LGP2 appears to facilitate recognition of viral RNA by MDA5 through interactions between the LGP2 CTD and RNA (41). Indeed, a recent study identified a specific picornaviral RNA ligand (in the antisense L region) to which LGP2 binds tightly, thereby stimulating MDA5 signaling (31).

The seemingly contradictory roles of LGP2 in RLR signaling remain an open question. The experimental approaches used to study LGP2 in relation to MDA5 and RIG-I have been different, potentially explaining some of the differences. As evidence accumulates for the opposing roles of LGP2 on RLR signaling, however, the emerging perspective is that LGP2 can control the balance between RIG-I and MDA5 responses during viral infection.

Pathogen evasion tactics against RLR-mediated immune response are extensive and occur at every level of signaling [reviewed in Ref. (57)]. A complete description of these tactics is beyond the scope of this review, so we highlight below a few representative examples of different modes of RLR evasion. MAVS is the primary target of viral factors for inhibiting RLR signaling. MAVS is cleaved by hepatitis C virus NS3/4A protease (28, 29), enterovirus 71 protease 2Apro (58), GB virus B NS3/4A (59), and coxsackie virus B 3C protease, which also cleaves TRIF (60). In a distinct mechanism of RLR signal inhibition, paramyxovirus V proteins disrupt the fold of MDA5 (61). Another major mechanism for evasion of the RLR innate immune response is masking or hiding of viral RNA ligands by viral proteins, such as VP35 from Ebola and Marburg viruses, which coat the ends and backbone of dsRNA to prevent RLR recognition (62–64). Similarly, nucleoproteins from arenaviruses bind to the ends of viral dsRNA and digest the RNA in a 30–5<sup>0</sup> direction, thereby making the RNA a weaker ligand for RLRs (65–68). Interestingly, MAVS was recently also shown to be under cellular control. A truncated variant of MAVS resulting from alternative translation initiation interferes with interferon production induced by full-length MAVS (69).

#### **CONCLUSION**

RIG-I and MDA5 are the principal sensors of viral dsRNA in the cytoplasm. The interferon-dependent inflammatory response triggered by RLR ligand binding is one of the first and most important lines of defense against infection. RIG-I and MDA5 recognize distinct and complementary sets of viral dsRNA ligands. The molecular signaling mechanisms of RIG-I and MDA5 differ in some respects but also share certain key features. Differences include the sequestration of CARDs by RIG-I but not by MDA5 in the absence of ligand, the much greater propensity of MDA5 to form filaments along dsRNA, and the different contribution of K63-linked ubiquitin chains, which remains poorly defined for MDA5. Common features in RLR signaling include proximity-induced assembly of CARD oligomers, which serve as platforms to nucleate MAVS CARD polymerization, and signal amplification through the amyloid-like properties of the MAVS CARD. Together, the recent advances reviewed here shift the paradigm of RLR signaling away from the prototypical linear signaling cascade to a model in which signaling is activated by the cooperative assembly of an oligomeric signaling platform. The signal output depends on the cellular localization of MAVS (mitochondria or perixosome), and signaling is finely regulated by a multitude of cellular and pathogen-derived factors. Key outstanding questions include when, where, and how ubiquitin chains potentiate RIG-I and MDA5 signaling, exactly how RLRs interact with MAVS, and how LGP2 and other factors modulate RLR signaling.

#### **OUTSTANDING QUESTIONS**

	- Is the mechanism of action of K63-linked ubiquitin chains the same for RIG-I and MDA5?
	- How do RIG-I CARD tetramers, stabilized by K63-linked ubiquitin, nucleate MAVS filament assembly?
	- How do MDA5 CARDs nucleate MAVS filament assembly? Does this process require K63-linked ubiquitin chains?

#### **ACKNOWLEDGMENTS**

Work on this article was supported by a Burroughs Wellcome Investigator in the Pathogenesis of Infectious Disease Award and NIH grant R01 GM102869.

#### **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.

*Received: 28 April 2014; accepted: 05 July 2014; published online: 23 July 2014.*

*Citation: Reikine S, Nguyen JB and Modis Y (2014) Pattern recognition and signaling mechanisms of RIG-I and MDA5. Front. Immunol. 5:342. doi: 10.3389/fimmu.2014.00342*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Reikine, Nguyen and Modis. 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.*

## Recognition of human oncogenic viruses by host pattern-recognition receptors

#### **Nelson C. Di Paolo\***

Lowance Center for Human Immunology, Division of Rheumatology, Departments of Pediatrics and Medicine, Emory University, Atlanta, GA, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Thomas A. Kufer, University of Cologne, Germany Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Nelson C. Di Paolo, Lowance Center for Human Immunology, Division of Rheumatology, Departments of Pediatrics and Medicine, Emory University, 1760 Haygood Drive, E269, Atlanta, GA, USA e-mail: ncdipaolo@emory.edu

Human oncogenic viruses include Epstein–Barr virus, hepatitis B virus, hepatitis C virus, human papilloma virus, humanT-cell lymphotropic virus, Kaposi's associated sarcoma virus, and Merkel cell polyomavirus. It would be expected that during virus–host interaction, the immune system would recognize these pathogens and eliminate them. However, through evolution, these viruses have developed a number of strategies to avoid such an outcome and successfully establish chronic infections.The persistent nature of the infection caused by these viruses is associated with their oncogenic potential. In this article, we will review the latest information on the interaction between oncogenic viruses and the innate immune system of the host. In particular, we will summarize the available knowledge on the recognition by host pattern-recognition receptors of pathogen-associated molecular patterns present in the incoming viral particle or generated during the virus' life cycle. We will also review the data on the recognition of cell-derived danger associated molecular patterns generated during the virus infection that may impact the outcome of the host–pathogen interaction and the development cancer.

**Keywords: PRRs, oncogenic viruses, cancer, innate immunity, innate sensors**

#### **INTRODUCTION**

Seven human viruses have been found so far to cause approximately 10–20% of human cancers worldwide (1). They include the herpesviruses, Epstein–Barr virus (EBV) and Kaposi's associated sarcoma virus (KSHV), the hepatitis B (HBV) and hepatitis C (HCV) viruses, high-risk human papillomaviruses (HPV) (the most clinically important ones being types 16 and 18, but most probably a few others will be found to be relevant to cancer development as well in the future), the human T-cell lymphotropic virus-1 (HTLV-1), and the recently discovered Merkel cell polyomavirus (MCPyV) (1). The mechanisms by which these viruses cause cancer are diverse. They have prolonged latency periods, during which viral factors combine with other environmental factors in the setting of the genetic background of each particular host (2). However, it could be proposed that these viruses have no intention of generating disease in their hosts, as evidenced by the overall rate of disease/infected humans worldwide for each virus (**Table 1**). Although exact numbers are not available for every region in the world, the number of humans that suffer a disease associated with each oncogenic virus, as compared to the number of people infected with each virus is evidently low. It appears that during evolution these viruses have found a balance of "live and let live" with their host. Until very recently in history, humans were not living long enough to considerably suffer from the diseases attributed to these viruses (3). Today, however, human longevity is greatly extended, and although the burden of diseases associated with oncogenic viruses is still low in comparison with the number of infected people, the goal of medicine is, of course, to eradicate diseases. Understanding the interactions of these viruses with the host will certainly help to achieve this goal. Of particular importance is their

interaction with the innate immune system, which functions to recognize non-self like microorganisms, and also plays a critical role in recognition of modified self that indicates damage or danger (4).

Germline-encoded pattern-recognition receptors (PRRs) recognize chemically distinct moieties in microorganisms or "pathogen-associated molecular patterns" (PAMPs) (12). PRRs can also recognize endogenous host molecules that in different ways signal danger ("damage" or "danger"-associated molecular patterns' or "DAMPs") (13, 14). It is noteworthy that the "D" in DAMPs is used interchangeably for "danger" or "damage." However, "danger" would seem to be more appropriate, as there could be danger without damage, and it would be more in line with the original "danger" theory proposed by Matzinger several years ago (15).

There are two families of transmembrane PRRs, namely tolllike receptors (TLRs) (16) and C-type lectin receptors (CLRs) (17). They are positioned to scan the extracellular and endosomal spaces. The families of *cytoplasmic* PRRs include the retinoic acid-inducible gene (RIG)-I-like receptors (RLRs) (18) and the nucleotide-binding, oligomerization domain (NOD)-like receptors (NLRs) (19, 20), as well as a large number of DNA sensors that converge in the adaptor for cytosolic DNA sensing stimulator of interferon genes (STING). An excellent very comprehensive review on nucleic acid sensing was recently published (21). The double-stranded (ds)RNA-dependent protein kinase R (PKR) and the 2<sup>0</sup> ,50 -oligoadenylate synthetases (OAS) are considered part of the cytoplasmic PRRs as well (22). Recently, a *nuclear* DNA sensor was identified, IFI-16, a PYHIN protein that, together with the cytoplasmic AIM-2 DNA sensor, was proposed to form a new family of innate DNA sensors ("AIM2-like receptors" or "ALRs")



\*Numbers are approximate, and may vary in different geographical regions.

HCC, hepatocellular carcinoma; MCC, Merkel cell carcinoma. See other abbreviations in text.

(23). Importantly, some of the members of the NLR and ALR families form a molecular complex termed "inflammasomes," molecular platforms that control the secretion of the pro-inflammatory cytokines interleukin-1b and -18 (14). Some of the members of the already mentioned families of receptors recognize DNA. However, there is a growing list of DNA sensors not belonging to these families, recognizing both pathogen's DNA as well as modified or displaced self DNA (24). Finally, although it is not be in the scope of this review, it is relevant to mention here that there is a particular set of immune proteins called "intrinsic antiviral factors." Unlike PRRs that function against viruses by triggering a cascade of antiviral signaling events, intrinsic antiviral factors directly block viruses at different points of their life cycle (25). Each of these families of proteins does work in concert in order to eradicate viruses. However, viruses have evolved a myriad of mechanisms to evade and subvert these host antiviral defenses in order to ensure their evolutionary survival (26).

There is abundant information on the mechanisms by which the seven oncogenic viruses block the molecular pathways of the innate immune system at the level of intracellular adaptors, and the reader is referred to the several extensive published reviews in the specific sections below. However, much less is known on the recognition of these viruses by the sensors that physically interact with viral PAMPs. Here, we will focus on the latest findings on the growing list of innate immune sensors that have been implicated in sensing each known human oncogenic virus (**Figure 1**). We believe that by combining this information in one single review, parallelisms and differences between these very distinct viruses, which trigger the same human disease, i.e., cancer, may be revealed.

#### **HEPATITIS C VIRUS**

Hepatitis C virus is a single-stranded RNA virus, with an enveloped nucleocapsid of about 50 nm. It is transmitted via parenteral route, and there are millions of people infected with HCV worldwide, for which there is no available vaccine (27). During its evolution with the host, it has developed a number of mechanisms to avoid being eliminated by the innate immune system, establishing chronic infection of the liver. This chronic infection triggers injury

to the liver, which is believed to be the basis for the development of liver cancer. HCV and its interaction with the adaptive and innate immune systems is a very active field of research, and many recent review articles have exhaustively discussed these topics (6, 27–32). However, the sensing of the virus and the innate pathways activated during the first days of infection in humans remain largely unknown (33). Understanding of these steps is critical, as they are likely to set the stage for the ultimate outcome of the infection.

#### **CELLULAR MEMBRANE AND ENDOSOME SENSING**

TLR2 has been proposed to sense HCV proteins at the cell surface (30). TLR3 has been shown to be relevant for the activation of the transcription factors IRF-3 and NF-κB in response to HCV–RNA (28). TLR7 was also shown to be relevant in HCV sensing (34), and the proposed mechanism suggested the existence of a cell–cell RNA transfer process where HCV-infected cells activated plasmacytoid dendritic cells (pDCs) in *trans*. This was shown to be the case as well by Dreux et al., who reported the transfer of HCV–RNA containing exosomes from infected cells to pDCs (35).

#### **INTRACELLULAR SENSING**

Hepatitis C virus recognition in the cytosol is mediated by the host RNA-dependent PKR, which identifies an internal ribosome entry site (IRES) in HCV genome. In contrast, the virus' 3<sup>0</sup> poly-U/UC sequence, short dsRNA regions, and 5<sup>0</sup> triphosphate of the uncapped HCV–RNA are recognized by RIG-I [reviewed in detail by Horner (28); Horner and Gale (29)]. A detailed analysis of the HCV–RNA that activates RIG-I was described by Schnell et al. (36), who reported a 34-nt poly-uridine"core"of the 5<sup>0</sup> -ppp poly-U/UC sequence as a critical structure for RIG-I activation. Recently, a new mechanism by which HCV controls interferon (IFN) induction was described, where RIG-I is ubiquitinated through the di-ubiquitin-like protein ISG15,one of the early interferon responsive genes (ISGs) (37). Other investigators, however, propose a different mechanism of RIG-I activation, where Riplet-mediated K63-linked polyubiquitination releases RIG-I RD autorepression, allowing the access of downstream signaling factors to the RIG-I

protein (38). These differences in the proposed models of RIG-I activation may be due to the use of different cell types and experimental conditions. More recent data also suggest that the STING may be relevant for HCV recognition (39, 40). The mechanism these investigators propose implicates direct interaction of HCV NS4B with STING, blocking IFN beta production downstream of both STING and RIG-I. Finally, although human biopsies provide limited opportunities for mechanistic studies, they are critical since they allow a snapshot view of the tissue that is infected in the actual host. Consistent with this concept, Mozer-Lisewska et al. reported that in liver from patients with chronic hepatitis C infection, the expression of TLR1, 2, 4, NALP, and RIG-I helicase was markedly increased, suggesting that these PRRs may be important for the pathogenesis of chronic viral hepatitis by HCV in humans (41).

#### **HEPATITIS B VIRUS**

Hepatitis B virus genome consists of partial dsDNA, its nucleocapside is enveloped, and is transmitted via the parenteral route; although there is a vaccine available, millions of people are infected (27). The major challenge for mechanistic analysis of HBV interaction with the innate immune system is the lack of a suitable animal model. Woodchuck infected with the woodchuck hepatitis virus (WHV) (42) is an accepted study model, but available immunological tools are limited. Researchers use transfected cells or mice hydrodynamically injected with HBV replicative plasmids, but they cannot faithfully recapitulate the *in vivo* infection process. Even with these caveats in mind, the field is advancing toward an understanding of the interaction between HBV and the human innate immune system. Until recently, it was believed that the virus was just a stealth pathogen that could not be detected by PRRs (43–45). However, it is becoming clear that HBV just have a number of very efficient strategies to block innate immunity, and they were recently reviewed in Ref. (5, 46). Indirect data seem to support the fact that PRR sensing of HBV is important for HVB

pathogenesis. For example, Guo et al. showed that transfection in cells with the plasmids expressing adaptors for PRRs signaling pathways (myeloid differentiation primary response gene 88, or MyD88), TIR-domain-containing adaptor-inducing beta interferon (TRIF), or the RIG-I/MDA5 adaptor, interferon promoter stimulator 1 (IPS-1), reduced HBV DNA and RNA levels (47). However, it is difficult to conclude that the data obtained in this *in vitro* system correlates with the behavior of the virus in naturally infected hosts.

#### **CELLULAR MEMBRANE AND ENDOSOME SENSING**

Using the HBV/WHV model, Zhang et al. described that addition of TLR2 ligands activate NF-κB, PI3K/Akt, and different arms of the MAPK signaling pathways to induce pro-inflammatory cytokines, leading to the reduction of WHV replication and gene expression in HepG2.2.15 cells and primary woodchuck hepatocytes (48). However, in a previous study using an HBV transgenic mice model, a single intravenous injection of exogenous ligands specific for TLR2, TLR3, TLR4, TLR5, TLR7, and TLR9 showed that all of the ligands except for TLR2 inhibited HBV replication in the liver non-cytopathically in an alpha/beta IFN-dependent manner (49). Differences in these results could easily be attributed to the different model systems used, and warrant further investigation. In a more relevant study model, i.e., the chimpanzee, Lanford et al. showed that the small molecule GS-9620, which activates TLR7 signaling in immune cells, provided long-term suppression of serum and liver HBV DNA (50). Based on these and other results, TLR ligands are being developed as drugs for the treatment of chronic viral infections, including HBV (51).

#### **INTRACELLULAR SENSING**

RIG-I and MDA5 are important PRRs responsible for recognition of viral RNAs produced during viral infection, and represent targets for immunosuppression during HBV infection. Lu and Liao demonstrated that in human Huh7 cells transfected and in the livers of mice hydrodynamically injected with HBV replicative plasmids, the expression of MDA5, but not RIG-I, was increased, and it was the critical protein for HBV detection (52). It is interesting that mice heterozygous for MDA5 also had an increase in HBV replication, indicating the existence of a possible threshold in MDA5 expression level necessary for its function as a HBV sensor. In another study, Zhao et al. proposed that RIG-I, and not MDA5, is the protein involved in HBV sensing (53). Although it is not clear as yet which specific sensor is involved, viral RNA sensing in the cytoplasm is clearly occurring during HBV infection. Studies using hepatocytes (54), 293 cells (55), or the cytoplasmic fraction of HBx transgenic mouse livers (56) showed that hepatitis B virus X (HBX) protein interacts with MAVS (also called IPS-1, a critical molecule in RNA signaling pathways) (57), and prevents the induction of IFN genes. DNA sensing mechanisms are also likely to be relevant, since in the cell line Huh7, Chen et al. showed that DAI can inhibit HBV replication, where the inhibitory effect was associated with activation of NF-κB, and was independent of IRF-3 or cytokines (58).

In summary, it is clear that many more studies identifying new mechanisms of HBV detection by the innate immune system are likely to follow. The true challenge will be to reconcile those *in vitro* identified pathways with the mechanisms of HBV control in more relevant infectious models, i.e., the chimpanzee, and translate this knowledge into human settings.

#### **HERPESVIRUSES: EPSTEIN–BARR VIRUS AND KAPOSI'S ASSOCIATED SARCOMA VIRUS**

There is a significant body of data demonstrating that herpesviruses can be sensed by the innate immune system at the cellular membrane, in the endosomes, and in the cytosol. Furthermore, recent studies showed that herpesviruses can also be sensed in the nuclei. A recent comprehensive review on herpesviridae was published by Paludan and Bowie (24). EBV and KSHV are the two members of this virus family that have been identified as having growth transforming potential, and therefore, we focus on these here.

#### **EPSTEIN–BARR VIRUS**

Epstein–Barr virus was discovered approximately 50 years ago. It is an enveloped virus with a dsDNA genome, for which there is extensive knowledge about its biology (59). The innate immune recognition of EBV was also reviewed in detail (60, 61).

#### **CELLULAR MEMBRANE AND ENDOSOME SENSING**

Epstein–Barr virus can be sensed by TLR2 in certain cells; however, the exact virion component being sensed is still unclear (62). Ariza et al. proposed that deoxyuridine triphosphate nucleotidohydrolase (dUTPase), a non-structural protein encoded by EBV, is sensed by TLR2 and initiates a MyD-88 dependent response (63). This group further extended their results to demonstrate that the protein was secreted in exosomes inducing NF-κB activation and cytokine secretion in primary DCs and peripheral blood mononuclear cells (PBMCs) (64). However, these results should be interpreted with caution given that the studies were done using an *in vitro* experimental system. EBV produces non-coding RNAs or "Epstein–Barr virus-encoded small RNA" ("EBER"). TLR3 is a

sensor of viral dsRNA. Very interestingly, it was discovered that a substantial amount of EBER was released from EBV-infected cells in exosomes that stimulated DCs to produce type-I IFN. Most importantly, they found EBER in sera from patients with EBVrelated diseases, suggesting that EBER could be responsible for immune activation by EBV, inducing type I IFN and proinflammatory cytokines (65). These results were further discussed by the same group (66). TLR7 has not been proposed as a direct sensor for EBV. However, Valente et al. reported that the aberrant activation of TLR7 in EBV-infected cells might induce the expression of the EBV-protein LMP1 (67). As LMP1 is known to prime cells to express IFN, and both TLR7 and IFNs are believe to be involved in the development of systemic lupus erythematosus (SLE, or simply "lupus"), the association of EBV infection and autoimmunity clearly warrants further investigation.

Interestingly, Severa et al. showed that EBV can activate pDCs through TLR9 and TLR7, in combination with functional autophagic machinery (68). However, these pDCs were not able to mature and induced an inefficient T-cell response, suggesting a new virus escape mechanism potentially related to EBV induced diseases. Another important finding reported by van Gent et al. showed that EBV encoded deubiquitinase, BPLF1, interferes with NF-κB activation mediated by TLR signaling (69). TLR9 can initiate a response by detecting EBV DNA in the endosomes. However, Fathallah et al. showed that EBV infection of human primary B cells results in the strong inhibition of TLR9 transcription by the EBV oncoprotein latent membrane protein 1 (LMP1) (70). The role of TLR9 in EBV infection has been exhaustively reviewed in Ref. (7).

#### **INTRACELLULAR SENSING**

In the cytosol, EBV EBERs are recognized by RIG-I (62). Moreover, RIG-I has been proposed to indirectly sense EBV DNA by recognizing the 5<sup>0</sup> -triphosphate transcribed by the host RNA polymerase III (71). However, there are conflicting results that need to be resolved by further experimentation to clarify the role of RNApol-III in EBV sensing mechanism (62). There are numerous DNA sensors in the cytosol, and although some of them have been shown to recognize other herpesviruses (62), the relevance of cytosolic DNA sensors to EBV remains unclear.

#### **KAPOSI'S ASSOCIATED SARCOMA VIRUS**

This virus, formally classified as human herpesvirus 8 (HHV-8), is associated with Kaposi's sarcoma (KS), among other pathologies (72). It is a big enveloped virus with a dsDNA genome (73). Employing many proteins and micro-RNAs, KSHV modulates the innate and adaptive immune system of the host at multiple levels. A number of excellent reviews on these topics have been recently published (8, 73–75).

#### **CELLULAR MEMBRANE AND ENDOSOME SENSING**

Only recently, researchers have started investigating the role of TLR-mediated sensing of KSHV. Although a direct interaction of KSHV with a TLR has not been reported, the virus downregulates the expression of TLR4 soon after infection in endothelial cells (76). West and Damania, however, showed that in monocytes TLR3 expression is upregulated after KSHV infection (77). Gregory et al. showed that agonists specific for TLR7/8 reactivated latent KSHV and induced viral lytic gene transcription and replication (78). Moreover, the same was accomplished by secondary infection with vesicular stomatitis virus (VSV), which also activates those same TLRs. More recently, pDCs were shown to respond to KSHV in TLR9-dependent manner (79). Finally, it has been shown that stimulation of the TLR3–TRIF axis increases the expression of the KSHV protein RTA (replication and transcription activator), only for RTA to degrade TRIF in order to block the innate immune response (80, 81). Collectively, although these results do not demonstrate a direct interaction between KSHV and TLRs, they clearly indicate that there is a physiologically relevant interplay between them.

#### **INTRACELLULAR SENSING**

The field of intracellular sensing of KSHV has recently seen a number of very exciting discoveries. Gregory et al. reported that KSHV Orf63 blocks NLRP1-dependent innate immune responses, including caspase-1 activation and processing of interleukin-1 beta (IL-1beta) and IL-18, and significantly reduces NLRP1-dependent cell death (82). Moreover, the inhibition of Orf63 expression resulted in increased expression of IL-1beta during the KSHV infection that could have an effect on KSHV induced pathologies. In a new development in the field of innate immune sensing, Unterholzner et al. reported that IFI-16 acts as a nuclear sensor for HSV-1 (23). Based on their findings, they proposed the existence of a new family of "AIM-2 like receptors" or ALRs. In the same line of research, Kerur et al. found that the same protein is responsible for KSHV sensing through an IFI-16/ASC inflammasome assembled in the nuclei (83). They reported that caspase-1 activation is IFI-16/ASC inflammasome dependent, and it leads to IL-1b secretion. Moreover, the same group proposed that latent KSHV genome is continuously sensed in the nuclei through IFI-16 sensing mechanism (84). Further studies will be needed to shed light on the biological significance of these very exciting findings. Finally, West et al. suggested a role for MAVS and RIG-I dependent signaling mechanisms during KSHV infection (85). Therefore, all of the families of cytosolic sensors have been implicated in the recognition of KSHV. These results clearly indicate that KSHV has a complex interaction with host innate immunity by activating several PRRs. It is conceivable that activation of this network of innate immune receptors is a necessary step in the virus pathogenesis to establish lifelong persistence of the virus infection.

#### **HUMAN PAPILLOMAVIRUSES**

The HPV family encompasses a large number of stable dsDNA viruses (86). Infections with high-risk HPVs are causally associated with the development of anogenital cancers (87). It has been proposed that HPVs evade the innate immune response of the host cells by deregulating immunomodulatory factors such as cytokines and chemokines, thereby creating a microenvironment that favors malignancy (88). The combination of knowledge from the fields of basic HPV virology and vaccinology was the driving force for the successful development of clinically effective vaccines against HPV (89). However, the developed vaccines are prophylactic, not therapeutic, and cover only a subset of HVP types. It is certainly clear that improving our understanding of the interaction of HPV with the innate immune system will improve the probability of success in developing better treatments. Similar to all other viruses described in this review, experimental systems that would be informative about HPV pathogenesis in humans are very limited. The vast majority of studies were performed using virus like particles (VLPs). This approach, and the differences between laboratories in their techniques for virus particles preparation, is partially responsible for the incomplete understanding of HPV biology. For example, the exact mechanism of virus entry into the cell remains incompletely defined (90). Along the same lines, the full spectrum of PRRs relevant to HPV recognition by the cell is yet to be determined.

#### **CELLULAR MEMBRANE AND ENDOSOME SENSING**

The current understanding of the interaction between HPV and PRRs is mostly based on studies aimed to potentiate immunological responses to HPV vaccines by modulating innate immunity. Therefore, research in the field has focused primarily on the role of TLRs. To date there are no publications on the involvement of cytosolic or nuclear sensors in HPV recognition. There is currently no evidence that any cellular PRRs interact with HPV directly [reviewed in Ref. (88)]. A comprehensive review on the role of TLRs in HPV infection has been recently published by Zhou et al. (91). Although TLR4 was suggested to bind HPV L1 directly, these studies were performed using VLPs, and although TLR9 may recognize HPV DNA in the endosomes, it is not clear whether the HPV DNA is exposed in the endosome during natural viral infections (91). More recently, it was described that an HPV16 transcriptional repressor complex associates with the TLR9 promoter, suggesting that blocking this TLR-mediated sensing pathway may be of significance for the virus pathogenesis (92). Collectively, these data indicate that although direct interaction between HPV and PRRs is yet to be shown, the virus does interfere with innate pathogen recognition machinery. In this regard, several recent publications describing how HPV may control cellular responses initiated by PRRs pathways should be mentioned. IL-1beta is a critical cytokine that mediates inflammation and is important for both innate and adaptive immunity. Using immortalized keratinocytes, it was shown that the highrisk HPV16 E6 oncoprotein can abrogate IL-1beta processing and secretion independently of the NALP3 inflammasome (93). The authors further demonstrated that pro-IL-1beta is degraded by a novel proteasome-dependent mechanism via the ubiquitin ligase E6-AP and p53. Moreover, in a panel of HPV-positive tissue samples, the authors found correlation between reduced amounts of IL-1beta and the stage of cellular progression toward cervical cancer (93). HPV was also shown to interfere with innate immune signaling pathways through virus-dependent upregulation of an intrinsic ubiquitin ligase, ubiquitin carboxyl-terminal hydrolase L1 (UCHL1). Upregulation of UCHL1 inhibited TRAF-3 dependent phosphorylation of interferon regulatory factor-3 (IRF-3), and the activation of NF-κB (94). However, the role of this ubiquitin ligase *in vivo* remains unclear as these studies were performed in HPV infected keratinocytes. Using an *in vitro* approach, Sunthamala et al. found that HPV E2 protein interferes with innate immune signaling pathways by downregulating STING and IFN-κ (9). Importantly, they also demonstrated in clinical specimens that

STING and IFN-κ are downregulated in HPV low grade lesions when compared to normal tissues. Conceptually and mechanistically interesting findings were made by Kumar et al., who showed that Langerhans cells from cervical tumors lack TLR9 expression and are functionally anergic to TLR7, TLR8, and TLR9 ligands (95). These data suggest that apart from directly interacting with cellular PRRs, HPV may interfere with innate signaling pathways in neighboring cells in an indirect paracrine manner leading to PRRs signaling inhibition.

#### **HUMAN T-CELL LYMPHOTROPIC VIRUS**

Human T-cell lymphotropic virus-1 belongs to the retroviridae family and is an enveloped, round shaped particle with a single-stranded RNA genome (96). The diseases that induce are diverse and this diversity in clinical manifestations in response to HTLV-1 is likely associated with genetic heterogeneity of the host. The pathologies induced by this virus include the aggressive, fatal T-cell malignancy adult T-cell leukemia (ATL) and a chronic, progressive neurologic disorder called HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP), among others. Unfortunately, the molecular mechanisms underlying the diversity in host responses to HTLV-1 remain unclear (96). The studies of host defense against HTLV-1 have largely focused on understanding the very strong CTL response against the virus. It is puzzling how the virus can establish a persistent infection in the face of such a response. One of the potential mechanisms to escape from the CTL response is the capacity of the virus to downregulate the expression of all but one viral protein (HBZ), thus directly reducing the immunogenicity of the infected cells (97). This capacity of the virus also makes its detection by the innate immune system very challenging. Several reviews have recently summarized the advances in the field of HTLV-1 interactions with the innate immune system (10, 97–99).

The fact that there is not an adequate animal model to study the virus interaction with innate immunity makes advancing in the field very challenging. Rabbits and monkeys models can be used; however, the available immunological tools are scarce. For HTLV-1, mice represent a very poor animal model. Finally, in contrast to the availability of human cervix samples for studies of HPV pathogenesis, access to central nervous system tissue of HTLV-1 infected individuals is not available (100, 101). Therefore, it is not surprising that as of yet there is no evidence of direct recognition of HTLV-1 by PRRs. Furthermore, the role of innate immunity in HTLV-1-associated diseases is not clear (99). Only recently, the induction of an innate immune response to HTLV-1 (102) was reported for the first time. The authors found that cell-free HTLV-1 stimulates pDCs to produce massive amounts of type-I IFN. The proposed mechanism of type-1 IFN induction was the degradation of the viral particles in the endosomal compartments, and consequent exposure of the ssRNA to TLR7. This model was supported by the indirect observations that an endosomal acidification inhibitor and a TLR7 specific blocker drastically inhibited pDC response to HTLV-1 measured by type-1 IFN production. Progress in understanding the innate immune responses to HTLV-1 may come from the use of humanized mouse models (100). For example, reconstitution of mice with WT or TLR7 deficient human cells may reveal the contribution

of the TLR7 innate immune signaling pathway to recognition of HTLV-1.

#### **MERKEL CELL POLYOMAVIRUS**

Merkel cell carcinoma (MCC) is a highly aggressive nonmelanoma skin cancer arising from epidermal mechanoreceptor Merkel cells. In 2008, a novel human polyomavirus, MCPyV, was identified and is now implicated in MCC pathogenesis. Polyomaviruses are small, non-enveloped dsDNA viruses [for a detailed review on polyomaviruses and MCPyV in particular see Ref. (11, 46, 103, 104)]. Although little is known about this newly identified virus, it is plausible that, as with other oncogenic viruses, MCPyV has an array of mechanisms to block the innate immune responses. There is limited information on the innate immune recognition of this virus, as the field is in its infancy. It was reported that MCPyV large T antigen (LT) expression downregulates TLR9 expression in epithelial and MCC-derived cells (105), but nothing is known regarding the direct recognition of the virus by PRRs. More data are clearly needed on the interaction of this virus with the innate immune system.

#### **CONCLUSION**

Over evolutionary times, the battle between the oncogenic viruses and their hosts has arrived at a balance that ensures the survival of both organisms. However, with the current advances in vaccinology and drug development, it is plausible to imagine that we are potentially getting closer to limiting the impact these seven viruses have on the population of the world. Although a complete understanding of all of the complexity of interactions with the native host for all of the oncogenic viruses discussed in this review is still lacking, it is clear that the innate immune system is able to recognize their presence through a network of sensors. Undoubtedly, the understanding of virus interactions with the innate immune system will aid in the development of effective treatments against these pathogens. More research is clearly warranted to devise effective approaches to harness the tools of the innate immune system for elimination of these viral pathogens without negatively affecting their hosts.

#### **ACKNOWLEDGMENTS**

I would like to sincerely thank Wendy Carter and Becky Kinkead (Emory University) for their help with editing the manuscript. Also, thank you to Dmitry Shayakhmetov for sharing ideas and generating very productive discussions. This work was supported by the National Institute of Health (AI065429) and by funding provided by the Children's Healthcare of Atlanta (CHOA) Research Trust.

#### **REFERENCES**


to suppress the keratinocyte's innate immune response. *PLoS Pathog* (2013) **9**:e1003384. doi:10.1371/journal.ppat.1003384


**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.

*Received: 20 April 2014; paper pending published: 01 May 2014; accepted: 09 July 2014; published online: 22 July 2014.*

*Citation: Di Paolo NC (2014) Recognition of human oncogenic viruses by host pattern-recognition receptors. Front. Immunol. 5:353. doi: 10.3389/fimmu.2014.00353 This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Di Paolo. 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.*

### Cancer: a tale of aberrant PRR response

#### **Raunaq Singh Nagi <sup>1</sup> , Ashish Shekhar Bhat <sup>1</sup> and Himanshu Kumar 1,2\***

<sup>1</sup> Laboratory of Immunology, Department of Biological Sciences, Indian Institute of Science Education and Research, Bhopal, India

\*Correspondence: hkumar@iiserb.ac.in

#### **Edited and reviewed by:**

Anton G. Kutikhin, Research Institute for Complex Issues of Cardiovascular Diseases, Russia

**Keywords: cancer, pattern recognition receptors, innate immunity, inflammation, regulation**

#### **INTRODUCTION**

Cancer is a disease of complex etiology and multistep progression, manipulating the regular routes to homeostasis. Any deviation from homeostasis alerts the innate immune system and provokes inflammation. Inflammation is generated by the signaling cascades launched by the pattern recognition receptors (PRRs), the germline encoded molecules dedicated to sense pathogen, or danger-associated molecular patterns (PAMPs or DAMPs) in case of pathogen/foreign matter invasion and intrinsic disturbances, respectively (1–3). Through inflammation, PRRs eliminate stress signals and re-establish homeostasis in the body, via drawing the required cellular machinery to the inflammatory sites. However, the same lympho-reticular infiltrate has been linked with incidence of cancer at the site of chronic inflammation, since 1863, by Rudolf Virchow (4). From 1990s vast amount of literature has accumulated associating soluble and cellular factors of innate immune system with prevalence and progression of cancer. Furthermore,in the past decade,several pathogens have been linked with cancer as well [Ref. (5, 6) and references therein].

Fascinatingly, it is remarkable how the tightly regulated sensory system for stress removal and maintenance of homeostasis functions anomalously and promotes occurrence and progression of cancers (7–9).

#### **PRR-MEDIATED RESPONSES AND CANCER PROGRESSION**

All PRR-dependent pathways activate a particular set of transcription factors to generate appropriate responses. The same factors govern cellular proliferation, apoptosis, tissue remodeling, or angiogenesis, and exhibit a perturbed activity during

cancer. One such key protein is nuclear factor κB (NFκB); up-regulation of which leads to production of pro-inflammatory cytokines. Additionally, it induces antiapoptotic proteins like Bcl2 or inhibitors of apoptotic proteins (IAPs) and angiogenic proteins, such as angiopoietin or vascular endothelial growth factor (VEGF). NFκB also induces nitrous oxide synthase-2 (NOS-2), thus producing nitrous oxide (NO) in the immune cells, which along with reactive oxygen species (ROS) eradicates infected cells by lipid per-oxidation and DNA damage (10–14). Conversely, genomic instability and free radicals thus produced act as DAMPs, leading to sensitization of neighboring PRRs and further immune activation, for instance, the DNA fragments released can activate local DNA sensors, resulting in production of Type I IFN by DAI-TBK1, and activate KRAS pathway of cellular proliferation via TBK1- Sec5 complex, which leads to further activation of NFκB and production of antiapoptotic proteins (15). That is, detouring regular anti-cancer pathway toward proliferation. Also, RONS induce DNA methylases, which lead to methylation and silencing of tumor suppressor and DNA damage repair genes (2, 16–18).

Another pathway crucial in immunity and cancer is the Janus kinases (JAK)-signal transducers and activators of transcription (STAT) pathway. Triggered primarily by interferons and some other mediators, this pathway stimulates various proliferative genes, such as IL-6-mediated induction of myc and CyclinD1/D2 through JAK; also TNFαmediated up-regulation of STAT-3 leading to activation of Ras-mitogen activated protein kinase (MAPK) pathway, which leads to the expression of transcription factor activating protein (AP)-1, and epidermal growth factor (EGFs) along with eukaryotic initiation factor (eIF)-4. AP-1 couples with NFκB, inducing matrix metalloproteinase (MMP)-9, a protein involved in tissue remodeling required during angiogenesis (19). Thus, the pro-inflammatory signal culminates in the production of proteins aiding tumor survival, proliferation, and development of tumor-associated vasculature (18).

Furthermore, NFκB is also involved in the expression of NLRP3, which assembles with apoptosis-associated speck-like protein containing a CARD (ASC) caspase-1 to form multi-protein complexes, the inflammasomes, and responds to DAMPs, especially nucleotides released from damaged or necrotic tissue (due to cytotoxicity of free radicals) (20). Likewise, absent in myeloma (AIM)-2 inflammasomes also organize in response to the formation of DNA adducts (DNA and cytosolic protein HMGB-1) from the dying tissue (21). These assemblies lead to activation of IL-1β–IL-1βR pair; a system found commonly over-activated in many cancers (2, 22). Additionally, NFκB also generates cyclo-oxygenase-2 (COX-2) enzyme, which converts arachidonic acid into prostaglandin-E2 (PGE-2), one of the dual (pro-inflammatory and/or anti-inflammatory) mediators of immune response. PGE-2 enhances Tcell activation and represses B-cell activity (23, 24). Another common enzyme, activation-induced deaminase (AID), also induced by NFκB, involved in somatic hypermutation and class switch recombination in B-cells, causes genome instability and releases additional DAMPs into the microenvironment (25). Thus, the immune mediators produced for protection can divert inflammation toward pro-tumor facet (26).

<sup>2</sup> Laboratory of Host Defense, WPI, Immunology Frontier Research Center, Osaka University, Osaka, Japan

A set of pro-inflammatory cytokines consisting of TNF-α and IL-1 and 6 is essentially tumor directing. TNF-α promotes tumor initiation and DNA damage. It also up-regulates hypoxia-inducible factor (HIF)-1α (attributed to the increasingly low oxygen levels due to multiplying cells) aiding in angiogenesis (27). IL-1β aids in tumor invasiveness and adhesion required during metastasis to new sites. IL-1α, the membrane bound form, induces IL-1 expression, associated with tissue damage, compensatory cell proliferation, and activation of JAK-STAT pathway, as seen in hepatocellular carcinomas and colitis-associated cancers (22, 28).

Cigarette smoking has long been associated with incidence of cancer. Cigarette smoke contains numerous compounds with known cytotoxicity, mutagenicity, and carcinogenicity, most of which are particulate. Stable ROS present in the smoke damage DNA and cause lipid per-oxidation, sensitizing the PRRs present from the buccal cavity to lungs leading to increased IL-8 and TNF-α (11). In addition, both, NFκB and AP-1 are up-regulated exaggerating the pro-inflammatory signal, at the same time homeostatic activity of both is reduced, compensating normal immune response. Such a response coupled with prolonged exposure can spontaneously lead to cellular transformations and their expansion (29, 30).

#### **ROLE OF CELLULAR COMPONENTS OF INNATE IMMUNITY IN CANCER PROGRESSION**

Specialized cells of the immune system are equipped with PRRs, and are responsible for clearance of diseased/damaged cells. Pro-inflammatory cytokines draw these cells toward the inflammatory site and direct them for removal of pathogens, particulates, or immune debris. These populations recede as the signal resolves. Since Virchow proposed their role at the site of chronic inflammation and cancer, a number of cellular populations and their effector responses have been ascertained for the same. In cases of prolonged exposure to PAMPs/DAMPs, infiltrating cells fail to withdraw and differentiate into M2 macrophages, identified as tumor-associated macrophages (TAMs), an integral population programed for tissue remodeling and tumor progression. Upon activation of their PRRs, TAMs promote various properties of cancer by releasing a range of inflammatory and angiogenic bio-chemicals. These cells stimulate proliferation of stromal tissue and macrophages by growth factors such as platelet-derived growth factor (PDGF) and colony stimulating factor (CSF)-1 respectively (31). Moreover, they organize a route to metastasis by digesting the substratum, basal lamina and release inactive growth factors via, MMPs. In addition, they assist in cellular movements by releasing cell adhesion molecules such as intercellular adhesion molecule (ICAM)-1 (32–34).

Another such population, the NK cells meant to carry out cytotoxic clearance of all the cells which do not express human leukocyte antigen (HLA) A/B/C and thus fail to activate the membrane-expressed inhibitory receptors (NKp30/44/46). NK cells are also responsible for killing any cell, irrespective of HLA tag, which presents them with stress/abnormality/tumor-associated antigens, via, activating receptor (NKG2D). In addition, they also participate in antibody-dependent cell-mediated cytotoxicity (ADCC), on cells tagged through FCRγIII (35, 36). A number of cytokines activate NK cells and turn them into lymphokine-activated killers, causing them to display their killing property at the site of recruitment (37). Tumor cells escape NK cells by blocking the activating receptor. Also, even if recruited, NK cells can only bring about killing of the outer cells in a solid tumor (38, 39). Furthermore, a reduction in NK cell cytotoxic function as well as NK cell dependent tumor surveillance is evident as a tumor directing effect of cigarette smoke (29, 30).

Another newly characterized population of cells called myeloid-derived suppressor cells (MDSCs) are recruited by inflammatory mediators, which inhibit the anti-tumor responses and release pro-tumor molecules, like, NOS-2 and TGF-β. Arginase-1 and indolamine-2,3 dioxygenase produced by MDSCs are involved in silencing the anti-tumor immunity by reducing Th1 activization (40, 41).

Soluble and cellular factors work in union. Cellular proliferation at an

enhanced rate at tumor sites gives rise to hypoxia, low concentration of oxygen that stabilizes HIF-1α, which activates NFκB to secrete an angiogenic protein, VEGF and also induces expression of IL-12 and TNF-α. These mediators collectively induce STAT-3 production that upregulates PGE-2 which further recruits NK cells and their cytotoxic function releases DAMPs in vicinity attracting TAMs which aid in shaping neo-vasculature and creating area for growing cell mass. Deregulation of Th1 responses by PGE-2 and MDSC/TAMs creates an imbalance (42– 44). HIF-1α which causes glycolytic environment, such conditions may tip the balance into a state when IL-12 production is replaced by IL-23, that is, a shift from anti-tumor to pro-tumor (45, 46). In this manner, the PRR triggered pathways to eliminate PAMPs/DAMPs, are rerouted in a manner that leads to exaggeration of the initial signal causing chronic inflammation; moreover eliciting other pathways concluding in tumor growth and metastasis (1, 47, 48).

#### **REGULATORY MECHANISMS AND PROGRESSION OF CANCER**

To maintain homeostasis in the body, all cellular processes, including PRR generated immune responses are regulated by various mechanisms. This control is exercised at various levels. At transcriptional level certain cytokines can inhibit transcription factors, such as IL-4/13 which hinder NFκB. Alternatively; some cytokines directly inhibit other cytokines at protein level, such as inhibition of TNF-α by IL-10. At post-transcriptional level microRNAs play a crucial role by either resolving inflammation or potentiating pro-tumor effects of cytokines (18). For instance, TNF-α/IL-1β induced mir-146a inhibits IRAK/TRAF6, that is, the downstream signaling of TLR pathway, thus resolving inflammation (23). In contrast IL-6 induced mir-21 targets tumor suppressor genes, such as phosphatase and tensin homolog (PTEN), programed cell death (PDCD)-4, and others, thereby hampering the antitumor effects (49–52). Another microRNA found commonly up-regulated in cancer, miR-155, induces NOS-2 and inhibits apoptosis by down-regulating TP53INP1 (a downstream molecule of p53 signaling)

(53). Thus, a slight imbalance in the proinflammatory and anti-inflammatory signals can shift the equilibrium toward oncogenic transformations (18).

Pattern recognition receptor elicited signaling cascades induce synthesis of several zinc-finger proteins that help in regulating the inflammatory signals. For example, ZC3H12a/c and Zfp36 proteins degrade the cellular mRNA coding for proinflammatory cytokines; while ZAP proteins directly degrade the viral RNA, to be precise, the PAMP itself (54, 55). Thus, these proteins degrade the PAMPs or signal generated by them to curb inflammation. In addition, certain PRRs themselves restrain the downstream signaling such as inhibition of IPS-1, adaptor molecule of RLR, by NLRX-1. Mutations within these genes or their regulatory elements or imbalance at cellular level can skew the balance toward tumorigenesis (6, 55) (**Figure 1**).

#### **NEW FACET IN THE FIELD**

A relatively new but noteworthy field in analyzing cancer biology is polymorphisms. Single nucleotide polymorphisms (SNPs) are single base changes in the DNA, which may produce totally drastic effects on the structure and function of the encoded molecules. PRRs and the associated machinery, such as the adaptors or receptors are also susceptible to such polymorphisms and many such SNPs have been reported. SNPs could promote the anti-tumor effect if they prevent the PRR cascade from commencing, or support the pro-tumor effect, if they cause spontaneous induction of signaling without any stimulus. The cytoplasmic domains of the PRRs and cytokine receptors are of prime importance in this context as they form the docking site for progression of inflammatory response. Many polymorphisms have been identified to be associated with cancers of various origins. Mostly CLRs and RLRs have been associated with sensing PAMPs of oncogenic origin, and polymorphisms in their genes have been correlated with cancers of mesoderm, endoderm, and also ectoderm origin. Also several genes and mutations have been correlated with cigarette smoke in association with cancers (56). Still, a rather comprehensive endeavor would be required to establish the integral role of these SNPs at the molecular level to outline their part in incidences and progression of cancers [Ref. (6) and references therein, Ref. (56)].

#### **CONCLUSION**

Recognition of pathogen/stress is one of the essential processes of the host. Tumor cells are in fact, abnormal cells which are steadily eliminated through PRR-mediated pathways. However, hyper or anomalous behavior of same pathways can divert the protective route toward malignancy, by contributing to abnormal proliferation, angiogenesis, or modifying tissue architecture. The origin of anomalous behavior maybe external and internal, such as pathogen/foreign insults tissue damage/necrosis or mutations and polymorphisms in vital signaling components, respectively. Unfolding the root of these irregularities and malfunctions shall help in better understanding of the disease and thus, create new and personalized prospects for treatment.

#### **ACKNOWLEDGMENTS**

This work is supported by research grants number SR/S2/RJN-55/2009, BT/PR6009/GBD/27/382/2012 from Department of Science and Technology (DST) and Department of Biotechnology (DBT), Government of India to Himanshu Kumar and Intramural Research Grant from IISER, Bhopal, India; Raunaq Singh Nagi would like to thank DST INSPIRE for doctoral fellowship support (IF 130017).

#### **REFERENCES**


the link? *Biochem Pharmacol* (2006) **72**:1605–21. doi:10.1016/j.bcp.2006.06.029


*Cell* (2008) **135**:1028–38. doi:10.1016/j.cell.2008. 09.062


creates local immune suppression by inducing indoleamine 2,3 dioxygenase. *Proc Natl Acad Sci U S A* (2008) **105**:17073–8. doi:10.1073/pnas. 0806173105


(2011) **31**:986–1000. doi:10.1161/ATVBAHA.110. 207449


is an RNase essential for controlling immune responses by regulating mRNA decay. *Nature* (2009) **458**:1185–90. doi:10.1038/nature07924


**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.

*Received: 26 March 2014; accepted: 27 March 2014; published online: 09 April 2014.*

*Citation: Nagi RS, Bhat AS and Kumar H (2014) Cancer: a tale of aberrant PRR response. Front. Immunol. 5:161. doi: 10.3389/fimmu.2014.00161*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Nagi, Bhat and Kumar. 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.*

### Pattern recognition receptors and autophagy

### **Ji Eun Oh and Heung Kyu Lee\***

Laboratory of Host Defenses, Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Sulev Kõks, University of Tartu, Estonia Mikhail A. Gavrilin, Ohio State University, USA Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Heung Kyu Lee, Laboratory of Host Defenses, Graduate School of Medical Science and Engineering, Korea Advanced Institute of Science and Technology, 291 Daehak-ro, Daejeon 305-701, South Korea e-mail: heungkyu.lee@kaist.ac.kr

#### **INTRODUCTION**

Innate immune signaling pathways are initiated when microorganism-specific pathogen-associated molecular pattern (PAMP) molecules are recognized by host pattern recognition receptors (PRRs) (1). PRRs can be classified based on their site of localization (e.g., plasma membrane, endosomal vesicles, and cytoplasm) or by molecular structural similarities. PRRs classified by structural similarity include toll-like receptors (TLRs), nucleotide oligomerization domain (NOD)-like receptors (NLRs), C-type lectin receptors (CLRs), and RIG-I-like receptors (RLRs).

The TLRs, which reside both within the cell surface membrane (TLR 1, 2, 4, 5, and 6) and in endosomal compartments (TLR 3, 7, 8, and 9), are the most well-characterized PRRs. After recognition of PAMPs, TLRs initiate downstream signaling pathways via myeloid differentiation primary response gene 88 (MyD88) or Toll/interleukin (IL)-1 receptor (TIR) domain-containing adapter-inducing interferon (IFN)-β (TRIF), ultimately activating the transcription factors nuclear factor (NF)-κB and activator protein-1 (AP-1) or IFN regulatory factor 3 (IRF3). Activation of NF-κB and AP-1 results in the production of proinflammatory cytokines, and activation of IRF3 results in the production of type I IFNs (2). NLRs are cytoplasmic members of the PRR family, and more than 20 NLRs have been identified in mammals. NOD1 and NOD2 – the first NLRs identified in mammals – recognize cytoplasmic bacterial cell wall components, eventually activating NF-κB to induce the production of proinflammatory cytokines. In addition, NLRs act as sensory proteins in inflammasomes (which serve as platforms for protein complexes involved in innate immunity) and activate inflammasome-associated caspase-1 for pro-IL-1β and pro-IL-18 processing. RLRs and other cytosolic sensors primarily recognize microbial nucleic acids in the cytosol.

The immune system senses exogenous threats or endogenous stress through specialized machinery known as pattern recognition receptors (PRRs).These receptors recognize conserved molecular structures and initiate downstream signaling pathways to control immune responses. Although various immunologic pathways mediated by PRRs have been described, recent studies have demonstrated a link between PRRs and autophagy. Autophagy is a specialized biological process involved in maintaining homeostasis through the degradation of long-lived cellular proteins and organelles. In addition to this fundamental function, autophagy plays important roles in various immunologic processes. In this review, we focus on the reciprocal influences of PRRs and autophagy in modulating innate immune responses.

**Keywords: autophagy, toll-like receptors, RIG-I-like receptors, NOD-like receptors, inflammasomes, cytosolic DNA sensors**

> RLRs composed of retinoic acid-inducible gene I (RIG-I) and melanoma differentiation-associated gene 5 (MDA-5) are caspaserecruiting domain (CARD)-containing RNA helicases that recognize double-stranded RNA and signal through IFN-β promoter stimulator-1 [IPS-1; also known as mitochondrial antiviral signaling (MAVS), virus-induced signaling adaptor (VISA), or Cardif] to subsequently activate IRF3 and NF-κB (3).

> Autophagy is a highly conserved homeostatic process in eukaryotic cells that degrades long-lived cellular proteins and organelles. There are at least three types of autophagy: microautophagy, chaperone-mediated autophagy, and macroautophagy (4). During microautophagy, continuous degradation of cytosolic constituents close to the lysosomes occurs through budding of the lysosomal membrane. In chaperone-mediated autophagy, proteins containing a "KFERQ" motif are transported into the lysosomal lumen via Lamp2a for subsequent degradation. During this process, cytosolic chaperones such as HSC70 recognize the KFERQ motif and facilitate importation of substrates into the lysosomes (5). Macroautophagy, which is the primary route of degradation, involves the formation of a double-membrane vesicle known as an autophagosome. During this process, long-lived cellular components are first surrounded by an elongated cup-shaped membrane that forms the autophagosome, which then matures and fuses with lysosomes for degradation of the internalized materials (6). Recent research has suggested that autophagy is a selective process, in which specific adaptors such as p62 target ubiquitinated substrates for selective degradation (7).

> The molecular processes involved in autophagy consist of three distinct stages. Initiation of isolation membrane formation requires complex interaction between autophagy-related gene *(Atg) 6* (also known as beclin-1) and type III [phosphatidylinositol

3-kinase (PI3K). Elongation of the isolation membrane and termination of autophagosome formation are regulated by at least two ubiquitin-like molecules: microtubule-associated protein 1 lightchain 3 (LC3; mammalian homolog of yeast Atg8) and Atg12 (8, 9). Atg12 is conjugated to Atg5 through the sequential actions of the E1- and E2-like enzymes Atg7 and Atg10. Association of Atg12–Atg5 conjugates with Atg16 in turn facilitates elongation of the isolation membrane and catalyzes LC3 conjugation. The C-terminal amino acids of LC3 are cleaved by Atg4 and then transferred to phosphatidylethanolamine (PE) in the newly formed isolation membrane by the E1- and E2-like enzymes Atg7 and Atg3. Upon completion of the autophagosome, LC3 remains in the autophagosomal lumen (thus serving as an autophagosomal marker), whereas the Atg12–Atg5–Atg16 complex dissociates from the outer autophagosomal membrane. The outer membrane of the autophagosome eventually fuses with the lysosome for degradation of the autophagosomal contents and membrane (10).

Autophagy was originally identified as a mechanism for maintaining homeostasis through the degradation of long-lived proteins and recycling of intracellular organelles (11). However, autophagy is now recognized as playing multiple roles in various biological processes. For example, dysregulation of autophagy has been linked to many diseases, including cancer. Recent studies have revealed that PRRs activate autophagy to enhance immune responses against pathogens and that PRR-induced signaling pathways are regulated by autophagy to prevent excessive inflammation. In this review, we focus on the interactive role of PRRs and autophagy in controlling innate immune responses.

#### **TLRs AND AUTOPHAGY**

Toll-like receptors, which bind to conserved microbial molecular structures and initiate downstream signaling pathways, are the most thoroughly characterized type of PRR (1). Xu et al. (12) were the first to report that TLR4 stimulation activates autophagy to enhance elimination of phagocytosed mycobacteria. The authors found that stimulation of TLR4 with lipopolysaccharide (LPS) induces autophagosome formation in primary human macrophages and RAW 264.7 murine macrophages. This pathway is mediated by the TRIF–p38 axis rather than MyD88 (**Figure 1A**). In their study, Xu et al. provide an evidence of close relationship between autophagy and TLR-mediated innate immunity. In addition to LPS-induced autophagy, ligands of TLR3 and TLR7 also activate autophagy. Two different ligands of TLR7, single-stranded RNA (ssRNA) and imiquimod, induce autophagosome formation, characterized by LC3 puncta formation in murine macrophages [**Figure 1A**; Ref. (13)]. This process occurs via MyD88 and ultimately results in the killing of intracellular mycobacteria, even though mycobacteria are normally not associated with TLR7 signaling.

Recently, several studies reported that TLR2 stimulation by various pathogens induces autophagy (14, 15). In response to *Listeria monocytogenes*, macrophages deficient in the TLR2 and NOD/receptor-interacting protein 2 (RIP2) pathways show defective autophagy induction and fail to colocalize bacteria within autophagosomes [**Figure 1B**; Ref. (14)]. Autophagy induction in this process was found to be dependent on the extracellular signal-regulated kinase (ERK) pathway. Another study showed that *Staphylococcus aureus*-mediated stimulation of TLR2 in RAW 264.7 mouse macrophages induces phagocytosis and autophagy. In particular, knockdown of *TLR2* was shown to attenuate *S. aureus*induced phosphorylation of macrophage c-Jun N-terminal kinase (JNK) but not phosphorylation of p38 or ERK (15). Collectively, these data indicate that TLR2 stimulated by invading microbes could mediate autophagy induction and promote the clearance of pathogens, despite the different pathways involved. Shi and Kehrl (16) revealed that various TLR agonists, including TLR1, TLR3, TLR5, TLR6, and TLR7, trigger autophagy induction through MyD88 and TRIF, which interacts with beclin-1. Beclin-1 is critical for the initiation of autophagosome formation. Interaction of beclin-1 with TLR-signaling pathway adaptor molecules partially inhibits the binding of beclin-1 to B cell lymphoma 2 (Bcl-2).

In addition to its role in autophagy induction, TLR-signaling is also utilized by Atg proteins to mediate phagosome maturation. Phagocytosis of the fungal cell wall component zymosan promotes the rapid recruitment of LC3 to phagosomes and facilitates their fusion with lysosomes (17). In RAW 264.7 cells, phagocytosis of Pam3CSK4-coated latex beads involves recruitment of LC3 to the phagosomes. This process is dependent on TLR2 but not MyD88 and requires both Atg5 and Atg7. However, LC3 translocation to phagosomal membranes is not associated with double-membrane structures, which is a unique feature of autophagosomes. Collectively, these results demonstrate a novel way in which the autophagic machinery is utilized for phagocytosis after TLR activation. Another recent study characterized the role of non-canonical autophagy in type I IFN secretion in response to DNA-immune complexes (DNA-ICs) (18). Upon stimulation of TLR9, which responds to double-stranded DNA (dsDNA) and facilitates the production of proinflammatory cytokines and type I IFNs, IFN-α is produced by the convergence of the phagocytic and autophagic pathways, a process termed LC3-associated phagocytosis (LAP). LAP occurs in response to DNA-ICs but not soluble ligands. In addition, LAP requires FcγR engagement, which controls TLR9 and LC3 recruitment (**Figure 1C**). The study of Henault et al. revealed the function of non-conventional autophagy in regulating type I IFN signaling in phagosomes. Moreover, their results suggest a mechanism for the uncontrolled production of type I IFNs induced by pathogenic DNA-ICs in systemic lupus erythematosus, which may lead to development of new therapeutic targets for treating this disease.

As previously discussed, induction of autophagy through TLR activation directly promotes pathogen clearance to enhance host protection. However, autophagy also enhances antiviral defenses by facilitating delivery of cytosolic viral PAMPs to endosomal TLRs. Viral nucleic acids endocytosed by host cells are recognized by endosomal TLR7 and TLR9. After recognition, signaling through NF-κB and IRF7 induces production of proinflammatory cytokines and type I IFNs, respectively. In response to vesicular stomatitis virus (VSV) infection in plasmacytoid dendritic cells (pDCs), endosomal TLR7 recognizes the replicating virus in the cytosol rather than the viral genome. How these cytosolic replication intermediates gain access to endosomal TLR7 was demonstrated by Lee and colleagues. These authors showed that autophagy facilitates the delivery of cytosolic PAMPs to the lysosomes, activating TLR7 signaling [**Figure 1D**; Ref.

**FIGURE 1 |TLRs and autophagy. (A)** TLR4 stimulated with LPS induces autophagosome formation via TRIF–p38 mitogen-activated protein kinase (MAPK) signaling axis. Similarly, TLR7 activation by ssRNA and imiquimod also promotes autophagy induction. These processes facilitate fusion of the autophagosomes with the lysosomes, which in turn finally result in the killing of intracellular mycobacteria. **(B)** TLR2 and NOD1/2 mediate autophagy induction in response to Listeria monocytogenes. Autophagy induction in this process requires the ERK pathway. **(C)** LAP mediates the production of type I IFNs induced by TLR9 activation in response to

DNA-ICs. Large DNA-ICs engulfed by phagocytosis are internalized using FcγR, which recruits LC3 and TLR9–UNC93B to the phagosomes. Unlike NF-κB-dependent inflammatory cytokine production, LC3 is required for the trafficking of TLR9 into a specialized IRF7 signaling compartment for type I IFNs secretion. In response to the endocytosed CpG-ODN, however, AP3 is required for delivery to the IRF7 signaling compartment. **(D)** Autophagy facilitates the viral sensing by delivery of cytosolic viral PAMPs to lysosomes, enabling endosomal TLRs to sense of virus and subsequently activating type I IFNs production.

(19)]. Consequently, pDCs lacking Atg5 cannot secrete IFN-α or IL-12p40 following VSV infection. Atg5-deficient mice are also susceptible to systemic VSV infection *in vivo*. Interestingly, in pDCs infected with herpes simplex virus-1 (HSV-1), which is recognized by TLR9, Atg5-deficient cells fail to produce IFNα, whereas the IL-12 response of these cells is not affected. Thus, the precise mechanisms by which the NF-κB and IFN-α signaling pathways are controlled by autophagy remain to be determined (20).

#### **NLRs AND AUTOPHAGY**

NOD-like receptors, which recognize bacterial cell wall components such as peptidoglycan in the cytosol, also play an important role in autophagy. Studies have shown that NOD1 and NOD2 activate autophagy in response to bacterial invasion (21, 22). In mouse embryonic fibroblasts (MEFs), NOD2 recruits Atg16L1 to the plasma membrane at the site of bacterial entry, in turn facilitating bacterial trafficking to the autophagosomes and fusion of the autophagosomes with the lysosomes to promote bacterial clearance and antigen presentation via MHCII [**Figure 2A**; Ref. (22)]. Another study using human DCs showed that stimulation of NOD2 with muramyl dipeptide induces autophagosome formation and consequently enhances MHCII-associated antigen presentation. In this process, autophagic proteins such as Atg5, Atg7, Atg16L1, and receptor-interacting serine–threonine kinase 2 are required [**Figure 2A**; Ref. (21)]. The intracellular bacterial sensors NOD1 and NOD2 link the autophagic machinery via Atg16L1, thereby enhancing both bacterial clearance and protective immunity. However, the role of Atg16L1 in NOD-derived inflammation remains unclear. A recent study demonstrated that Atg16L1 suppresses NOD-induced inflammatory responses in an autophagy-independent manner (23). Atg16L1 blocks the activation of RIP2 by reducing the level of RIP2 polyubiquitination and diminishing the incorporation of RIP2 into NOD signaling complexes. This process appears to be specific to Atg16L1, given that knockdown of *Atg5* or *Atg9a* does not affect the NOD response. In addition, autophagy-incompetent truncated forms of Atg16L1 retain the capacity to regulate NOD-driven cytokine responses. Interestingly, NOD2 mutations and single-nucleotide polymorphisms in *Atg16L1* are well-known features of Crohn's disease. Collectively, the above-mentioned studies suggest that a functional relationship exists between NOD2 and Atg16L1 in Crohn's disease.

Inflammasomes are protein complexes in which NLRs serve as sensory proteins that promote innate immunity by enabling the maturation of pro-IL-1β and pro-IL-18 through activation of pro-caspase-1. Many studies have described regulation of inflammasomes by autophagy and vice versa. Suppression of inflammasomes by autophagy was first reported in 2008 by Saitoh et al. (24), who showed that Atg16L1-deficiency results in increased production of IL-1β and IL-18 following LPS stimulation. Atg16L1 is an

essential component of the autophagosome, forming a complex with Atg5–Atg12 conjugates, resulting in LC3–PE conjugation. Thus, Atg16L1-deficient macrophages impaired in autophagosome formation induce TRIF-dependent activation of caspase-1, leading to excessive production of IL-1β in response to LPS. Considering that *Atg16L1* is an important gene in the development of Crohn's disease, endotoxin-induced inflammasome activation in Atg16L1-deficiency could be involved in the occurrence of Crohn's disease. Although the above-mentioned data suggest that inflammatory responses are regulated by autophagy, the mechanism by which autophagy regulates cytokine secretion is not clear. Two fascinating studies have provided evidence indicating that mitochondria play a critical role controlling innate immunity mediated by NLRP3 inflammasomes (25, 26). Zhou and colleagues demonstrated that blocking autophagy, especially mitophagy (mitochondrial autophagy), results in the accumulation of damaged, reactive oxygen species (ROS)-generating mitochondria, which in turn activate NLRP3 inflammasomes. Of note, inhibition of mitochondrial activity suppresses both ROS generation and inflammasome activation [**Figure 2B**; Ref. (26)]. Similarly, Nakahira et al. (25) showed that depletion of the autophagic proteins LC3B and beclin-1 induce excessive secretion of IL-1β and IL-18, which is mediated by accumulation of dysfunctional mitochondria and cytosolic translocation of mitochondrial DNA (mtDNA) following LPS and adenosine triphosphate (ATP) stimulation. The NALP3 inflammasome, which is critical for the activation of caspase-1 in response to LPS and ATP stimulation, contributes to the release of mtDNA into the cytosol [**Figure 2B**; Ref. (25)]. Together, these studies indicate that regulation of NLRP3-induced inflammatory processes by autophagy is dependent on mitochondrial integrity.

Autophagy also limits the inflammatory responses resulting from inflammasome activation in another way. A recent study

**FIGURE 2 | Interactions between NLRs or inflammasomes and autophagy. (A)** Activation of NOD1 and NOD2 by bacteria induces autophagosome formation, which leads to facilitating bacterial clearance and MHC class II-associated antigen presentation. In this process, autophagy proteins such as Atg5, Atg7, and Atg16L1 are required. **(B)** Autophagy (especially mitophagy) regulates NLRP3 inflammasomesinduced inflammatory responses by quality control of mitochondrial integrity. Blocking mitophagy leads to the accumulation of damaged, ROS-generating mitochondria, which in turn activates NLRP3 inflammasomes. The NLRP3 inflammasome also contributes to the

release of mtDNA into the cytosol, enhancing further activation of NLRP3 inflammasomes in a feed-forward circuitry. This process finally activates caspase-1 and results in the excessive production of IL-1β and IL-18. **(C)** Autophagy induced by inflammatory signals targets ubiquitinated inflammasomes, thereby limiting IL-1β production by destruction of inflammasomes. Induction of AIM2 or NLRP3 inflammasomes triggers the activation of RalB to bind to Exo84, which serves as a platform for the formation of isolation membranes. Autophagy engulfs ubiquitinated, assembled inflammasomes through autophagic adaptors such as p62, in turn limiting inflammasome activity.

showed that autophagy induced by inflammatory signals targets ubiquitinated inflammasomes, thereby limiting IL-1β production through inflammasome destruction (27). Induction of absent in melanoma 2 (AIM2) or NLRP3 inflammasomes triggers nucleotide exchange on RalB and autophagosome assembly through Exo84, which serves as a platform for the formation of isolation membranes (28). During autophagy, ubiquitinated assembled inflammasomes are engulfed through autophagic adaptors such as p62, which contain both ubiquitin-associated domains and LC3-interacting regions that recognize ubiquitinated molecules and assist their entry into the autophagy pathway (**Figure 2C**). Thus, activation of inflammasomes induces autophagy, which in turn limits inflammasome activity via autophagic engulfment in order to maintain homeostasis as it pertains to inflammation.

Conversely, NLRs also negatively regulate autophagy. NLRC4, NLRP3, NLRP4, and NLRP10 interact with beclin-1, and NLRP4 in particular has a strong affinity for beclin-1. Following invasion by bacteria such as group A streptococci (GAS), NLRP4 recruits GAS-containing phagosomes and transiently dissociates from beclin-1, enabling the initiation of beclin-1-mediated autophagy. Moreover, NLRP4 physically interacts with the class C vacuolar protein-sorting complex, resulting in inhibition of autophagosome and endosome maturation (29). Taken together, the available data indicate that homeostasis is maintained through reciprocal regulation of NLR activation and autophagy.

#### **OTHER CYTOSOLIC SENSORS AND AUTOPHAGY**

Viral recognition in most cell types is mediated by cytosolic sensors such as RIG-I and MDA-5. RIG-I and MDA-5, both of which are RLRs, signal through IPS-1 to activate the transcription factors IRF3 and NF-κB, leading to cytokine production. Several studies have revealed that the RLR signaling pathway might be controlled by autophagy (30, 31). In Atg5- or Atg7 deficient MEFs, which lack Atg5–Atg12 conjugates, type I IFNs are overproduced following VSV infection. In contrast, overexpression of Atg5 or Atg12 results in suppression of IFN signaling. The Atg5–Atg12 conjugates directly interact with the CARD domains of RIG-I and IPS-1, inhibiting subsequent RLR signaling [**Figure 3A**; Ref. (30)]. These data indicate that autophagyrelated proteins act as negative regulators of RLR-mediated antiviral responses. Similarly, Tal and colleagues revealed that Atg5 deficient cells overproduce IFNs through enhanced RLR signaling in response to VSV infection (31). However, the authors explained that dysfunctional mitochondria and mitochondriaassociated IPS-1 that accumulate in the absence of autophagy enhance RLR signaling. Data suggest that ROS associated with dysfunctional mitochondria are the primary inducers of these responses, as increased mitochondrial ROS production following treatment with rotenone, which is independent of autophagy, also results in amplification of RLR signaling [**Figure 3A**; Ref. (31)]. Consequently, autophagy contributes to homeostatic regulation of antiviral responses through control of RLR signaling pathways.

The cytosolic DNA sensor stimulator of IFN genes (STING) is also associated with autophagy. In a study to determine the mechanism of mycobacterial clearance, ubiquitin-mediated autophagy targeting *M. tuberculosis* was shown to be activated by the STING-dependent cytosolic sensing pathway (32). In case of wild-type *M. tuberculosis*, which expresses the virulence factor extra-embryonic spermatogenic homeobox 1 (ESX-1) secretion system, mycobacterial DNA may be exposed to the host through ESX-1-mediated permeabilization of the phagosomal membrane.

**(A)** Autophagy negatively regulates type I IFNs production after viral infection. The Atg5–Atg12 conjugates directly interact with the CARD domains of RIG-I and IPS-1, inhibiting subsequent RLR signaling pathway and type I IFNs production. In another way, autophagy regulates RLR signaling by acting as a scavenger of dysfunctional mitochondria as well as mitochondria-associated IPS-1. Following dsDNA stimulation, STING is translocated from the ER to the Golgi apparatus and assembled with TBK1, which phosphorylates the transcription factor IRF3. During this process, Atg9a colocalizes with STING in the Golgi apparatus and controls the assembly of STING. **(B)** During mycobacterial clearance, ubiquitin-mediated autophagy targeting M. tuberculosis is shown to be activated by the SITNG-dependent cytosolic sensing pathway. Mycobacterial extracellular

permeabilization of the phagosomal membrane, is recognized by the STING-dependent cytosolic pathway. The ubiquitinated bacterial DNA, which binds to the autophagosome-associated protein LC3 via adaptor protein p62 and NDP52, is targeted to the selective autophagy pathway. **(C)** Cytosolic DNA-sensing cGAS produces cGAMP, which binds to and activates the adaptor protein STING, thus leading to the production of type I IFNs. Direct interaction between cGAS and Beclin-1 suppresses cGAMP synthesis. Moreover, this interaction activates PI3K III-induced autophagy, enhancing the autophagy-mediated degradation of pathogen DNA. cGAMP generated by cGAS initially activate STING-dependent type I IFN responses. However, they subsequently trigger negative-feedback control of STING activity through phosphorylation of STING by serine/threonine ULK1 (ATG1).

The released DNA may in turn be recognized by the STINGdependent cytosolic pathway. The bacteria are consequently surrounded by ubiquitin chains, and the ubiquitin and LC3-binding autophagic adaptors p62 and nuclear dot protein 52 recruit autophagy components that target the bacilli to the selective autophagy pathway (**Figure 3B**). Other studies involving dsDNA viruses such as HSV-1 or human cytomegalovirus revealed that STING plays a role in autophagy induction (33, 34).

Conversely, autophagy may also negatively regulate STINGdependent IFN responses. After dsDNA stimulation, Atg9a colocalizes with STING in the Golgi apparatus, where it controls the assembly of STING (35). The loss of Atg9a, but not that of Atg7, promotes the translocation of STING from the Golgi apparatus and its assembly with TBK1, thus inducing aberrant activation of type I IFN responses (**Figure 3A**). Collectively, these findings demonstrate the reciprocal regulation of autophagy and STING-dependent cytosolic pathways.

Recently, cyclic guanosine monophosphate–adenosine monophosphate (GMP–AMP) synthase (cGAS) was shown to be a cytosolic DNA sensor that activates the type I IFN pathway (36). Cytosolic DNA-sensing cGAS produces cyclic GMP– AMP (cGAMP), which binds to and activates the adaptor protein STING, thus leading to the production of type I IFNs. A very recent study showed that direct interaction between cGAS and beclin-1 suppresses cGAMP synthesis, leading to dampened type I IFN responses following dsDNA stimulation or HSV-1 infection. Moreover, this interaction activates PI3K III-induced autophagy through release of Rubicon, a negative regulator of autophagy, thus enhancing the autophagy-mediated degradation of pathogen DNA to prevent excessive immune stimulation [**Figure 3C**; Ref. (37)]. Similarly, cyclic dinucleotides contribute to the negative regulation of the STING pathway by activating UNC-51-like kinase (ULK1/Atg1). Cyclic dinucleotides generated by cGAS initially activate STING-dependent type I IFN responses; however, they subsequently trigger negative-feedback control of STING activity through phosphorylation of STING by serine/threonine ULK1/Atg1 [**Figure 3C**; Ref. (38)]. Taken together, these data suggest that autophagy controls the excessive and persistent immune responses mediated by cytosolic DNA-sensing pathways.

#### **CONCLUSION**

In this review, we describe the close interaction between PRRs and autophagy in various immunologic conditions. PRRs are not only involved in autophagy induction but also in the promotion of phagosomal maturation mediated by Atg proteins when pathogenic bacteria invade host cells. In addition, autophagy facilitates the delivery of viral PAMPs and TLR9 trafficking for type I IFN production. Autophagy regulates PRR-induced inflammation in various ways to prevent excessive inflammatory responses, and conversely, PRR signaling also controls autophagy. Collectively, the available data indicate that targeting autophagy would allow us to enhance pathogen clearance or suppress PRR-mediated inflammatory conditions, such as those associated with autoimmune diseases. Therefore, a more detailed analysis of how we could control autophagy is recommended.

#### **ACKNOWLEDGMENTS**

We thank Sang Eun Oh and Jeongsu Park for their help with the figure. This work was supported by the National Research Foundation (NRF-2012R1A1A2046001, NRF-2012M3A9B4028274, NRF-2013R1A1A2063347) and the Converging Research Center Program (2011K000864) funded by the Ministry of Science, ICT, and Future Planning of Korea.

#### **REFERENCES**


membrane at the site of bacterial entry. *Nat Immunol* (2010) **11**:55–62. doi:10.1038/ni.1823


cytomegalovirus or herpes simplex virus 1. *J Virol* (2011) **85**:4212–21. doi:10. 1128/JVI.02435-10


**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.

*Received: 21 April 2014; accepted: 13 June 2014; published online: 25 June 2014. Citation: Oh JE and Lee HK (2014) Pattern recognition receptors and autophagy. Front. Immunol. 5:300. doi: 10.3389/fimmu.2014.00300*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Oh and Lee. 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.*

## Pattern recognition receptors and DNA repair: starting to put a jigsaw puzzle together

#### **Anton G. Kutikhin1,2,3\*, Arseniy E.Yuzhalin<sup>4</sup> , Eugene A. Tsitko<sup>5</sup> and Elena B. Brusina2,6**


#### **Edited by:**

Masoud H. Manjili, Virginia Commonwealth University Massey Cancer Center, USA

#### **Reviewed by:**

Gregory B. Lesinski, The Ohio State University Comprehensive Cancer Center, USA

**Keywords: toll-like receptors, NOD-like receptors, C-type lectin receptors, RIG-I-like receptors, pattern recognition receptors, DNA repair, innate immunity, inflammation**

The group of pattern recognition receptors (PRRs) includes families of toll-like receptors (TLRs), NOD-like receptors (NLRs), C-type lectin receptors (CLRs), RIG-I-like receptors (RLRs), and AIM2 like receptors (ALRs) (1–7). Conceptually, receptors constituting these families are united by two general features. Firstly, they directly recognize common antigen determinants of virtually all classes of pathogens (so-called pathogen-associated molecular patterns, PAMPs) and initiate immune response against them via specific intracellular signaling pathways (1–7). Secondly, they also recognize endogenous ligands released in cells under stress, which are known as damage-associated molecular patterns (DAMPs). Therefore, a subset of PRR-mediated immune response can be activated without an influence of infectious agents (1–7).

Long-standing data implicate that PRRs play a key role in innate and adaptive immune responses (1–7). Besides their effect on immunity, many PRRs may have a crucial impact on almost all vital cellular processes, such as cell growth, survival, apoptosis, cell cycle control, cell proliferation and differentiation, autophagy, angiogenesis, cell motility, and migration (8–14). In recent years, the evidence of the involvement of PRRs in the processes of DNA repair started to emerge. A recent comprehensive review by Harberts and Gaspari (15) has shed light on this issue; nevertheless, a number of newer investigations were performed after the publication of their paper.

One of the most investigated TLRs is TLR4, which is a transmembrane protein with an ectodomain located on the cell surface (16). The two most known TLR4 ligands are lipopolysaccharide (LPS), one of the main components of Gramnegative bacteria outer membrane, and high-mobility group protein B1 (HMGB1), which is known to be an important chromatin protein (16). It is well known that X-ray repair cross-complementing group (XRCC)5/KU80 and XRCC6/KU70 are the key non-homologous end-joining (NHEJ) repair pathway proteins (17, 18). Wang et al. observed that a diminishment of TLR4-mediated immune response may lead to reduced expression of XRCC5/KU80 and XRCC6/KU70 in mouse liver tissue and cells in response to the diethylnitrosamine, therefore, being a cause of the DNA repair impairment and reactive oxygen species (ROS) accumulation (17, 18). However, when TLR4−/<sup>−</sup> mice and wild-type mice were locally exposed to ultraviolet B (UVB, shortwave radiation), the expression of DNA repair gene xeroderma pigmentosum, complementation group A (*XPA*), and production of interleukins (ILs) 12 and 23 were significantly higher (19). Further, cyclobutane pyrimidine dimers were repaired more efficiently in the skin and bone

marrow-derived dendritic cells (DCs) of TLR4−/<sup>−</sup> mice (19). The addition of anti-IL-12 and anti-IL-23 antibodies to bone marrow-derived DCs of TLR4−/<sup>−</sup> mice before UVB exposure inhibited repair of cyclobutane pyrimidine dimers along with a decline of *XPA* gene expression; similarly, the addition of TLR4 agonist to wildtype bone marrow-derived DCs lowered *XPA* gene expression and diminished repair of cyclobutane pyrimidine dimers (19). Hence, the activation of TLR4 signaling by ultraviolet radiation may launch a specific pathway and result in decrease of IL-12 and/or IL-23 production, thereby reducing the expression of genes encoding DNA repair enzyme such as *XPA* (19). According to these studies (17–19), TLR4 may both upregulate and downregulate distinct DNA repair proteins, and possibly does it in different ways in distinct cell types, so its exact role in DNA repair remains unclear.

Certain TLRs are located on the endoplasmic reticulum membrane (in a resting state) or on the endosomal/lysosomal membrane (upon ligand stimulation and trafficking) (20). Among these are TLR7, TLR8, and TLR9 (20). The main ligands for TLR7 and TLR8 are imidazoquinolines, ssRNA, and antiphospholipid antibodies, while the main ligands for TLR9 are bacterial and viral CpG DNA and IgG-chromatin complexes (20). However, all these receptors signal via the protein encoded by myeloid differentiation primary response

gene 88 (*MyD88*) (20). Tsukamoto et al. found that 8-mercaptoguanosine (8SGuo) induces the activation-induced cytidine deaminase (AID) expression and doublestrand breaks (DSBs) through TLR7– MyD88-dependent pathway in cluster of differentiation (CD)38- or B cell receptor (BCR)-activated B cells (21). Nevertheless, imiquimod, a TLR7/8 agonist, which is used in the treatment of certain non-melanoma skin cancer, increased an expression and nuclear localization of *XPA* gene and other DNA repair genes in a MyD88-dependent manner (22). In addition, as it was detected by Fishelevich et al. imiquimod enhanced DNA repair and accelerated the resolution of cyclobutane pyrimidine dimers after an exposure of bone marrow-derived cells to ultraviolet light (22). Imiquimod-activated cutaneous antigen presenting cells were characterized by better DNA repair in comparison with resting antigen presenting cells under the exposure to both non-ionizing and ionizing radiation (22). Moreover, topical application of imiquimod before the exposure to ultraviolet light had a protective effect and reduced the number of cyclobutane pyrimidine dimers-positive antigen presenting cells (22). Therefore, the role of TLR7 and TLR8 in DNA repair may differ depending on their influence on the specific DNA repair proteins or on the cell type, as in the case with TLR4.

In the study of Zheng et al., TLR9 stimulated CD4 T cells demonstrated an increased capacity to repair ionizing radiation-induced DSBs, whereas the treatment of irradiated CD4 T cells with TLR9 ligands along with checkpoint kinase (Chk)1/2 inhibitors or along with ataxia telangiectasia mutated/ataxia telangiectasia and Rad3 related (ATM/ATR) inhibitor wortmannin abrogated the improvement of DNA repair observed previously (23). In addition, TLR9 stimulation did not elevate DNA repair rates after an exposure to ionizing radiation in TLR9−/<sup>−</sup> and MyD88−/<sup>−</sup> CD4 T cells; thus, TLR9-induced DNA repair may be mediated by Chk1/2 and ATM/ATR via MyD88-dependent pathway (23). Klaschik et al. performed a global gene expression analysis on mouse splenic cells and revealed that CpG DNA, a ligand for TLR9, may cause the activation of genes responsible for DNA repair 3– 5 days after an intraperitoneal injection, so the long-term enhancement of DNA repair after TLR9 stimulation is possible (24). Sommariva et al. carried out an *in silico* analysis of DNA repair genes in data sets obtained from murine colon carcinoma cells in mice injected intratumorally with synthetic oligodeoxynucleotides expressing CpG motifs (CpG–ODN, a TLR9 agonist) and from splenocytes in mice treated intraperitoneally with CpG–ODN (25). According to their results, CpG– ODN downregulated DNA repair genes in tumors, but upregulated them in immune cells (25). Moreover, «CpG-like» expression pattern of CpG–ODN modulated DNA repair genes was associated with a better outcome of patients with breast and ovarian cancer treated by DNA-damaging agents than «CpG-untreated-like» expression pattern, so these genes may determine tumor cell response to genotoxic drugs (25). It seems to be that the exact role of TLR9 in DNA repair substantially depends on the cell type.

It was found that MyD88 mediates the optimal activation of the Ras/mitogenactivated protein kinase (MAPK) pathway by binding to extracellular signal-regulated kinase (ERK) and protecting it from dephosphorylation (26–29). In accordance with the data obtained by Kfoury et al., MyD88 inhibition may lead to defective excision repair cross-complementing rodent repair deficiency, complementation group 1 (ERCC1)-dependent DNA repair and to accumulation of DNA damage (29, 30). In addition, abrogation of *MyD88* gene expression sensitizes cancer cells to genotoxic agents such as platinum salts *in vitro* and *in vivo* (29, 30). It is worthy of note that platinum-based chemotherapeutic agents (cisplatin, carboplatin, and oxaliplatin) cause DNA damage that is preferentially repaired by the nucleotide excision repair (NER) pathway, which is implicated in the repair of DNA single-strand breaks (SSBs), and ERCC1 predominantly functions as NER enzyme via Ras-MAPK pathway (29, 30). So, MyD88-dependent Ras-MAPKmediated activation of ERCC1 may play a major role in DNA repair (29, 30). However, Lai and Egan reported that early induction of DSBs in mouse colonic epithelial cells by ionizing radiation was independent of the presence and absence of *MyD88* gene expression (31). Notwithstanding, they observed a later loss of DSBs

and an enhanced activation of DSB repair pathways in MyD88−/<sup>−</sup> mice compared to control mice (31). It seems to be that MyD88 has no specific inhibitory effects regarding the pathways of DSB repair since both the NHEJ and homologous recombination (HR) repair pathways were over-activated in the absence of MyD88 (31). Possibly, MyD88-mediated signaling pathway may regulate the repair of SSBs and DSBs in a distinct way via activation or inhibition of the proteins specific for each of pathways responsible for the repair of SSBs and DSBs.

The only study investigating the role of NLRs in DNA repair was carried out by Licandro et al. regarding NLR family, pyrin domain containing 3 (*Nlrp3*) gene (32). The ectodomain of NLRP3 recognizes certain DAMPs that may lead to the assembly of inflammasome and, hence, to the development of aseptic inflammation (33). The authors exposed murine DCs to monosodium urate, rotenone, and γ-radiation, and detected a lesser level of DNA fragmentation in Nlrp3−/<sup>−</sup> DCs compared to wild-type DCs (32). Moreover, Nlrp3−/<sup>−</sup> DCs experienced significantly less ROS-mediated DNA damage, and a significantly lower expression of several genes involved in DSB and base excision repair (BER) was revealed in wild-type DCs (32). These genes included *XRCC1*, *RAD51*, 8-oxoguanine–DNA glycosylase 1 (*OGG1*), breast cancer 1, early onset (*BRCA1*), DNA polymerase beta (*POLB*), and thymidylate synthase (*TYMS*) (32). It was demonstrated that DSB and BER enzymes responsible for repair of 8-oxoguanine, which is a DNA adduct formed as a result of oxidation, and therefore, is considered a marker of oxidative stress, were more active in Nlrp3−/<sup>−</sup> cells in comparison with wild-type DCs (32). In addition, Nijmegen breakage syndrome 1 (NBS1), another protein involved in DNA repair, was highly phosphorylated in Nlrp3−/<sup>−</sup> DCs compared with wildtype DCs, indicating greater efficacy of DNA repair in the absence of *Nlrp3* gene expression (32).

Taken together, these reports strongly implicate PRRs, in particular TLRs (TLR4, TLR7, TLR8, and TLR9) and NLRs (NLRP3), as major regulators of DNA repair (Table S1 in Supplementary Material). According to the above-mentioned

**growth factors-mediated signaling pathway**. There are three main TLR-mediated pathways of DNA repair. The protein encoded by myeloid differentiation primary response gene 88 (MyD88) and its downstream signaling proteins (not shown) may inhibit mitogen-activated protein kinase phosphatase 3 (MKP3), which hinders phosphorylation of extracellular signal-regulated kinase (ERK), and therefore, prevents further factor, which promotes transcription of certain DNA repair genes. Finally, IL-12 and IL-23, which enhance DNA repair and whose transcription is also amplified by MyD88-regulated transcription factors, bind to their receptors, activate Janus kinase-signal transducer and activator of transcription (JAK-STAT) signaling pathway, and increase further transcription of their own encoding genes.

findings, these five receptors may affect the expression of at least eight enzymes (XRCC1, XRCC5, XRCC6, XPA, BRCA1, POLB, TYMS, OGG1, and RAD51) and two ILs (IL-12 and IL-23) involved in various mechanisms of DNA repair. Further, PRRs are responsible not only for

the initiation of one specific DNA repair pathway, but a number of such pathways repairing different types of DNA damage, i.e., oxidation, alkylation, and hydrolysis of bases, bulky adducts, SSBs, DSBs, and crosslinks. Interestingly, the effect of PRRs on DNA repair may vary between cell

types and cell lines, which address a number of questions to be answered in future studies.

Nowadays, we are only beginning to put the pieces of this puzzle together. Current vision of this topic is blurred, although a preliminary picture based on

recent research can be drawn (**Figure 1**). Both TLRs located on the cell surface and thus responsible for the recognition of the pathogen envelope molecular patterns (TLR4) and TLRs located on the endoplasmic reticulum, endosomal, or lysosomal membrane, and therefore, responsible for the recognition of pathogen nucleic acids (TLR7, TLR8, and TLR9) are involved in DNA repair. Therefore, other TLRs belonging to any of these groups may also participate in such processes. Definitely, the cytokine-mediated DNA repair feedback loop is not restricted to IL-12 and IL-23, and might consist of much greater number of cytokines, possibly TLRregulated cytokines [IL-1, IL-2, IL-6, IL-8, IL-10, IL-13, IL-27, macrophage inflammatory protein-1 (MIP-1), monocyte chemotactic protein-1 (MCP-1), regulated on activation, normal T-cell expressed and secreted (RANTES), suppressor of cytokine signaling (SOCS), granulocytemacrophage colony-stimulating factor (GM-CSF), tumor necrosis factor-α (TNFα), interferon (IFN)-α, IFN-β, IFN-γ, and IFN-inducible proteins]. Furthermore, the exact composition of the growth factors-mediated DNA repair signaling pathway is still elusive; importantly, this pathway may have a particular importance since it includes both MyD88 and Ras-MAPK pathways, representing an interesting example of a crosstalk between canonical TLR MyD88-mediated signaling pathway and Ras-MAPK signaling pathway. In addition, there are no studies on the feasible influence of CLRs, RLRs, and ALRs on DNA repair. The improvement of our understanding of the role of PRRs in DNA repair may find implications for clinical medicine; peculiarities of PRRs functioning should definitely be considered when assessing the possibility of the use of PRR agonists in therapy of various diseases such as cancer. No doubt, further in-depth investigations are needed for deciphering the role of PRRs in sophisticated mechanisms of DNA repair.

#### **SUPPLEMENTARY MATERIAL**

The Supplementary Material for this article can be found online at http://www. frontiersin.org/Journal/10.3389/fimmu. 2014.00343/abstract

#### **REFERENCES**


hepatocellular tumorigenesis and progression by regulating expression of DNA repair protein Ku70 in mice. *Hepatology* (2013) **57**:1869–81. doi:10. 1002/hep.26234


and cancer cell resistance to genotoxic drugs. *J Natl Cancer Inst* (2013) **105**:937–46. doi:10.1093/jnci/ djt120


DNA damage responses after oxidative and genotoxic stress in dendritic cells. *Eur J Immunol* (2013) **43**:2126–37. doi:10.1002/eji.201242918

33. Chen GY, Nuñez G. Sterile inflammation: sensing and reacting to damage. *Nat Rev Immunol* (2010) **10**:826–37. doi:10.1038/nri2873

**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.

*Received: 07 May 2014; accepted: 05 July 2014; published online: 23 July 2014.*

*Citation: Kutikhin AG, Yuzhalin AE, Tsitko EA and Brusina EB (2014) Pattern recognition receptors and DNA repair: starting to put a jigsaw puzzle together. Front. Immunol. 5:343. doi: 10.3389/fimmu.2014.00343 This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Kutikhin, Yuzhalin, Tsitko and Brusina. 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.*

## Toll-like receptor 4 in inflammation and angiogenesis: a double-edged sword

#### **Sheeba Murad \***

Molecular Immunology Lab, Health Care Biotech Department, Atta-Ur-Rahman School of Applied Biosciences, National University of Sciences and Technology, Islamabad, Pakistan

\*Correspondence: sheebamall@yahoo.com; s.mall@asab.nust.edu.pk

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Fulvio D'Acquisto, Queen Mary University of London, UK Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

**Keywords: LPS,TLR4, PAMPS, DAMPs, angiogenesis**

Toll-like receptors (TLRs) primarily known for the pathogen recognition and subsequent immune responses are being investigated for their pathogenic role in various chronic diseases. The recent reports correlating the microbial infections with chronic disorders such as atherosclerosis have lead to questions in relation to the role of microbial sensors such as TLR4 in an intriguing phenomenon of the *inflammation-induced angiogenesis*. This article focuses on the possible mechanisms involved in it.

Toll-like receptors comprise a large family of the pathogen-pattern recognition receptors (PPRR) originally identified in *Drosophila* in the mid 1990s as a *Toll protein* (1). In *Drosophila,* it was found to be involved in the resistance against fungal infections (2)*.* The first human homolog for the Toll protein was described in 1997 (3). Since then, 13 mammalian homologs of the TLR family have been identified; including 12 in mice (TLR1-9 and TLR11-13) and 10 in humans (TLR1- 10). TLR 10 is a pseudogene in mice, but is functional in humans (4). The membrane expressed TLRs recognize the pathogenassociated molecular patterns (PAMPs) either directly on the plasma membrane or within the endosomal compartment after the phagocytosis. In addition to the foreign molecules, a range of various endogenous ligands are also detected by TLRs, which suggests a role beyond that of simple pathogen recognition. Endogenous ligands released from the damaged, apoptotic, or fibrotic cells during inflammation, are termed *danger-associated molecular patterns* (DAMPs). A significant number of DAMPs have been reported for TLR4 (5, 6).

TLR4 is one of the best characterized and the first member of the TLR family to be discovered as a PPRR. TLR4 signaling is implicated in the innate immune responses against a wide-range of microbes, including Gram-negative and -positive bacteria, mycobacteria, spirochetes, yeasts, and some viruses such as respiratory syncytial viruses (RSV) and mammary tumor viruses (4). TLR4 is a type I transmembrane protein characterized by an extracellular domain containing leucine-rich repeats (LRRs) and a cytoplasmic tail harboring a conserved region known as Toll/IL-1 receptor (TIR) domain. TLR4, along with its two coreceptors, the myeloid differentiation antigen (MD2) and the LRR protein CD14, forms a trimeric receptor that is involved in the recognition of lipopolysaccharide (LPS). The TLR4 ligand binding causes the C termini of the ectodomains to move close to each other, thus triggering signaling and inflammation. The diverse interactions between TLRs with their ligands converge into either the MyD88-dependent or MyD88-independent pathways, resulting in the: (1) activation of lymphocytes, (2) up-regulation/expression of costimulatory signals, and (3) release of proinflammatory cytokines/chemokines (7). As sentinels in the innate immunity, TLR expression was thought to be confined to the immune cells such as macrophages, monocytes, and dendritic cells. However, an increasing number of reports show a more diverse expression of TLRs; including epithelial cells, endothelial cells (8), neural and glial cells, thereby playing an important role in tissue-specific inflammation (9).

TLR4 is implicated in a diverse range of pathological processes associated with or induced by angiogenesis including autoimmune diseases such as psoriasis, diabetic retinopathy, thrombosis, and inflammatory disorders including arthritis and atherosclerosis and cancer (10, 11). It has been proposed that TLR4 contributes to these diseases through *inflammationinduced angiogenesis*. The recent association between bacterial infections and atherosclerosis has intensified the search for the biological functions of TLRs especially TLR4 in blood vessel formation (12). The exact mechanism needs to be elucidated.

Angiogenesis is the normal process required for the development of an extensive vasculature. With its over 60 trillion endothelial cells, the vascular network is the first and the largest organ to develop in the human body (13). It mainly occurs during embryonic development. In adults, angiogenesis is a highly regulated process only occurring during the retinal development, in the adult intestinal villi and in the female reproductive organs (14). The postnatal angiogenesis may take place through one of the two possible mechanisms; (1) vasculogenesis – the *de novo* generation of blood vessels from endothelial progenitor cells (EPCs) or mesoderm and more commonly (2) angiogenesis, which is the sprouting/branching of the pre-existing blood vessels – together they are called neoangiogenesis. Angiogenesis is a highly complex series of sequential events orchestrating various molecular events involving multiple cell populations, cytokines, and chemokines. It takes place in two important steps; (1) formation of a nascent

Murad TLR4 in inflammation and angiogenesis

vascular network and (2) its subsequent maturation. The degradation of extracellular matrix (ECM) allows the sprouting of EPCs from old vessel into an avascular space and differentiation into nascent vasculature under the influence of proangiogenic factors. The maturation process involves the recruitment of supporting cells (mural cells) and vessel remodeling. Mural cells include vascular smoothmuscle cells (VSMC) in arteries, arterioles, and veins; pericytes in capillaries (15, 16). They provide structural integrity to the developing vasculature and may also interact with the endothelial cells, through paracrine signaling. Pro-angiogenic factors such as the vascular endothelial growth factor (VEGF); the basic fibroblast growth factor (bFGF); the transforming growth factor beta (TGF-β); the platelet-derived growth factor (PDGF); the tumor necrosis factor alpha (TNF-α); the insulin-like growth factor-1 (IGF-1); the monocyte chemotactic protein (MCP)-1; interleukin (IL)-6 and 8 all help in the recruitment of cells, ECM degradation, and with vessel development and maturity (14). An important empirical role played by TLR4 in the lymphocytic activation, recruitment, and release of cytokines is evident in TLR4 deficient mice. Such mice are reported to display significantly impaired expression of pro-inflammatory cytokines after reperfusion triggered by retinal ischemia injury (17). The process of lymphangiogenesis was shown to be affected in TLR4-deficient mice through lack of macrophage recruitment by TLR4<sup>+</sup> lymphatic endothelial cells (LEC) (7).

As one of the two main sources of cytokines, macrophages play a critical role in the leukocyte trafficking and the postnatal angiogenesis. TLR4-mediated LPSactivated macrophages have been shown to be an important source of proangiogenic factors. Accumulating evidence shows that antigenic stimulation and the surrounding cytokine environment can have profound effects on the activation status and the functional capabilities of macrophages. Although there are various schools of thought regarding the macrophage activation status, here, we focus on two; the M1 and M2 phenotypes. The classical activation or M1 phenotype of macrophages contributes substantially toward anti-microbial immune responses via the production of proinflammatory cytokines such as IL-6, IL-8, IL-12, inducible nitric oxide synthase (iNOS), and interferons (IFNs) (18) (**Figure 1**). The alternate activation of macrophages may lead to the M2 phenotype, which is reported to be involved in the wound repair and fibrosis by contributing toward angiogenesis through the VEGF production (19). The strong mitogenic effect on the endothelial cells and the induction of vascular permeability are the pro-angiogenic effects, which makes VEGF the most potent simulator of angiogenesis. In murine macrophages and other TLR4<sup>+</sup> cell populations, a strong synergism is reported to significantly influence the production of VEGF. Endotoxins (including LPS) together with the growth factors and cytokines such as IFNγ, TGF-β, IL-1, and IL-6 have been implicated in a significant augmentation in VEGF levels (20–24). In this regard, the synergism reported between TLR4 and adenosine receptor 2A (A2AR) in the murine macrophages (M2) is noteworthy (**Figure 1**) (25). Adenosine receptor signaling plays an important role in inflammation. Adenosine is produced by many different cell types and is elevated in conditions such as hypoxia, ischemic conditions, and stress. So far, four adenosine receptors have been reported, i.e., the A1, A2A, A3B, and A<sup>3</sup> receptors (26). The synergistic effect of A2AR is not restricted to TLR4, but TLR2, 7, and 9 also lead to high VEGF production in the presence of adenosine signaling (22). Both TLR4 and A2AR were shown to signal through hypoxia inducible factor (HIF)-1α and hypoxia response element (HRE) (27). Although the TLR4 along with its co-receptors are known to be expressed on the endothelial cells, it is not yet known whether the endothelial cells share the synergistic effect of TLR4 with A2AR. The transcriptional expression of A2AR has been reported on the endothelial cells; however, there are limited number of studies in this context. Many groups have demonstrated potent endothelial responses to LPS *in vitro* (28–32). However, there are reports supporting the *in vivo* role of LPS in postnatal angiogenesis. A study conducted in murine tumor model (metastatic) demonstrated the proangiogenic effects of LPS. The LPS-induced

growth and metastasis of 4T1 experimental lung metastases model was shown to take place through increased angiogenesis, vascular permeability, and tumor cell migration (33). The LPS-mediated angiogenic effects can be reversed through TLR4 downregulation. While studying the anti-inflammatory affects of a compound known as *Baicalein*, its anti-angiogenic effects were shown to be carried out through the downregulation of TLR4 and its downstream mitogen-activated phosphate kinase (MAPK) pathway (34).

The ubiquitous and abundantly expressed DAMPs are often found in association with different anomalies. One such commonly expressed protein is high mobility group chromatin protein B1 (HMGB-1). It is a nuclear DNA binding protein released by injured or necrotic cells. Resting, non-activated inflammatory cells, such as monocytes or macrophages, contain HMGB-1 in their nuclei. When these cells are activated by LPS or inflammatory cytokines, HMGB-1 translocates in the cytoplasm, undergoes acetylation, and is exocytosed. It is evident that excreted HMGB-1 acts like a pro-inflammatory cytokine, therefore, HMGB-1 can be regarded as a signal of tissue injury and a mediator of inflammation (35). Macrophage-derived HMGB-1 has been shown to increase the endothelial cell proliferation, sprouting, and chemotaxis by stimulating the migration of adherent cells, such as fibroblasts and smooth-muscle cells. In a recent study, HMGB-1-TLR4 signaling was reported to be an important mediator in retinal neoangiogenesis in an oxygen-induced retinopathy murine model (36). HMGB-1 is an important marker for tumor endothelial cells and was shown to be necessary for the sustained expression of pro-angiogenic genes. A positive feedback mechanism has been suggested for the HMGB-1 expression and that of its cognate receptors, i.e., TLR4 and receptor for advanced glycation end products (RAGE) on the endothelial cells. Thus HMGB-1 may prove to be a promising target for interfering with cancer-related angiogenesis (37). However, there is some disagreement in relation to the HMGB-1 as an endogenous ligand for TLR4. The lack of an LPS-free *in vitro* system makes it difficult to study the signaling resulting exclusively

from the TLR4-ligands other than LPS. Even small traces of LPS can upregulate TLR4 and can affect the interpretation of results.

Ischemic diseases are one of the major causes of morbidity and mortality. Treatment of such disorders requires angiogenesis. It is therefore the prime goal of *therapeutic angiogenesis* to achieve this. However, the close association between angiogenesis and inflammation presents an obstacle to the success of the therapy. Most of the pro-angiogenic factors are also pro-inflammatory. Therefore, the reperfusion of ischemic tissues often results in injury due to the microvascular dysregularities and inflammation (edema) associated with it. The activated endothelial cells lead to an imbalance between oxygen radicals and nitric oxide causing the release of inflammatory mediators (38, 39). The TLR4-deficient mice have been a valuable tool for studying the role of TLR4 in tissue-related ischemia–reperfusions *in vivo*. A recent study reported the role of TLR4-mediated responses contributing to the oxygen-induced neovascularization in ischemic neural tissue (retina). The TLR4-dependent responses, proposed to be mediated through HMGB-1 release in the ischemic neural tissue were found to be impaired in TLR4-deficient mice, revealing an important angiogenic role of TLR4 in neural tissues (36). On the other hand, there are several studies highlighting the inflammatory role of TLR4 in various reperfusion–ischemic models in tissues such as liver, lung, and intestine. Most of these studies showed reduced inflammation in relation to the injury induced by the reperfusion of various organs after a period of ischemia in TLR4-deficient mice, thus, highlighting the inflammatory role of TLR4 in reperfusion-related injury models, without significant compromise in angiogenesis (40–43). Considering these reports, the dual role of TLR4 in angiogenesis and inflammation comes to light, which seems to be governed by an intricate balance between the inhibitory or stimulatory factors that may be tissue-specific. Nevertheless, TLR4 remains a promising target for suppressing the undesired and prolonged inflammatory responses. In this regard, various synthetic and plantderived therapies are currently being tested.

TLR4-blocking through small molecule inhibitors and antibodies are being evaluated in pre-clinical trials for their efficacy in various inflammatory conditions. Novimmune is a humanized counterpart of rat anti-TLR4 monoclonal antibody; 1A6, found to reduce inflammation in a murine colitis model. It is undergoing preclinical evaluation for the treatment of the inflammatory bowel diseases (44–46). Various plant-derived drugs such as wogonoside and celastrol have shown promising results against TLR4-mediated LPSinduced angiogenesis in pre-clinical drug testing (47, 48).

In conclusion, it can be said that the close association between inflammation and angiogenesis makes the therapeutic modulation of TLR4 somewhat challenging and can lead to potential side effects. Therefore, the fine tuning of TLR4 and its associating proteins is required in order to circumvent the undesired inflammatory or angiogenic responses associated with TLR4 targeting in various pathologies. For that purpose, further insight into its *in vivo* networking and the effects of TLR4 targeting in various pathologies through the use of closely related animal disease models is required.

#### **REFERENCES**


VEGF promoter. *Mol Biol Cell* (2007) **18**:14–23. doi:10.1091/mbc.E06-07-0596


injury in Toll-like receptor 4-deficient mice. *Circulation* (2004) **109**(6):784–9. doi:10.1161/01.CIR. 0000112575.66565.84


but impairs mucosal healing in murine colitis. *Am J Physiol Gastrointest Liver Physiol* (2009) **296**:G1167–79. doi:10.1152/ajpgi.90496.2008


**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.

*Received: 21 April 2014; accepted: 22 June 2014; published online: 07 July 2014.*

*Citation: Murad S (2014) Toll-like receptor 4 in inflammation and angiogenesis: a double-edged sword. Front. Immunol. 5:313. doi: 10.3389/fimmu.2014.00313*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Murad. 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.*

## Toll-like receptors in esophageal cancer

### **Joonas H. Kauppila1,2,3,4\* and Katri S. Selander 5,6**


#### **Edited by:**

Anton G. Kutikhin, Research Institute for Complex Issues of Cardiovascular Diseases under the Siberian Branch of the Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Anton G. Kutikhin, Research Institute for Complex Issues of Cardiovascular Diseases under the Siberian Branch of the Russian Academy of Medical Sciences, Russia Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Joonas H. Kauppila, Department of Pathology, University of Oulu, PO-Box 5000, 90014 Oulu, Finland e-mail: joonas.kauppila@oulu.fi

Esophageal squamous cell carcinoma and esophageal adenocarcinoma are cancers of high mortality. EAC develops through Barrett's esophagus (BE) and columnar dysplasia, preceded by gastro-esophageal reflux disease. The risk of esophageal squamous cell carcinoma is increased by smoking and alcohol consumption. New treatment options for esophageal cancer are desperately needed. Toll-like receptors (TLRs) play a central role in mammalian immunity and cancer. TLRs are activated by microbial components, such as lipopolysaccharide, flagellin, DNA, and RNA, as well as endogenous ligands, including heat-shock proteins and endogenous DNA. This review summarizes the studies on TLRs in esophageal squamous cell carcinoma and EAC. It has been shown that TLRs 1–10 are expressed in the normal esophagus. In esophageal squamous cell carcinoma, TLRs3, 4, 7, and 9 have been studied, showing associations to aggressive disease properties. In BE and EAC, only TLRs4, 5, and 9 have been studied. In the review, we discuss the implications of TLRs in esophageal cancer.

**Keywords:Toll-like receptors, microbiome, esophageal cancer, esophageal adenocarcinoma, esophageal squamous cell carcinoma**

#### **INTRODUCTION**

Toll-like receptors (TLRs) are evolutionarily conserved receptors of the innate immune system (1). The 13 TLRs that have been identified so far recognize their unique pathogen-associated molecular patterns (PAMPs), such as bacterial lipopolysaccharide (LPS) (TLR4), DNA (TLR9), or flagellin (TLR5) (1, 2). TLR stimulation induces down-stream activation of various signaling molecules and this ultimately results in the innate immune response, which also activates the adaptive immune system (1– 3). The aim of this review is to explore the role and function of TLRs in esophageal adenocarcinoma (EAC) and in squamous cell carcinoma.

#### **ESOPHAGEAL CANCER**

Esophageal cancer is the eighth most common cancer in the world with estimated 482,000 new cases worldwide in 2008. The incidence of esophageal cancer was 70/100,000 in 2008 in the world. The majority of esophageal cancers are esophageal squamous cell carcinomas (ESCC), but the incidence of EAC is rising rapidly (4, 5).

As with oral squamous cell cancer, tobacco and alcohol, low socioeconomic status, poor oral health, and betel nuts, as well as the autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy (APECED)-syndrome have been listed as risk factors for ESCC (6–10). With regard to pathologic anatomy, esophagus could be considered as an extension of the oral cavity, as it is lined by squamous epithelium and it encounters swallowed oral bacteria before they enter the stomach.

For EAC, the most important risk factor is Barrett's esophagus (BE), determined by columnar metaplastic cells, which replace the normal squamous epithelium after long-lasting gastro-esophageal reflux, or gastro-esophageal reflux disease (GERD). Patients with BE have a 30- to 125-fold risk for EAC compared to normal population (11, 12). A recent study, however, concluded that only 0.12% of patients with BE develop EAC (13). Other minor risk factors include obesity, smoking, hiatal hernia, and low socioeconomic status (10, 14–18). Furthermore, both types of esophageal cancers develop through dysplasia to cancer via genetic alterations (19, 20).

The 5-year survival rate for esophageal cancer varies between 10 and 16% (4). After esophageal resection, the 5-year survival rate was 20.6% in a meta-analysis of Western population (21). Furthermore, these cancers are often diagnosed late because at the time of the diagnosis, more than half of the patients have an inoperable disease (22).

The most important prognostic determinant for both esophageal cancers is the WHO TNM-classification (23). The histologically defined grade of differentiation is also a predictor of prognosis (24).

#### **TOLL-LIKE RECEPTORS IN NORMAL ESOPHAGUS**

Esophageal epithelial cells have been shown to express TLRs. The human esophageal epithelial cell-line TE-1 was shown to express TLRs2, 3, 4, and 7, with up-regulation of beta-defensin 2 as a response to stimulation with their cognate, synthetic ligands (25). In 2009, Lim and colleagues demonstrated the expression of TLRs 1–10, but not TLR4 at the mRNA level in the normal human esophageal epithelial cell-line EPC-2. Furthermore, they demonstrated TLR1, 2, 3, and 5 mRNA expression in biopsies taken from esophageal mucosa. IL-8 was up-regulated in the EPC-2 cells by stimulation of the respective TLR-ligands. TLR3 stimulation was the most effective in inducing IL-8 expression synergistically with TLR2 and this effect was dependent on NF-kB activation (26).

TLR3 was later demonstrated also to mediate the induction of IL-8 mRNA via NF-kB by necrotic cell supernatants in the EPC-2 cells (27). TLR2 and TLR3 protein expression was demonstrated in esophageal epithelial cells, but not in cultured primary esophageal epithelial cells (28). The expression of TLR3, 4, 5, 7, and 9 proteins in normal esophagus has been characterized using immunohistochemistry in clinical samples (29–31). These studies have demonstrated that TLRs 1–10 are expressed in normal esophagus.

#### **TOLL-LIKE RECEPTORS AND ESOPHAGEAL SQUAMOUS CELL CARCINOMA**

Esophageal squamous cell carcinoma develops to squamous epithelium via dysplasia. A variety of TLRs, including TLR3, 4, 7, and 9, have been shown to be overexpressed in esophageal squamous cell carcinoma, when compared to normal esophagus (30, 31). We demonstrated an increased TLR9 expression in esophageal squamous dysplasia and in squamous cell carcinoma, suggesting a possible role for TLR9 in esophageal carcinogenesis (31).

High TLR3, 4, and 9 expression in esophageal squamous cell carcinoma cells have been associated with lymph node metastasis and TLR7 and 9 expression to worse histological grade (30, 31). TLR9 expression in the fibroblastoid cells of the tumor was, however, associated with decreased invasion depth and a smaller prevalence of lymph node metastasis at the time of diagnosis (30). TLR4 stimulation by LPS has been shown to increase migration and adhesive properties of esophageal squamous cell carcinoma cells via p38 and selectin (32). No studies thus far have evaluated the anti-cancer efficacy of TLR-agonists or inhibitors in the treatment of ESCC.

#### **TOLL-LIKE RECEPTORS, BARRETT'S ESOPHAGUS, AND ESOPHAGEAL ADENOCARCINOMA**

Esophageal adenocarcinoma is developed through the metaplasia– dysplasia–carcinoma sequence. Normal or inflamed esophageal epithelium is believed to transform to BE or columnar metaplasia through continuous exposure to acidic gastric contents, but also transformation of esophageal microbiome occurs during these changes (33, 34).

It was shown in BE and in normal human esophageal cell lines, that stimulation of TLR4 with LPS resulted in NF-κB activation and an increase of IL-8 secretion, this response was more significant in BE. *Ex vivo* culture demonstrated increased cyclooxygenase-2 (COX-2) activation by LPS stimulation of TLR4 in BE (35). TLR5 was recently analyzed in the metaplasia–dysplasia– adenocarcinoma sequence, with high expression potentially differentiating between BE and columnar dysplasia (29).

The increased expression of TLR5 and 9 has been shown in EAC. TLR5 expression had no associations to clinico-pathological variables or prognosis, but TLR9 expression was associated with metastasis, poor grade of differentiation and poor prognosis in EAC (29, 36). Stimulation of EAC cells with CpG-oligonucleotides that either have the physiological phosphodiester DNA-backbone or the nuclease-resistant phosphothioate backbone, induced cellular invasion and matrix metalloproteinase-9 and -13 mRNA expression (37).

At the current moment, there are no published clinical studies on TLRs in EAC.

#### **TOLL-LIKE RECEPTOR GENETICS AND ESOPHAGEAL CANCER**

Genetic studies have been performed on Toll-like receptor polymorphisms in esophageal cancer. Unlike in gastric cancer, polymorphisms in *TLR4* + *896A* > *G* and *TLR9-1237T/C* genes were not associated to esophageal cancer risk (38, 39). However, genetic up-regulation of CD14, a co-receptor of TLR4, was observed in families with history of esophageal cancer (40).

#### **DISCUSSION**

The treatment of esophageal cancer is overshadowed by its poor prognosis. New options for early diagnosis and treatment are desperately needed. The esophageal epithelium encounters bacteria from oral cavity and in the case of reflux disease, also from the stomach and possibly also from the duodenum. TLRs act by recognizing bacteria-derived molecular patterns which results in a pro-inflammatory reaction in the epithelium.

The role of TLRs in esophageal cancer has been studied sparsely. However, there is evidence that the function of TLRs is procarcinogenic and pro-inflammatory as the overexpression of many of the TLRs have been linked with esophageal cancer and with poor prognosis. Inflammation is a known important factor in the pathogenesis of various cancers. It was demonstrated by Yang et al. that the microbiome of distal esophagus frequently undergoes changes during esophagitis and BE. During these processes, the microbiome is switched from aerobic to gram-negative anaerobic bacteria (33, 34). This finding together with abnormal TLR expression, particularly those of TLRs4, 5, and 9, in esophageal cancer supports the hypothesis of bacteria contributing to the carcinogenesis of esophageal cancer. These findings further suggest that TLRs may be important mediators for bacteria in oncogenesis (37, 40, 41).

In addition to microbes, TLRs can also detect molecular patterns that are derived from the host itself. TLRs3, 4, and 9 are known to be activated by endogenous ligands from dead or damaged host cells (42, 43). The combination of cellular damage by alcohol, tobacco, and acidic contents of the stomach results in the loss of epithelial wall integrity, through epithelial cell death and by disruption of the cell-to-cell contacts. Especially TLR3 and TLR9 (but also other TLRs) can recognize particles from dead cells (43). This can result in an inflammatory wound reaction through the activation of interleukins, NF-kB, and matrix metalloproteinases. This wound reaction could facilitate the passage of bacteria through epithelium and result in the loss of hostmicrobiome homeostasis, further leading to abnormal activation of for example TLR2, 4, 5, and 9 by bacterial components. Inflammation and wound reaction then could produce a vicious cycle of cellular damage, which might be a major player in esophageal metaplasia and carcinogenesis. This role of bacteria and TLR4 in genesis of BE has been discussed earlier by Yang et al. (33). Cell-to-cell junctions become dysfunctional in exogenous damage to the epithelium as discussed earlier. Thus, a similar effect can also

be observed in dysplasia and cancer (44). This dysfunction may lead to Toll-like receptor activation in cancer by exogenous and endogenous ligands. The hypothesis is summarized in **Figure 1**.

Finally, Toll-like receptor expression is up-regulated in both squamous cell carcinoma and adenocarcinoma of the esophagus. This may indicate that cancer cells are sensitized to bacteria- and host-derived ligands. Poor prognosis in strongly TLR-expressing tumors could then be an indicator of increased level of tumor– stroma interaction.

#### **CONCLUSION**

There seems to be a connection between TLRs and esophageal cancer development. The fact that bacterial flora changes during esophageal metaplasia and inflammation, as well as observed upregulation of TLRs in esophageal cancers support the hypothesis that bacteria as well as TLRs have a role in esophageal cancer.

#### **ACKNOWLEDGMENTS**

We thank Dr. Kevin Harris for his valuable comments on the manuscript.

#### **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.

*Received: 10 April 2014; paper pending published: 19 April 2014; accepted: 23 April 2014; published online: 07 May 2014.*

*Citation: Kauppila JH and Selander KS (2014) Toll-like receptors in esophageal cancer. Front. Immunol. 5:200. doi: 10.3389/fimmu.2014.00200*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Kauppila and Selander. 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.*

REVIEW ARTICLE published: 22 July 2014

### Pattern-recognition receptors and gastric cancer

#### **Natalia Castaño-Rodríguez, Nadeem O. Kaakoush and Hazel M. Mitchell \***

School of Biotechnology and Biomolecular Sciences, The University of New South Wales, Sydney, NSW, Australia

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Masaaki Murakami, Hokkaido University, Japan Arseniy E. Yuzhalin, University of Oxford, UK Si Ming Man, St. Jude Children's Research Hospital, USA

#### **\*Correspondence:**

Hazel M. Mitchell, School of Biotechnology and Biomolecular Sciences, Laboratory 301A, Biological Sciences Building, The University of New South Wales, Sydney, NSW 2052, Australia e-mail: h.mitchell@unsw.edu.au

Chronic inflammation has been associated with an increased risk of several human malignancies, a classic example being gastric adenocarcinoma (GC). Development of GC is known to result from infection of the gastric mucosa by Helicobacter pylori, which initially induces acute inflammation and, in a subset of patients, progresses over time to chronic inflammation, gastric atrophy, intestinal metaplasia, dysplasia, and finally intestinal-type GC. Germ-line encoded receptors known as pattern-recognition receptors (PRRs) are critical for generating mature pro-inflammatory cytokines that are crucial for both Th1 and Th2 responses. Given that H. pylori is initially targeted by PRRs, it is conceivable that dysfunction within genes of this arm of the immune system could modulate the host response against H. pylori infection, and subsequently influence the emergence of GC. Current evidence suggests thatToll-like receptors (TLRs) (TLR2,TLR3,TLR4,TLR5, andTLR9), nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) (NOD1, NOD2, and NLRP3), a C-type lectin receptor (DC-SIGN), and retinoic acid-inducible gene (RIG)-I-like receptors (RIG-I and MDA-5), are involved in both the recognition of H. pylori and gastric carcinogenesis. In addition, polymorphisms in genes involved in theTLR (TLR1,TLR2,TLR4, TLR5, TLR9, and CD14) and NLR (NOD1, NOD2, NLRP3, NLRP12, NLRX1, CASP1, ASC, and CARD8) signaling pathways have been shown to modulate the risk of H. pylori infection, gastric precancerous lesions, and/or GC. Further, the modulation of PRRs has been suggested to suppress H. pylori-induced inflammation and enhance GC cell apoptosis, highlighting their potential relevance in GC therapeutics. In this review, we present current advances in our understanding of the role of the TLR and NLR signaling pathways in the pathogenesis of GC, address the involvement of other recently identified PRRs in GC, and discuss the potential implications of PRRs in GC immunotherapy.

**Keywords: stomach neoplasm, Helicobacter pylori, inflammation, pattern-recognition receptors,Toll-like receptors, NOD-like receptors, genetic polymorphism, therapeutics**

#### **INTRODUCTION**

Of the three main types of stomach cancer, gastric adenocarcinoma (GC), non-Hodgkin's lymphoma, and gastrointestinal stromal tumors, approximately 95% are GC, which remains one of the most commonly diagnosed cancers in the world (1). In 2012, stomach cancer was the fifth most common cancer worldwide, with 952,000 new cases diagnosed, accounting for 6.8% of the total cancer cases (1). Furthermore, it is the third leading cause of cancer-related deaths worldwide, accounting for 8.8% of total deaths from cancer, with 5-year relative survival rates lower than 30%, except in Japan where mass screening has been undertaken for several years (2).

Gastric cancer is a heterogeneous pathology with respect to anatomical location and histological subtypes (**Figure 1A**). In relation to location, GC may occur in the cardia or noncardia region of the stomach. Cardia GC has been associated with gastro-esophageal reflux, *Helicobacter pylori* infection, and atrophic gastritis, male gender, smoking, and diet (3). Epidemiological studies assessing the worldwide incidence of GC by anatomical location have shown an increase in the incidence of cardia GC, however, in high GC risk areas, non-cardia GC remains the most frequent pathology (4). Further, even though cardia and non-cardia GC have been considered etiologically different phenomena, it has been demonstrated that cancer of the cardia among individuals from areas with a high risk of GC represents a subset of cardia GC that is associated with *H. pylori*-related atrophic gastritis and resembles non-cardia GC pathogenesis (5, 6).

According to the Lauren Classification, non-cardia GC has been further subdivided into the two histological variants intestinaltype and diffuse-type. Intestinal-type GC is characterized by the formation of gland-like structures, distal stomach localization, and a predilection for older individuals. It is also more frequent in males (2:1 ratio) and in subjects of lower socioeconomic status (10). This type of GC is often preceded by a precancerous phase that starts with the transition of normal mucosa into multifocal atrophic gastritis. This initial histological alteration is followed by intestinal metaplasia, dysplasia, and finally adenocarcinoma (11). On the other hand, diffuse-type GC is poorly differentiated, affects younger individuals, and has been highly associated with genetic susceptibility (the variant hereditary diffuse GC, which is associated with germ-line mutations in *CDH1*, a gene encoding E-cadherin) (12, 13). Additionally, it is not associated with the formation of precancerous lesions and has been found to affect the

**FIGURE 1 | Gastric cancer classification and etiology**. **(A)** Stomach cancer comprises gastric adenocarcinoma (GC), non-Hodgkin lymphomas, including mucosa-associated lymphoid tissue (MALT) lymphoma, and the rare gastrointestinal stromal tumors (GIST), leiomyosarcoma, and carcinoid tumors. The most common type, GC, has been classified as cardia and non-cardia GC according to anatomical location. Cardia GC is divided into two different etiological entities, esophageal-like cardia GC, which is associated with gastro-esophageal reflux, smoking, and diet, and is frequent in areas with a low risk of GC and distal stomach-like cardia GC, which is associated with the presence of H. pylori and gastric atrophy, and is the most frequent cardia GC variant in areas with a high risk of GC. Non-cardia GC is further subdivided into two histological variants called intestinal-type and diffuse-type GC. Intestinal-type GC, according to the widely accepted

Correa's cascade (7), is a biological continuum that commences as chronic gastritis and progresses to atrophic gastritis, intestinal metaplasia, dysplasia, and finally, GC. \*Stomach cancer subtypes that have been associated with Helicobacter pylori infection. (**B**) H. pylori infection causes chronic inflammation of the gastric mucosa of all infected individuals, and in combination with host and environmental factor, leads to the development of GC in a subset of infected individuals (1–3%). In these subjects, inflammation represents the seventh hallmark of cancer and an enabling characteristic that facilitates the acquisition of the other established hallmarks that collectively dictate malignant growth (tissue invasion/metastasis, limitless replicative potential, sustained angiogenesis, evasion of programed-cell death (apoptosis), self-sufficiency in growth signals, and insensitivity to growth-inhibitory signals) (8, 9).

entire surface of the stomach. This type of GC is present equally between the two sexes and is associated with a worse prognosis in comparison to intestinal-type GC (10, 12).

Most GC cases are sporadic and arise due to the combination of a permissive environment interacting with a susceptible host. Several factors that contribute to the development of GC have been identified; these include bacterial (*H. pylori*), host, and environmental factors (12).

*Helicobacter pylori* is a Gram-negative bacterium that infects nearly 50% of the human population (14). In the gastric mucosa, the majority of *Helicobacter pylori* are found within the mucus layer but they can also be attached to epithelial cells leading to the maintenance, spread, and severity of the infection (15). *H. pylori* infection has been associated with the development of a range of diseases, including peptic ulcer disease (10%), non-cardia GC (1–3%), and gastric mucosa-associated lymphoid tissue (MALT) lymphoma (<0.1%) (14, 16–18). Furthermore, this bacterium has

been associated with three distinct phenotypes in the infected host: (1) a corpus-predominant gastritis, which has the potential to lead to atrophic gastritis, hypochlorhydria, and to the development of GC; (2) a duodenal ulcer phenotype in which an antrum-predominant gastritis leads to increased gastric acid secretion; and (3) a benign phenotype in which the bacterial infection causes a mild mixed gastritis that has a minor effect on gastric acid production (19).

*Helicobacter pylori* infection is transmitted by direct humanto-human transmission, via either the oral–oral route, fecal–oral route, or both (14). *H. pylori* is acquired early in life, the majority of individuals being infected before the age of 10 years with close family members being a common source of infection (20–22). It has been postulated that early acquisition of infection might be associated with the broad pathological spectrum associated with *H. pylori* infection and the highly persistent GC incidence rates in genetically susceptible populations who have migrated to developed countries. In the absence of antibiotic therapy, *H. pylori* infection generally persists for life (23).

Natural colonization by *H. pylori* is restricted to humans, primates, and domestic animals such as cats (23–25). *H. pylori* is considered to be the dominant microorganism in the human stomach as the majority of bacteria cannot survive in the low gastric pH (26). Several other factors make the human stomach an unfavorable environment for bacterial colonization including peristalsis, poor nutrient availability, and host innate and adaptive immunity (23). The ability of *H. pylori* to survive and colonize the stomach relates to a number of mechanisms. Most importantly *H. pylori*, unlike other bacteria, produces large amounts of the enzyme urease, which hydrolyzes urea to ammonia, which subsequently interacts with hydrogen ions in the stomach to form ammonium (27, 28). In addition, *H. pylori* is able to regulate gene expression in response to changes in pH (29). Further, *H. pylori* expresses multiple paralogous outer membrane proteins, including the blood-group antigen-binding adhesin (BabA), the sialic-acid binding adhesin (SabA), and the outer inflammatory protein (OipA), which appear to bind to receptors on the surface of gastric epithelial cells, which reduces the rate of bacterial elimination as a result of peristalsis (30, 31). *H. pylori* counteracts the lack of nutrients by inducing tissue inflammation and using specific systems that facilitate the transport and uptake of nutritional resources (23). In addition, *H. pylori* has been reported to produce antibacterial peptides that might decrease competition from other microorganisms (32).

Further, a number of other factors have been shown to help *H. pylori* evade the host immune system. For example, the vacuolating cytotoxin (VacA) produced by some strains of *H. pylori* has been shown to inhibit T-cell proliferation as well as antigen presentation by B cells and to alter the normal functions of CD8<sup>+</sup> T cells, mast cells, and macrophages (33–36). In addition, gammaglutamyl transpeptidase, another immunosuppressive factor of *H. pylori*, has been associated with inhibition of T-cell proliferation by induction of a cell cycle arrest in the G<sup>1</sup> phase (37). Furthermore, *H. pylori* has been shown to use arginase to downregulate the production of inducible nitric oxide synthase by macrophages (38).

The fact that more than one *H. pylori* strain can colonize the gastric mucosa provides the opportunity for *H. pylori* to acquire new genetic sequences and to undergo recombination events (23). One of the most remarkable differences among *H. pylori* strains is the presence or absence of a 40-kb DNA insertion element known as the cytotoxin-associated gene pathogenicity island (*cag* PAI) (39). This region contains between 27 and 31 genes flanked by 31-bp repeats and encodes the most widely investigated *H. pylori* virulence factor, the cytotoxin-associated antigen A (CagA) (40, 41). *H. pylori* strains expressing CagA represent 60–70% of Western strains and approximately 100% of East Asian strains (39, 42). CagA is a 120- to 140-kDa protein that is translocated into host cells through a type IV secretion system following attachment to gastric epithelial cells (43). Following translocation, CagA is tyrosine phosphorylated at the EPIYA (glutamate–proline– isoleucine–tyrosine–alanine) motifs by members of the host cell kinase families known as proto-oncogene proteins Abl and Src

(18). In Western populations strains, EPIYA-A, EPIYA-B, and varying numbers of EPIYA-C motifs have been reported, while in *H. pylori* strains from East Asian populations, EPIYA-A and EPIYA-B with EPIYA-D motifs, are found (44). Both phosphorylated and non-phosphorylated CagA result in alterations in the gastric epithelium including: (1) the activation of the protein tyrosine phosphatase, non-receptor type 11 (SHP-2), (2) alterations in cell scattering and proliferation, (3) alterations in cell structure and cell motility, (4) perturbation of epithelial cell differentiation and polarity, (5) alteration of tight junctions, and (6) aberrant activation of β-catenin (45–47). Furthermore, numerous studies have shown that *cag* PAI-positive *H. pylori* strains are associated with an increased risk of gastric diseases including peptic ulcer disease, premalignant gastric lesions and GC (48–51). Further details of the interplay between *H. pylori* virulence factors and gastric epithelial cells and GC, can be found in an excellent review by Posselt et al. (44).

In the last two decades, a large number of epidemiological studies have established the association between *H. pylori* and the subsequent risk of developing both intestinal-type and diffusetype GC (52–57). This finding has been consistent among different populations. For example, in the study by Parsonnet et al. (57), conducted in Caucasian,African-American, and Asian individuals, subjects infected with *H. pylori* who had antibodies against CagA were shown to be more likely than uninfected subjects to develop both intestinal-type and diffuse-type GC (OR: 5.1, 95% CI: 2.1– 12.2 and OR: 10.1, 95% CI: 2.2–47.4, respectively). Consistently, a further study conducted in a Japanese population showed that, although the association was stronger in cases with intestinal-type GC (OR: 3.2, 95% CI: 1.8–5.8), there was also a positive association between *H. pylori* infection and diffuse-type GC (OR: 3.0, 95% CI: 1.0–8.8) (53). Further, a study conducted in a Spanish population showed no differences in *H. pylori* infection between the two GC histological subtypes (58). Similarly, a recent study in German individuals showed that *H. pylori* prevalence was comparable in patients with intestinal-type (82.1%) and diffuse-type (77.9%) GC (59).

Interestingly, more recent studies, assessing *H. pylori* infection through Western blot (CagA) for the detection of past infection, have shown an unprecedented association between *H. pylori* and GC that can be explained by a reduction of the misclassification that might take place when samples are analyzed with the enzymelinked immunosorbent assay (ELISA) alone (60, 61). For example, Ekstrome et al. (60) conducted a population-based study, comprising 298 GC patients and 244 controls, in which the OR for *H. pylori* infection among non-cardia GC was 21.0 (95% CI: 8.3– 53.4). Further, Siman et al. (61) showed that *H. pylori* significantly increased the risk of non-cardia GC showing an OR of 17.8 (95% CI: 4.2–74.8).

While *H. pylori* infection has been established as the most important risk factor for GC and was classified as a class 1 carcinogen by the World Health Organization in 1994, the etiology of GC also involves host and environmental factors. This is evidenced by the fact that only 1–3% of *H. pylori*-infected patients develop GC, and that progression to GC in some subjects occurs even after eradication of the bacterium (18).

Given that *H. pylori* is initially targeted by germ-line encoded receptors known as pattern-recognition receptors (PRRs), it is conceivable that dysfunction within genes of this arm of the immune system would affect the magnitude and direction of the host inflammatory response against the infection, resulting in an increased risk of GC development. Recent studies clearly show that PRRs are critical for generating mature pro-inflammatory cytokines that are crucial for both Th1 and Th2 responses during *H. pylori* infection, and these immune responses have been directly associated with gastric immunopathology. In this review, we present current advances in the understanding of the role of PRRs, mainly the Toll-like receptor (TLR) and nucleotide-binding oligomerization domain (NOD)-like receptor (NLR) signaling pathways, in the pathogenesis of GC, and discuss future directions for continued research in this area. In the first section, we highlight the relevance of inflammation in GC. In subsequent sections, we address new developments in the TLR and NLR signaling pathways in GC, the role of other PRRs in GC, and the new frontier of therapeutic application of these concepts.

#### **INFLAMMATION IN GASTRIC CANCER**

It is well established that most cancer cell genotypes are the manifestation of six essential alterations in cell physiology that collectively dictate malignant growth: (1) self-sufficiency in growth signals, (2) insensitivity to growth-inhibitory signals, (3) evasion of programed-cell death (apoptosis), (4) limitless replicative potential, (5) sustained angiogenesis, and (6) tissue invasion/metastasis (8). Recently, inflammation has been considered the seventh hallmark of cancer and an enabling characteristic that facilitates the acquisition of the other hallmarks (**Figure 1B**). Inflammation initiated by innate immune cells, mainly macrophage subtypes, mast cells, myeloid progenitors, and neutrophils (62–65), designed to fight infections and heal wounds, can instead result in unintentional support of multiple cancer hallmark functions, thereby manifesting the widely accepted tumor-promoting consequences of inflammatory responses (9). In addition, active evasion by cancer cells from attack and elimination by immune cells, mainly CD8+ cytotoxic T lymphocytes, CD4+ Type 1 helper T cells, and natural killer (NK) cells, highlights the dual role of an immune system that both antagonizes and promotes cancer development and progression (9).

In the context of tumor enhancement, it has been proposed that once inflammation is initiated, tissue integrity is compromised leading to the multistage process of carcinogenesis by altering targets and pathways that are pivotal for normal tissue homeostasis (66). The mechanisms that are connected to these alterations include production of mutagenic reactive oxygen and nitrogen species as well as synthesis of cytokines and growth factors that favor tumor cell growth (67). In addition, inflammation provides a source of other bioactive molecules to the tumor microenvironment, including survival factors that limit cell death, pro-angiogenic factors, extracellular matrix-modifying enzymes that facilitate angiogenesis, invasion, and metastasis, and inductive signals that lead to activation of the epithelial–mesenchymal transition (a developmental regulatory program that enables epithelial cells to invade, resist apoptosis, and disseminate) (9). Interestingly, inflammation can be considered a "perigenetic alteration" of

cancer cells because it may promote growth, expansion, and invasion of tumors even without the involvement of further genetic mutations or epigenetic alterations (68).

In 1988, Correa proposed a human model of intestinal-type gastric carcinogenesis (7). The model hypothesized a sequence of events progressing from acute inflammation to chronic inflammation, to atrophy, to intestinal metaplasia, to dysplasia, to carcinoma *in situ*, and finally to invasive GC. A subsequent study by Correa evaluated the gastric precancerous process in a Colombian population (7). The results of this cross-sectional study led to the widely accepted conclusion that the severity of atrophy correlates with the prevalence of metaplasia and that the severity of metaplasia correlates with the prevalence of dysplasia, suggesting that the process is indeed a biological continuum (69).

Given that inflammation is a hallmark of gastric carcinogenesis, polymorphisms in genes encoding pro-inflammatory cytokines/chemokines have been the focus of much research in recent years. To date, polymorphisms in the interleukin (IL)-1 family genes have been the most widely studied, including polymorphisms in *IL1A*, *IL1B*, and *IL1RN* that encode IL-1α, IL-1β, and their endogenous receptor antagonist IL-1RA, respectively. In particular, IL-1β, a potent endogenous pyrogen and an important component in the development of Th2-mediated immunity (70, 71), has been associated with lipid peroxidation, DNA damage, inhibition of gastric acid secretion, increased *H. pylori* colonization, and induction of gastric atrophy and dysplasia in the presence or absence of *H. pylori* (72). Global meta-analyses have shown that the *IL1B-511* T allele is significantly associated with an increased risk of developing GC in Caucasians but not Asians or Mestizos (73, 74). Furthermore, IL-1 receptor signaling is known to induce the production of genes that not only stimulate tumor growth but are also involved in angiogenesis and metastasis such as matrix metalloproteinases, basic fibroblast growth factor, vascular endothelial growth factor, vascular cell adhesion molecule 1, intercellular adhesion molecule 1, monocytic chemotactic protein 1, and CXCL-2 (75). To date, only one study has addressed the role of *IL1R1* (also known as *CD121A*) in GC and *H. pylori* infection. The study, conducted in a Caucasian population, showed an increased risk of *H. pylori* infection in those harboring the *IL1R1* Hinfl A allele (OR: 2.01, *P*-value: 0.009) but failed to show an association with GC (76). In addition, a recent meta-analysis on the endogenous receptor antagonist IL-1RA has shown the *IL1RN\*22* genotype to increase the risk of gastric precancerous lesions, supporting a role for this polymorphism in the early stages of gastric carcinogenesis (OR: 2.27, 95% CI: 1.40–3.70) (77). A further meta-analysis that included 39 case–control studies, showed statistically significant associations between the *IL1RN\*22* genotype and both intestinal-type and diffuse-type GC, showing ORs of 1.83 and 1.72, respectively (78). Further examples of polymorphisms in pro-inflammatory cytokines/chemokines that play an essential role promoting inflammation in the context of gastrointestinal carcinogenesis are IL-4 (*IL4*-590C/T and -168T/C) (79), IL-6 (*IL6*-174 G/C) (80–82), IL-8 (*IL8*-251 A/T, +396 T/G, and +781 C/T) (79, 83), IL-10 (*IL10*-1082 A/G, −819 C/T, and −592 C/A) (84–86), IL-12 (*IL12A*-701 C/A, −798 T/A, +277 G/A, and −504 T/G) (87), IL-17 (*IL17*-197 G/A and +7488 T/C) (79), IL-18 (*IL18*-137 G/C) (88), and TNF-α (*TNFA* −238 G/A, −308 G/A, and −857 C/T) (89). In addition to this, a recent comprehensive review on this topic recommended the investigation of other polymorphisms in *IL1B* (3954 C/T and −1473G/C), *IL4* (–168T/C), *IL6* (572 G/C and 597 G/A), and *IL17* (+7488A/G and −197G/A), given their potential relevance in GC (79).

While extensive evidence supports the important role of proinflammatory cytokines/chemokines in gastric carcinogenesis, given that PRRs, mainly TLRs and NLRs, are important modulators of intestinal epithelial barrier function, epithelial repair, and immune homeostasis in the gastrointestinal tract (90), and that signal transduction from these receptors converges upon a common set of signaling molecules, including the activation of the transcription factors nuclear factor kappa-light-chainenhancer of activated B cells (NF-κB) and the activator protein 1 (AP-1) that lead to the production of pro-inflammatory cytokines/chemokines (e.g., IL-1α, IL-1β, IL-6, IL-8, IL-10, and TNF-α) as well as members of the interferon (IFN) regulatory transcription factor family that mediate type I IFN-dependent responses, defects in PRRs function could be even more important than defects in pro-inflammatory cytokines/chemokines *per se* in the instauration of an inflammation-related disorder such as GC.

#### **PATTERN-RECOGNITION RECEPTORS IN GASTRIC CANCER**

Innate immunity refers to responses that do not require previous exposure to an immune stimulus and represents the first line of host defense in the response to pathogens. PRRs are part of the innate immune system and are pivotal for the detection of invariant microbial motifs. PRRs have been divided into five distinct genetic and functional clades: TLRs, NLRs, C-type lectin receptors (CLRs), retinoic acid-inducible gene (RIG)-I-like receptors (RLRs), and absent in melanoma 2 (AIM2)-like receptors (ALRs) (91, 92). PRRs are commonly expressed by cells of the innate immune system such as monocytes, macrophages, dendritic cells (DCs), neutrophils, and epithelial cells, as well as cells of the adaptive immune system (93).

Toll-like receptors and CLRs scan the extracellular milieu and endosomal compartments for pathogen-associated molecular patterns (PAMPs), which are highly conserved microbial structures that are essential for microbial survival (94), while intracellular PRRs, including NLRs, RLRs, and ALRs, cooperate to provide cytosolic surveillance (92, 93).

In *H. pylori* infection, the first physical–chemical barriers for the pathogen are the mucus layer,gastric epithelial cells,autophagy, and PRRs (TLRs, NLRs, CLRs, and RLRs) (**Figure 2**).

#### **TOLL-LIKE RECEPTORS AND HELICOBACTER PYLORI-RELATED GASTRIC CANCER**

#### **TOLL-LIKE RECEPTORS RECOGNITION OF HELICOBACTER PYLORI**

The involvement of the TLR signaling pathway in infectious, autoimmune, and inflammatory diseases is well accepted (95). During *H. pylori* infection, TLRs on gastric epithelial and immune cells recognize diverse PAMPs such as flagellin/unknown PAMP (TLR5), unmethylated CpG motifs (TLR9), and lipopolysaccharide (LPS) (TLR4 and TLR2).

TLR4 was initially identified as the potential signaling receptor for *H. pylori* LPS on gastric epithelial cells (96–99). After forming a complex with the LPS-binding protein (LBP), LPS interacts with the monocyte differentiation antigen CD14 (CD14), and subsequently with the myeloid differentiation protein-2 (MD-2) (100). Together with TLR4, this complex induces the TLR4-mediated MyD88-dependent signal transduction pathway, which leads to the rapid activation of transcription factors, mainly NF-κB, and cytokines such as TNF-α, IL-1β, IL-6, and IL-12 (95). On the other hand, stimulation of TLR4 by LPS also facilitates the activation of a MyD88-independent pathway that activates IFN-regulatory factor (IRF) 3 and involves the late phase of NF-κB activation, both of which lead to the production of IFN-β and the expression of IFN-inducible genes (101, 102). In addition to LPS, the *H. pylori* secretory protein HP0175, through its ability to bind to TLR4, was shown to transactivate the epidermal growth factor receptor (EGFR) and stimulate the EGFR-dependent vascular endothelial growth factor production in the GC cell line AGS, which have been linked to *H. pylori*-associated gastroduodenal diseases, ulcerogenesis, and carcinogenesis (103).

Although early studies concluded that TLR4 is the first innate immune response against *H. pylori* (104, 105), later studies suggested that TLR4 had a limited role, given that *H. pylori* LPS appeared to bind poorly to LBP, resulting in it being inefficiently transferred to CD14 (106). Consequently, recent studies addressing the role of other TLRs during *H. pylori* infection, have found TLR2 to be the initial barrier against *H. pylori* infection (107–112). A potential explanation for these inter-study differences in relation to the TLRs response to *H. pylori* might be attributed to cell type (i.e., epithelial versus immune cells), origin of the cell studied (i.e., peritoneal versus bone marrow derived macrophages), and the type of inflammatory response measured (i.e., type of cytokines), and thus, currently any conclusions regarding the role of TLR4 must be treated with caution.

In contrast, there is strong evidence supporting an important role for TLR2 in *H. pylori* infection, with both animal and cell culture experiments suggesting that TLR2 ligands (LPS or other) exist in *H. pylori* and related *Helicobacter* species (112–114), and that TLR2 may be involved in the innate immune sensing of these bacteria by epithelial cells (113). Furthermore, an interesting publication by Smith et al. (115) showed that *H. pylori* LPS functions as a classic TLR2 ligand and induces a discrete pattern of chemokine expression in epithelial cells, which involves modulation of the expression of the signaling protein tribbles 3 (TRIB3), a molecule implicated in the regulation of NF-κB.

Yet, the most likely scenario is that both TLR4 and TLR2 are involved in the early immune response against *H. pylori* as has been demonstrated by a number of investigators (116–118). For example, Obonyo et al. (116) showed that both TLR2 and TLR4 were crucial signaling receptors for *H. pylori* activation of the host immune response leading to the secretion of cytokines. Further, Yokota et al. (118) not only showed that *H. pylori* LPS was initially targeted by TLR2 as described by others, but, for the first time, showed that this TLR2 activation leads to cell proliferation and TLR4 expression via the MEK1/2-ERK1/2 pathway. The final outcome of this signaling pathway is increased proliferation of gastric epithelial cells and the instauration of a strong inflammatory reaction. Once this response is instaurated, *H. pylori* could then enhance inflammatory reactions mediated by TLR4 agonists such as other bacterial LPS, which would also contribute to gastric

**FIGURE 2 | Pattern-recognition receptors involvement in Helicobacter pylori infection**. H. pylori is recognized by the Toll-like receptors (TLRs) (TLR2, TLR4, TLR5, and TLR9), NOD-like receptors (NLRs) (NOD1, NOD2, NLRP3, and possibly, NLRP12 and NLRX1), RIG-I like receptors (RLRs) (RIG-I and possibly, MDA-5), and C-type lectin receptors (CLRs) (DC-SIGN). TLR4 poorly recognizes H. pylori lipopolysaccharide (LPS) to generate pro-inflammatory cytokines (e.g., IL-1α, IL-1β, IL-6, IL-8, IL-10, and TNF-α) and interferons (IFNs) through the myeloid differentiation primary response gene 88 (MyD88)-dependent and -independent pathways, respectively. TLR2 recognizes H. pylori LPS/peptigoglycan/unknown pathogen-associated molecular pattern (PAMP) while TLR5 poorly recognizes H. pylori flagella and TLR9 recognizes H. pylori DNA (unmethylated CpG motifs). H. pylori recognition by these three TLRs leads to nuclear factor-κB (NF-κB) activation. NOD1 and NOD2 recognize H. pylori peptidoglycan-derived peptides [γ-d-glutamyl-meso-diaminopimelic acid (iE-DAP) and muramyl dipeptide (MDP)], leading to the activation of both transcription factors NF-κB and activator protein (AP)-1. The NLRP3 inflammasome, comprising NLRP3, apoptosis-associated speck-like protein containing a CARD (ASC) and caspase-1, recognizes a yet unknown H. pylori PAMP and/or damage-associated molecular pattern (DAMP), and through caspase-1

cleavage, leads to the maturation and secretion of interleukin (IL)-1β and IL-18. NLRX1 and NLRP12, two known negative regulators of NF-κB, appear to be significantly down-regulated during H. pylori infection in vitro, however, their exact role during H. pylori infection remains unclear. RIG-I recognizes H. pylori 5 0 -triphosphorylated RNA (5<sup>0</sup> -PRNA) while MDA-5 possibly recognizes H. pylori dsRNA. The dendritic cell-specific intercellular adhesion molecule-3 grabbing non-integrin (DC-SIGN) recognizes H. pylori fucosylated ligands and this interaction appears to counteract the pro-inflammatory immune response to H. pylori. Only one generic cell type depicting all TLRs, NLRs, RLRs, and CLRs involved in H. pylori recognition is shown here for simplicity. MAL, MyD88 adaptor-like protein, also named TIRAP; TRAM, translocating chain-associating membrane protein; TRIF, TIR domain containing adaptor inducing interferon-beta protein; TBK-1, TANK-binding kinase 1; IRF3, IFN-regulatory factor 3; TRAF6, TNF receptor-associated factor 6; IRAK, interleukin 1 receptor-associated kinase; RAS, proto-oncogene ras; c-RAF, proto-oncogene protein ras; RIP2, receptor-interacting serine/threonine-protein kinase 2, also known as RICK; CARD9, caspase activation and recruitment domain; MD-2, myeloid differentiation protein-2; ILs: interleukins. Names in orange correspond to molecules with a probable but not established role in the host response to H. pylori.

inflammation and subsequent carcinogenesis (118). Further, the heat-shock protein 60, an immune-potent antigen of *H. pylori*, has been shown to activate NF-κB and induce IL-8 production through TLR2 and TLR4 pathways in gastric epithelial cells, a

phenomenon that is likely to contribute to the development of gastric inflammation caused by *H. pylori* infection (117).

In addition, TLR9 appears to play an important role in *H. pylori* recognition. Interestingly, Rad et al. (112) identified TLR9-mediated recognition of *H. pylori* DNA as a main *H. pylori*induced intracellular TLR signaling pathway in DCs. Further, a study using a murine model of *H. pylori* infection has suggested that TLR9 signaling is involved in the suppression of *H. pylori*-induced gastritis in the early phase of infection via downregulation of Th1-type cytokines modulated by IFN-α (119). In addition, a recent study has shown that the gastric epithelia of children respond to *H. pylori* infection by increasing the expression of TLR2, TLR4, TLR5, and TLR9, as well as the cytokines IL-8, IL-10, and TNF-α (120).

Although TLR5 interaction with *H. pylori* induces only weak receptor activation (121), TLR5 has been involved in the inflammatory response to *H. pylori*. An interesting publication by Smith et al. (107), using HEK293 cells transfected with specific TLR expression constructs and MKN45 cells expressing dominant negative versions of TLR2, TLR4, and TLR5, which block the activity of wild-type forms of these receptors, has demonstrated that live *H. pylori* induces NF-κB activation and chemokine gene expression due to ligation of TLR2 and TLR5. A further study that aimed to explore the involvement of TLR2 and TLR5 in THP-1 cells and HEK293 cell lines (stably transfected with TLR2 or TLR5) during *H. pylori* infection, has indicated that *H. pylori*-induced expression of TLR2 and TLR5 can qualitatively shift *cag* PAI-dependent to *cag* PAI-independent pro-inflammatory signaling pathways with possible impact on the outcome of *H. pylori*-associated diseases (122). Given the established TLR5 evasion of α and ε *Proteobacteria* including *H. pylori* (123), the TLR5-mediated inflammatory responses during *H. pylori* infection described by Smith et al. (107) and Kumar Pachathundikandi et al. (122) are likely to be flagellinindependent, and therefore, a still unknown *H. pylori* factor might be responsible for this.

The importance of TLRs recognition during *H. pylori* infection and GC development is further supported by the acquired characteristics that enable *H. pylori* to survive in the human stomach and cause chronic inflammation. For example, *H. pylori* LPS is characterized by a modification of the lipid A component of LPS that makes it less pro-inflammatory (124) and has been reported to exhibit a 1000-fold reduction in bioactivity as compared to *Escherichia coli* LPS (125). Also, the flagellin of this bacterium has been shown to be poorly recognized due to modifications in the TLR5 recognition site of the N-terminal D1 domain of flagellin (123).

#### **TOLL-LIKE RECEPTORS AND GASTRIC CARCINOGENESIS**

While TLR2, TLR4, TLR5, and TLR9 appear to be important for *H. pylori* recognition, their role in the evolution of gastritis to more advanced lesions remains unclear. Interestingly, Schmausser et al. (126) showed that TLR9 was not detectable in intestinal metaplasia or dysplasia and was only focally detected in 6 out of 22 gastric carcinomas, while TLR4 and TLR5 were strongly expressed by gastric carcinomas. Consistently, a study by Pimentel-Nunes et al. (127) showed a statistically significant trend for a progressive increase of TLR2, TLR4, and TLR5 expression from normal mucosa to gastric dysplasia (mean expression in normal mucosa: 0.1, gastritis: 1.0, metaplasia: 2.2, and dysplasia: 2.8, *P*value <0.01), with dysplasia presenting more than 90% positive epithelial cells showing strong expression (2.8, 95% CI: 2.7–3). In addition, these authors showed a significant trend for decrease in TOLLIP and PPARγ, two TLR signaling pathway inhibitors, which was associated with increasing levels of CDX-2, a marker for adenocarcinoma, from normal mucosa to carcinoma (*P*-value <0.05) (128). Fernandez-Garcia et al. (129) have also reported increased expression of TLR3, TLR4, and TLR9 in GC, and furthermore, these authors noted that TLR3 expression by cancer cells was significantly associated with a poor overall survival in patients with resectable tumors, which lead them to suggest that TLR3 might be an indicator of tumor aggressiveness. Similarly, Yakut et al. (130) investigating the association between serum IL-1β, TLR4 levels, pepsinogen I and II, gastrin 17, vascular endothelial growth factor, and *H. pylori* CagA status in patients with a range of gastric precancerous lesions, concluded that serum TLR4 levels could be used as a biomarker to differentiate individuals presenting with dysplasia from those with other gastric precancerous lesions, the mean TLR4 level in patients with dysplasia (0.56 ± 0.098 ng/mL) being significantly higher than in patients with *H. pylori* positive chronic non-atrophic gastritis (0.10 ± 0.15 ng/mL), chronic atrophic gastritis (0.06 ± 0.07 ng/mL), and intestinal metaplasia (0.12 ± 0.18 ng/mL). Furthermore, while TLRs have been shown to be expressed at the apical and basolateral pole of both normal gastric epithelial cells and in *H. pylori* gastritis, in metaplasia, dysplastic, and neoplastic epithelial cells all TLRs are expressed diffusely and homogeneously throughout the cytoplasm, with no apparent polarization, which may suggest an increased activation of these diffusely over-expressed receptors during gastric carcinogenesis (126, 128).

In recent years, TLRs have been associated with tumor development and progression processes including cell proliferation, epithelial–mesenchymal transition, angiogenesis, metastasis, and immunosuppression. Interestingly, Chochi et al. (104) not only showed that *H. pylori* augmented the growth of GC via the LPS-TLR4 pathway but also found that this bacterium attenuated the antitumor activity and IFN-γ-mediated cellular immunity of human mononuclear cells. In addition, Song et al. (131) have suggested that flagellin-activated TLR5 enhances the proliferation of GC cells through an ERK-dependent pathway. Furthermore, Tye et al. (132) have proposed a novel role for TLR2 in promoting gastric tumorigenesis independent of inflammation, whereby up-regulation of TLR2 within epithelial tumor cells, rather than infiltrating inflammatory cells, by the uncontrolled activation of the oncogenic transcription factor STAT3, promoted gastric tumor cell proliferation, and survival via upregulation of anti-apoptotic genes [e.g., BCL2-related protein A1 (*BCL2A1*),baculoviral IAP repeat containing 3 (*BIRC3*),and B-cell CLL/lymphoma 3 (*BCL3*)]. Further, two processes that facilitate carcinogenesis and involve TLRs have recently been described by Li et al. (133). Using LPS-treated CD14-knockdown GC cells, these authors showed that CD14, an important co-receptor in the TLR4 complex, promotes tumor cell epithelial–mesenchymal transition and invasion through TNF-α (133).

In addition, the expression of tumor-associated molecules known to be important in gastric carcinogenesis has been linked to the activation of the TLR signaling pathway. For example, prostaglandin-endoperoxide synthase 2 (PTGS2), which is also termed cyclooxygenase 2 (COX2), a key enzyme that catalyzes

the conversion of arachidonic acid to prostaglandins, has been shown to play a pivotal role in gastric inflammation and carcinogenesis (134). For example, a study by Chang et al. (108), using clinical *H. pylori* isolates, has shown that *H. pylori* acts through TLR2/TLR9 to activate both the PI-PLCγ/PKCα/c-Src/IKKα/β and NIK/IKKα/β pathways, resulting in the phosphorylation and degradation of IκBα, which in turn leads to the stimulation of NF-κB and the expression of PTGS2.

Further, as compared with normal cells, cancer cells are more metabolically active and generate more reactive oxygen species (ROS), which affects cell survival. Several studies have suggested that ROS can act as secondary messengers and control a range of signaling cascades, leading to sustained proliferation of cancer cells (135, 136). In the context of gastric carcinogenesis, *H. pylori*infected gastric epithelial cells have been shown to generate ROS (137). Interestingly, Yuan et al. (138) recently suggested that TLR4 expression in GC correlated with tumor stage and that activation of TLR4 contributed to GC cell proliferation via mitochondrial ROS production through up-regulation of phosphorylated Akt and NF-κB p65 activation and nuclear translocation.

However, the involvement of TLRs in GC might be more complex than initially suspected as TLRs not only recognize antigenic determinants of viruses, bacteria, protozoa, and fungi, but are also involved in the detection of damage-associated molecular patterns (DAMPs) (e.g., extracellular adenosine triphosphate, hyaluran, extracellular glucose, monosodium urate crystals) (139). Release of DAMPs, which are especially targeted by TLR2 and TLR4 (140–145) during cancer progression may cause chronic inflammation leading to down-regulation of the ζ chain of the T-cell and NK cell activating receptors [for comprehensive information on this topic see the review by Baniyash et al. (146)], which entails T-cell and NK cell dysfunction, a phenomenon observed in some malignancies such as GC (147, 148), colon (149), prostate (150), cervical (151), and pancreatic cancer (152). In addition to immunosuppression, DAMPS appear to facilitate other processes during gastric carcinogenesis. For example, Wu et al. (153) have recently showed that hyaluronan, derived from malignant cells, induced long-lived tumor-associated neutrophils and subsequent malignant cell migration in gastric carcinomas via a TLR4/PI3K interaction.

Collectively, TLRs might be involved in both gastric carcinogenesis mediated by *H. pylori* infection (a tumor-promoting consequence of inflammatory responses) and in GC perpetuation associated with immunosuppression (active evasion by cancer cells from attack and elimination by immune cells) and increased metastasis.

#### **GENETIC POLYMORPHISMS INVOLVED IN THE TOLL-LIKE RECEPTOR SIGNALING PATHWAY AND GASTRIC CANCER**

In recent years, a number of investigations have attempted to establish the relationship between polymorphisms in molecules of the TLR signaling pathway and risk of GC. Recent studies, conducted in several populations, have shown associations between the polymorphisms *TLR1* rs5743618 (Ile602Ser) (154), *TLR2* −196 to −174del (155–158), *TLR2* rs3804099 (157), *TLR4* rs4986790 (Asp299Gly) (155, 157, 159), *TLR4* rs4986791 (Thr399Ile) (160), *TLR4* rs10116253 (161), *TLR4* rs10983755

(162), *TLR4* rs11536889 (+3725G/C) (155), *TLR4* rs1927911 (161), *TLR5* rs5744174 (158), *TLR9* rs187084 (−1486 T/C) (163), and *CD14* rs2569190 (−260 C/T) (155, 164–167), and risk of GC development in an ethnic-specific manner (**Table 1**). In addition, three polymorphisms located in the *TLR4* mRNA promoter region (sites −2081, −2026, and −1601) and *TLR4* Thr135Ala at the leucine-rich repeat (LRR), have been associated with poorly differentiated GC (168, 169).

Interestingly, some of these polymorphisms including *TLR4* Asp299Gly (159, 184), *TLR4* Thr399Ile (184, 185), *TLR4* rs10759932 (186), *CD14*-260 C/T (187), and *TLR2* −196 to −174del (157), appear to be involved in the biological continuum that results in intestinal-type GC as they have also been associated with gastric precancerous lesions (**Table 2**).

Given that some authors have failed to show specific associations between polymorphisms in the TLR signaling pathway, especially in *TLR2*, *TLR4*, and *CD14*, and gastric precancerous lesions/GC (157, 160, 162, 164, 172, 174–178, 180–183, 185, 188, 189), we performed the first global meta-analysis to assess the role of *TLR2*, *TLR4*, and *CD14* polymorphisms in gastric carcinogenesis (155), in an attempt to clarify the limited and current conflicting evidence, and to establish the true impact of the TLR signaling pathway in GC. Our meta-analysis, which included 18 case–control studies conducted in Caucasian, Asian, and Latin American populations, showed that *TLR4* Asp299Gly was a definitive risk factor for GC in Western populations (pooled OR: 1.87, 95% CI: 1.31–2.65). In addition, there was a potential association between *TLR2* −196 to −174 and GC in Japanese (pooled OR: 1.18, 95% CI: 0.96–1.45) (155). Interestingly, a recent metaanalysis on *TLR2* −196 to −174 and the risk of GC, conducted by Cheng et al. (190), failed to reproduce the findings in our metaanalysis, however, their stratification by ethnicity analyses included subjects from both Japan and China, which might explain the different outcomes. A further meta-analysis conducted by Chen et al. (191) that included 21 case–control studies showed an overall increased risk of GC in individuals harboring *TLR4* Asp299Gly (Allele analysis, OR: 1.84, 95% CI: 1.41–2.39) and *TLR4* Thr399Ile (Allele analysis, OR: 1.97, 95% CI: 1.22–3.18). Consistently, in stratified analyses by ethnicity, these authors only found an association between *TLR4* Asp299Gly (Allele analysis, OR: 1.90, 95% CI: 1.43–2.51) and *TLR4* Thr399Ile (Allele analysis, OR: 2.84, 95% CI: 1.56–5.15) in Caucasian individuals (191). Further, Zhao et al. (192) in an updated version of a meta-analysis that was initially conducted by Zhang et al. (193), on the risk of *TLR4* polymorphisms and risk of cancer in general, found a significant association with GC after stratifying by cancer type (OR: 2.00, 95% CI: 1.53–2.62). In addition, Zou et al. (194), through a metaanalysis that included 10 case–control studies, not only found that *TLR4* Asp299Gly was associated with GC (OR: 1.87, 95% CI: 1.44– 2.44), especially non-cardia GC (OR: 2.03, 95% CI: 1.51–2.72), but also gastric precancerous lesions (OR: 2.47, 95% CI: 1.57–3.88), especially in *H. pylori*-infected individuals (OR: 3.43, 95% CI: 1.92–6.13).

Given limited evidence regarding the association between polymorphisms in other molecules of the TLR signaling pathway and the risk of GC, and the fact that 42% of cases of GC worldwide occur in the Chinese population, we conducted a case–control

#### **Table 1 | Genetic polymorphisms in theToll-like receptor signalling pathway that have been studied in relation to gastric cancer (170)**.


(Continued)

#### **Table 1 | Continued**


GC, gastric cancer; OR, odds ratio; CI, confidence intervals; NS, not specified.

<sup>a</sup>OR and 95% CI correspond to allele or genotype analysis, depending on available information in the article.

<sup>b</sup>The control group included individuals with high risk gastritis (pangastritis, corpus-predominant gastritis with or without the presence of gastric atrophy, and intestinal metaplasia in either antrum or corpus).

<sup>c</sup>Compared to gastric atrophy controls.

<sup>d</sup>The study population is from Poland.

<sup>e</sup>The study population is from the United States. No significant association was found with cardia GC.

<sup>f</sup>Subjects from Germany, Lithuania and Latvia.

<sup>g</sup>Subjects from France, Italy, Spain, United Kingdom, The Netherlands, Greece, Germany, Sweden, Denmark and Norway.

<sup>h</sup>Effect size for intestinal-type GC, diffuse type: OR: 0.99, 95% CI: 0.78–1.26.

study comprising 310 ethnic Chinese individuals (87 non-cardia GC cases and 223 controls with functional dyspepsia), in which 25 polymorphisms involved in the TLR signaling pathway were investigated (170). Seven polymorphisms showed significant associations with GC (*TLR4* rs11536889, *TLR4* rs10759931, *TLR4* rs1927911, *TLR4* rs10116253, *TLR4* rs10759932, *TLR4* rs2149356, and *CD14* −260 C/T). In multivariate analyses, *TLR4* rs11536889 remained a risk factor for GC even after adjustment (OR: 3.58, 95% CI: 1.20–10.65). Further, *TLR4* rs10759932 decreased the risk of *H. pylori* infection (OR: 0.59, 95% CI: 0.41–0.86) (170). Strikingly, statistical analyses assessing the joint effect of *H. pylori* and the selected polymorphisms revealed that *H. pylori*-infected individuals harboring *TLR2* rs3804100, *TLR2* −196 to −174del, *TLR4* rs11536889, *MD-2* rs11465996, *MD-2* rs16938755, *LBP* rs2232578, and *TIRAP* rs7932766 were at most risk of developing GC (**Table 1**) (170).

The functional relevance of a number of these polymorphisms has already been established. For example, two polymorphisms in *TLR4*, Asp299Gly, and Thr399Ile, have been shown to disrupt the normal structure of the extracellular domain of TLR4, and thus, as a result, may reduce responsiveness to *H. pylori* by diminishing the binding affinity of the bacterial ligands (195). In addition, the *TLR4* rs11536889 polymorphism, which is located in the center of the 2818-bp *TLR4* 3 <sup>0</sup> untranslated region (UTR), has recently been shown by Sato et al. (196) to contribute to the translational regulation of *TLR4*, possibly by binding to microRNAs. Further, these authors elegantly demonstrated that subjects harboring *TLR4* rs11536889 exhibited higher levels of TLR4 receptors on monocytes and secreted higher levels of IL-8 in response to LPS (196). In addition, *TLR4* rs10759932 has been shown to decrease the expression of forkhead box protein P3 (FOXP3), the most specific marker for natural regulatory T (Treg) cells (197). FOXP3+ Treg cells, which suppress the immune response of antigen-specific T cells, have been demonstrated to play a key role in immunologic tolerance (198). Notably, recent studies have not only shown

that *in vivo* depletion of FOXP3+ Treg cells in *H. pylori*-infected mice leads to increased gastric inflammation and reduced bacterial colonization (199), but also recruitment of FOXP3+ Treg cells is increased in *H. pylori*-related human disorders including gastritis (200, 201), duodenal ulcer (202), and GC (200, 203, 204), suggesting that FOXP3+ Treg cells might contribute to lifelong persistence of *H. pylori* infection. Also, *TLR1* rs5743618 appears to impair the surface expression of TLR1 of NK cells and NK cellsderived IFN-γ production (154). Further, *TLR2* −196 to −174 has been associated with decreased transcriptional activity of *TLR2* (205, 206). Similarly, it has been demonstrated that*TLR9* rs187084 down-regulates TLR9 expression (207).

Further, CD14 has been shown to activate macrophages/ monocytes to release Th1-type cytokines including IL-12, thus, establishing the chronic inflammation stimulated by *H. pylori* infection (208–210). A Th1 predominant response has been extensively associated with the pathogenesis of *H. pylori*-related gastric disease (211–213). Currently, however, controversy exists regarding the influence of *CD14* −260 on expression of soluble CD14 (sCD14). According to a number of studies, the *CD14* −260 T allele is believed to increase sCD14 production and therefore, serum sCD14 levels (214–217). In contrast, it has been reported that elevated sCD14 levels are associated with *H. pylori* infection, especially in subjects with the *CD14* −260 CC genotype (167). Alternatively, others have argued that this polymorphism has no effect on transcription (218). Since the evidence to date is conflicting, more functional studies are required to clarify this issue.

Overall, it is clear that genetic variability in genes of the TLR signaling pathway plays an important role in GC pathogenesis. Investigations of polymorphisms in different molecules of this pathway among different populations could provide novel insights into targeted treatment in genetically susceptible individuals, and thus, improve primary and secondary prevention of *H. pylori*-related GC in high risk populations.


#### **Table 2 | Genetic polymorphisms in theToll-like receptor signalling pathway that have been studied in relation to gastric precancerous lesions**.

CG, chronic gastritis; AG, atrophic gastritis; IM, intestinal metaplasia; OR, odds ratio; CI, confidence intervals.

<sup>a</sup>OR and 95% CI correspond to allele or genotype analysis, depending on available information in the article.

<sup>b</sup>Only individuals with corpus-predominant chronic gastritis were included in the meta-analysis (individual presenting antrum-predominant gastritis were excluded). <sup>c</sup>Analyses including only H. pylori seropositive individuals.

<sup>d</sup>Only patients with chronic gastritis were included in the meta-analysis (patients presenting duodenal ulcer were excluded).

<sup>e</sup>Cases were GC patients' relatives with gastric atrophy and infected with H. pylori from a Scotland population.

#### **NOD-LIKE RECEPTORS AND HELICOBACTER PYLORI-RELATED GASTRIC CANCER**

#### **NOD-LIKE RECEPTORS RECOGNITION OF HELICOBACTER PYLORI**

The NLR family not only recognizes PAMPs but also DAMPs in the cytoplasm (93). The NLRs characteristic structure includes a central nucleotide-binding and oligomerization (NACHT) domain that is present in all NLR family members, a C-terminal LRRs and an N-terminal caspase recruitment (CARD) or pyrin (PYD) domain.

Based on phylogenetic analysis of NACHT domains, the NLR family has been shown to comprise three subfamilies: (1) the NOD family which includes NOD1-2, NOD 3 (NLRC3), NOD4 (NLRC5), NOD5 (NLRX1), and CIITA, (2) the NLRPs including NLRP1-14 (also known as NALPs), and (3) the IPAF subfamily, which consists of IPAF (NLRC4) and NAIP (93).

The NACHT domain belongs to a family of P-loop NTPases known as the signal transduction ATPases with numerous domains (STAND) (219). This domain permits activation of the signaling complex via adenosine ATP-dependent oligomerization (94). NACHT domain oligomerization is essential for the activation of NLRs, forming high molecular weight complexes, probably hexamers or heptamers that characterize inflammasomes (molecular complexes involved in the activation of inflammatory caspases for the maturation and secretion of IL-1β, IL-18, and possibly IL-33) and NOD signalosomes (complexes that are assembled upon oligomerization of NOD1 or NOD2 and lead to NF-κB activation through the receptor-interacting protein-2) (94). CARD and PYD are death domains that mediate homotypic protein–protein interactions for down-stream signaling (93, 94). These domains are characterized by six α helices that form trimers or dimers with other members of the same subfamily (94). The third domain, the LRR region, has been implicated in ligand sensing and autoregulation of not only NLRs but TLRs (93, 94). The LRR is formed by tandem repeats of a structural unit consisting of a β strand and an α helix and is composed of 20–30 amino acids that form a horse-shoe shaped structure rich in the hydrophobic amino acid leucine (220). The NLRPs LRR gene is made up of tandem repeats of exons of exactly 171 nucleotides, which encode one central LRR and two halves of the neighboring LRRs (221). This particular modular organization possibly allows extensive alternative splicing of the LRR region leading to maximum variability in the ligand-sensing unit (94). However, a recent publication by Tenthorey et al. (222) analyzing a panel of 43 chimeric NAIPs, showed that LRR was unnecessary for NAIP/NLRC4 inflammasome ligand specificity, leading them to propose a model in which NAIP activation is instead triggered by ligand binding to NACHTassociated helical domains. This recent evidence suggests that the ligand-sensing function of the LRR domain in NLRs, which has been supported primarily by analogy to the well-established ligand-sensing function of the LRR region in TLRs, needs to be re-examined.

The most widely studied NLRs during *H. pylori* infection are NOD1 and NOD2, which are expressed in epithelial and antigen-presenting cells, and are known to specifically recognize peptidoglycan-derived peptides (γ-d-glutamyl-mesodiaminopimelic acid and muramyl dipeptide, respectively). An early study, attempting to determine the mechanism whereby *H. pylori* delivers peptidoglycan to cytosolic host NOD1, demonstrated that *H. pylori* peptidoglycan is delivered to the host cell via a type IV secretion system (223). More recently, Hutton et al. (224) showed, for the first time, that cholesterol-rich microdomains called lipid rafts, were important for the type IV secretion system-dependent peptidoglycan delivery and subsequent NF-κB activation and IL-8 production, mediated by NOD1. Interestingly, Kaparakis et al. (225) reported a novel mechanism in Gram-negative bacteria, including *H. pylori*, for the delivery of peptidoglycan to cytosolic NOD1 in host cells that involves outer membrane vesicles that enter epithelial cells through lipid rafts. In addition, Necchi et al. (226) demonstrated the formation of a particle-rich cytoplasmic structure (PaCS) in *H. pylori*-infected human gastric epithelium having metaplastic or dysplastic foci, where VacA, CagA, urease, outer membrane proteins, NOD1 receptor, ubiquitin-activating enzyme E1, polyubiquitinated proteins, proteasome components, and potentially oncogenic proteins like SHP-2 and ERKs colocalized, inferring that this structure is likely to modulate inflammatory and proliferative responses during *H. pylori* infection.

The recent finding that NF-κB and AP-1 complexes can be physically translocated to the nucleus in response to NOD1 activation has led to the view that NOD1 is likely to be essential for the induction of both NF-κB and AP-1 activation during *H. pylori* infection (227). A number of studies have shown up-regulation of *NOD1* expression in diverse human cell lines challenged with *H. pylori* in a *cag* PAI-dependent manner (228–230). Further, *H. pylori cag* PAI-positive strains have recently been shown to activate the NOD1 pathway through two components of the IFN-γ signaling pathway, STAT1 and IRF1 (228). Similarly, expression of NOD2 was shown to significantly sensitize HEK293 cells to *H. pylori*induced NF-κB activation in a *cag* PAI-dependent manner (231). Further, NOD2, but not NOD1, seems to be required for induction of pro-IL-1β and NLRP3 in *H. pylori*-infected DCs (232).

A limited number of studies have assessed the interaction between NLRPs and other inflammasome-associated molecules, and *H. pylori*. NLRPs represent the largest NLR subfamily (14 genes have been identified in humans) and are believed to be the scaffolding proteins of inflammasomes (221, 233). NLRPs interact and recruit the adaptor apoptosis-associated speck-like protein (ASC) via PYD-PYD interaction (94). ASC (also known as PYCARD), a key component required for inflammasome formation, is formed by an N-terminal PYD and a C-terminal CARD (234, 235). This interaction leads to the recruitment of caspase-1, an intracellular aspartate specific cysteine protease, which subsequently leads to the maturation and release of pro-inflammatory cytokines (236).

An early study by Tomita et al. (237) demonstrated that in *H. pylori* positive patients antral IL-18 mRNA expression was increased as compared with *H. pylori* negative patients, however, mature IL-18 protein and active caspase-1 were found to be present in both infected and non-infected gastric mucosa. Interestingly, in the following year, Potthoff et al. (238) reported activation of caspase-3, -8, and -9, but not caspase-1, in AGS cells challenged with *H. pylori*. However, this finding is in contrast with subsequent studies, which have demonstrated an important role for NLRPs and inflammasome-related molecules in *H. pylori* infection. For example, Basak et al. (96) demonstrated that *H. pylori* LPS could activate caspase-1 through Rac1/PAK1 signaling, and that activated caspase-1 played a role in LPS-induced IL-1β maturation (96). Further, ASC-deficient mice challenged with *H. pylori* have been shown to exhibit higher bacterial loads and significantly lower levels of gastritis, when compared with wild-type mice, and were incapable of producing IL-1β or IL-18 and produced less INF-γ in response to *H. pylori* infection (239). Later, Hitzler et al. (240) showed in both cultured DCs and *in vivo* that *H. pylori* infection activates caspase-1, leading to IL-1β/IL-18 processing and secretion. Consistently, three studies, using human GC cell lines, gastric tissue, and murine models, confirmed increased

expression of caspase-1, IL-1β, and IL-18 in *H. pylori*-infected cells (171, 241, 242). Further, Jiang et al. (243), also using a murine model, have reported the expression of NLRP3 inflammasomerelated molecules as well as serum IL-1β, IL-18, and IL-33 levels to be significantly increased in *H. pylori*-infected mice. More recently, a study by Kim et al. (232) has shown that secretion of IL-1β by DCs infected with *H. pylori* requires TLR2, NOD2, and the NLRP3 inflammasome.

Given that little is known about the role of NLRPs, inflammasomes, or other molecules involved in the NLR signaling pathways in response to *H. pylori* infection, we recently assessed the gene expression of 84 different molecules involved in the NLR signaling pathways, through quantitative real-time PCR, using THP-1-derived macrophages infected with two strains of *H. pylori*, GC026 (GC) and 26695 (gastritis) (173). Our gene expression analyses showed five genes encoding NLRs to be significantly regulated in *H. pylori*-challenged cells (*NLRC4*, *NLRC5*, *NLRP9*, *NLRP12*, and *NLRX1*) (173). Interestingly, *NLRP12* and *NLRX1*, two known NF-κB negative regulators, were markedly downregulated, while *NFKB1* and several NF-κB target genes encoding pro-inflammatory cytokines (*IFNB1*, *IL12A*, *IL-12B*, *IL6*, and *TNF*), chemokines (*CXCL1*, *CXCL2*, and *CCL5*) and molecules involved in carcinogenesis (*PTGS2* and *BIRC3*) were markedly upregulated, in THP-1 cells infected with a highly virulent *H. pylori* strain isolated from a GC patient. These findings highlight the relevance of the NLR signaling pathways in gastric carcinogenesis and its close interaction with NF-κB (173).

Overall, current evidence clearly shows that, in response to *H. pylori*, members of the NOD and NLRP subfamilies are critical for generating mature pro-inflammatory cytokines/chemokines that are crucial for Th1 responses and lead to *H. pylori*-related gastric disorders.

#### **NOD-LIKE RECEPTORS AND GASTRIC CARCINOGENESIS**

The role of the NLR signaling pathways in the biological continuum that characterizes GC remains relatively unexplored as a very limited number of studies have addressed this issue. For example, Allison et al. (228) have shown that NOD1 expression was significantly increased in human gastric biopsies displaying severe gastritis, when compared with those without gastritis, as well as in gastric tumor tissues, as compared with paired non-tumor tissues. In contrast, Jee et al. (244), who analyzed human GC tissues and GC cell lines, showed that a significant decrease in the expression of caspase-1 was associated with poor survival and was inversely correlated with p53 expression.

Given the reported interaction of *H. pylori* with NLRs and the importance of this in the development of gastric inflammation and subsequent carcinogenesis, as well as the production of DAMPs during tumor formation (245), further comprehensive studies of the functional relevance of NLRs activation during chronic gastritis, atrophic gastritis, intestinal metaplasia, dysplasia, and GC are clearly warranted.

#### **GENETIC POLYMORPHISMS INVOLVED IN THE NOD-LIKE RECEPTOR SIGNALING PATHWAY AND GASTRIC CANCER**

The majority of studies examining the association between polymorphisms involved in the NLR signaling pathways and the risk of GC have focused on *NOD1* and *NOD2* polymorphisms. Studies, conducted in a number of populations, have investigated the association between the polymorphisms *NOD1* rs2907749 (246), *NOD1* rs7789045 (246), *NOD1* rs2075820 (E266K) (179, 247), *NOD1* rs5743336 (180), *NOD2* rs7205423 (246), *NOD2* rs7202124 (164), *NOD2* rs2111235 (164), *NOD2* rs5743289 (164), *NOD2* rs2066842 (P268S) (248, 249), *NOD2* rs2066844 (R702W) (250), *NOD2* rs2066845 (G908R) (184), and *NOD2* rs2066847 (L1007insC) (184, 250), and risk of gastric precancerous lesions and GC (**Table 3**). Further, a recent meta-analysis by Liu et al. (251) that included six case–control studies has shown consistent associations between *NOD2* R702W, G908R, and L1007insC, and risk of GC.

Given the documented relevance of other NLRs in *H. pylori* infection and related GC, and that polymorphisms in genes such as *NLRP3* (252–255) and *CARD8* (255, 256) have been associated with inflammatory gastrointestinal disorders, we addressed, for the first time, the association between 51 polymorphisms in six genes (*NLRP3*, *NLRP12*, *NLRX1*, *CASP1*, *ASC*, and *CARD8*) involved in the NLR signaling pathways and risk of GC in a high risk Chinese population (173). In this study, we found novel associations between *CARD8* rs11672725 and the risk of GC, and *NLRP12* rs2866112 and the risk of *H. pylori* infection (**Table 3**). Further, we showed that the concomitant presence of polymorphisms involved in the NLR signaling pathways (*CARD8*, *NLRP3*, *CASP1*, and *NLRP12*) and *H. pylori* infection dramatically increased the risk of GC in Chinese (**Table 3**) (173).

The functional relevance of a number of these polymorphisms has been examined. For example, the introduction of *NOD2* R702W, a polymorphism located in the LRR of NOD2, into the HEK293 cell line, resulted in abrogation of *H. pylori*-induced activation of NF-κB signaling (231). Further, Maeda et al. (257) observed increased NF-κB activation in response to muramyl dipeptide in mice harboring a *NOD2* mutation that is homologous to *NOD2* rs5743293 (3020insC) in humans. However, it is worth noting that the conclusions described by Maeda et al. (257) must be interpreted with care given that the authors subsequently found a duplication of the 3' end of the wild-type *Nod2* locus, including exon 11, which was targeted by the mutation, and therefore, they are currently working to recreate a mutant strain without such a duplication.

Given that investigation of the role of polymorphisms involved in the NLR signaling pathways in GC is a relatively recent field of research, further studies are required to assess the association between these polymorphisms and GC in a range of human populations, especially those at high risk of GC.

#### **OTHER PATTERN-RECOGNITION RECEPTORS AND HELICOBACTER PYLORI-RELATED GASTRIC CANCER**

A further two PRR subfamilies, RLRs and CLRs, have been studied in relation to *H. pylori* infection and gastric carcinogenesis. It is well known that RLRs (RIG-I, MDA-5, and LGP2) induce type I IFN in response to different RNA viruses, however, investigation on the role of RIG-I-like receptors in the recognition of RNA derived from intracellular bacteria is very limited. Interestingly, a study by Rad et al. (112), which used mice lacking simultaneously up to four different TLRs, apart from identifying TLR2 and


**Table 3 | Genetic polymorphisms in the NOD-like receptor signalling pathway that have been studied in relation to gastric precancerous lesions and gastric cancer**.

GC, gastric cancer; IM, intestinal metaplasia; AG, atrophic gastritis; GC, chronic gastritis; OR, odds ratio; CI, confidence intervals.

<sup>a</sup>OR and 95% CI correspond to allele or genotype analysis, depending on available information in the article.

<sup>b</sup>Only in H. pylori-infected individuals.

<sup>c</sup>Significant only in non-cardia H. pylori CagA negative individuals.

<sup>d</sup>Results obtained through a Fisher's exact probability test (two-tailed P-values) conducted in the current review using the information provided in the original article.

TLR9 to be important *H. pylori* recognizing PRRs, also showed that *H. pylori* 5 0 -triphosphorylated RNA can be sensed by RIG-I and can contribute to the TLR-independent type I IFN response to this bacteria in DCs. Further, Tatsuta et al. (258) have recently shown that MDA-5 expression was significantly increased in the human gastric antral mucosa of *H. pylori*-infected individuals. In addition, these authors showed that increased MDA-5 levels correlated with atrophy and intestinal metaplasia in the corpus of these individuals (258).

C-type lectin receptors bind to carbohydrates (mannose- or fucose-containing glycans) present on pathogens to tailor immune responses to viruses, bacteria, and fungi. DC-specific intercellular

adhesion molecule-3-grabbing non-integrin (DC-SIGN) is a CLR expressed on the surface of both macrophages and DCs. Interestingly, it has been shown that *H. pylori* harbors fucosylated ligands that can be recognized by DC-SIGN (259). Further, *H. pylori* DC-SIGN ligands appear to actively dissociate the signaling complex down-stream of DC-SIGN (KSR1–CNK–Raf-1) to suppress pro-inflammatory cytokine production (259). In addition, *H. pylori* LPS Lewis blood-group antigens can bind to DC-SIGN in a fucose or galactose-dependent manner (260, 261) and this interaction appears to inhibit a Th1 response in DCs (262). It has also been demonstrated that *H. pylori*-induced IL-10 production in monocyte-derived DCs is significantly suppressed by the addition of anti-DC-SIGN, TLR2, or TLR4 antibodies, either alone or in combination, before *H. pylori* stimulation (263). Further, *in vitro* and *in vivo* experiments have shown that the expression of DC-SIGN is significantly higher in *H. pylori*-infected individuals as compared with that in their uninfected counterparts (264, 265).

To date, no studies have been conducted to determine the association between genetic polymorphisms involved in the RLR and CLR signaling pathways and GC, however, Kutikhin and Yuzhalin (266) have comprehensively analyzed the oncogenic potential of both RLRs and CLRs, suggesting that future oncogenomic investigations should focus on polymorphisms in *MRC1* (rs1926736, rs2478577, rs2437257, and rs691005), *CD209* (rs2287886, rs735239, rs4804803, and rs735240), *CLEC7A* (rs16910526), and *RIG-I* (rs36055726, rs11795404, and rs10813831).

Given the limited but consistent current evidence suggesting a role of RLRs and CLRs in *H. pylori* infection, and the documented interaction between these signaling pathways and other important PRRs in GC such as TLRs (267, 268) and NLRs (269, 270), further studies assessing the implications of RLRs and CLRs in *H. pylori*related inflammation and subsequent carcinogenesis need to be conducted.

#### **PATTERN-RECOGNITION RECEPTORS AS THERAPEUTICS TARGETS IN GASTRIC CANCER**

Pattern-recognition receptors are increasingly recognized as important players in immunotherapy as PRRs-specific agonists elicit a potent immune response to cancers, allergic diseases, and chronic viral infections, while reducing the risk of an uncontrolled and detrimental systemic inflammatory response (for comprehensive information on this topic refer to the reviews by Hedayat et al. (271) and Paul-Clark et al. (272).

In the context of gastric carcinogenesis, Tye et al. (132), using a GC murine model (gp130F/F) displaying elevated gastric TLR2 expression levels, have elegantly shown that genetic and antibodymediated therapeutic targeting of TLR2 leads to a substantial reduction in stomach size and overall tumor burden, including the number of gastric tumors. A further example is presented in the study by Gradisar et al. (273), which suggested that MD-2 is one of the important targets of curcumin (diferuloylmethane), the main component of the spice turmeric (*Curcuma longa*) that is widely used for gastric disorders in the Indian subcontinent, in its suppression of the innate immune response to bacterial infection. Furthermore, curcumin was recently shown to polarize myeloid-derived suppressor cells, extracted from a human GC xenograft mouse model, toward a M1-like phenotype with an increased expression of CCR7 and decreased expression of the CLR dectin 1, being both observed *in vivo* (tumor tissue) and *in vitro* (splenic myeloid-derived suppressor cells from tumorbearing mice) (274). In addition, a study by Yang et al. (171) demonstrated that the combination of catechins and sialic acid is effective in suppressing the inflammatory responses mediated by the inflammasome/caspase-1 signaling pathway in gastric epithelial cells during *H. pylori* infection. Also, poly(I:C), an agonist of TLR3 and RLRs, has been shown to have a pro-apoptotic effect *in vitro*, and has significantly inhibited xenograft growth of human GC in a mouse model, through up-regulation of RLRs (RIG-I, MDA-5, and LGP2) as well as an increased expression of Bcl-2 family members, suggesting that it may be a promising chemotherapeutic agent against GC (275).

Given that modulation of PRRs has been proven to be relevant in gastric carcinogenesis through diverse mechanisms, including suppression of *H. pylori*-induced inflammation and enhancement of cancer cell apoptosis, this approach should be considered a new and promising angle of immunotherapy in GC.

#### **CONCLUSION**

In conclusion, abundant evidence supports the pivotal role of PRRs in gastric carcinogenesis as these receptors of the innate immune system, including TLRs, NLRs, CLRs, and RLRs, have been shown to recognize diverse components of *H. pylori*, the major risk factor of GC. In addition, PRRs are also involved in gastric carcinogenesis *per se* as these receptors are known to exert tumor-promoting functions (cell proliferation, epithelial– mesenchymal transition, angiogenesis, and metastasis) as well as immunosuppression during cancer. Given that host genetic variability in the TLR and NLR signaling pathways are known to be associated with an increased risk of *H. pylori* infection, the development of gastric precancerous lesions and GC, this knowledge has the potential to allow better prevention of GC through selective treatment and surveillance of individuals harboring high risk genetic profiles. Finally, given that PRRs are increasingly being used as a target for immunotherapy against both cancer and infectious diseases, the established relevance of PRRs in *H. pylori* infection and GC, could suggest that PRR agonists and/or antagonists may potentially improve the outcome of GC. Based on the extensive evidence presented in the current review, we propose a synergistic interaction between PRRs and *H. pylori*, which over time, could facilitate the sequence of events that characterizes GC development including inflammation, atrophy, intestinal metaplasia, dysplasia, and finally, GC (**Figure 3**).

#### **ACKNOWLEDGMENTS**

Nadeem O. Kaakoush is supported by an Early Career fellowship from the National Health and Medical Research Council, Australia.

#### **REFERENCES**


systematic review and meta-analyses. *Am J Epidemiol* (2011) **173**:259–70. doi:10.1093/aje/kwq370


receptor polymorphisms in a Japanese population. *Clin Exp Allergy* (2004) **34**:177–83. doi:10.1111/j.1365-2222.2004.01839.x


through phase-variable interaction between lipopolysaccharide and DC-SIGN. *J Exp Med* (2004) **200**:979–90. doi:10.1084/jem.20041061


**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.

*Received: 15 April 2014; paper pending published: 25 May 2014; accepted: 03 July 2014; published online: 22 July 2014.*

*Citation: Castaño-Rodríguez N, Kaakoush NO and Mitchell HM (2014) Pattern-recognition receptors and gastric cancer. Front. Immunol. 5:336. doi: 10.3389/fimmu.2014.00336*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Castaño-Rodríguez, Kaakoush and Mitchell. 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.*

## Multiple roles of toll-like receptor 4 in colorectal cancer

#### **DhanushaYesudhas† , Vijayakumar Gosu† , Muhammad Ayaz Anwar and Sangdun Choi \***

Department of Molecular Science and Technology, Ajou University, Suwon, South Korea

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Amedeo Amedei, University of Florence, Italy Rajesh Kumar Sharma, University of Louisville, USA Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Sangdun Choi, Department of Molecular Science and Technology, Ajou University, Suwon 443-749, South Korea

e-mail: sangdunchoi@ajou.ac.kr

†Dhanusha Yesudhas and Vijayakumar Gosu have contributed equally to this work.

#### **INTRODUCTION**

The innate immune system possesses a robust mechanism in the form of evolutionarily conserved toll-like receptors (TLRs) that can detect the signature pattern of invading microorganisms for the protection of the host. TLRs are a class of type I transmembrane glycoproteins. Human and mouse cells comprises of 13 types of TLRs that can detect different kinds of bacterial and viralassociated patterns (1–3). TLR1–9 are highly conserved in both species; while the mouse TLR10 is non-functional due to retroviral insertion, TLR11–13 are undetected in the human genome. Examples of TLR-specific ligands are: lipopeptides for TLR1/2 and 2/6 (4–6), dsRNA for TLR3 (7), lipopolysaccharide (LPS) for TLR4 (8), flagellin for TLR5 (9), ssRNA for TLR7/8 (10, 11), and CpG DNA for TLR9 (12–14). TLRs not only detect invading microbes but also recognize intracellular anomalies and mount an immune response, thereby playing a cardinal role in the homeostasis of the human immune system (15, 16). The abnormal activation of TLRs can jeopardize normal physiological processes and cause several inflammatory diseases, cancers, and autoimmune diseases (17, 18).

Toll-like receptors are ubiquitously expressed, although their expression level may vary according to the circumstances and the tissues. In addition, induced expression of TLRs has been observed when ligands bind to their cognate TLRs (19). Research in the last decade has focused on elucidating various functions, intermediate molecules, and ligands associated with TLRs. There is a well-established link between TLR-induced inflammation and the development and progression of cancer (20, 21). Similarly, TLRs are also known to play a vital role in colorectal cancer (CRC) that affects the large intestine and the rectum. This region is heavily populated by intestinal microbes, highlighting the crucial role of TLRs in CRC pathogenesis (17, 22).

Toll-like receptor (TLR) signaling has been implicated in the inflammatory responses in intestinal epithelial cells (IECs). Such inflammatory signals mediate complex interactions between commensal bacteria and TLRs and are required for IEC proliferation, immune response, repair, and homeostasis. The upregulation of certain TLRs in colorectal cancer (CRC) tissues suggests that TLRs may play an essential role in the prognosis of chronic and inflammatory diseases that ultimately culminate in CRC. Here, we provide a comprehensive review of the literature on the involvement of the TLR pathway in the initiation, progression, and metastasis of CRC, as well as inherited genetic variation and epigenetic regulation. The differential expression of TLRs in epithelial cells has also been discussed. In particular, we emphasize the physiological role of TLR4 in CRC development and pathogenesis, and propose novel and promising approaches for CRC therapeutics with the aid of TLR ligands.

**Keywords: colorectal cancer, immune response, inflammation, ligand, toll-like receptor 4**

Colorectal cancer is one of the most complex diseases and causes death in many cases in the United States (23). Globally, more than one million new cases of CRC are reported annually (24, 25). The complexity of CRC is primarily attributed to environmental factors, while genetic factors play a minor role. The known risk factors for CRC are food-borne mutagens, pollution, certain commensal bacteria, and chronic intestinal inflammation (25). Commonly, CRC occurs in the right ascending colon with the most common symptom being blood in the stool or rectal bleeding. Genetically, inherited colon polyps also contribute to the development of CRC (26). Since CRC can damage the host immune system during their proliferation period, stimulating it against CRC promises to be an attractive approach for drug discovery (27).

In this review, we discuss the role of TLRs in the maintenance of homeostasis and the development of CRC in intestinal epithelial cells (IECs). Improved techniques to detect dysfunctional TLR signaling in carcinogenesis may stimulate the development of novel therapies to prevent or treat CRC. Recent studies have improved the understanding of TLR-targeted applications such as identifying their differential expression, their role in tumor progression, potential use as immune modulating agents, and development of novel TLR ligands in anti-cancer therapies.

#### **TLR SIGNALING: AN OVERVIEW**

The localization of TLRs is heterogeneous and varies from the cell surface (TLR1, 2, 4, 5, 6, 10, and mouse TLR11, 12) to the endosomes (TLR3, 7, 8, and 9) (28), depending on the localization of pathogen-associated molecular patterns (PAMPs). TLRs comprises of the following three domains: ectodomain [contains leucine rich repeats (LRR)] that recognizes PAMPs, a trans-membrane region, and a cytosolic toll/interleukin-1 (IL-1) receptor (TIR) domain that interacts with adaptor molecules (such as MyD88/MAL and TRIF/TRAM) to propagate downstream signaling. Ligand binding triggers the dimerization of TLRs, facilitating the binding of adaptor molecules, which subsequently activate the IL-1 receptor-associated kinase (IRAK) family (29). Upon IRAK recruitment, IRAK4 phosphorylates IRAK1 at key serine and threonine residues, and enables IRAK1 to eventually activate tumor necrosis factor receptor-associated factor 6 (TRAF6) (30) that subsequently activates transforming growth factor-βactivated protein kinase 1 (TAK1), a member of the mitogenactivated protein (MAP) kinase kinase kinase (MAP3K) family. TAK1 forms a complex with TGF-β-activated kinase 1/MAP3K7 binding protein 1 (TAB1), TAB2, and TAB3 and then activates nuclear factor (NF)-κB by phosphorylating IKK that in turn phosphorylates IκB for proteasomal degradation. Following the degradation of IκB, NF-κB translocates into the nucleus and induces inflammatory mediators. Moreover, TAK1 activates members of the MAP kinase kinase 3 (MKK3) and MKK6 to activate an alternative closely related pathway that phosphorylates c-Jun N-terminal Kinase (JNK) and p38. TLR signaling can also activate extracellular signal-regulated kinase (ERK) via the activation of MEK1/2. In response to various TLR ligands, reduced activity of NF-κB, JNK, and p38 was observed in B cells and embryonic fibroblasts derived from TAK1-deficient mice (31). In the TRIF-dependent pathway triggered by TLR3 and TLR4, TRIF recruits TRAF3, TAB1, and IKK and activates the type I IFN. The TRIF-dependent pathway also activates TRAF6 and TAB1, which regulate the delayed activation of NF-κB and MAP kinases (32) (**Figure 1**).

#### **TLRs AND THEIR EXPRESSION PATTERNS IN IECs**

The human intestinal tract plays a crucial role in maintaining the complex ecosystem of commensal bacteria and also physically isolates the countless resident bacteria from the lamina propria (33). It was originally believed that IECs prevent bacteria from invading the body. However, IECs have a complicated and common beneficial link with the microorganisms in the intestinal gut flora. The commensal bacteria metabolize carbohydrates and the IECs break down the short-chain fatty acids produced as a result of bacterial fermentation of undigested carbohydrates and use them as an energy source (34). IEC membranes express TLRs that detect the commensal PAMPs and mediate signaling to maintain epithelial cell integrity and tight junctions, cell proliferation, immunoglobulin A (IgA) production, and antimicrobial peptide expression (34). In addition, they can also induce a pro-inflammatory response by interacting with the immune cells in the lamina propria (35, 36). Therefore, tight regulation of TLRs is imperative to prevent adverse effects since anomalous or dysregulated TLR signaling can mediate cancer induction and propagation.

Colorectal cancer pathogenesis is governed by TLR expression that is difficult to detect due to the heterogeneous nature of IECs (33). To elucidate the expression profile of TLR2–5 in epithelial cells, small intestinal, and colonic biopsy specimens from patients with inflammatory bowel disease (IBD) were assessed by immunofluorescence histochemistry using polyclonal antibodies against TLR2, 3, 4, and 5. This study showed that TLR3 and TLR5 are ubiquitously expressed while TLR2 and TLR4 are expressed at a very low level in normal cells (37). Conversely, the diseased tissue specimens demonstrated significant overexpression of TLR4 and a decline in TLR3 expression. The expression pattern of TLR2 and TLR5 remained unaltered between the normal and diseased specimens. Furthermore, in normal human IECs, TLR2, and TLR4 were marginally expressed, while TLR3 expression was relatively high. While TLR2 was expressed in the colonic tissue from the epithelium and lamina propria, TLR3 was expressed in the mature epithelial cells of the crypts. Furthermore, TLR5 was moderately overexpressed in a basolateral fashion in the epithelial cells of normal human tissues (38). Tissues from CRC patients demonstrated increased expression of TLR7, 8, 9, and 10 (39); this study also showed that TLR8 expression is an independent marker for CRC.

#### **TLRs AND INTESTINAL HOMEOSTASIS**

Toll-like receptor activation is responsible for fighting against microbial infections, while leaving the host cell intact. This is usually accomplished by producing antimicrobial peptides, inflammatory mediators, adenomatous polyposis coli (APCs) maturation, and triggering of cell survival and tissue repairing pathways (40). TLRs are marginally expressed on IECs and are primarily localized on the basolateral surface or in the endosomal vesicles (41). Moreover, regulatory mechanisms such as the expression of TLR inhibitors like single immunoglobulin IL-1-related receptor (SIGIRR), toll-interacting protein (TOLLIP), A20, and IRAK3 are involved in the regulation of TLR signaling (42); these inhibitory molecules prevent TLRs from mounting an immune response even during continuous interaction (34, 43) and nurturing the anti-inflammatory phenotype of homing leukocytes (44). SIGIRR-deficient mice demonstrate defective intestinal homeostasis, and these defects are associated with the microbiota and hyper-expression of inflammatory mediators. Notably, these defects also render the azoxymethane (AOM)-dextran sodium sulfate (DSS)-treated SIGIRR−/<sup>−</sup> mice prone to colitis and colitisassociated CRC. Interestingly, the rescue of SIGIRR expression in the IECs of SIGIRR−/<sup>−</sup> mice restored the immune tolerance and abolished the risk of tumor development in these mice (45).

Although elevated TLR activity disrupts the recognition of intestinal microbes by TLR2 and TLR4, TLR signaling is necessary for maintaining homeostasis and regulation of tissue repair in IECs. MyD88-deficient mice, which hamper signaling through IL-1 family members including TLRs, possess profound abnormalities in the mucosa with higher proliferation rates in the crypts (46). Cumulatively, this leads to defects in repair of the intestinal barrier following injury, and increased risk of colitis and CRC (46, 47). Moreover, mice in which normal flora is disrupted by antibiotics display a similar phenotype to mice lacking MyD88, as well as decreased expression of factors [i.e., tumor necrosis factor (TNF), CXC-chemokine ligand 1, IL-6, and heat shock proteins] required for normal intestinal homeostasis (48).

#### **RELATIONSHIP BETWEEN INFLAMMATION AND CRC**

The direct link between intestinal inflammation and CRC prognosis is well-established and is also supported by numerous genetic,

**FIGURE 1 |TheTLR4 signaling pathway**. TLR4 is activated by LPS, whereas CD14 and MD2 act as accessory proteins for LPS/TLR4 binding. Upon ligand binding, TLR4 dimerizes, and recruits downstream adaptor molecules such as MyD88/MAL and TRIF/TRAM to mount an inflammatory response. The activated MyD88/MAL then activates IRAK4, TRAF6, TAK1, and IKK complexes, while TRIF/TRAM signals through RIP1 to TRAF6/TAK1 and IKK. After this, both these pathways converge at NF-κB. The cytoplasmic NF-κB

complex is maintained in the inactive state by IκB, which is in turn degraded by proteasomes, resulting in the translocation of NF-κB into the nucleus. Besides activating NF-κB, TAK1 also phosphorylates MAPKs to further reinforce the inflammatory response. The TRIF/TRAM pathway not only activates NF-κB but also triggers IRF3 to mount an antiviral response. Cumulatively, all these signaling pathways assist in eradicating infection as well as play an important role in sustaining the normal physiological functions in IECs.

pharmacological, and epidemiological studies conducted during the last decade (49). Recent reports demonstrate the complex interplay between distinct immune cells, and also show that proinflammatory mediators influence almost all the steps of CRC progression. However, the mechanisms by which inflammation stimulates the development of cancer remain elusive and are expected to vary from colitis-associated CRC to other forms of CRC (25, 50). The relationship between inflammatory responses caused by multiple factors such as the microbiota, IBD, and CRC has been demonstrated by comparative experiments conducted in wild type and *Il10*−/<sup>−</sup> mice. When treated with AOM, *Il10-/-* mice were found to show an increased risk of colon tumor development, spontaneous colitis, and CRC, while AOM-WT mice were devoid of colitis and rarely progressed to adenomas. In addition,mice with *Bacteroides vulgatus* or dual knockout mice (*Il10-* and *MyD88* deficient mice) treated with AOM showed reduced transcription of *Il12p40* and *TNF-*α and remained tumor-free (51).

TLR-induced inflammation is a well-established phenomenon and is perpetuated by several cytokines, ILs, and TNF-α, all of which are known to substantially regulate immune cells and inflammatory responses against cancer (48, 52). Among these, TNF-α is of particular importance and is now recognized as a pro- as well as anti-tumorigenic protein (53). The activation of the TLR4 signaling pathway induces TNF-α and NF-κB, leading to the promotion of CRC (17,54–56); TNF-α knockout mice treated with AOM/DSS show significantly less tumor formation, representing the pro-tumorigenic role of TNF-α (57). Immunohistochemistry analyses of mononuclear cells in the lamina propria and colons of patients with advanced stage CRC demonstrate the expression of TNF-α (57). TNF-α also promotes the activation of NF-κB, which reinforces inflammation by inducing cyclooxygenase-2 (COX-2), IL-6, IL-8, and TNF-α to favor tumorigenesis (55, 58, 59). However, inflammation alone is not sufficient for colon cancer and the contribution of other risk factors is equally essential to the pathogenesis of this complex disease.

#### **CONTRIBUTION OF TLR4 TO CRC DEVELOPMENT**

Although IECs are in close proximity to LPS, they do not mount an immune response on the commensal bacteria under normal circumstances. However, in the diseased state, disruption of the coexistence between IECs, and bacteria leads to an inflammatory response. This raises an important question: when and how much inflammation should have to be raised in order to equilibrate the bacterial threat (**Figure 2**). Numerous studies have been conducted to address this dilemma (60–64). For instance, IFN-α and IFN-γ are known to increase the LPS response in IECs, which is directly linked to the expression of TLR4 and MD2 (63,65). Moreover, continuous LPS stimulation culminates in reduced TLR4 expression and increased expression of inhibitory proteins (62). However, a conflicting report demonstrated that long-term LPS exposure does not alter TLR4 expression (66). Moreover, hypoxia and numerous endotoxins are known to be prevalent in the inflamed intestinal lining, possibly causing induced TLR4 expression (60). Hung et al. observed an increase in the TLR4 expression from the mucosa of CRC patients of different ages and sexes as well as from a variety of CRC cell lines (HT29, SW480, and KM20) (67). In addition, Maria and colleagues showed that TLR4 expression is required

proximity to and may be stimulated by commensal bacteria. Therefore, it is extremely necessary to regulate their functions. Under normal conditions, homeostasis between bacterial induction and TLR activation is maintained to ensure a disease-free status. On the other hand, if TLRs are inappropriately activated or if they mount an exaggerated immune response to a low level stimulus, they may culminate in bacterial infection and inflammatory disease/cancer, respectively.

for dysplasia and polyp formation. This finding is consistent with results of experiments performed in TLR4 gene knockout mice (56, 68). Collectively, these data present a clear association between TLR4 and CRC development.

In CRC, elevated TLR4 expression is observed in all tumor components such as the epithelial, endothelial, and stromal layers (69). However, the level of this expression varies depending on the type of cancer. Although all TLRs are expressed at the minimal basal mRNA level, IECs can upregulate the TLR expression, based on the inflammatory signals or other stimuli (70). An alternate study demonstrated a low level of TLR4/MD2 expression in normal human colonic epithelial cells and the lamina propria, which is consistent with the level of TLR4/MD2 expression detected in various epithelial cell lines (71). These studies establish the fact that any alteration in the prevalent inflammatory conditions or the population of luminal bacteria may influence the strength and nature of TLR signaling, paving the way for initiation of inflammatory responses in IECs. Besides this negative role, studies in TLR4- and TLR9-deficient mice have shown that TLR signaling in IECs is essential for protecting the host from inflammation-related damage and for homeostasis (46, 72).

#### **TLR4 CROSSTALK IN CRC PROGRESSION**

TLR4 is overexpressed in the liver metastasis of CRC (73). In response to LPS binding, over-stimulation of the TLR4/MD2 complex enhances the phosphorylation of protein kinase B (also known as AKT), which in turn activates the function of β1 integrin. This complex interplay between multiple pathways promotes the adhesiveness and metastatic behavior of CRC (74). The enhanced AKT phosphorylation can be blocked by eritoran (a TLR4 antagonist), PI 103 [a phosphatidylinositide 3-kinases (PI3K) inhibitor], or anti-β1 integrin antibodies that are known to ameliorate CRC and its metastatic behavior (75–77), indicating that the PI3K/AKT signaling pathway is induced by TLR4 in response to LPS binding and plays a central role in the growth and progression of CRC. Furthermore, LPS is known to induce the expression of the urokinase plasminogen activator (uPA) system through TLR4 and NF-κB in human colorectal cell lines. During tumor progression, vital extracellular matrix (ECM) interactions occur, in which uPA and the expression and activity of its receptor facilitate the growth and metastasis of CRC (78). Conversely, inhibition of TLR4, NF-κB, or the uPA system can attenuate CRC progression. Although NF-κB is known to impair apoptosis in tumor cells (55, 79, 80), NF-κB activation through TNFR signaling also protects cells from apoptosis. Studies performed in the Saos-2 cell line reveal that p53-induced cell death is dependent on NF-κB, and the ablation of NF-κB leads to the abrogation of p53-dependent cell death (81). Thus, the TNF-α/NF-κB interaction plays a vital role in CRC and IBDrelated diseases and manipulation of this interaction may improve the treatment of CRC.

TLR4 is overexpressed during inflammation-associated colorectal neoplasia in humans and mice. Similarly, mice lacking TLR4 are largely protected from colon carcinogenesis (56). A dissection of this mechanism reveals that TLR4 triggers elevated production of prostaglandin E2, increases Cox-2 induction, and influences epidermal growth factor receptor signaling (EGFR) in chronic colitis. TLR4 can thus manipulate numerous pathways and cause further deterioration of the neoplastic situation. A recent comparative immunohistochemistry analysis between normal mucosa and adenomas showed that TLR4 and MD2 are overexpressed in 20 and 23% of the adenomas, respectively (82), further substantiating the involvement of the TLR4 pathway in CRC. Furthermore, mutations in the *APC* gene cause pre-disposition to CRC. A correlation between the TLR/MyD88 signaling pathway and *APC* mutations was recently proposed (82, 83) since MyD88 signaling was found to facilitate the growth of intestinal polyps while the ablation of MyD88 restricted polyp growth in *Apc*min/+/*Myd88*−/<sup>−</sup> mice, but not in *Apc*min/<sup>+</sup> mice (83, 84). In addition, MyD88 induces ERK to block the degradation of the oncoprotein c-Myc, and such cells with continued activation of c-Myc are prone to neoplastic transformation (85).

Similarly, c-Myc is also important for APC-mediated tumorigenesis (86), since knocking out c-Myc in IECs of *Apc*min/<sup>+</sup> mice impedes tumor growth (84). Furthermore, reduced expression of c-Myc has been reported in *Apc*min/+/*Myd88*-/- of both normal and tumor mice (84, 87). Treatment of *Apc*min/<sup>+</sup> mice with PD03259012, an inhibitor of MEK1/2, which is the kinase directly upstream of ERK, also inhibits tumor growth. These data indicate that a complex interplay of protein signaling brings about tumor proliferation in the IECs of various transgenic mouse models. Moreover, heritable changes in the *APC* gene frequently lead to familial adenomatous polyposis (FAP). FAP is the most dominant inherited syndrome of CRC (88, 89) and *Apc*min/<sup>+</sup> mice show increased propensity for the development of adenomatous polyps after the loss of the wild type *APC* allele (88). Up to 80% of sporadic CRCs are known to be initiated by DNA damage of the genes involved in the APC signaling pathway (87).

#### **CORRELATION BETWEEN CRC DEVELOPMENT AND INHERITED GENETIC VARIATIONS OF TLR4**

The human *TLR4* gene is located on the long (q) arm of chromosome 9 at position 33.1, and contain four exons. The dominant expression of TLR4 has been observed in lymphocytes, monocytes, leukocytes, and splenocytes (90). Besides CRC, many human pathologies and carcinomas are associated with the polymorphisms of TLR4 (91–93). The *TLR4* gene contains two single-nucleotide polymorphisms (SNPs), namely, Asp299Gly and Thr399Ile that are significantly important in tumor development (94, 95). Both these SNPs are located in the coding sequence for the TLR4 ectodomain and mediate an amino acid substitution. These Asp299Gly and Thr399Ile SNPs in *TLR4* are known to attenuate cytokine expression, leading to an increased propensity for the development of gastric cancer and CRC (94, 96–99). The detection of these two SNPs was carried out using allele-specific polymerase chain reaction and the primer extension method (SNaPshot) for gastric cancer and CRC, respectively. For gastric cancer, only Thr399Ile showed a significant correlation, while both the SNPs were significantly correlated to CRC (94, 100, 101). In addition, the association of the TLR3 (rs3775291) polymorphism and IL-10 promoter variation (rs1800872) to CRC pathogenesis was evaluated in a large cohort of German CRC patients. This study found that the IL-10 promoter variant is significantly associated with an increased risk of lymph node metastasis (for carriers of the TT genotype). Interestingly, a *TLR3* gene polymorphism was found to correlate with patient survival, and the TT genotype was responsible for increased mortality. This TLR3 variation was limited only to stage II patients who were devoid of adjuvant therapy (102, 103).

The LPS-sensing complex is comprises TLR4, MD2, LPS binding protein, and CD14. A positive link between CD14–260 polymorphisms and the occurrence of CRC in the Chinese Han population was demonstrated (104, 105), in which the CD14 polymorphism C/C, but not C/T, was significantly correlated to CRC; no correlation between TLR4 Asp299Gly and CRC was found. However, it is possible that the polymorphism in TLR4 was associated with the population under study (106). A multi-racial study (22 Malays, 20 Chinese, and 18 Indians) conducted in Malaysia showed that there is no correlation between TLR4 polymorphisms (Asp299Gly; Thr399Ile) and the risk of CRC (107). However, a study on Russian population revealed that IL1B\_1473G/C and TLR4\_896A/G SNPs are involved in rectal cancer development (108). A conflicting report validated the potential link between TLR4 polymorphisms (Asp299Gly and Thr399Ile) and the digestive tract cancer and CRC (101). This study retrieved and analyzed extensive data from various databases and concluded that Asp299Gly is significantly correlated with an increased risk of gastric cancer, while there was no correlation between this polymorphism and digestive tract cancer and CRC. Moreover, it was also observed that the T allele of Thr399Ile does not influence digestive tract, gastric, or CRC. It is evident that additional studies are necessary to support these findings.

#### **EPIGENETIC REGULATION OF TLR4 IN CRC**

Intestinal epithelial cells are stimulated by the commensal bacteria in the intestinal lumen with the help of TLRs for the maintenance of homeostasis. This stimulation from the commensal bacteria is finite, should not trigger an excessive inflammatory response, and is known to influence epigenetic modification in the host cells (109). These epigenetic modifications involve DNA methylation and histone deacetylation that suppress and promote the transcription process, respectively, and in turn regulate gene expression (110).

The *TLR4* gene is methylated in the 5<sup>0</sup> region; also, the degree of methylation in epithelial cells is higher than that in the splenic cells, caused by the interaction of the commensal bacterial with IECs. Takahashi and colleagues showed that commensal bacteria modulate the epigenetic regulation in IECs by DNA methylation of *TLR4* (111). In their study, the authors compared the methylation levels in the IECs from the small and large intestine obtained from conventional (CV) mice with commensal bacteria and germ-free (GF) mice without commensal bacteria. The methylation level of CpG motifs in the 5<sup>0</sup> region of *TLR4* from the large intestine was lower in the GF mice compared with CV mice, while in the small intestine, the methylation levels remained unchanged between the GF and CV mice. The frequency of methylation is also found to depend on the MyD88 adaptor molecule. Results from *in vivo* experiments show that the frequency of CpG methylation is less in the GF mice (MyD88 knockout mice) compared to CV mice (111).

Environmental factors also play a crucial role in regulating epigenetic modifications. In the presence of factors such as myriad food habits and increasing pollution, intestinal commensal bacteria produce short-chain fatty acids known as butyrates that inhibit histone deacetylation (112, 113). Besides *TLR4*, *MD2* can also be downregulated to attenuate the LPS response. IECs are known to poorly express *MD2*, which directly correlates to DNA hypermethylation (114). In IBD, IECs exhibit elevated expression of *MD2* and *TLR4* mRNA, while in normal cells; *TLR4/MD2* transcription is reduced due to DNA methylation. The deacetylation and blocking of methylation enables cells to express higher amounts of *TLR4* and *MD2* mRNA. This study demonstrates how epigenetic regulation of *TLR4* and *MD2* prevents dysregulation of inflammation in IECs and thus provides a novel approach to target CRC.

#### **THERAPEUTIC TARGETING OF TLR4**

Synthetic TLR4 ligands are potential targets for therapeutic applications for cancer, allergies, and viral infections (115). By virtue of their cell surface location, quick induction, and the ability to mount a wide array of inflammatory responses, TLRs are one of the most promising targets for therapeutics (91). The clinical trials of various TLR4 ligands are enlisted in **Table 1**.

TLR4 agonists have immune regulatory applications as adjuvants in vaccines and in the treatment of chronic viral infection and cancer therapy. LPS was the first microbial product identified as a potential TLR4 agonist and implemented for therapeutic applications (116). LPS is very toxic since it induces excessive inflammatory cytokines. However, low-dose LPS combined with non-steroidal anti-inflammatory drug ibuprofene was proved to be safe, with higher levels of TNF-α and IL-1 in all patients (117, 118). Marginal to encouraging results were observed when ibuprofene combined with *Salmonella abortus*


#### **Table 1 |TLR4 agonists in clinical trials**.

*equi* LPS for non-small cell lung carcinoma (NSCLC) and CRC patients, respectively (119). Currently, a few clinical trials are being conducted for oncological indications involving cell-based vaccination to treat Ewing sarcoma, neuroblastoma, and rhabdomyosarcoma patients (NCT00923351). Besides, peptide-pulsed dendritic cells (DCs) were combined with LPS to treat hematological malignancies (NCT00923910), and to treat melanoma patients (NCT01585350), LPS along with oil-based adjuvant and a peptide vaccine are being investigated (119). A less toxic TLR4 agonist, monophosphoryl lipid A (MPLA), is an immunity modulating agent that activates MyD88-independent pathway in TLR4 signaling, triggers the induction of IFN-γ, and regulation of CD80/86, which forms the crucial aspect of adjuvancy (27, 120). MPLA adjuvant plays a dual role in defending the host from pathogens by stimulating the innate immune system, and induces the long-term adaptive immune system (115, 121, 122). Food and Drug Administration has approved MPLA to use as a vaccine against HPV associated cervical cancer (Cervarix). It also enhances the inflammatory behavior of immune cells, which may be useful in a variety of cancers to overcome the cancerinduced immune suppression. However, this may not be helpful in case of CRC, where TLR should emphasize the tolerance of immune system, not the over-activation. Furthermore, it is established that TLRs can act as double-edge sword that may be exploited in pathologies-dependent circumstances to avoid the undesirable consequences (123). OM-174 is a triacyl lipid A analog that activates TLR4 and culminates in tumor growth regression by increasing the IFN-γ production (124, 125) This is well-tolerated at biological concentrations with strong antitumor effects (NCT01800812) (126). A new anti-cancer vaccine, BLP25 liposome vaccine (Stimuvax), can identify and destroy the cancer antigen MUC1, thereby inducing an immune response against cancer cells (127, 128). However, this could not significantly improve the NSCLC (128). Now, stimuvax is being investigated for the treatment of rectal and prostate cancers (NCT01507103 and NCT01496131). Group A *Streptococcus pyogenes* [in lyophilized form OK-432 (Picibanil)] is shown to stimulate TLR4, which is used to treat gastric, cervical, and oral cancers (119). This compound is currently being examined to treat pancreatic cancer patients in pre-operative settings using intra tumoral injection of DCs (NCT00795977), combined with chemotherapy (cyclophosphamide + docetaxel) for head and neck squamous cell carcinoma (HNSCC) patients (NCT01149902), and via intracystic injection at cystic malformation (NCT01699347). Currently, most of the TLR4 antagonists are being evaluated against cancer-unrelated symptoms.

#### **CONCLUSION AND FUTURE PERSPECTIVES**

In this review, we highlight the correlation between CRC and TLRs, in particular, TLR4. We also propose that a beneficial link exists between commensal bacteria and TLRs in order to maintain intestinal homeostasis. In IECs, TLRs are involved in epithelial cell proliferation, IgA production, regulating the permeability of the intestinal barrier, antimicrobial peptide expression, and defense against invading pathogens. Over-stimulation of TLRs in response to minor signals (due to dysregulation) may result in colitis and CRC. Several studies suggested the relationship of TLR4 signaling with CRC, therefore therapeutic benefit can be achieved by targeting TLR4. However, the development of CRC is highly complex. Experimental studies supported that the gut microbiota contributed to CRC. The studies involving human subjects and considering their microbiota composition revealed the vivid differences in microbial density and population. Therefore, modulating the microbial population, usage of probiotics to favor the growth of certain bacteria, and delineating the interaction of microbiota with the epithelial cells can potentially be used to limit the CRC development.

Furthermore, inflammation is central to the development of cancer, and there are few clinical trials being conducted for antiinflammatory drugs, but by combining molecular approaches with CV therapies, i.e., chemo- or radiotherapy, anti-inflammatory drugs would increase the efficacy to treat CRC. Additionally, targeting the downstream molecules in TLR4 pathway involved in CRC is also expected to have a tremendous impact on CRC therapeutics. Moreover, differential expression of TLRs leads to tumor development, in which the contribution of TLR4 is considerably higher than in the other TLRs. We hope that extensive studies involving the TLR4 pathway will eventually provide therapeutic targets to treat CRC. Recently developed techniques may also prove helpful in the analyses of differential expression levels of TLRs, their mutations, and epigenetic modifications. These analyses would further aid in the design and development of novel therapeutic approaches for CRC treatment.

#### **ACKNOWLEDGMENTS**

This work was supported by the Mid-Career Researcher Program through the National Research Foundation of Korea, funded by the Ministry of Education, Science, and Technology (2012R1A2A2A02016803), and partly supported by a grant from the Priority Research Centers Program (NRF 2012-0006687).

#### **REFERENCES**


and colorectal cancer patients with advanced stage and lymph node metastasis. *Tumour Biol* (2013) **35**(1):545–51. doi:10.1007/s13277-013-1075-6


metabolites formed in the colon. *J Nutr Biochem* (2008) **19**(9):587–93. doi:10.1016/j.jnutbio.2007.08.002


**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.

*Received: 13 April 2014; paper pending published: 30 April 2014; accepted: 01 July 2014; published online: 15 July 2014.*

*Citation: Yesudhas D, Gosu V, Anwar MA and Choi S (2014) Multiple roles of toll-like receptor 4 in colorectal cancer. Front. Immunol. 5:334. doi: 10.3389/fimmu.2014.00334 This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Yesudhas, Gosu, Anwar and Choi. 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.*

#### **Kimberly A. Luddy <sup>1</sup> , Mark Robertson-Tessi <sup>2</sup> , Narges K. Tafreshi <sup>1</sup> , Hatem Soliman<sup>3</sup> and David L. Morse<sup>1</sup>\***

<sup>1</sup> Department of Cancer Imaging and Metabolism, Imaging and Technology Center of Excellence, H. Lee Moffitt Cancer Center, Tampa, FL, USA

<sup>2</sup> Department of Integrated Mathematical Oncology, H. Lee Moffitt Cancer Center, Tampa, FL, USA

<sup>3</sup> Don and Erika Wallace Comprehensive Breast Program, Center for Women's Oncology, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Alessio Paone, Sapienza University of Rome, Italy Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **\*Correspondence:**

David L. Morse, Department of Cancer Imaging and Metabolism, Imaging and Technology Center of Excellence, H. Lee Moffitt Cancer Center, Tampa, FL, USA e-mail: David.Morse@moffitt.org

Toll-like receptors (TLRs) are expressed by immune cells, intestinal epithelium, and tumor cells. In the homeostatic setting, they help to regulate control over invading pathogens and maintain the epithelial lining of the large and small intestines. Aberrant expression of certainTLRs by tumor cells can induce growth inhibition while others contribute to tumorigenesis and progression. Activation of these TLRs can induce inflammation, tumor cell proliferation, immune evasion, local invasion, and distant metastasis.TheseTLR-influenced behaviors have similarities with properties observed in leukocytes, suggesting that tumors may be hijacking immune programs to become more aggressive. The concept of epithelial to leucocytic-transition (ELT) is proposed, akin to epithelial to mesenchymal transition, in which tumors develop the ability to activate leucocytic traits otherwise inaccessible to epithelial cells. Understanding the mechanisms of ELT could lead to novel therapeutic strategies for inhibiting tumor metastasis.

**Keywords: colorectal cancer, toll-like receptors, epithelial to leucocytic transition, ELT, metastasis, immune evasion, cell plasticity, inflammation**

#### **INTRODUCTION**

Toll-like receptors (TLRs) are a diverse family of pattern recognition receptors expressed by immune cells from both the innate and adaptive arms of the immune system (1–4). Hence, TLRs stimulate both innate and adaptive immune responses to invading pathogen-associated molecular patterns (PAMPs) as well as danger-associated molecular patterns (DAMPs) from damaged epithelial cells. Ligand specificity, downstream signaling, and subsequent immune stimulation vary greatly among the TLR subtypes (5, 6). The complex expression patterns of TLRs differ among cell types, physiological location, and are controlled by the microenvironment (7–9). In the healthy gastrointestinal (GI) tract, TLRs are expressed by intestinal epithelial cells (IEC) and play a role in immune modulation and tissue homeostasis by stimulating the immune response to bacterial pathogens, attenuating the immune response against favorable microbes, sensing breakdown of the protective intestinal barriers, and triggering proliferative signaling (6, 10, 11). The lumen of the gut is subjected to continual interactions with the microbiome and would be in a constant state of inflammation were it not for the controlled expression and normal function of these TLRs (2, 6).

Inflammation has been implicated as an underlying factor in tumorigenesis and cancer progression, leading transformed cells to develop the "hallmarks of cancer" (12, 13). Some of the same TLRs (TLR 2–4) that normally regulate inflammation in the gut are also found to be aberrantly expressed in colorectal cancers (CRCs) (14, 15). Overexpression of TLR4 in CRC is associated with poor survival (16). Deletion of TLR4, its signaling partner Myd88, or absence of its ligand LPS in the colon can lead to increased or reduced inflammation, depending on the cancer subtype, which can then lead to either increased or decreased tumorigenesis and tumor progression (17, 18). It is notable that the role of TLRs in CRC reflects a "natural history" of selection events that lead from normal TLR function in unaffected colon tissue and throughout all stages of CRC progression. More specifically, the normal role of TLR immune modulation in the gut involves the controlled release of cytokines and danger signals that stimulate the immune response to bacterial pathogens and attenuate the immune response against favorable microbes (6, 10). Microbial imbalance and/or dysregulation of these responses leads to chronic inflammation of the bowel (15). Chronic inflammation is correlated with initiation of cancer development and in the progression of cancer into more aggressive forms of malignancies (2). Aberrant TLR signaling and the resulting cytokine imbalance leads to increased epithelial proliferation and decreased cell death (19). Additionally, an active immune environment creates selection pressures for initiating cancer cells resulting in the evolution of an immuneevasive tumor phenotype (14). Furthermore, TLR dysregulation is implicated in cancer invasion and metastasis (2, 19–21). Understanding the role of TLRs in the natural evolution of metastatic disease is crucial for developing new therapies and optimizing current treatments.

#### **ROLE OF TLRs IN INFLAMMATION-MEDIATED TUMORIGENESIS**

The intestines house approximately 70% of the body's immune cells under normal conditions (10). Signaling between these immune cells, commensal bacteria, and IECs is critical for normal digestion and protection against invading pathogens (6). TLRs are key modulators of the immune system of the GI tract. In order to maintain homeostasis and suppress immune responses to commensal bacteria (11), TLR expression and signaling are tightly controlled in this environment (6, 20). However, these controls are disrupted in diseases such as Crohn's disease and ulcerative colitis, resulting in chronic inflammation (11, 15). Inflammation is linked to cancer through two pathways: extrinsic inflammation induced by non-transformed cells (e.g., invading pathogens or autoimmune disease), and intrinsic inflammation induced by transformed cells (22). In CRC, TLRs are involved in both. Autoimmune diseases cause chronic, smoldering levels of inflammation that predispose individuals into developing CRC (22). Once initiated, tumors can intrinsically activate inflammation through TLR binding by cancer-related DAMPs. Intrinsically and extrinsically induced TLR activation results in tumor-promoting inflammation through NF-κB signaling, leading to expression of the inflammatory cytokines IL-1β, TNFα, and IL-6 (17). This aberrant expression by tumor cells in early carcinogenesis can recruit tumor-promoting immune cells, leading to inflammation and protection from cytotoxic immune cells. Additional data from Kim et al. links mutations in p53 and PTEN to SOCS-mediated activation of IL-6 signaling, leading to intrinsic inflammation (23). Since p53 mutation is a very common event in the natural history of CRC, this is likely a major mechanism of tumor-induced inflammation. Additionally, inflammation can drive genetic and epigenetic changes in cells as well as possible alterations in lineage differentiation programs leading to increased plasticity. This process is also thought to involve NF-κB signaling; however, further studies are needed (22, 24).

#### **ROLE OF TLRs IN INFLAMMATION-MEDIATED PROLIFERATION AND SURVIVAL**

Inflammatory pathways are tightly linked to aberrant proliferation and resistance to cell death, which are key cancer hallmarks that can be mediated through TLR activation (14). IECs are the barrier layer that protects the interstitial layers from the changing exterior environment of the GI tract. Infiltrating bacteria and the resulting immune response can cause tissue damage. To prepare for this, IECs utilize TLR4 signaling as an early warning system to initiate proliferation, maintain tissue integrity, and protect the interstitial compartments (6). Tumors can co-opt this system, allowing cells to proliferate unchecked (25).

Tumor growth is further fueled through an overabundance of growth factors (e.g., TGF-β, IL-8, CXCR4, and VEGF) (15), a decline in immune surveillance, and the evolution of mobile and invasive phenotypes. TLR expression on tumor cells stimulates the release of cytokines that recruit favorable immune cells further driving proliferation. Additionally, the release of cytokines and chemokines due to TLR signaling generates an autocrine loop that further stimulates tumor cell growth. The cumulative result is tumor control over its own environment.

#### **ROLE OF TLRs IN IMMUNE EVASION**

An active immune environment selects for the natural evolution of cancer cells with decreased immunogenic phenotypes. TLR expression in tumors can confer the advantages of both immune evasion and immunosuppression (26). Often pro-inflammatory signals reduce elements of the adaptive immune response. TLR signaling causes a shift in this response from anti-tumor to pro-tumor by affecting the balance toward inflammation and suppression of anti-tumor immunity. Direct TLR activation results in production of immunosuppressive cytokines IL-10 and TGF-β (14, 27), as well as increased expression of immune modulating surface markers PD-L1 and HLA-G (19,20,28). These secreted and surface proteins have a tolerizing effect on immune cells. TLR-activated IECs induce the transformation of dendritic cells (DC) into an antigen-specific CD103+ phenotype. These DC promote contactdependent antigen-specific regulatory T cells (Tregs) that express gut-homing integrins, which further attenuates the anti-tumor immune response (10). Each of these mechanisms are used in the healthy gut to avoid food hypersensitivity or auto-immune diseases. However, dysregulation through abnormal TLR expression can lead to malignant progression.

#### **ROLE OF TLRs IN INVASION AND METASTASIS**

The most dangerous effect of tumoral TLR signaling is the acquisition of invasive and metastatic tumor phenotypes (29). Ninety percent of patients who succumb to their disease have metastatic lesions (30). TLR expression in tumors is linked to increased grade and distant metastasis (2, 18, 21, 31). The ability of a tumor cell to detach from its epithelial neighbors, break through the basement membrane, and invade nearby tissues is, in part, the result of a long history of aberrant TLR signaling. In CRC, TLR-mediated alterations of the immune system components in the tumor microenvironment can change intracellular signaling (NF-κB), integrin expression (B1 integrin), and motility (29, 32). Activation of TLR4 by LPS *in vitro* and *in vivo* induces epithelial to mesenchymal transition (EMT) and invasive phenotypes in certain cell lines (29, 33).

Immune cells are educated by tumor-secreted factors and then actively migrate through the lymphatic vessels and secondary lymphoid organs. These tightly gated organs allow entry and passage to soluble antigens and select immune cell phenotypes, and yet lymph nodes are often the first site of metastasis (34). While it was once thought that tumors cells passively filter into draining lymph nodes, it has recently been shown that tumor cells require chemokine-mediated (CCR7 and CCR8) active transport through the subcapsular sinus epithelium (35,36). Furthermore, it has been shown that tumor-mediated lymphatic remodeling of peritumoral lymph vessels and draining lymph nodes facilitates metastasis (37– 40). TLRs may play a role in this metastatic process, since TLR activation leads to increased expression of CCR7 and CCR8 (41), which are key molecules expressed by leukocytes to access lymphatics (35, 42). This suggests that the tumor cells can harness existing leucocytic mechanisms to begin the metastatic cascade through the lymph nodes.

Lymphocytes typically traffic throughout the body to sites of inflammation, using chemokines, selectins, and integrins as homing signals (43). Many metastatic tumors have been shown to use the expression of these same molecules to colonize distal sites (44, 45). As an example, CXCR4 is a well-characterized bone marrow homing receptor expressed by T cells (46); research has found that both prostate cancers (47) and breast cancers (48) that metastasize to the bone commonly express CXCR4. CRC typically metastasizes to the liver or lung. Aberrant expression of CXCR3, CXCR4/CXCR7, and CCR6 are commonly found in liver and lung metastasis of colon cancer (49–55). Ligands for these receptors (CXCL19, SDF-1, and CCL20, respectively) are highly expressed in the liver and lungs of metastatic CRC patients (53, 56–58). Local inflammation in these organs induces ligand expression and preferential organ metastasis is determined by their expression (59, 60).

Alteration in integrin signaling is another metastatic mechanism induced by TLR signaling (26). Integrin signaling is used in healthy systems to aid immune cell trafficking (61). Aberrant expression of these integrins via TLR signaling allows circulating tumor cells to respond to the same trafficking mechanisms that an immune cell uses to migrate to distal sites (2, 32, 62, 63). Similar examples have been shown with integrins in colon cancer (64), breast cancer (65), and melanoma (66). These expressed surface markers are a natural part of the lymphocytic trafficking system, and their expression on tumor cells could be evidence that tumor cells use leucocytic trafficking mechanisms to metastasize.

#### **EPITHELIAL TO LEUCOCYTIC TRANSITION**

The co-opting of immune cell signaling and migration mechanisms by tumor cells is well documented, with many citing the plasticity of tumor cells and inappropriate gene expression as the underlying cause of treatment resistance and metastatic growth (13, 67–70). Pressures from cytotoxic immune cells, abundant inflammation, cytotoxic drugs, and targeted therapies push tumor cells into plastic states where they may begin to access programed mechanisms outside of their usual function (68). The survivors of these selection pressures are adaptive and dynamic cells, many of which express patterns of proteins found in other normal cell types (70, 71). These protein expression patterns have been used to define and detect EMT, for example. An increasing number of publications suggest that although EMT is important in locally invasive disease, it is not enough to allow tumor cells access to lymph nodes, lymphatic and vascular systems, as well as entrance and settlement into distant tissues (35, 69, 72, 73). Others hypothesize a myeloid lineage expression pattern gained from horizontal gene transfer and Lamarckian inheritance, tumor cell myeloid cell fusion, or a possible myeloid cell origin (69, 74–76). Here, we build on these observations and propose a new concept, the transition from epithelial phenotype to leucocytic phenotype.

Immune cells of myeloid and lymphoid origins house a diverse set of mechanisms that make them perfect trafficking cells. They can shift their metabolism easily, survive in low oxygenated areas, roll along the endothelium in the presence of high shear forces, read integrin codes, and facilitate tissue specific extravasation (77– 79). As described above, the aberrant expression of TLRs by CRC cells results in the acquisition of a number of tumor-promoting mechanisms. At the same time, these mechanisms are key properties of the immune system, as is TLR expression. In a broad

sense, immunosuppression, migration through tissue, intra- and extravasation through lymph and blood vessels, rapid proliferation, altered metabolism, and homing to specific tissues are key hallmarks of both cancer and the immune system.

Pathogenic EMT has its roots in normal embryogenesis. In cancer, this transition results in epithelial cells with a range of mesenchymal protein expression. These alterations increase motility and invasive capability of tumor cells, but do not necessarily explain immunoevasion, lymphatic access, and metastatic spread (35, 69, 72). We therefore propose the parallel concept of epithelial to leucocytic transition (ELT) as a framework, akin to EMT, with which to understand the metastatic properties of cancer cells. **Figure 1** illustrates the primary properties gained by tumor cells that undergo ELT. We consider ELT to be a partial transition in which epithelial cells retain their epithelial origin while at the same time acquiring a set of leucocytic traits. Tumor cells co-opt many mechanisms of the immune system for their own transport and these mechanisms are activated by proteins typically reserved for

**FIGURE 1 | Epithelial to leucocytic transition (ELT) is the acquisition of immune properties by tumor cells**. Epithelial tumor cells (purple box) can make transition to a mesenchymal phenotype (orange box), which enhances local motility and remodeling of the extracellular matrix (ECM). Tumor cells undergoing ELT (green box) can gain the ability to (1) evade the immune system at the primary tumor site, (2) access the lymphatic system, (3) circulate through the vasculature, home to favorable sites of metastasis, and extravasate into a metastatic niche, and (4) avoid destruction by the immune system at the site of metastasis.

the immune response. A leucocytic tumor cell expresses proteins that allow for regulation and co-opting of the immune system such as PD-L1, CD80/86, TLR, TGF-β, CCL4, and CCL5 (80) (**Figure 1**, properties 1 and 4). Additional leucocytic proteins (CXCR4,CCR7, CCR8) facilitate invasion and proliferation within lymph nodes (**Figure 1**, property 2) (35, 42, 81). Processes critical to survival in circulation, homing to tissue specific sites, and successful extravasation are mediated by E/P-selectins, L-selectin ligands, α4β1, ICAM-1, and VCAM-1 (61, 73, 82–86) (**Figure 1**, property 3). By harnessing mechanisms usually reserved for immune cells, tumor cells gain the ability to become more aggressive. In the case of TLRs, a cycle of overexpression and resulting inflammation promotes plasticity of the epithelial phenotype. This plasticity permits tumor cells to undergo ELT, accessing immune programs that facilitate invasion and metastasis of the cancer. ELT, as with other plastic states, is likely transient, making the evaluation of these phenotypes a significant challenge. TLR-mediated evolution of CRC may be a good model to study how ELT occurs, since TLRs are primarily seen in immune cells and the overexpression of TLRs appears to promote an immune-like phenotype in CRC.

Understanding the acquisition of the leucocytic phenotype could reveal key targets that would prevent CRC cells from accessing dangerous invasion and trafficking mechanisms through a plastic transition. Simply antagonizing TLRs and associated molecules may not be enough, since resistance is likely to develop. However, if the mechanisms of plasticity induced by TLRs are understood, new targets may be developed to inhibit ELT.

It is important to note that the key functional activities of immune cells, specifically the CD8 cytotoxic T-cell phenotype and the antibody producing activated B-cell phenotypes, have not yet been described in tumor cells. However, other cytotoxic mechanisms utilized by immune cells have been seen in normal and neoplastic epithelial cells. Tumor cell cannibalism, resembling phagocytosis, of neighboring apoptotic cells as well as infiltrating immune cells has been seen during times of metabolic stress (87). During mammary involution, epithelial-derived FAS plays a role in FASL-mediated cell death (88). Tumor cells can secrete FAS, TNFa, and TGFb, proteins capable of promoting and inhibiting epithelial cell death (89–91). Additionally, PD-L1 proteins on tumor cells result in T-cell anergy and apoptosis (92,93). Although, none of these represent the antigen-specific killing of the adaptive immune system, it is our opinion that further exploration is needed to determine how far epithelial cells can evolve to obtain immune-like processes and that cell killing can not yet be included or excluded from that hypothesis.

#### **CONCLUSION**

Originally, lymphatic dissemination into draining lymph nodes was considered a clear indicator of prognosis and was attributed to tumor chronology based on the correlation of tumor volumes and lymph node metastasis. However, later larger studies often showed conflicting results. Jatoi et al. (94, 95) and others attributed these differences to the tumor phenotype as opposed to a simple passage of time. This means that tumor phenotypes can exist on a continuum from slow growing with late lymph node metastasis to aggressive early disseminators much more capable of exiting the lymph node and establishment at distant sites (94, 95). While lymph node positivity is a useful tool for treatment decisions understanding the complexities of these aggressive phenotypes is key to halting the lymphatic dissemination of cancer.

Many host parameters contribute to natural progression of tumor metastasis and the extent of tumor cell plasticity is not yet fully appreciated. In an opinion article on tumor and immune cell plasticity, Holzel et al. (68) recognize the similarity between cancer cells and immune cells by linking inflammation and evolutionary pressures to the creation of plastic phenotypes. We think that this idea needs to be taken further to include a plastic transition to an immune-like phenotype, i.e., ELT, in the context tumor development, invasion, metastasis, and resistance to therapies. Specifically in CRC, the tumor cells acquire many hallmarks of the immune system, and this transition is intimately tied to aberrant TLR expression. By considering TLR expression in the context of ELT, the transition to a migratory immune-like and therefore metastatic phenotype might be better understood, and therefore, lead to better therapeutic strategies.

#### **REFERENCES**


prognosis. *World J Gastroenterol* (2012) **18**(11):1249–56. doi:10.3748/wjg.v18. i11.1249


**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.

*Received: 22 April 2014; accepted: 22 August 2014; published online: 20 October 2014. Citation: Luddy KA, Robertson-Tessi M, Tafreshi NK, Soliman H and Morse DL (2014) The role of toll-like receptors in colorectal cancer progression: evidence for epithelial to leucocytic transition. Front. Immunol. 5:429. doi: 10.3389/fimmu.2014.00429 This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Luddy, Robertson-Tessi, Tafreshi, Soliman and Morse. 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.*

#### **Shu Zhao1,2† ,Yifan Zhang<sup>1</sup>† , Qingyuan Zhang<sup>2</sup> , FenWang<sup>1</sup> and Dekai Zhang<sup>1</sup>\***

1 Institute of Biosciences and Technology, Texas A&M University Health Science Center, Houston, TX, USA <sup>2</sup> Department of Medical Oncology, Affiliated Tumor Hospital of Harbin Medical University, Harbin, China

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

F. Ronchese, Malaghan Institute of Medical Research, New Zealand Alessio Paone, Sapienza University of Rome, Italy Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Dekai Zhang, IBT-CIID, Texas A&M University Health Science Center, 2121 West Holcombe Blvd., Houston, TX 77030, USA

e-mail: dzhang@ibt.tamhsc.edu

†Shu Zhao and Yifan Zhang have contributed equally to this work. Prostate cancer is the second leading cause of cancer-related death in men after lung cancer. Immune responses clearly play a critical role in the tumorigenesis and in the efficacy of radiation therapy and chemotherapy in prostate cancer; however, the underlying molecular mechanisms are still poorly understood. Toll-like receptors (TLRs) are a well-known family of pattern recognition receptors that play a key role in host immune system. Recent studies demonstrate that there are links betweenTLRs and cancer; however, the function and biological importance ofTLRs in prostate cancer seems complex.To elucidate the role ofTLRs and innate immunity in prostate cancer might provide us with a better understanding of the molecular mechanisms of this disease. Moreover, utilizing the agonists or antagonists of TLRs might represent a promising new strategy against prostate cancer. In this review, we summarize recent advances on the studies of association between TLR signaling and prostate cancer, TLR polymorphisms and prostate cancer risk, and provide some insights about TLRs as potential targets for prostate cancer immunotherapy.

**Keywords: toll-like receptor,TLR signaling, prostate cancer, innate immunity, immunotherapy**

#### **INTRODUCTION**

Based on the latest cancer statistics, prostate cancer predictably ranks first among all the cancers in men and second in cancerrelated deaths in the United States in 2014 (1). Treatments against prostate cancer, including chemotherapy and radiotherapy, could improve survival; however, many patients will endure relapse and metastasis, which eventually leads to death. These treatments also destroy cancer cells and normal cells alike. Therefore, a more effective and less toxic therapy needs discovery. A promising strategy for dramatically preventing cancer development and improving cancer treatment might rely on immunotherapy. Immune evasion is a hallmark of cancer pathogenesis. Cancer cells escape from immune attack through a variety of mechanisms. A compromised immune system and chronic inflammation increase the incidence of cancer development. Inflammation has been proposed as the seventh hallmark of cancer (2) and an excellent review has elegantly summarized the role of inflammation in prostate cancer development and potential underlying mechanisms (3). Immunotherapy, which utilizes host immune system to fight cancer, has been recently highlighted with several advantages including specificity, less side effects, and less likely to develop resistance. It could be achieved in two ways: stimulating immune system to attack cancer cells or taking away the inhibitory machinery to the immune system in cancer. One potential approach to modulate immune system is targeting pattern recognition receptors (PRRs) in innate immune system, among which toll-like receptors are most well studied.

#### **TOLL-LIKE RECEPTOR: A WELL-KNOWN FAMILY OF PATTERN RECOGNITION RECEPTORS IN INNATE IMMUNITY**

Toll-like receptors are a family of transmembrane receptors that play a key role in the innate immunity. TLRs prevent invading pathogens by recognizing pathogen-associated molecular patterns (PAMPs), which are highly conserved components derived from bacteria, viruses, fungi, and parasites (4, 5). It can also recognize endogenous damage-associated molecular patterns (DAMPs) in different disorders and diseases such as cancer (4, 5). At present, 10 TLRs have been identified in human. TLR1s, TLR2, TLR4, TLR5, and TLR6 are expressed on cell surface; however, TLR3, TLR7, TLR8, and TLR9 are found exclusively within endosomes (**Figure 1**). Different TLRs exhibit specificity for ligand recognition. TLR2 recognizes bacterial lipoproteins, TLR3 recognizes double-stranded RNA/polyinosinic–polycytidylic acid [poly (I:C)], TLR4 recognizes lipopolysaccharides (LPS), TLR5 recognizes flagellin, TLR7 recognizes single-stranded RNA, and TLR9 recognizes CpG-containing DNA (CpG-ODN) (6–11). TLR10 is so far an orphan receptor and highly expressed in the human spleen (12) and B cells (13). Upon activation,TLRs transmit signals through one or more of four adaptor proteins: myeloid differentiation factor 88 (MyD88), TICAM1 (also known as TRIF), TIRAP (also known as MAL), and TICAM2 (also known as TRAM and TIRP). All TLRs (except for TLR3) and IL-1 receptor family members signal through MyD88. TLR3 signals through TRIF pathway; TLR4 signals through both the MyD88 and the TRIF pathways (4). Stimulation of TLRs leads to activation of NF-κB, MAPKs, Jun N-terminal kinases (JNKs), p38, and ERKs, as well as interferon regulatory factor (IRF3, IRF5, and IRF7) signaling pathways, which results in the production of inflammatory cytokines (14). Activation of TLRs in antigen-presenting cells (APC) also triggers adaptive immunity. TLRs have also been shown to regulate cell death and increase expression of the anti-apoptotic proteins Bcl-2-related protein A1 (BCL2A1), inhibitor of apoptosis 1 (cIAP1), cIAP2, XIAP, and Bcl-2 family members (15).

#### **TLR EXPRESSION AND FUNCTION IN PROSTATE CANCER**

Toll-like receptors are predominantly expressed in innate immune cells such as dendritic cells, macrophages, and natural killing

(NK) cells. Activation of TLRs in these cells leads to the activation of innate immunity and results in the production of pro-inflammatory cytokines, chemokines, as well as adhesion molecules, and then facilitates the activation of adaptive immunity (16). Intriguingly, growing evidence has demonstrated that TLRs are also expressed in tumor cells. TLR activation in tumor cells and its activation in tumor microenvironment such as in typical innate immune cells lead to a complex scenario (**Figure 2**); therefore, the activation of TLRs might play a "double-edged sword" role in the influence of tumor progression.

In most cases, it is difficult to figure out a specific pathogen to activate TLR signaling in prostate cancer. An endogenous TLR ligand, DAMPs released from damaged and/or necrotic tissues, might play a pivotal role. In term of endogenous TLR ligands in cancer, HMGB1 can activate TLR2 and TLR4 (17), and versican acts as a TLR2 agonist (18). Peroxiredoxin 1 (Prx1) appears to be an agonist of TLR4 in prostate cancer development (19). Perhaps, there are more endogenous TLR ligands that need to be further identified and verified.

The activation of some TLRs might prevent the tumor growth of prostate cancer (**Figure 2**). It has been shown that TLR3 is expressed in prostate cancer cells (20–25). TLR3 mRNA is detected in three prostate cancer cells lines including LNCaP, PC3, and DU-145. TLR3 mRNA level was clearly enhanced in prostate cancer cells by stimulating with poly (I:C), which suggests a functional role of TLR3 in prostate cancer (20). TLR3

protein was also expressed at similar levels in LNCaP and DU-145 cells, with a slightly lower expression in PC3 cells. Treatment with poly (I:C) rapidly triggered NF-κB-dependent expression of inflammatory molecules. Condition medium from poly (I:C)-treated LNCap and DU145 cells recruited leukocyte subpopulation, indicating that TLR3 activation might influence early immune responses in tumor microenvironment (25). Stimulation with poly (I:C) strongly suppressed prostate tumor growth *in vivo*, perhaps due to increased infiltration of T lymphocytes and NK cells in a type I IFN-dependent manner (24). In human prostate cancer patients, 85 in 112 prostate carcinomas samples showed positive expression of TLR3. High TLR3 expression level was significantly associated with high probability of the recurrence of prostate cancer (23). Paone and colleagues found that TLR3 could regulate the process of angiogenesis and apoptosis in prostate cancer cells through hypoxia-inducible factor 1α (HIF-1α) and PKC-dependent mechanism (21, 22). TLR5 is expressed in LNCap and DU-145 by which stimulation triggers the production of chemokines that recruit immune cells, including NK cells and cytotoxic CD8 cells, which most likely contribute to tumor inhibition (25).

The activation of other TLRs might play a different role in the tumor growth of prostate cancer (**Figure 2**). The expression of TLR4 in prostate cancer has been demonstrated in several animal models. Studies revealed a constitutive expression of TLR4 in the epithelial cells of rat ventral prostate as well as in a rat

to either Th1 and T cytotoxic responses or Th2 and Treg responses. The activations of TLR2, 4, and 9 in prostate cancer cells appear to promoter tumor growth, but the activation of TLR3, 4, 5, and 7 might inhibit prostate cancer.

adenocarcinoma cell line and in prostate primary culture cells (26, 27). TLR4 is also expressed in DU-145, PC3, and normal prostate gland in both stroma and epithelium (28, 29). In addition, TLR4 has also been shown to be expressed in clinical samples of prostate cancer. Initially, TLR9 expression was thought to be restricted to immune cells, but recent studies have showed that a variety of tumor cell types including prostate cancer also express functional TLR9 (23, 30, 31). A clinical study demonstrated that TLR9 is expressed in prostate cancer specimens (23). Joanna et al. found that TLR9 is expressed in human prostate cancer cell lines LnCaP, C4-2B, Du-145, PC3, and in clinical samples of prostate cancer through immunohistochemistry and western blotting, but not in MDA Pca2b and stromal cells of the clinical adenocarcinoma samples (32). TLR9 expression was also statistically significantly increased in prostate cancer epithelium and stroma, compared with the same cellular compartments in benign hyperplasia, especially in the most poorly differentiated forms (30).

The function and biological importance of TLRs in prostate cancer seems complex (**Figure 2**). Perhaps the distinct and unidentified TLR signaling pathways are activated in cancer cells or innate immune cells during tumor progression; or, the first activation of TLR in cancer cells or innate immune cells markedly affect the subsequently activation and induced effectors. The mystery will be further investigated and will affect the potential of TLR agonists or antagonists as anti-tumor therapeutic agents.

#### **MicroRNA REGULATE TLRs IN PROSTATE CANCER**

MicroRNAs (miRNAs) are a class of small non-coding RNAs (~22 nt in length), which negatively regulate gene expression at the post-transcriptional level (33). By binding to target sequences within the 3<sup>0</sup> UTR of mRNA, miRNAs induce gene silencing by either inhibiting translation or leading to degradation of mRNA. MiRNA alterations are shown to be involved in both initiation and progression of human cancer (34–39). Deregulation of miR-NAs is implicated as an important mechanism in tumorigenesis and several miRNAs have been proposed as oncogenes or tumor suppressors (40–42).

MicroRNAs are emerging as a fundamental mechanism in the regulation of TLR signaling (43–47). Recent works have linked miRNAs and TLRs in prostate cancer. MiR-29a has been shown as a potential tumor suppressor miRNA to regulate TRAF-4 expression in metastatic prostate cancer (48). TLR3 activation by poly (I:C) induces upregulation of miRNAs including miR-29b, -29c, -148b, and -152, which target DNA methyltransferases and leads to reexpression of oncosuppressor RARβ in prostate cancer cells (49). TLRs activation facilitates either prostate cancer inhibition or progression. MiRNAs are likely to act as important regulators to control TLRs expression and signaling, thus contribute to prostate cancer development.

#### **TLR SIGNALING IN PROSTATE CANCER**

Toll-like receptor signaling pathway has been well defined in innate immune cells. TLR ligation recruits one or more adaptor proteins such as MyD88, TRIF, Mal, and TRAM though TIR domain interactions. Most TLRs except TLR3 go through a MyD88-dependent signaling pathway. MyD88 engagement activates IL-1 receptor associated kinase (IRAK), which interacts with tumor necrosis factor receptor associated factor 6 (TRAF6), resulting in the activation of MAPK and NF-κB signaling. TLR3 and TLR4 activate a MyD88-independent signaling pathway. TRIF is recruited upon stimulation and leads to the activation of NF-κB and type I IFN signaling.

Although TLR3 can be activated in prostate cancer cells, the molecular signaling pathway has not been fully elucidated. A recent study in human prostate cancer cells suggests that TLR3 signaling triggers apoptosis and growth arrest of LNCaP cells partially through inactivation of the PI3K/Akt pathway. CyclinD1, c-Myc, p53, and NOXA are indicated to play a role in poly (I:C)-treated LNCaP cells (20). In other studies, HIF-1α facilitates apoptosis through a PKC-dependent mechanism in poly (I:C)-treated prostate cancer cells. TLR3 activation by poly (I:C) activates JNK and p38 through PKC-α and triggers apoptosis in a caspase-8 dependent manner (21, 22). In LNCap cells, poly (I:C) treatment upregulates a pattern of chemokines, including CCL3, CCL4, CCL5, CCL8, CXCL9, and CXCL10, which could induce massive NK cell and CD8 T cell chemotaxis. Moreover, poly (I:C) induced the expression of inflammatory molecules such as IL-6 and IL-12, which are NF-κB signaling dependent (25). In TRAMP tumor model, poly (I:C) treatment recruits NK cells and T lymphocytes through a type I IFN dependent mechanism, resulting in suppression of tumor growth (24). TLR5 agonist flagellin can activate NF-κB signaling in LNCaP and DU145 cells, and lead to the production of pro-inflammatory molecules (25).

Stimulation of TLR4 in DU145 by LPS activates NF-κB signaling pathway, which leads to production of pro-inflammatory cytokines such as IL-6 and IL-1β through MyD88-dependent pathway (29). In addition, TLR4 activation increases expression of VEGF and TGF-β1 in PC3 cells, which promote tumor development (28). Also, knockdown of TLR4 using siRNA in PC3 cells reduces tumor cell migration and invasion (50). TLR9 stimulation by CpG-ODN plays an important role in prostate cancer invasion. This effect is mediated by activating NF-κB and upregulation of COX-2 (31). TLR9 expression in prostate cancer cells has similarly been found to enhance invasiveness via induction of MMP-13 *in vitro* (32). In both studies, CpG-ODN stimulation did not affect cellular proliferation, which suggests TLR9 signaling plays a role in cancer progression and metastasis.

These defined TLR signaling pathways seem difficult to help understand why the activation of some TLRs such as TLR3 inhibits tumor growth but the activation of other TLRs such as TLR2 promotes tumor growth (**Figure 2**). Some distinct TLR signal pathways must exist to determine the specific effectors in the different TLR activations leading opposite consequences.

#### **TLR GENE POLYMORPHISMS AND PROSTATE CANCER RISK**

Polymorphisms in TLR genes are reportedly related to susceptibility of a large spectrum of infectious and inflammatory diseases. Growing evidence suggest that chronic intra-prostatic inflammation contribute to prostate cancer progression. It was suggested that TLR gene polymorphisms might alter TLR signaling, thus affecting inflammation and prostate cancer risk. A number of studies have been done to investigate whether there is a connection between TLR gene polymorphisms and prostate cancer risk, and the results are controversial (51, 52).

Single nucleotide polymorphisms (SNPs) in TLR4 were reported to be associated with prostate cancer risk in several studies (53–58). Sequence variants in TLR gene cluster (TLR6-TLR1- TLR10) were also reported to be associated with prostate cancer risk (51, 52). However, controversial results were also obtained. Shui and colleagues investigated 10 SNPs in TLR4 and found no significant correlation between TLR4 genetic variation and prostate cancer risks (59). Chen et al. reported that sequence variants of gene cluster TLR6-TLR1-TLR10 were not associated with the risk of prostate cancer (60). A meta-analysis by Lindström et al. did not show clear correlation between TLR gene polymorphisms and prostate cancer risks.

The discrepancies among these results might be due to multiple factors including detection method, the race of population, and sample size. It is important to clarify this issue because it will determine not only whether the TLR polymorphisms can be used as a diagnosis/prognosis marker but also whether we can develop a novel strategy to treat prostate cancer by targeting TLRs and their signaling pathway. A more comprehensive study including a sufficient sample size should be performed to investigate the association between TLR gene polymorphisms and prostate cancer risk.

#### **TARGETING TLRs FOR PROSTATE CANCER IMMUNOTHERAPY**

The ability of TLRs to manipulate prostate cancer development has raised the interests in developing immunotherapy against prostate cancer with the TLR agonists or antagonists. Actually, three drugs targeting TLRs have been approved by FDA for use in cancer patients: the bacillus Calmette–Guérin (BCG), monophosphoryl lipid A (MPL), and imiquimod (61). BCG is prepared from an attenuated strain of *Mycobacterium bovis* and activates TLR2/4. BCG is used as a vaccine in prevention of tuberculosis, but also for treatment of *in situ* bladder carcinoma. Derived from LPS as a potent TLR4 agonist, MPL is an active component of Cervarix, which is used against cancer-causing human papillomavirus (HPV) (62, 63). Imiquimod, one of the most successful drugs targeting TLRs, is a synthetic imidazoquinoline that signals though TLR7 and is commonly used in the treatment of skin cancer such as basal cell carcinoma and Bowen's disease (64–66). Imiquimod induces the proinflammatory cytokines including IFNα, IL-6, and TNF-α (67). The activation of TLR7/8 leads to a Th1 response and an anti-tumor activity, which depends on IFNγ (68). In prostate cancer, to support this concept, Han et al., reported that Imiquimod can inhibit both human and mouse prostate cancer growth by inducing apoptosis (69, 70).

A number of preclinical and clinical studies are ongoing to investigate the immunotherapeutic potency utilizing TLRs against prostate cancer. TLR3 activation directly triggers apoptosis of human prostate cancer cells (21); therefore, TLR3 agonists have potential to be developed as anti-tumor therapeutic agents. Indeed, Ampligen, composed of poly (I:C) (a TLR3 agonist), has been shown to inhibit a variety of tumor growth in early clinical trials (71, 72). Hiltonol, a particular formulation of poly (I:C), is currently in Phase I/II clinical trial to evaluate its safety and efficacy (71). Meanwhile, a phase 2 clinical study (NCT00514072) utilizing a BCG vaccine to treat prostate cancer is ongoing. A multi-peptide, dual-adjuvant telomerase vaccine (GX301) in which Imiquimod is an active component showed less toxic and highly immunogenic in prostate cancer patients, but requires future studies to determine its clinical efficacy (73). Furthermore, TLR4 stimulation by LPS is shown to contribute to chemoresistance to docetaxel in prostate cancer cells (74).

#### **CONCLUDING REMARKS**

Toll-like receptors play a critical role in innate immunity. TLRs are expressed not only in innate immune cells, but also in nonimmune cells including cancer cells. Functional expression of TLRs has been linked to prostate cancer development. TLRs may serve as a double-edged sword in prostate cancer tumorigenesis by promoting malignant transformation of epithelial cells and tumor growth, or on the contrary, inducing apoptosis, and inhibiting tumor progression. The consequences might be dependent on complex signaling networks triggered by TLRs activation and tumor microenvironment. Genetic variations and polymorphisms of TLRs have been associated with prostate cancer; however, the results are inconclusive and need further validation (75, 76). The ability of boosting immune responses but with less serious side effect makes TLRs a good target to treat cancers. A wave of preclinical and clinical studies showed the potential of developing treatment targeting TLRs against prostate cancer. Based on these researches, one of the most probable approaches is to use agents targeting TLRs as adjuvants along with other treatments (67, 68, 71, 77, 78). Above all, elucidation of the mechanisms of cancer cell

TLR signaling and crosstalk with other signaling pathways as well as the mechanisms of cancer progression will definitely provide a promising novel strategy for cancer treatment.

#### **ACKNOWLEDGMENTS**

We thank Navella Richard and Mar'Kiffany Lane for critical reading of the manuscript. This work is supported partially by NIH grant R21CA176698 and TAMHSC development grant (to Dekai Zhang).

#### **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.

#### *Received: 28 March 2014; paper pending published: 08 April 2014; accepted: 09 July 2014; published online: 23 July 2014.*

*Citation: Zhao S, Zhang Y, Zhang Q, Wang F and Zhang D (2014) Toll-like receptors and prostate cancer. Front. Immunol. 5:352. doi: 10.3389/fimmu.2014.00352*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Zhao, Zhang , Zhang , Wang and Zhang . 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.*

## Toll-like receptors in ovarian cancer as targets for immunotherapies

#### **Maria Muccioli <sup>1</sup> and Fabian Benencia1,2\***

<sup>1</sup> Molecular and Cell Biology Program, Ohio University, Athens, OH, USA

<sup>2</sup> Department of Biomedical Sciences, Heritage College of Osteopathic Medicine, Ohio University, Athens, OH, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Stephan Gasser, National University of Singapore, Singapore Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Fabian Benencia, Heritage College of Osteopathic Medicine, Ohio University, 228 Irvine Hall, Athens, OH 45701, USA e-mail: benencia@ohio.edu

In the last decade, it has become apparent that toll-like receptor (TLR) signaling can play an important role in ovarian cancer (OC) progression. Interestingly, TLR activation in immune cells can help activate an anti-tumor response, while TLR signaling in tumor cells themselves is often associated with cancer-promoting inflammation. For example, it has been shown that TLR activation in dendritic cells can result in more effective antigen presentation toT cells, thereby favoring tumor eradication. However, aberrantTLR expression in OC cells is associated with more aggressive disease (likely due to recruitment of pro-tumoral leukocytes to the tumor site) and has also been implicated in resistance to mainstream chemotherapy. The delicate balance of TLR activation in the tumor microenvironment in different cell types altogether help shape the inflammatory profile and outcome of tumor growth or regression. With further studies, specific activation or repression of TLRs may be harnessed to offer novel immunotherapies or adjuvants to traditional chemotherapy for some OC patients. Herewith, we review recent literature on basic and translational research concerning therapeutic targeting of TLR pathways for the treatment of OC.

**Keywords: ovarian cancer, toll-like receptors, tumor microenvironment, immunotherapy, pattern-recognition receptors**

#### **INTRODUCTION**

Ovarian cancer (OC) has the most devastating death rate of gynecological cancers with only 44% of women surviving 5 years after diagnosis (1–4). The low long-term survival statistics are in part due to lack of efficient screening technology; by the time symptoms occur, most patients exhibit advanced-stage disease (over 60% of OC is diagnosed after distant metastasis). The survival rates decrease with each later stage of diagnosis with only a 27% 5-year relative survival rate for distantly metastasized tumors, highlighting the need for more efficacious treatments for advanced OC (4). Today, the standard of therapy includes surgery (hysterectomy and bilateral salpingooophorectomy) and several rounds of platinum- or taxane-based chemotherapy (1–3). Chemotherapy typically produces significant side effects, such as nausea, weight loss, fatigue, and alopecia, largely a result of toxicity of the treatment to healthy cells (5). Moreover, many cancers become resistant to treatment, further warranting the development of additional and more tumor-specific therapies. Thus, although chemotherapy remains the gold-standard of OC management, replacement as well as adjuvant treatments are in process of intensive investigation. As it is well-established that the immune system (if properly functioning) can fight tumor growth, tumor immunology research and immunotherapy clinical trials are taking center-stage in the quest for better clinical outcomes for late-staged OC (6–9).

As aggressive OC often correlates with an immunosuppressive leukocyte population in the tumor environment, efforts to modulate these cells to potentiate an anti-tumor immune response are ongoing (10–12). Of particular interest to the processes of tumorinfiltration by immune cells and their activation are the toll-like receptors (TLRs), pattern-recognition receptors (PRRs) that ligate conserved pathogen-associated molecular patterns (PAMPs) such as bacterial lipopolysaccharide (LPS) or viral dsRNA (13– 20). TLR expression is well-established in immune cells, such as macrophages and dendritic cells (DCs), where upon PAMP recognition, an inflammatory response occurs, activating numerous transcription factors,such as NF-kB and IRF 3/7 (21–23). Cytokine and chemokine secretion subsequently ensues, further activating inflammation and stimulating the adaptive immune response. In fact, TLR activation in leukocytes (e.g., DCs) can trigger a shift in the inflammatory profile of the tumor site by decreasing immunosuppression and activating immune cells that can actively fight tumors (11, 24, 25). However, in addition to their expression in leukocytes, TLRs are found in multiple tumor types, including in OC, where their activation can have tumor-promoting effects (26– 29). In fact, high levels of different TLRs in cancer cells have been associated with disease aggressiveness, treatment resistance, and poor clinical outcome. Most likely, this is a result of cytokine and chemokine-induced (e.g., as a result of NF-κB activation) recruitment of immunosuppressive and pro-angiogenic leukocytes to the tumor site (13, 30–32). In this mini review, we summarize recent studies and clinical trials aimed at exploiting TLR signaling pathways for OC immunotherapy.

#### **TOLL-LIKE RECEPTOR SIGNALING IN LEUKOCYTES**

Toll-like receptors expressed in leukocytes (e.g., macrophages) serve a crucial function at the start of the immune response, activating numerous pro-inflammatory pathways resulting in cytokine secretion, and further activation of immune cells, including the adaptive immune response (21–23). It is known that the white blood cell population infiltrating the tumor environment differs between cancers and it has been established that the specific leukocyte profile at the site has a profound effect on tumor progression or regression (8, 30, 33). As the microenvironment of OC is typically immunosuppressive, efforts are ongoing to stimulate the immune population to effectively recognize and clear the tumor cells (12). In this regard, TLR activation in immune cells can favor the anti-tumor immune response, by increasing the capability of professional antigen-presenting cells (APCs) and facilitating the activation of anti-tumoral T cells (natural killer, NK cells; cytotoxic T lymphocytes, CTLs). In fact, the last decade of cancer immunology research has brought about several examples of the benefits of TLR activation in the immune cells surrounding the ovarian tumor milieu. Several clinical trials have been performed in an attempt to stimulate TLRs for OC therapy, including using TLR agonists in combination with other immunostimulating agents, such as DC vaccines (34, 35). Overall, these studies point to the potentially promising effects of TLR stimulation for OC patients with few efficacious treatment options available, especially if integrated with mainstream treatments or as adjuvants to other immunotherapies on a case-by-case basis.

In 2005, Adams et al. first described the rationale for TLR3 agonist therapy for advanced OC (35). In 2009, it was reported that TLR3 activation in DCs enhanced antigen processing and presentation by the APCs (24). Specifically, the authors described the inability of tumor-localized DCs to successfully activate anti-tumor immunity. Instead, they suppressed T cell function, although they were shown to be capable of processing tumor antigens. However, after stimulation with dsRNA (TLR3 ligand), with co-stimulation of CD40, DC function improved to trigger the desired tumor-eliminating inflammatory response. In these studies, this was indicated by the increase of interleukin 12 (IL-12) and type I IFN secretion by the DCs, as well as higher levels of co-stimulatory molecule expression and enhanced antigenprocessing capability in both mouse and human OC samples. Furthermore, the treatment augmented the migratory capabilities of the DCs (to lymph nodes) and increased their antigenpresentation capability. These results point to the promising potential to re-structure the immunosuppressive OC environment to facilitate a robust anti-tumor response.

Earlier this year, Bellora and colleagues demonstrated that TLR activation in tumor-associated macrophages (TAMs) obtained from OC patients resulted in a shift from an M2 to an M1 polarization phenotype (36). This is significant, as M2-activated macrophages in the tumor environment are implicated in cancer growth, whereas M1-type (classically activated) macrophages are associated with better clinical outcome (37). M1-polarization is primarily immunostimulatory, characterized by the secretion of IL-12 and production of cytotoxic factors, such as nitric oxide (NO). M2-type or alternative macrophage activation largely results in immunosuppressive functions, and can be differentiated from M1-type activation by high levels of interleukin 10 (IL-10) secretion, as well as expression of specific markers, such as the mannose receptor (MR). In fact, the authors demonstrated that upon M1-polarization, the macrophages were able to induce cytolytic activity of NK cells (36). Thus, TLR activation in TAMs may be of clinical benefit by shifting the M2-polarized, immunosuppressive macrophages to a more immunostimulatory, anti-tumor phenotype.

Recently, a TLR8-specific agonist, VTX-2337 (Venti-RX Pharmaceuticals), entered Phase II clinical trials for OC patients with chemotherapy-resistant and recurring disease (38). The Phase I clinical trial with this agent was conducted in 2011 and was shown to be well-tolerated while exhibiting a dose-dependent therapeutic activity (39). The rationale for the therapy is to activate TLR8 in immune cells, whereby its signaling has been shown to have a suppressive effect on Tregs (40). Although the mechanism for the TLR8-dependent inhibition of this immune cell population is unclear, it is known to occur independently of DCs (41). In addition, TLR8 signaling appears to affect the morphology of NK cells, increasing their IFN-γ secretion, thereby strengthening the innate immune response (42). Furthermore, there have been implications for the potential of TLR7 stimulation for OC treatment (41, 43). In 2010, Geller and colleagues were the first to administer a selective small-molecule TLR7 agonist, 852A, to a small group of breast, ovarian, and cervical cancer patients with recurrent disease (43). Although significant side effects were observed with ~30% of those enrolled in the study discontinuing the therapy prior to completion, the authors showed immune activation and stabilization of disease in 2 of the 15 patients.

TLR9 ligands have similarly received interest as potential treatments for OC, specifically in combination with other immunomodulatory agents (44). In 2009, it was reported that a combinational treatment of CpG oligodeoxynucleotides (CpGODN), TLR9 ligand, and LL-37 (cathelicidin peptide) resulted in a better therapeutic outcome in mice. The authors demonstrated that the dual treatment increased the uptake of the TLR9 agonist CpODN (as TLR9 is endosomal). It was shown that the treatment increased the expansion and activation of NK cells in the murine peritoneal space, indicating an activation of innate immunity. Furthermore, studies assessing the potential role of the NK cells in the tumor environment revealed that they were heavily implicated in the observed anti-cancer effects of the therapy.

#### **TOLL-LIKE RECEPTOR SIGNALING IN OVARIAN CANCER CELLS**

In 2009, Zhou and colleagues reported on the expression of TLRs in human ovarian tissue samples, including both normal and neoplastic (benign and malignant) tissue (26). It was concluded that TLR2, TLR3, TLR4, and TLR5 were found on the epithelium of healthy ovary tissue. Additionally, this subset of TLRs was also expressed in a variety of human epithelial tumors and in numerous OC cell lines. The authors also found differential expression of TLR6 and TLR8 on all the samples, as well as low levels of TLR1, TLR7, and TLR9. It was demonstrated that the TLRs expressed in the epithelial cells were functional and it was suggested that their activation may constitute a mechanism by which the cancerous epithelial cells can manipulate inflammatory pathways to encourage tumor growth. The last decade of research on TLRs in tumor cells indicates that TLR activation in cancer cells generally results in increased production of cell survival and angiogenic molecules, as well as up-regulation of T-cell-suppressive factors, facilitating immune evasion. TLR signaling in ovarian has been

attributed with more aggressive disease, potential for metastasis, and poorer end results in the clinic. Thus, specific inhibitors of TLRs (delivered to tumor cells) may be explored as potential therapeutic targets for some patients, especially in late-stage disease with fewer therapeutic options available (18). Recent research highlights the detrimental effects of TLR engagement in OC cells, indicating that inhibition of this receptor may be of benefit to the patient if targeted specifically in the cancer cells that overexpress the molecule.

The effects of TLR signaling in cancer cells have been extensively investigated for TLR4, perhaps the best-studied PRR. In 2005, Huang and colleagues reported on its expression and activation in numerous mouse cancer cell lines (45). They determined that TLR4 stimulation by LPS in tumor cells increased production of numerous soluble factors, such as IL-6, and ultimately inhibited the ability of CTLs to recognize and kill the cancer cells. It was also found that LPS treatment of the murine tumor cell supernatants impeded the proliferation of T cells and inhibited NK cell activity. Further, the authors demonstrated that inhibition of TLR4 signaling in tumor cells significantly increased survival in animal studies. The menacing effects of TLR4 activation specifically on human OC progression have also been reported (46). Kelly et al. demonstrated that TLR4 is upregulated in numerous ovarian epithelial tumors and that high expression correlates with increased tumor progression and likelihood of developing chemo-resistance to Paclitaxel. Additionally, TLR4 (and subsequent NF-κB) activation has been demonstrated for human ovarian granulosa tumor cells (47). Thus, TLR4 inhibition in several types of OC cells may be therapeutically beneficial in conjunction with standard chemotherapy in an effort to decrease the likelihood of drug resistance.

Similarly, TLR9 signaling by OC cells (as well as breast cancer cells) has been associated with disease aggressiveness and poor clinical outcome (48). Berger and colleagues determined that higher levels of TLR9 expression correlated with more severe tumor grade. Consistently, *in vitro* scratch essays revealed the increased migratory capabilities of tumor cells expressing higher TLR9 levels (in both ovarian and breast tumor cells). It was also reported that higher TLR9 expression was more common in poorly differentiated tumors (hormone-receptor-negative tumor cells were found to have more TLR9); thus, these tumors have fewer targeted therapeutic options. In addition, it was found that OC patients with metastatic disease had elevated levels of hypo-methylated DNA (TLR9 ligand) in their serum. Further, the authors offered even more evidence of the detrimental effects of TLR9 signaling in OC cells, showing the co-localization of TLR9 and its ligand, as well as NF-κB activation, which was proportional to the levels of TLR9 expression. Significantly, NF-κB appears to be constitutively activated in numerous cancer types, whereby it is associated with highly aggressive disease and poor disease outcome, highlighting the potential of TLR targeting to inhibit this important inflammatory switch in tumor cells (28, 49).

#### **ENDOGENOUS TLR LIGANDS AND IMPLICATIONS FOR CANCER THERAPY**

In addition to the PAMPs that can activate TLRs (e.g., LPS, viral RNA, etc.), endogenous ligands for these molecules have also been identified (50). For instance, TLR2 and TLR4 can be triggered by biglycan and endoplasmin, while nucleic acid-sensing TLRs can bind to mRNA (TLR3), as well as siRNA (TLR7, TLR8). Additionally, damage-associated molecular patterns (DAMPs), molecules induced during cell stress or damage (e.g., HMGB1) can activate TLRs (51–53). As discussed, attempts to harness TLRs to promote cancer regression have been attempted in numerous trials, where the treatments are often used in combination with standard chemotherapy or radiation practices in an effort to maximize patient response. In fact, it appears likely that cell death (e.g., necrosis from standard therapy) can result in release of endogenous TLR ligands, which may activate nearby leukocytes, potentially improving the anti-tumor response (50). Continued characterization of ligands and determining downstream signaling will help elucidate the full function of TLRs in cancer progression and give more direction for novel therapeutic strategies for specific cancer types.

#### **CONCLUDING REMARKS**

The last decade of research on TLR activity and its implications in OC progression indicate that inhibition of certain TLRs in cancer cells and/or TLR stimulation in immune cells may be of therapeutic benefit in some patients. While immune activation by means of TLR stimulation can generate an anti-cancer effect, the cytokine profile following TLR activation in tumor cells typically favors an immunosuppression that can potentiate immune-tolerance and promote angiogenesis, furthering tumor growth. **Figure 1** summarizes the differential effects of TLR signaling by OC cells and immune cells. Undoubtedly, TLR targeting is a promising area of research for OC and other malignancies, although these pathways can produce such varying effects that exploitation of TLR pathways for cancer therapy has frequently been referred to as a "doubleedged sword" (54, 55). Therefore, TLR targeting for OC therapy must be pursued with care and stimulating or inhibiting agents be delivered in a cell-specific manner. Given the complex nature of the effects of TLR activation in various cells, much remains to be investigated, including the multiple regulators of TLR expression and activation in the different cell types. For instance, miRNAs have recently been shown to be "fine-tuning" regulators of TLR signaling pathways; thus further research in this exciting area of study may yield even more targeting opportunities for TLR regulation that could be applied in cancer therapy (56, 57). Finally, future therapeutic strategies may be realized more effectively in conjunction with novel drug delivery mechanisms that allow for more cell-specific drug targeting.

#### **ACKNOWLEDGMENTS**

This work was supported in part by the NIH Grant R15 CA137499- 01, a startup fund from OU and the RSAC (Grant RP1206) from the Heritage College of Osteopathic Medicine, OU (Fabian Benencia). Maria Muccioli was supported by the MCB program (OU) and the 2011 OU Student Enhancement Award (Grant 010-0500-30200-xxxxxx-IA1011018).

#### **REFERENCES**

1. Cannistra SA. Cancer of the ovary. *N Engl J Med* (2004) **351**:2519–29. doi:10.1056/NEJMra041842


schedule in heavily pretreated recurrent breast, ovarian, and cervix cancers. *Cancer Immunol Immunother* (2010) **59**:1877–84. doi:10.1007/s00262-010-0914-1


**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.

*Received: 28 March 2014; accepted: 04 July 2014; published online: 22 July 2014. Citation: Muccioli M and Benencia F (2014) Toll-like receptors in ovarian cancer as targets for immunotherapies. Front. Immunol. 5:341. doi: 10.3389/fimmu.2014.00341 This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Muccioli and Benencia. 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.*

#### **Jouko Sandholm<sup>1</sup> and Katri S. Selander 2,3,4\***


#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Sophia Ran, Southern Illinois University, USA Seth B. Coffelt, Netherlands Cancer Institute, Netherlands Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Katri S. Selander, Division of Hematology-Oncology, Department of Medicine, University of Alabama at Birmingham, SHEL 514, 1825 University Blvd, Birmingham, AL 35294-3300, USA e-mail: katriselander@uabmc.edu

Toll-like receptor 9 (TLR9) is a cellular DNA receptor of the innate immune system. DNA recognition via TLR9 results in an inflammatory reaction, which eventually also activates a Th1-biased adaptive immune attack. In addition to cells of the immune system, TLR9 mRNA and protein are also widely expressed in breast cancer cell lines and in clinical breast cancer specimens. Although synthetic TLR9-ligands induce cancer cell invasion in vitro, the role of TLR9 in cancer pathophysiology has remained unclear. In the studies conducted so far, tumor TLR9 expression has been shown to have prognostic significance only in patients that have triple-negative breast cancer (TNBC). Specifically, high tumorTLR9 expression predicts good prognosis amongTNBC patients. Pre-clinical studies suggest that TLR9 expression may affect tumor immunophenotype and contribute to the immunogenic benefit of chemotherapy. In this review, we discuss the possible contribution of tumor TLR9 to the pathogenesis and treatment responses in breast cancer.

**Keywords:TLR9, breast cancer, invasion, inflammation, prognosis**

#### **INTRODUCTION**

Toll-like receptor 9 (TLR9) is a DNA receptor that recognizes microbial and vertebrate DNA (1–5). Initially, TLR9 was thought to recognize specifically the CpG sequence in DNA (1, 6). The sequence-requirement may, however, be relevant only for the synthetic, oligonucleotide TLR9-ligands in the phosphorothioate backbone, and also CpG sequence-independent TLR9 activation by DNA has been reported (6–8). Like the other TLRs that recognize nucleic acids (TLR3, TLRs 7–8, and TLR13), TLR9 is located at the endoplasmic reticulum in resting cells (9, 10). When DNA enters the cell, TLR9 translocates to the endosomal/lysosomal compartment where ligand recognition and binding takes place (9, 11). DNA recognition by TLR9 initiates a downstream signaling cascade, which includes the adaptor molecule MyD88 (12, 13). As an effector of the innate immune system, stimulation of TLR9 induces a NF-κBmediated rapid inflammation, characterized by increased expression of various interleukins and cytokines. A common feature for the nucleotide-sensing TLRs is the induction of both antiviral and antitumoral type I interferons (IFNs) from plasmocytoid dendritic cells (pDCs) (14). Eventually, this inflammation also activates the adaptive immune system, which then results in the clearance of the invading pathogens and the infected cells (2, 15). A similar inflammatory response, mediated via TLRs, also takes place during sterile tissue damage (16–19). In addition to DNA, other biological molecules have also been suggested to induce TLR9-mediated responses. Such molecules include the malaria pigment hemozoin and histone proteins (17, 20–22). TLR9 was recently shown also to recognize RNA–DNA hybrids (23).

#### **TLR9 EXPRESSION IN BREAST CANCER**

Toll-like receptor 9 expression has been detected in cells of breast milk (TLR9 mRNA) and also in normal epithelial cells of the mammary gland (TLR9 protein) (24, 25). TLR9 mRNA and protein are also widely expressed in various human cancer-cell lines as well as in clinical cancer specimens, including breast, prostate, brain, gastric, renal cell carcinoma, and esophageal tumors (24, 26–33). Specifically in breast cancer, TLR9 protein expression has been detected both in the epithelial cancer cells as well as in the fibroblast-like cells associated with the tumors (24, 26, 29, 34). Consistent with the endosomal/lysosomal localization of TLR9 at the subcellular level, in breast cancer cells *in vitro*, TLR9 appeared punctate in intracellular fluorescence staining, located especially in the perinuclear region, where these organelles are located (35). Of the five human TLR9 isoforms (A–E), mRNA expression of the TLR9 A and B isoforms has been studied and detected in breast cancer specimens (36, 37).

#### **TLR9 AS A PROGNOSTIC FACTOR IN BREAST CANCER**

The prognostic significance of TLR9 in cancers appears to be bimodal. In some cancers, such as glioma, prostate cancer, and esophageal adenocarcinoma, high tumor TLR9 expression has been associated with poor survival whereas in others, such as triple-negative breast cancer (TNBC) or renal cell carcinoma, low tumor TLR9 expression upon diagnosis predicts poor prognosis (27, 30, 32, 33, 38, 39). We demonstrated recently that although widely expressed in all clinical subtypes of breast cancer, TLR9 expression has significant, prognostic significance only in TNBC that lack the expression of estrogen (ER), progesterone (PR), and HER2 receptors. More specifically, low tumor TLR9 expression upon diagnosis was associated with a significantly shortened disease-free-specific survival (29, 32). Furthermore, although we demonstrated that low-TLR9–TNBC cells become highly invasive in hypoxic conditions, it is currently unclear whether this mechanism contributes to the poor survival of the breast cancer patients that have hypoxic, low-TLR9–TNBC tumors. The mechanism for the increased invasion in hypoxia when TLR9 is absent is also not known (32). In addition to the actual tumor cells, the TLR9 expression status of tumor-associated fibroblast-like cells has also been shown to be of prognostic value in breast cancer. In this context, high TLR9 expression was associated with better prognosis (34). This study did not, however, assess triple-negative status of the cancers, and the exclusion of metastatic and neoadjuvant-treated patients probably counter selected against patients of the TNBC subtype.

#### **EFFECTS OF TLR9 STIMULATION ON CELLULAR INVASION**

Synthetic TLR9-ligands, the CpG sequence-containing oligonucleotides (CpG–ODNs, such as ODN M362) that mimic bacterial DNA, are strong inducers of inflammation in cells of the immune system (40, 41). These oligonucleotides mimic bacterial DNA based on their high CpG content and unmethylated cytosines. CpG–ODNs are taken up into cells via DEC-205, a multilectin cell surface receptor, which is expressed in various cell types (42). These same compounds induce cellular invasion in macrophages, mesenchymal stem cells, and in cancer cells of various origins *in vitro* (24, 28, 43, 44). In breast cancer cells, such synthetic TLR9 ligand-induced invasion has been detected both in ER-positive and ER-negative breast cancer cells (24, 28, 35). This invasive effect is mediated via TLR9, and it is blocked by chloroquine, an inhibitor of endosomal acidification and an inhibitor of TLR9 signaling. Downstream of TLR9, such invasion is mediated via TRAF6, but not MyD88 (24, 28, 35). At the proteolytic level, CpG–ODN-induced invasion is associated with down-regulation of tissue inhibitor of matrix metalloproteinases-3 (TIMP-3) and activation of matrix metalloproteinase-13 (MMP-13) (24, 28, 35, 44). Interestingly, although methylation of cytosines in CpGs has been shown to decrease their pro-inflammatory effects, the invasive effects of these molecules are independent of their methylation status (35, 40, 45). CpG–ODNs can form various secondary structures, including homopolymer duplexes and hairpins, containing stem loop structures. The stem loop secondary structure appears important for the invasive effects of the CpG–ODN (35). Furthermore, the invasive effects can also be seen with non-CpG sequence-containing ODNs that in inflammatory experiments act as TLR9 antagonists (24, 46). The synthetic, phosphorothioatebackbone-modified CpG–ODNs do not exist in nature. Thus, for this invasion to have physiological significance, it would have to be caused also by natural DNA in the phosphodiester backbone. In prostate cancer cell lines and in gastrointestinal cancer cell lines, bacterial DNA (purified from *Escherichia coli* or *Helicobacter pylori*, respectively) also has similar, stimulatory effects on invasion (28, 43). Whether microbe-derived DNA similarly induces invasion in breast cancer cells is not known. We, however, demonstrated recently that self-DNA, which is derived from chemotherapy-treated, dead cancer cells is rapidly taken up into surviving cancer cells, where it serves as an invasion-inducing

TLR9 ligand (47). This cellular uptake is possibly endocytosis or pinocytosis-mediated, since fluorescently labeled, dead cancercell-derived DNA, which was added to cell culture medium, was seen inside the recipient cells rapidly, within 15 min. However, similar to other reported TLR9-mediated effects of cell-derived self-DNA, complex formation of such cell-derived DNAs with the cationic antimicrobial peptide LL-37 enhanced DNA uptake into viable breast cancer cells, and was a requirement for the invasioninducing effects (47, 48). This scenario may be physiologically relevant since LL-37 is expressed also in breast cancers (49, 50). Interestingly, the effects of cell-DNA on invasion are mediated via cathepsins and surprisingly, not via MMPs, which are the mediators for CpG–ODN-induced invasion (44, 47, 51, 52). DNA that was derived from intact, proliferating cancer cells did not induce invasion. This suggests that the invasive effect requires a certain DNA-structure, either alone or in complex with LL-37. Such DNAstructures could possibly be formed upon DNA degradation by nucleases. Whether self-DNA-induced and TLR9-mediated cancer cell invasion takes place *in vivo* in breast or any cancer is currently unknown. In principle, however, such DNA-induced and TLR9-mediated cancer cell invasion could represent a novel mechanism of treatment resistance. Since tumor growth is the sum of local proliferation and local invasion, such treatment resistance could theoretically manifest as no change or even increase in tumor size despite treatment. Finally, TLR9 appears to have also ligand-independent invasive activity. Down-regulation of TLR9 in MDA-MB-231 breast cancer cells through siRNA results in decreased *in vitro* invasion in the absence of exogenous DNA. The decreased invasion of the TLR9 siRNA cells was associated with decreased MMP activity and increased expression of TIMP-3 (32). Similar effects were also detected by TLR9 siRNA in brain cancer cells *in vitro* (53). These TLR9 expression-induced changes in the cellular invasive machinery suggest that TLR9, as a DNAbinding protein, might also have effects on gene transcription. TLR9 expression has indeed been detected in the nuclei of renal cell carcinoma tumor samples (30), but whether or not it can directly affect gene expression, requires further experimenting.

#### **EFFECTS OF TLR9 STIMULATION ON INFLAMMATION**

Toll-like receptor 9 agonists have various well documented proinflammatory effects in cells of the immune system (40, 41, 48, 54). Whether synthetic TLR9 agonists also induce the expression of inflammatory mediators in breast cancer cells, is not known. In cells of the immune system, a key characteristic of the TLR9-induced innate immune response is the promotion of a strong type I T helper cell (Th1) adaptive immune response. This includes both CD8<sup>+</sup> T-cell responses and antigenspecific antibody responses (55). Since CD8<sup>+</sup> T-cells are capable of immunologic tumor cell destruction, CpG–ODNs have been tested both as monotherapy and as an adjuvant for cancer vaccines, against various cancer types in pre-clinical cancer models, including breast cancer (55). In mouse models of breast cancer, CpG–ODN treatment resulted in the eradication of orthotopic tumors (56, 57). CpG–ODN treatment also induced an immunologic memory against tumor challenge, which was associated with an up-regulation of IFN-γ-positive CD4<sup>+</sup> and CD8<sup>+</sup> T-cells (56, 57). CpGs, when given as an adjuvant with a peptide vaccine, also prevented the formation of spontaneous tumors in a mouse model of HER2-positive breast cancer (58). Although the direct growth inhibitory effects of CpG–ODNs on cancer cells are quite weak *in vitro*, certain modifications in the CpG structure have resulted in increased tumor growth inhibition, also in nude mouse models *in vivo*, suggesting direct tumor effects of these compounds (24, 59–61). Furthermore, when given in a combination, the immunomodulatory ODN was also shown to potentiate the efficacy of trastuzumab, an anti-HER2-antibody,in a mouse model of breast cancer (59). In conclusion, these pre-clinical experiments suggest that TLR9 ligands can directly inhibit the growth of breast cancer cells *in vitro* and *in vivo*, and they can enhance antitumor immunity, possibly via inducing a Th1 adaptive immune response. These studies have not, however, addressed the role of TLR9 expression in tumors vs. host in these responses. Despite the successful pre-clinical results, CpG treatment has demonstrated anti-tumor activity only in select patients in clinical trials. There are, however, no reports on their efficacy in breast cancer trials (55). Finally, the discrepancies between the *in vitro*-observed, unwanted tumor invasion-promoting effects and the favorable, most likely immune system-mediated anti-tumor effects of the synthetic TLR9-ligands are likely explained by the pharmacokinetics of these compounds. After s.c. and i.v. administration, highest concentrations of TLR9 ligands are detected in plasma, kidneys, and organs of the reticuloendothelial system, and much less so in tumor tissues (59).

Self-DNA has been shown to have TLR9-mediated inflammatory effects in other cell types, especially when complexed with LL-37, which is expressed in various tissues (16, 48, 52, 62). We demonstrated recently that self-DNA, which is derived from doxorubicin-killed breast cancer cells, induces mRNA expression of various inflammatory mediators in living, TLR9-expressing cells. Furthermore, while assessing treatment responses to doxorubicin in a mouse model of TNBC, we discovered that although the tumor response to treatment was similar in TLR9 siRNA and control siRNA TNBC groups, mice bearing TLR9 siRNA tumors lost significantly less weight than similarly treated mice with control siRNA tumors. Similar weights of the vehicle-treated mice suggested to us that TLR9 expression in the tumors may be an important determinant of chemotherapy-induced inflammation and activation of anti-tumor immunity (47). Inflammatory response to chemotherapy is gaining acceptance as an important mediator of treatment responses to standard cancer therapy (63). More specifically,we hypothesize that the tumor TLR9-dependent, post-treatment weight loss is actually a surrogate marker for self-DNA-induced and TLR9-mediated inflammation that takes place at the tumor site. Such tumor TLR9-mediated inflammation might then amplify the anti-tumor immune response, eradicate microscopic disease and through this mechanism, translate into cure (47). We predict that the lack of such immunogenic effect in tumors that have low-TLR9 expression indeed contributes to the described poor disease-specific survival in triplenegative disease (32). This hypothesis requires a detailed analysis of tumor TLR9-dependent immune response to chemotherapy in immune-competent pre-clinical cancer models. However, if true, it would mean that patients with low-TLR9–TNBC could especially benefit from adjuvant cancer immunotherapy. It is also

possible that TLR9 expression changes tumor immunophenotype independent of treatment and this aspect also requires further investigation.

#### **TLR9 REGULATION IN BREAST CANCER**

Several cancer-associated viruses have been shown to downregulate TLR9 expression through their oncoproteins. For example, human papillomavirus (HPV), Epstein–Barr virus, and hepatitis B virus inhibit the expression and impair the function of TLR9 in infected target cells (64–66). Patients with chronic hepatitis B virus have decreased levels of TLR9 mRNA in peripheral blood mononuclear cells (67). The Merkel cell polyomavirus large T antigen down-regulates TLR9 expression in epithelial cells and in cells derived from Merkel cell carcinomas (68). For the HPV16, the mechanism behind TLR9 suppression was recently shown to involve the viral oncoprotein E7-induced formation of transcriptional inhibitory complex that includes NF-κB p50–p65, ERα, and chromatin modifying enzymes. This complex induces epigenetic changes at the TLR9 promoter area (69). It is likely that these viral effects on TLR9 expression and function play an important role in viral persistence, through inhibition of host immune responses (64, 65, 67, 70, 71). Nevertheless, also opposite effects on microbial TLR9 regulation have been suggested (72, 73). Although breast cancer is not currently considered to have viral etiology, several viruses, including human papilloma viruses, have been detected in normal and cancerous human breast tissues (74–77). Whether or not these viral effects have a role in breast cancer development or pathophysiology is currently unknown.

Tumor microenvironment oxygen concentration is also an important regulator of TLRs. Similar with the effects of hypoxia on other TLRs in other cell types, hypoxia also up-regulates TLR9 expression in breast cancer cells *in vitro* and in orthotopic breast tumors *in vivo* (32, 51, 78). These hypoxia effects on TLR9 mRNA and protein expression were mediated via HIF-1α in breast cancer cells *in vitro* (32). TNBCs are typically hypoxic (79). Therefore, understanding the mechanism on why tumor TLR9 expression levels remain low despite hypoxia in some TNBCs might open novel therapeutic possibilities that might also apply to renal cell carcinoma (30). It was also demonstrated recently that TLR9 expression is under the control of the circadian molecular clock (80). The significance of this finding for breast and other cancers is currently open.

Although TLR9 is expressed in all clinically relevant subtypes of breast cancer, we and others have discovered that there is an inverse correlation between tumor TLR9 and ER expression: ERpositive breast cancers have significantly lower levels of TLR9 expression, as compared with TNBCs (26, 29, 32, 36). The basal TLR9 expression is also significantly lower in human ER-positive breast cancer cells, as compared with human ER-negative breast cancer cell lines *in vitro*. Furthermore, transfection of ERα cDNA into TNBC cells suppresses TLR9 expression of the recipient cells (36). Both estradiol and testosterone induced TLR9 expression via their cognate receptors in breast cancer cells *in vitro*. Testosterone also augmented the pro-invasive effects of CpG–ODNs. Finally, bicalutamide, a commonly used hormonal treatment in prostate cancer, increased TLR9 expression in ER-positive breast cancer cells (36). This effect of bicalutamide on TLR9 expression might

be of therapeutic interest since a proportion of TNBC tumors express the androgen receptor that bicalutamide targets (81).

#### **TLR9 POLYMORPHISM IN BREAST CANCER**

The TLR9 gene is located on human chromosome 3 (82). Although TLR9 gene polymorphisms have been studied in other diseases, including infectious and autoimmune diseases and some cancers, very little is known about TLR9 gene polymorphism in breast cancer (83–86). A study conducted by Resler and coworkers using over 800 case and control samples, found that the single nucleotide polymorphism (SNP, rs352140) in *TLR9*, which does not alter protein amino acid sequence but might alter protein function or stability, was associated with breast cancer risk (OR 0.85, 95% CI 0.74–0.97) (87). The patients in this study were all post-menopausal (65–79 years) and 80% of the cases had hormone receptor-positive breast cancer. These results were in contrast to those of Etokebe et al., who found no association in the same TLR9 SNP with breast cancer risk in a small Croatian cohort, consisting of 130 breast cancer cases and 101 controls (88).

#### **CONCLUSION**

Although TLR9 is widely expressed in breast cancers, it appears that tumor TLR9 expression has prognostic significance only in TNBC. Especially, TNBC patients that have low tumor TLR9 expression upon diagnosis have a significantly shortened diseasespecific survival, as compared with TNBC patients that have high tumor TLR9 expression. These findings, however, need to be repeated in larger and more diverse patient populations. TNBC tumors are typically hypoxic and low oxygen concentrations upregulate TLR9 expression in TNBC cells in pre-clinical models. Understanding why TLR9 expression levels remain low in some TNBC tumors in the hypoxic tumor microenvironment might reveal novel therapeutic opportunities. It has been demonstrated recently that viral oncoproteins down-regulate TLR9 expression in various cancer tissues. Although breast cancer is not currently considered to have viral etiology, various viruses known to be capable of down-regulating TLR9 expression have also been detected in breast cancers. The contribution of these viral infections to low tumor TLR9 status in TNBC should therefore be addressed in future studies. Finally, the mechanisms how the lack of tumor TLR9 expression results in poor prognosis are unknown. Studies from pre-clinical TNBC models suggest that tumor TLR9 expression might affect tumor immunophenotype or be required for chemotherapy-induced anti-tumor immune response. If this is the case, then patients with low-TLR9–TNBC tumors might benefit from anti-cancer immune therapy. The specificity of the immune therapy requires, however, a clear understanding of how TLR9 expression affects tumor immunity. Synthetic TLR9 agonists, CpG–ODNs have demonstrated promising direct and immune system-mediated anti-cancer effects against breast cancer in preclinical models but they have not been studied in clinical breast cancer trials. It is clear that synthetic CpG–ODNs induce cancercell invasion *in vitro*.Whether this finding is relevantfor the clinical situation, where such agonists are given in order to boost the antitumor immune response, remains to be resolved. Finally, aiming to increase tumor TLR9 expression prior to chemotherapy should

be considered a therapeutic opportunity in the TNBC patients that have low tumor TLR9.

#### **ACKNOWLEDGMENTS**

Dr. Johanna Tuomela (University of Turku, Finland) and Dr. Kevin Harris (University of Alabama at Birmingham, AL, USA) are kindly acknowledged for carefully reviewing the manuscript.

#### **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.

*Received: 01 May 2014; paper pending published: 08 May 2014; accepted: 30 June 2014; published online: 22 July 2014.*

*Citation: Sandholm J and Selander KS (2014) Toll-like receptor 9 in breast cancer. Front. Immunol. 5:330. doi: 10.3389/fimmu.2014.00330*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Sandholm and Selander. 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 role of pattern-recognition receptors in graft-versus-host disease and graft-versus-leukemia after allogeneic stem cell transplantation

#### **Simon Heidegger <sup>1</sup> , Marcel R. M. van den Brink <sup>2</sup> ,Tobias Haas <sup>1</sup> and Hendrik Poeck 1,2\***

1 III. Medizinische Klinik, Klinikum Rechts der Isar, Technische Universität München, Munich, Germany <sup>2</sup> Department of Medicine and Immunology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA

#### **Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

#### **Reviewed by:**

Leonardo Nimrichter, Federal University of Rio de Janeiro, Brazil Rena Feinman, Hackensack University Medical Center, USA Arseniy E. Yuzhalin, University of Oxford, UK

#### **\*Correspondence:**

Hendrik Poeck, III. Medizinische Klinik, Klinikum Rechts der Isar, Technische Universität München, Ismaningerstr. 22, 81675 München, Germany e-mail: poeckh@mskcc.org

Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is the only treatment with curative potential for certain aggressive hematopoietic malignancies. Its success is limited by acute graft-versus-host disease (GVHD), a life-threatening complication that occurs when allo-reactive donor T cells attack recipient organs. There is growing evidence that microbes and innate pattern-recognition receptors (PRRs) such as toll-like receptors (TLR) and nod-like receptors (NLR) are critically involved in the pathogenesis of acute GVHD. Currently, a widely accepted model postulates that intensive chemotherapy and/or totalbody irradiation during pre-transplant conditioning results in tissue damage and a loss of epithelial barrier function. Subsequent translocation of bacterial components as well as release of endogenous danger molecules stimulate PRRs of host antigen-presenting cells to trigger the production of pro-inflammatory cytokines (cytokine storm) that modulate T cell allo-reactivity against host tissues, but eventually also the beneficial graft-versusleukemia (GVL) effect. Given the limitations of existing immunosuppressive therapies, a better understanding of the molecular mechanisms that govern GVHD versus GVL is urgently needed.This may ultimately allow to design modulators, which protect from GvHD but preserve donor T-cell attack on hematologic malignancies. Here, we will briefly summarize current knowledge about the role of innate immunity in the pathogenesis of GVHD and GVL following allo-HSCT.

**Keywords: graft-versus-host disease, allogenic hematopoietic stem cell transplantation, pattern-recognition receptors, inflammsome, microbiota, danger molecules**

#### **INTRODUCTION**

Allo-HSCT is an established treatment modality for aggressive hematological malignancies and is performed in more than 30,000 patients annually worldwide (1). Donor-derived T cells in the graft can maintain remission after induction therapy by attacking residual tumor cells in a process known as graft-versus-leukemia (GVL). Unfortunately, beneficial GVL effects are tightly associated with the pathogenesis of acute graft-versus-host disease (GVHD). Allogeneic donor T cells recognize mismatches in major or minor histocompatibility antigens present in non-malignant host tissues and subsequently induce immune-mediated damage to target organs such as the gastrointestinal tract, skin, liver, and lungs (2). Acute GVHD occurs in 40–50% of all allo-HSCT patients and accounts for considerable morbidity and mortality (3). Depletion of T cells from the allograft can decrease the incidence of acute GVHD, but comes at the cost of greater risk of graft failure, reduced GVL activity, and increased incidence of leukemic relapse (4). As a current standard of care for GVHD, glucocorticoids and other immunosuppressive drugs are used to inhibit T-cell activation and proliferation, which similarly affects GVL activity. A better understanding of the underlying molecular mechanisms may help to design measures to prevent GVHD but preserve donor T-cell responses and GVL activity, thus allowing for a broader

application of allo-HSCT in the future. Here, we discuss how the innate immune system and its environmental triggers shape the clinical course and outcome of allo-HSCT in patients and corresponding animal models.

The biology and function of pattern-recognition receptors (PRRs) is reviewed in detail within this research topic issue (5, 6). In brief, PRRs are germ line-encoded receptors that detect conserved molecular structures that are specific to invading microbes but are absent on host cells under homeostatic conditions. Ligation of such pathogen-associated molecular patterns (PAMPs) leads to activation and maturation of antigen-presenting cells (APCs), release of pro-inflammatory cytokines and, eventually, the initiation of an adaptive immune response. PRRs are expressed on different cell types of the innate and adaptive immune systems as well as non-hematopoietic cells such as endo- and epithelial cells.

#### **IMPORTANCE OF HOST MICROBIOTA AND THE EMERGING ROLE OF INNATE IMMUNITY IN GVHD**

Primary target organs of acute GVHD such as the gastrointestinal tract, skin, liver, and lungs all form epithelial linings that constantly interact with commensal and pathogenic bacteria, either through the epidermis, intestinal, or airway mucosa or the portal circulation. Consistently, there is growing evidence that bacteria and innate PRRs are critically involved in the pathogenesis of acute GVHD. Landmark studies by van Bekkum and colleagues in mice demonstrated that bacterial decontamination or utilization of germ-free mice lead to less severe intestinal GVHD (7, 8). Reduction of intestinal microbiota by antibiotic treatment not only mitigated intestinal but also skin GVHD, suggesting a systemic effect of gut decontamination (9). Similarly, antibiotic decontamination in patients undergoing allo-HSCT seemed to confer robust protection from acute GVHD (10, 11). Lipopolysaccharide (LPS) derived from Gram-negative bacteria was identified as a driver of GVHD pathogenesis. In experimental models, allo-HSCT recipients that were treated either with anti-endotoxin neutralizing antibodies (12, 13) or an oral LPS inhibitor (14) showed reduced GVHD severity associated with preserved GVL effects and improved overall survival. These findings launched widespread use of prophylactic antibiotic treatment to reduce the bacterial burden prior to allo-HSCT, now routinely performed in many transplantation centers worldwide (15). Interestingly, modification of the intestinal microbiota using the probiotic microorganism *Lactobacillus rhamnosus* resulted in reduced translocation of enteric bacteria to the mesenteric lymph nodes, associated with improved survival and reduced acute GVHD in mice (16). Furthermore, intestinal inflammation during GVHD in mice and humans is associated with major shifts in the composition of the intestinal microbiota. In one report, GVHD-associated loss of paneth cells resulted in reduced production of antimicrobial peptides and a loss of microbial diversity with outgrowth of *Escherichia coli.* Antibiotic treatment prevented outgrowth of *E. coli* and ameliorated the course of GVHD (17). Another study showed a marked expansion of *Lactobacillales* in murine GVHD. Elimination of this species from the flora of mice before allo-HSCT aggravated GVHD, whereas its reintroduction mediated significant protection, indicating that the microbiota can modulate the severity of intestinal inflammation (18). A recent study suggested that not only bacteria but also host fungal communities (mycobiome) can critically shape acute GVHD (19). Patients colonized with candida species suffered from more severe GVHD and showed more frequent intestinal involvement (33 versus 19%). Interestingly, candida colonization was more frequent in patients bearing a loss-of-function single nucleotide polymorphism (SNP) that is associated with impaired function of the innate PRR Dectin-1, a member of the C-type lectin family of receptors that detect carbohydrates constituent of fungal cell walls, thus playing an important role in the initiation of antifungal immunity (20).

With increasing knowledge on how PRRs detect conserved microbial and danger-associated molecular patterns (DAMPs) and initiate adaptive immune responses, their role in the pathogenesis of acute GVHD has become a focus of intense research. A widely accepted model (depicted in **Figure 1**) postulates that intensive chemotherapy and/or total-body irradiation (TBI) during pre-transplant conditioning results in tissue damage and loss of epithelial barrier function. Bacterial components translocated across the barrier as well as endogenous danger molecules released from damaged cells are sensed by PRRs on host and/or donor APCs such as dendritic cells (DCs), which produce pro-inflammatory cytokines and prime allo-reactive donor-derived T cells (21). This model is supported by mouse studies, which demonstrate that intensified TBI increases epithelial damage and is associated with more severe GVHD (14, 22). Intriguingly, innate lymphoid cellderived IL-22 protects both the intestinal stem cell compartment and the mature intestinal epithelium from inflammatory tissue damage (23) in line with the general concept that IL-22 can maintain epithelial integrity under inflammatory conditions (24). The enhanced intestinal barrier function thus may limit LPS translocation and subsequent PRR activation. Consistently, genetic deficiency for IL-22 results in impaired gut epithelial integrity and increased tissue damage and mortality from acute GVHD (23). Along these lines, prophylactic treatment with recombinant keratinocyte growth factor protected mice from the development of lethal acute GVHD, presumably via reduction of intestinal epithelial apoptosis and diminished LPS-mediated pro-inflammatory cytokine release (25). However, administration of the recombinant human keratinocyte growth factor palifermin before and after allo-HSCT in a phase I/II placebo-controlled clinical trial had no significant effect on the incidence and severity of acute GVHD and short-term survival (26), presumably due to pleiotropic effects of palifermin.

#### **TOLL-LIKE RECEPTORS IN GVHD PATHOGENESIS**

Toll-like receptors (TLRs) constitute a family of transmembrane PRRs that are broadly expressed in hematopoietic and nonhematopoietic cells (27). TLR ligation by a variety of microbial components leads to activation of APCs, production of pro-inflammatory cytokines, and release of chemokines. One of the best-studied TLRs in the context of GVHD is TLR4, which detects LPS in the cell wall of Gram-negative bacteria. The importance of LPS translocation and subsequent release of pro-inflammatory cytokines such as TNF-α for the pathogenesis of acute GVHD have been clearly documented (14). Moreover, genetic deficiency for TLR4 in either donor or recipient cells resulted in reduced DC activation, dampened allogenic T-cell proliferation, and less severe acute GVHD (28). However, signaling through TLR4 seems not to be absolutely required for the development of GVHD in all cases. Accordingly, in another study TLR4-deficient recipient mice showed GVHD severity comparable to wild-type mice (29), suggesting that alternative pathways in the absence of TLR4 signaling can lead to the activation of host APCs and subsequent donor T-cell stimulation. Genetic association studies in patients undergoing allo-HSCT have shown inconsistent results concerning the role of TLR4 in the pathogenesis of GVHD. Patients showed reduced frequency of severe GVHD when they or their sibling donors carried at least one of two SNPs that are associated with reduced TLR4 responsiveness to LPS (odds ratio of 0.63 and 0.88, respectively) (30). A second study showed that if both patient and donor carry the SNP Thr399Ile, the incidence of severe acute GVHD was significantly increased but overall survival was not influenced (31). These contrasting results may be attributable to differences in patient cohorts, conditioning regimens and antimicrobial treatment routines.

Other members of the TLR family have been associated with immunomodulatory capacities and suppression of GVHD. Pretreatment of mice with the TLR5 ligand flagellin resulted in reduced GVHD and improved overall survival (32). Interestingly,

translocation of luminal bacteria as well as the release of endogenous danger molecules such as adenosine triphosphate (ATP) and uric acid result in the

in a clinical study of adoptively transferred immunosuppressive regulatory T cells to allo-HSCT recipients, patients who developed GVHD showed significantly increased TLR5 mRNA expression in peripheral blood mononuclear cells (33), whereas patients that did not show GVHD had reduced TLR5 mRNA expression. These results in the human system are difficult to interpret but may indirectly suggest a pro-inflammatory role of TLR5 in allo-HSCT recipients, contrary to the mouse study cited above.

Furthermore, it was shown that tissue inflammation induced by TLR ligation can modulate the development of GVHD at a local level (34). In this regard, the authors created mixed chimeras by transplanting B6 bone marrow cells into lethally irradiated BALB/c mice. After establishment of the B6 allograft, they transferred additional B6 donor T cells, which mimic the clinical use of donor lymphocyte infusions. Transplantation of donor T cells into established mixed chimeras did not induce GVHD, as donor T cells did not enter target tissues despite undergoing allo-activation, expansion, and up-regulation of homing molecules. Strikingly, topical application of R-848, a synthetic TLR7 agonist, unleashed massive skin infiltration of donor T cells, and development of localized GVHD. Using a different TLR7 ligand (3M-011), another group demonstrated that the timing of TLR activation has important consequences for the pathogenesis of GVHD. While repetitive applications of 3M-011 after allo-HSCT aggravated GVHD severity (35), a single treatment timed between TBI and allo-HSCT induced expression of the immunoinhibitory enzyme indoleamine 2,3-dioxygenase (IDO) in host APCs, which resulted in reduced lethal intestinal GVHD (36).

NACHT, LRR, and PYD domains-containing protein 3.

In addition, signaling via TLR9 that detects microbial CpG-DNA motifs has been implicated in the pathogenesis of acute GVHD. Studies in TLR9 deficient mice showed reduced GVHD and improved survival (29, 37). Repetitive application of CpG-DNA following allo-HSCT results in increased GVHD mortality (35). This effect was dependent on TLR9 signaling and subsequent IFN-γ release in host hematopoietic cells. Less consistent results come from human studies: Transplant patients who carry gene variants associated with reduced TLR9 expression showed GVDH occurrence similar to control patients (38). A recent report analyzed two alternative SNPs that have been described to interfere with the TLR signaling pathway (39). While patients receiving stem cells from an unrelated donor with the A1174G variant experienced severe acute GVHD more frequently (49.5 vs. 20.7%), the T1635C variant in donor cells was associated with protective effect against severe acute GVHD (16.7 vs. 49.1%).

Taken together, TLR signaling can both aggravate and attenuate the development of local and systemic GVHD; critical factors seem to be the cell type primarily affected (e.g., hematopoietic versus non-hematopoietic) and the time point of TLR ligation. Thus, the role of TLRs in the pathophysiology of GVHD remains controversial. Recipient mice that are genetically deficient for either the TLR signaling adaptor molecules MyD88 or TRIF were found to show less severe intestinal GVHD (37). In a contrasting report, bone marrow chimeric recipient mice deficient for MyD88 and/or TRIF only in hematopoietic cells developed GVHD comparable to wildtype controls (40). Other than by differences in the experimental setting between institutions (e.g., microbiota and conditioning regime), these differences might be explained by alternative (non-TLR) pathways in APCs or epithelial cells,leading to allo-activation and proliferation of donor T cells in the absence of TLR signaling.

#### **NOD-LIKE RECEPTORS IN GVHD**

Another family of PRRs with relevance to GVHD is the cytoplasmic NOD-like receptors (NLRs). NOD1 and NOD2 detect peptidoglycans as components of the bacterial cell wall (6). Both receptors have been extensively studied in the context of Crohn's disease, a chronic inflammatory bowel disease that shares several immunopathogenic features with intestinal GVHD. Reduced NOD2 activity was found to be associated with impaired epithelial barrier function and aggravated intestinal inflammation (41). Similarly, following allo-HSCT, *NOD2*-deficient mice showed signs of exacerbated GVHD (42). Another study with bone marrow chimeric mice that lacked NOD2 activity only in hematopoietic cells showed that NOD2 negatively regulates the development of GVHD through its inhibitory effect on host APCs. The presence of different SNPs in the *NOD2* coding region resulting in impaired downstream signaling via the pro-inflammatory transcription factor NF-κB in either the patient, donor or both was associated with more severe GVHD (43). Two follow-up reports confirmed *NOD2* mutations as independent risk factor for transplant-related mortality (44, 45). However, several studies proposed contrasting data as they could not find an impact of *NOD2* polymorphisms on GVHD severity and outcome after allo-HSCT (46–48).

Several members of the NLR family not only detect microbial invaders but also survey cellular homeostasis and sense endogenous danger signals (6). Examples of such DAMPs are adenosine triphosphate (ATP), uric acid crystals, and double-stranded DNA released from dying cells. Activation of specific members of the NLR family by DAMPs results in the formation of cytosolic multi-protein complexes called inflammasomes, whose exact composition depends on the activator initiating their assembly (49). Inflammasome activation leads to the cleavage of pro-caspase-1 and the subsequent processing of the bioactive form of IL-1β and IL-18. These downstream effector molecules have been shown to modulate GVHD as antibody-mediated neutralization of IL-1β resulted in less severe acute GVHD in mice (50, 51). In a phase I/II clinical trial, blockade of IL-1 signaling attenuated GVHD in 8 out of 14 patients with glucocorticoid-refractory disease (52). In contrast, a larger randomized study showed no effect of a recombinant IL-1 receptor antagonist on GVHD severity and overall survival (53). However, timing and way of administration of IL-1 receptor blockade may be critical. Novel IL-1β specific antibodies await clinical testing in the setting of allo-HSCT.

The NLRP3-inflammasome is an essential platform for caspase-1 activation in response to multiple distinct exogenous and endogenous danger signals (6) and its function can be regarded as a guardian of intracellular homeostasis. NLRP3 utilizes the adapter protein ASC for activation of caspase-1 and subsequent cleavage of the precursor protein pro-IL-1β into its active form. Binding of the endogenous danger molecule ATP to the purinergic receptor P2X<sup>7</sup> leads to potassium efflux and subsequent activation of the NLRP3-inflammasome. In mice and humans undergoing allo-HSCT, increased extracellular levels of ATP were found after TBI and during the development of GVHD (54). ATP released from damaged or dying cells induces activation of host APCs and priming of allo-reactive donor T cells. Pharmacological metabolization of ATP using apyrase resulted in less severe GVHD (54). Chimeric mice that were genetically deficient for the purinoceptor *P2X*<sup>7</sup> in hematopoietic cells were partially protected from GVHD. Reconstitution with wild-type DCs resulted in restored GVHD development, demonstrating a critical role for host DCs in sensing ATP and the subsequent induction of GVHD. However, significantly reduced overall survival but no alterations in GVHD severity were found in patients or corresponding donors with a loss-of-function SNP in the *P2X<sup>7</sup> receptor* gene (55). After conditioning therapy in mice, intestinal commensal bacteria and uric acid contribute to NLRP3-inflammasome-mediated IL-1β processing, and gastrointestinal decontamination or enzymatic uric acid depletion led to reduced GVHD severity (51). NLRP3 and the adapter protein ASC, which are both required for pro-IL-1β cleavage, were critical for the full manifestation of GVHD. In transplanted mice, IL-1β exerted its effects on both DCs and T cells, which preferably differentiated into IL-17A-producing Th17 cells (51), a CD4<sup>+</sup> T-cell subpopulation that has been causally linked to instances of aggravated GVHD after allo-HSCT (56). Donors carrying one of two genetic alterations in the non-coding regions of the *NLRP3* gene are associated with increased disease relapse and reduced overall survival but no alterations in GVHD severity in allo-HSCT patients (57). Thus, directed therapies targeting the NLRP3-inflammasome or depletion of specific DAMPs remain promising therapeutic options to reduce the level of systemic inflammation in the setting of allo-HSCT, but data reported so far are somewhat controversial and await further clarification.

In summary, NOD2 signaling in hematopoietic cells appears to protect from acute GVHD. Conflicting data from genetic association studies in humans are most likely attributable to differences in frequency of *NOD2* SNPs between patient cohorts, and differences with conditioning, immune suppression, and antibiotic protocols (44). We refer to Ref. (58) for a more detailed discussion of NOD2 in GVHD. Data on inflammasomes in allo-HSCT are not yet abundant, but NLRP3 and possibly other inflammasomes that sense endogenous danger signals such as ATP and uric acid and induce IL-1β release seems to have a role in the pathogenesis of acute GVHD.

#### **INNATE PATTERN-RECOGNITION RECEPTORS AND THE GRAFT-VERSUS-LEUKEMIA EFFECT**

Many studies have highlighted the fact that innate PRRs contribute to the inflammatory processes that lead to activation of allo-reactive T cells and the pathogenesis of GVHD. In contrast, the molecular details that shape the beneficial GVL effect remain poorly understood. Yet, only a detailed molecular understanding of the GVL effect will allow for the discrimination between GVLpathways and allo-immune reactions that drive clinical GVHD, a prerequisite for broader application of allo-HSCT in the future. Unspecific depletion or proliferative inhibition of donor T cells is believed to come at the cost of increased relapse of the underlying malignant disease (59). However, recent data challenge that view, since T-cell depletion via selection of CD34<sup>+</sup> cells in the allograft was found to be associated with markedly reduced GVHD but no differences in the rate of leukemic relapse (60, 61). Yet, data on the role of PRRs in GVL remain scarce. Studies that showed an association between loss-of-function SNPs in the *NOD2* gene and the severity of GVHD found no impact on the relapse rate by these same mutations (43, 45). Thus, NOD2 would seem to be an attractive pharmacological target to attenuate GVHD without interfering with the GVL effect. However, other studies that investigated the same *NOD2* SNPs in transplant patients and corresponding donors could not confirm their effect on GVHD pathogenesis (52), or showed an increased risk of relapse and death if recipients and/or donors were carrying such an alteration in the *NOD2* gene (62, 63). These contrasting results emphasize that data on differential regulation of GVHD versus GVL by PRRs on a systemic level are still premature and do not yet allow for systemic modulation of PRRs as a general treatment approach. In contrast, as PRRs can control the development of GVHD at a local level (34), their pharmacological manipulation in specific immune compartments seems to be a more promising approach. Interfering with PRR signaling in GVHD target tissues, such as intestine and skin, but sparing lymphoid organs and bone marrow, where residual hematologic malignancies reside, may allow to efficiently target GVHD but leaving GVL intact.

#### **CONCLUSION AND FUTURE DIRECTIONS**

Toll-like receptors and NLRs respond to a variety of microbial and endogenous danger signals and there is increasing evidence that they influence the development of acute GVHD. Yet, the role of TLRs in the pathophysiology of GVHD remains controversial, as studies with TLR4- and MyD88-deficient mice demonstrated that TLR signaling may not be absolutely required for the development of GVHD. Loss-of-function mutations in the *NOD2* gene, on the other hand, correlated in some studies with adverse allo-HSCT outcome in humans, suggesting a protective role of NOD2. Furthermore, activation of the NLRP3-inflammasome during early conditioning in mice contributes to the development of acute GVHD. Other receptors involved in the local control of microbiota will be the focus of future studies. Type I interferon has been shown to play an important role in defining the balance between GVHD and GVL responses (64). Thus, PRRs that detect cytosolic nucleic acids and lead to the production of large amounts of type I interferon such as the family of RIG-I-like helicases (5) or the recently discovered cytosolic DNA receptor cyclic GAMP synthase (cGAS) and its adapter STING (65) are of particular interest. Unraveling their role in acute GVHD will not only boost our understanding of this major complication after allo-HSCT, but may allow for novel therapeutic approaches to GVHD and related disorders like inflammatory bowel disease.

In light of the contradicting data regarding the role of some PRRs in acute GVHD, we would like to point out some of the major obstacles in the field of allo-HSCT research. Mouse models of GVHD are heterogeneous, with different subsets of immune cells being the main drivers of respective GVHD pathologies. In addition, innate and adaptive immunity are influenced by intestinal microbiota, which can vary critically between different breeding facilities. The effect of a given genetic alteration or therapeutic intervention may therefore differ between models and breeding facilities, and interpretation of such data must be undertaken with caution. Parts of the existing data may have to be revised in light of these new perceptions. Awareness of these difficulties together with increasing knowledge of graft and host immune and microbial physiology will, however, make this task easier in the future.

#### **ACKNOWLEDGMENTS**

This study was supported by the Deutsche Forschungsgemeinschaft (PO 1575/3-1 to H. Poeck) and the Else-Kröner-Fresenius-Stiftung (A61 to H. Poeck). We apologize to all authors whose work could not be cited due to space restrictions.

#### **REFERENCES**


active repression of peripheral blood toll-like receptor 5 mRNA expression. *Biol Blood Marrow Transplant* (2014) **20**(2):173–82. doi:10.1016/j.bbmt.2013.10.022


myeloid leukemia in complete remission undergoing HLA-matched sibling allogeneic hematopoietic cell transplantation. *J Clin Oncol* (2012) **30**(26):3194–201. doi:10.1200/JCO.2012.41.7071


**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.

*Received: 21 April 2014; paper pending published: 01 May 2014; accepted: 03 July 2014; published online: 18 July 2014.*

*Citation: Heidegger S, van den Brink MRM, Haas T and Poeck H (2014) The role of pattern-recognition receptors in graft-versus-host disease and graft-versusleukemia after allogeneic stem cell transplantation. Front. Immunol. 5:337. doi: 10.3389/fimmu.2014.00337*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Heidegger, van den Brink, Haas and Poeck. 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.*

## Toll-like receptor-4 modulation for cancer immunotherapy

#### **Shanjana Awasthi \***

Department of Pharmaceutical Sciences, University of Oklahoma Health Sciences Center, Oklahoma City, OK, USA \*Correspondence: shanjana-awasthi@ouhsc.edu

**Edited by:**

Anton G. Kutikhin, Russian Academy of Medical Sciences, Russia

**Reviewed by:**

Justin Lathia, Cleveland Clinic Lerner Research Institute, USA

**Keywords: toll-like receptor 4, inflammation, immune response, cancer, immunomodulation**

#### **INTRODUCTION**

Toll-like receptors (TLRs) are evolutionarily conserved pattern recognition molecules. Since the discovery of the Toll pathway cascade (1, 2), our knowledge about the structure, function, and mechanics of TLRs in infectious and inflammatory conditions has increased remarkably. The role of TLR4 as a pathogenpattern recognition receptor has been studied extensively. We now know that TLR4 recognizes pathogen-associated molecular patterns (PAMPs), such as Gramnegative bacterial lipopolysaccharide (LPS) and endogenous damage-associated molecular patterns (DAMPs) like fibronectin and hyaluronan, which are released during infectious and non-infectious inflammatory conditions. Some chronic infections and inflammatory conditions are known to promote carcinogenesis. For example, *Helicobacter pylori* (3) and viral hepatitis (4) infections lead to gastric and liver cancers, respectively. Also, in inflammatory bowel disease, non-infectious inflammation promotes the development of colorectal cancer (5). Evidence from recent reports suggests that increased expression and activity of TLR4 in chronic infectious and inflammatory conditions is associated with cancer progression (6–8). At the same time, additional studies suggest the protective role of TLR4 in cancer (9–14). The role of TLR4 in cancer has only recently been studied. This review article provides a brief summary of the current understanding of TLR4-signaling, its pro- and anticancer effects, and the therapeutic potential of TLR4 immunomodulation in the prevention and treatment of cancer.

#### **LIGAND RECOGNITION AND ACTIVATION OF TLR4**

Activation of TLR4 and downstream intracellular signaling involves interaction with TLR4 ligands, dimerization, and assembly of the TLR4-complex with its adaptor and co-receptor molecules. Our understanding of TLR4-signaling is based on the results of the studies focused on the interaction of TLR4 with *Escherichia coli*-derived LPS. It has been demonstrated that the LPS binding protein (LBP) transfers the LPS to CD14 that is present in soluble form or linked to the cell surface by a glycosylphosphatidylinositol anchor. LPS is then transferred from CD14 to the myeloid differentiation (MD2) protein (6). CD14 and MD2 do not have cytoplasmic tails and are unable to transduce signals on their own. The fatty acyl chains of lipid A of LPS are integrated into the hydrophobic pocket of MD2, and negatively charged phosphate groups on the diglucosamine backbone of lipid A interact with the charged residues at the opening of the binding pocket of MD2. The LPS–MD2 interaction with TLR4 then causes dimerization of the TLR4–MD2 in 1:1 ratio and assembly of TLR4–MD2–LPS complex. The crystal structure of human TLR4–MD2–LPS complex shows the formation of an "m" shaped oligomer made up of two molecules of TLR4 and two molecules of MD2 (15). **Figure 1A** provides a pictorial representation of the steps involved in the formation of the TLR4–MD2–LPS complex.

#### **TLR4-SIGNALING AND HOST DEFENSE MECHANISMS**

After the TLR4–MD2–LPS complex formation, TLR4 signals through myeloid differentiation primary response protein (MYD88)-dependent and Toll/IL-1R domain-containing adaptor inducing interferon-beta (TRIF)-dependent pathways. Dimerization of TLR4–MD2 recruits TIRAP (Toll/IL-1R domain-containing adaptor protein) and MYD88, leading to intracellular signaling, activation of transcription factors, and production of pro-inflammatory cytokines. TLR4 also recruits additional proteins, TRIF-related adaptor molecule (TRAM) and TRIF, and induces production of type I interferons. Type I interferons are associated with immune responses elicited by T and NK cells, important effector cells of adaptive immunity (16). During the late phase of signaling, activation of TRIF can also induce NF-κB transcription factor.

Immune response to any pathogenic stimuli includes activation of innate immunity, inflammation, and adaptive immunity. TLR4-signaling can eventually lead to a multitude of cellular effects (17). It is well-established that during the innate phase of immune response, TLR4 recognizes its ligands (pathogens, PAMPs, or DAMPs), and facilitates their uptake, intracellular processing, and the inflammatory response (18–21). After the TLR4 ligands are internalized and processed, the antigens are loaded onto the major histocompatibility complex (MHC) molecules for presentation to naïve lymphocytes. Published reports support the role of TLR4 in antigen-presentation and activation of cellular and humoral immune responses (22–25). **Figure 1B** summarizes the recognition of ligands by TLR4, TLR4-signaling through MYD88 and TRIF, and its role in inflammation and antigen-presentation. Thus, it is apparent that TLR4 is involved directly or indirectly with different arms of the host defense system (21, 26).

#### **TLR4 AND CANCER**

TLR4 is associated with cancer in several ways. Diverse cell lines and tissue samples derived from patients with head and neck, esophageal, gastric, colorectal, liver, pancreatic, skin, breast, ovarian, cervical, and breast cancer have been shown to express increased amounts of TLR4 (27).

Constitutive expression of some TLR4 genetic variants has also been linked to cancer (28–32). These characteristics are therefore being considered for their prognostic value in cancer treatment (32–34). In these scenarios of established cancer, TLR4 facilitates an environment that is suitable for continued cancer cell proliferation. Pro-cancer mechanisms could include the evasion of cancer cells from immune surveillance (35–38).

Persistent activation of TLR4-induced inflammatory signaling in chronic inflammatory conditions can also contribute to carcinogenesis (39). Experimental evidence suggests that cancer cell migration and invasion are induced by triggering of TLR4-NF-κB under inflammatory conditions (40–42). LPS-induced TLR4 signaling also promotes cancer cell survival and proliferation in hepatocellular carcinoma (43, 44). Moreover, the blockade of TLR4 by siRNA and NF-κB inhibitors decreases the invasive ability of cancer cells. Correspondingly, TLR4 silencing has been shown to decrease tumor burden in a murine model of colorectal metastasis and hepatic steatosis (45).

At the same time, published data suggest that TLR4 is required for protective immune response and killing of cancer cells. For example, TLR4-deficient mice developed more tumors after oral gavage with polyaromatic hydrocarbon 7,12-dimethylbenz(a)anthracene than did wild-type mice (46). Similarly, silencing of TLR4 increased breast cancer metastasis (47). Although mechanism is not fully understood, TLR4 can induce an efficient cancer antigen-specific cytotoxic T cell immune response (48). The cytotoxic T cells will eventually kill the cancer cells. The dynamics of the TLR4-induced immune parameters in the tumor microenvironment could be complex, and is not well studied. It is possible that TLR4 exerts pro- or anti-cancer effects, depending on the prevailing conditions in the tissue microenvironment during different phases of cancer development or metastasis.

#### **CURRENTLY AVAILABLE TLR4 IMMUNOMODULATORY AGENTS**

A number of immunomodulators, which target TLR4 have been developed. These modulators (antagonists or agonists) have been grouped based on their binding and sequestration of LPS, antagonizing LBP and CD14/LPS interactions, and targeting of MD2, TLR4–MD2, or TLR4.

Monophosphoryl lipid A (MPLA), a chemically modified derivative of LPS, is less toxic, and retains most of the immunostimulatory activity of LPS. MPLA serves as a TLR4 agonist. It has been approved in Europe as a vaccine adjuvant, and is a component of Hepatitis B and Human Papillomavirus Virus vaccines (49). Another lipid-based agonist, E6020 (Eisai/Sanofi Pasteur), has

also been developed as a vaccine adjuvant (50, 51). Other lipid molecules are being investigated for their potential to target the CD14–LPS interaction and antagonistic activity (52). Eritoran (E5564), developed by Eisai (Tokyo, Japan), directly binds to the hydrophobic pocket of MD2, competitively inhibits LPS from binding to MD2, and prevents the dimerization of TLR4, as well as TLR4-signaling (53). TAK-242, a cyclohexene derivative, was later developed by Takeda Pharmaceuticals (Tokyo, Japan) to target the TLR4 on the cellular membrane. Both TAK-242 and Eritoran (E5564) have been investigated in clinical trials as possible treatments for sepsis (54). Ibudilast (AV4II), a TLR4 antagonist, has been shown to suppress proinflammatory cytokines, such as TNF-α and IL-6, in neuroinflammation (55). Antibodies that target TLR4, NI-0101, and IA6 (NovImmune, Geneva, Switzerland), are being investigated for the treatment of acute and chronic inflammation. Glucopyranosyl lipid adjuvant-stable emulsion (GLA-SE; Immune design, Seattle, WA, USA), is also being studied (http:// www.clinicaltrials.gov). Although Eritoran and TAK-242 did not show efficacy for treatment of sepsis, a complicated clinical problem, studies with these modulators have clearly improved our understanding of the structural aspects of TLR4-complex formation and signaling.

#### **POTENTIAL OF TLR4 IMMUNOMODULATION FOR THE PREVENTION OR TREATMENT OF CANCER**

Agents with TLR4-antagonistic activity have been shown to reduce inflammationinduced carcinogenesis by suppressing the TLR4-induced NF-κB signaling. Curcumin, the main constituent of the spice turmeric, has been found to most likely bind to MD2, thus competing with LPS (56). A number of synthetic curcuminoids, such as EF24, have also been found to have anti-inflammatory activity (57–60). Our lab recently developed TLR4-interacting surfactant protein-A (SP-A) peptide, called SPA4, which binds to TLR4 protein in complex with MD2, and is effective therapeutically in cell culture systems and in a mouse model (61, 62). In the initial studies, our results showed that the TLR4 interacting SPA4 peptide suppresses LPS– TLR4-induced migration and invasion of colon cancer cells (63). More studies are warranted to understand the mechanism of SPA4 peptide activity. Other agents, including resveratrol (64),NI-0101 antibody (65), and paeoniflorin (66), have also shown suppression of inflammation-induced carcinogenesis.

While TLR4 antagonists could help reduce progression of inflammationinduced carcinogenesis or metastasis,TLR4 agonists have been shown to induce antitumor immunity in patients and models of established cancer. Lipid A-based TLR4 agonists, known as OM-174 and AS15, exhibit anti-cancer effects (67–70). Incorporation of the LPS and E6020 to Paclitaxel, whole cell tumor cell vector, and Trastuzumab improved the antitumor immunity in mouse models (71– 73). Picibanil (OK-432) targets both TLR2 and TLR4 and suppresses cancer (74). Vacchelli et al. recently published a detailed review of the ongoing clinical trials on TLR modulators, including TLR4 agonists. While the results from the ongoing clinical trials are pending, there is currently a significant emphasis on the design and development of novel TLR4 immunomodulators.

Although the potential of TLR4 immunomodulation for cancer immunotherapy has not been explored extensively, initial results from pre-clinical and clinical studies look promising. It is reasonable

to imagine a TLR4 immunomodulatory agent that reduces inflammatory response, but promotes anti-tumor immunity. This could be beneficial in controlling multiple stages of cancer. Comprehensive studies are therefore needed to understand the mechanism of action of TLR4 immunomodulators in appropriate *in vitro* and *in vivo* models of cancer.

#### **INFORMATION ABOUT PATENT APPLICATIONS PERTAINING TO TLR4 IMMUNOMODULATION BY SURFACTANT PROTEIN-A (SP-A) DERIVED PEPTIDES**

Patent applications have been filed on the concept of TLR4-interacting SP-A peptides for immunomodulation with United States Patent and Trademark Office (USPTO), World Intellectual Property Organization, European, Canadian, and Australian Patent agencies. A patent was recently issued by the USPTO (US 8,623,832; Inventor: Shanjana Awasthi; Assigned to the Board of Regents of the University of Oklahoma, Norman, Oklahoma).

#### **REFERENCES**


role of IL-17-producing T cells. *J Immunol* (2006) **177**(11):7980–9. doi:10.4049/jimmunol. 177.11.7980


presentation. *J Immunol* (2012) **188**(2):686–93. doi:10.4049/jimmunol.1102214


*J Pharmacol* (2007) **563**(1–3):1–17. doi:10.1016/j. ejphar.2007.02.018


74. Akeda T, Yamanaka K, Kitagawa H, Kawabata E, Tsuda K, Kakeda M, et al. Intratumoral injection of OK-432 suppresses metastatic squamous cell carcinoma lesion inducing interferon-gamma and tumour necrosis factor-alpha. *Clin Exp Dermatol* (2012) **37**(2):193–4. doi:10.1111/j.1365-2230. 2011.04151.x

**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.

*Received: 01 May 2014; accepted: 27 June 2014; published online: 25 July 2014.*

*Citation: Awasthi S (2014) Toll-like receptor-4 modulation for cancer immunotherapy. Front. Immunol. 5:328. doi: 10.3389/fimmu.2014.00328*

*This article was submitted to Tumor Immunity, a section of the journal Frontiers in Immunology.*

*Copyright © 2014 Awasthi. 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.*

### ADVANTAGES OF PUBLISHING IN FRONTIERS

FAST PUBLICATION Average 90 days from submission to publication

COLLABORATIVE PEER-REVIEW

Designed to be rigorous – yet also collaborative, fair and constructive

RESEARCH NETWORK Our network increases readership for your article

#### OPEN ACCESS

Articles are free to read, for greatest visibility

#### TRANSPARENT

Editors and reviewers acknowledged by name on published articles

GLOBAL SPREAD Six million monthly page views worldwide

#### COPYRIGHT TO AUTHORS

No limit to article distribution and re-use

IMPACT METRICS Advanced metrics track your article's impact

SUPPORT By our Swiss-based editorial team

EPFL Innovation Park · Building I · 1015 Lausanne · Switzerland T +41 21 510 17 00 · info@frontiersin.org · frontiersin.org