Edited by: Luis Graca, University of Lisbon, Portugal
Reviewed by: Bin Li, Chinese Academy of Sciences, China; Lennart T. Mars, National Institute of Health and Medical Research, France
†Anna Inkeri Lokki and Jenni Heikkinen-Eloranta have contributed equally to this work.
This article was submitted to Immunological Tolerance, a section of the journal Frontiers in Immunology.
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.
Preeclampsia (PE) is a common disorder of pregnancy originating in the placenta. We examined whether excessive activation or poor regulation of the complement system at the maternal–fetal interface could contribute to the development of PE. Location and occurrence of complement components and regulators in placentae were analyzed. Cryostat sections of placentae were processed from 7 early-onset PE (diagnosis <34 weeks of gestation), 5 late-onset PE, 10 control pregnancies, and immunostained for 6 complement activators and 6 inhibitors. Fluorescence was quantified and compared between PE and control placentae. Gene copy numbers of complement components
Preeclampsia (PE) is a serious complication of human pregnancy. It can lead to multi-organ dysfunction and, rarely, to a life-threatening convulsive condition, eclampsia (
The complement (C) system is a phylogenetically ancient means of self–non-self discrimination. It plays a central role in innate immune defense, clearance, and as a mediator of the adaptive immunity (
Disturbances in C activity can predispose to infections or to a systemic lupus erythematosus (SLE)-like immunoinflammatory syndrome. The latter has been related to an inadequate waste disposal function of the classical pathway (
The depth of placentation required for a healthy human pregnancy presents a unique challenge to regulation of the maternal immunological processes (
A well regulated C system is a prerequisite for a healthy pregnancy (
Two forms of
To test the involvement of C in PE, we have analyzed immunohistochemically the deposition and expression of key activating components and regulators of the C system in preeclamptic placentae in relation to disease onset and in comparison to healthy placentae. The results favor the hypothesis that an insufficient complement function is linked to an inability to clear away trophoblast material from the placenta. As a consequence, the material deposits in fibrinoid clusters and could cause an endothelial–vascular disorder in the maternal circulation.
For this study, we chose randomly 12 women with PE and 10 controls without PE (Table
Controls | Late-onset | Early-onset | |
---|---|---|---|
PE, |
PE, |
||
Age | 30.6 ± 3.1 | 33.8 ± 4.1 | 31 ± 6.5 |
Gravidity | 1.8 ± 0.8 | 1.2 ± 0.5 | 1.9 ± 1.1 |
Parity | 0.6 ± 0.8 | 0.2 ± 0.5 | 0.6 ± 1.0 |
Maternal BMI (kg/m2) | 22.9 ± 2.7 | 21.0 ± 2.2 | 22.3 ± 2.4 |
Hypertension before pregnancy | 1/10 | 1/5 | 2/7 |
Celiac disease | 0/10 | 1/5 | 0/0 |
Thrombophilia | 0/10 | 0/5 | 1/7 |
PE in previous pregnancy | 0/10 | 1/5 | 1/7 |
Early pregnancy systolic BP (mmHg) | 114 ± 8 | 117 ± 7 | 129 ± 10 |
Early pregnancy diastolic BP (mmHg) | 73 ± 8 | 77 ± 5 | 82 ± 9 |
Highest systolic BP (mmHg) | 128 ± 13 | 166 ± 10 |
167 ± 16 |
Highest diastolic BP (mmHg) | 87 ± 10 | 109 ± 8 |
118 ± 9 |
Highest proteinuria (g/24 h) | – | 1.8 ± 0.4 | 5.7 ± 3.8 |
Gestational weeks at birth | 40 ± 2 | 38 ± 2 |
33 ± 4 |
Birth weight (g) | 3646 ± 282 | 2938 ± 423 |
1842 ± 544 |
Complications | |||
IUGR | – | – | 3/7 |
Placental insufficiency | – | – | 2/7 |
HELLP | – | 1/5 | 1/5 |
Preeclampsia was defined as hypertension and new-onset proteinuria occurring after 20 weeks of gestation. Hypertension was defined as systolic blood pressure of 140 mmHg or more, and/or a diastolic blood pressure of 90 mmHg or more after 20 weeks of gestation. Proteinuria was defined as the urinary excretion of ≥0.3 g protein in a 24-h specimen, or 0.3 g/l or, in the absence of concurrent quantitative measurement, at least a “2+” or more, or two “1+” proteinuria dipstick readings with no evidence of urinary tract infection. PE was considered severe if blood pressure was ≥160/110 mmHg, or proteinuria exceeded 5 g/24 h, or symptoms like cerebral or visual disturbances or abdominal pain appeared. Intrauterine growth restriction (IUGR)/placental insufficiency was defined as birth weight below −2SD and/or umbilical artery resistance ≥+2SD according to gestational age specific standards (
Placental samples were chosen preferentially from patients with severe and early-onset PE. Chronic hypertension (an elevated blood pressure that predated the pregnancy or detected before mid-pregnancy) was observed in three PE women and one woman in the control group (Table
Approximately 1 cm wide tissue samples from the placentae were dissected using a scalpel and scissors and placed in a cryotube for preservation. Following the nine-site procedure, the placenta was visually divided into nine pre-specified regions and one sample was taken from each region. Within 2 h of the delivery of the placenta, the cryotubes containing samples were placed into the inner compartment of a nested metal holder. Approximately 150 ml of liquid nitrogen was added to the outer compartment to cool down the 150 ml isopropanol poured into the inner compartment. According to a standardized tissue-preserving collection procedure, the samples were left for 20 min to freeze slowly through the isopropanol pool. When isopropanol reached a floury frozen state, the cryotubes were stored at −80°C. For our study, one region (no. 5) was immunohistochemically analyzed from all samples.
The antibodies used are listed in Table
Antibody | Type | Dilution | Source |
Role |
---|---|---|---|---|
C1q | Rabbit pAb | 1:1000 | DAKO | CP component |
C4c | Rabbit pAb | 1:400 | DAKO | CP component |
C4bp | Sheep pAb | 1:200 | The Binding Site | CP regulator |
CRP | Mouse mAb | 1 μg/ml | Fitzgerald | CP activator |
C3c | Rabbit pAb | 1:1000 | DAKO | AP component |
C3d | Rabbit pAb | 1:1000 | DAKO | AP component |
Factor H | Goat pAb | 1:400 | Calbiochem | AP regulator |
C9 | Goat pAb | 1:400 | Quidel | TP component |
MCP (CD46) | Mouse mAb | 1 μg/ml | IBGRL | AP and CP regulator |
Bric 230 (CD55) | Mouse mAb | 1 μg/ml | IBGRL | AP and CP regulator |
Bric 229 (CD59) | Mouse mAb | 1:200 | IBGRL | TP regulator |
CR1 | Mouse mAb | 1 μg/ml | AbD Serotec | AP and CP regulator |
s-Endoglin | Mouse mAb | 2 μg/ml | Santa Cruz | PE indicator |
Anti-C3c antibody was used to detect C3b and iC3b, which are the products of alternative pathway activation and amplification and of subsequent C3b inactivation. C3d fragment was separately stained for because the C3d antibody recognizes the C3dg fragment, which remains surface bound after the release of C3c. FH binds to the C3b molecule on the self-cell surface, where it can be detected by the FH antibody. The C1q antibody recognizes several different structures of the classical pathway activating C1q molecule. The C4c antibody recognizes the native C4 molecule (both C4A and C4B) as well as the activation product C4b and its inactivated form iC4b. For detection of C4bp, a cofactor for C4b in activation, a polyclonal sheep antibody was used. Membrane-bound DAF (CD55) and MCP (CD46) were analyzed by specific mouse monoclonal antibodies.
The frozen tissue samples were cryosectioned at 5 μm and when possible two or three serial sections were laid per each slide. The dried sections were rinsed with phosphate-buffered saline (PBS) and moist samples were blocked against non-specific binding with 1% bovine serum albumin (BSA) in PBS for 15 min in a humid chamber. Excess liquid was discarded and the first antibody was pipetted to the sample in 1% BSA/PBS as detailed in Table
One slide from each sample was used for standard automated hematoxylin and eosin (HE) staining to ensure diagnostic-level consistency and to obtain a histological reference point for the immunofluorescence (IFL) analyses.
Data were collected using standardized fluorescence microscopy settings, where all slides were photographed with 10×, 20×, and 40× magnifying objectives. Exposure times per each magnification were 20, 50, and 83.3 ms. Images were collected using Olympus DP Manager (ver. 2.2.1.195) and Olympus DP Controller (ver. 2.2.1.227) image capture softwares with Olympus BX51 fluorescence microscope camera.
The same protocol and machinery for imaging was used for histochemistry preparations. Images captured from HE stained samples were used to identify key structures and to characterize the placenta. The structures to be identified included stem villi, small villi, villous fibrinoid (i.e., as a part of a villus, often replacing syncytium), fibrinoid necrosis, syncytiotrophoblast (STB), cytotrophoblast, fetal arterial endothelium, and syncytial bodies (incl. syncytial sprouts and syncytial knots). The occurrence of syncytial bodies and fibrinoid structures was semiquantified by grading (0 = none observed, 1 = counted 1–3, 2 = >3, 3 = all over, cannot be counted). Structural integrity was measured by intactness of syncytium and special attention was paid on shedding of syncytial cells. Signs of nuclei of apoptotic cells were looked for and noted.
To correlate the C4 deposition observed in the placentae with the functional
ImageJ 1.46 and Fiji-win32 softwares were used to quantify the intensity of fluorescence in the fixed magnification images. These were chosen to minimize the variation of staining quality and tissue quality between individuals, which was more apparent at the highest levels of magnification. To correct for false positive readings resulting from background autofluorescence, mean intensity +1 SD
Activating components and regulators of the C system as well as s-endoglin were found to be deposited in the placenta in a structure-specific manner showing differences between patient groups and controls. In the following, the results are presented according to C pathways (Figure
C1q was observed at the STB layer in 5/10 controls but in none of the early- onset PE patients and only in one of the late-onset PE patients. Overall, we found less C1q in the PE patients when compared with normal controls (Figures
C4A or C4B deficiency |
C4A deficiency |
C4B deficiency |
||||
---|---|---|---|---|---|---|
Maternal | Fetal | Maternal | Fetal | Maternal | Fetal | |
PE (pooled) ( |
0.667 | 0.700 | 0.417 | 0.300 | 0.333 | 0.400 |
Early-onset PE ( |
0.714 | 0.667 | 0.429 | 0.333 | 0.286 | 0.333 |
Late-onset PE ( |
0.600 | 0.750 | 0.400 | 0.250 | 0.400 | 0.500 |
Control ( |
0.375 | 0.500 | 0 | 0.375 | 0.375 | 0.250 |
C4 deposits were observed mainly in the STB layer, either in the apical membrane or throughout the syncytium. There was no clear difference between the patient groups and controls in this pattern (Tables
Type of PE onset | C3b/iC3b |
C1q |
C4bp |
C4 |
|||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Control |
Control |
Control |
Control |
||||||||||
df | df | df | df | ||||||||||
Sum | Late | 1.735 | 0.106 | 13 | |
−0.487 | 0.635 | 13 | 0.969 | 0.352 | 12 | ||
Early | −0.671 | 0.512 | 15 | −0.530 | 0.604 | 15 | −0.667 | 0.516 | 14 | 0.486 | 0.635 | 14 | |
Mean | Late | 1.718 | 0.110 | 13 | −0.374 | 0.714 | 13 | 0.703 | 0.495 | 12 | |||
Early | −0.607 | 0.553 | 15 | −0.091 | 0.929 | 15 | −0.507 | 0.620 | 14 | 0.432 | 0.672 | 14 | |
Sum | Late | 0.059 | 0.954 | 13 | −0.217 |
0.836 | 5.463 | 0.825 | 0.424 | 13 | |||
Early | −0.913 |
0.378 | 12.533 | −0.365 |
0.723 | 9.333 | −0.129 | 0.899 | 15 | 1.524 | 0.150 | 14 | |
Mean | Late | 0.067 | 0.947 | 13 | 0.097 | 0.924 | 13 | 0.722 | 0.454 | 13 | |||
Early | −0.812 |
0.431 | 12.927 | −0.009 | 0.993 | 15 | −0.215 | 0.833 | 15 |
Type of PE onset | C3b/iC3b |
C1q |
C4bp |
C4 |
|||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Late |
Late |
Late |
Late |
||||||||||
df | df | df | df | ||||||||||
Sum | Early | 1.219 | 0.251 | 10 | 0.176 | 0.864 | 9 | 0.639 | 0.537 | 10 | |||
Mean | Early | 1.297 | 0.224 | 10 | 0.127 | 0.902 | 9 | 0.364 | 0.723 | 10 | |||
Sum | Early | 0.773 | 0.457 | 10 | 0.074 | 0.943 | 10 | 0.857 | 0.412 | 10 | 0.588 | 0.569 | 10 |
Mean | Early | 0.722 | 0.487 | 10 | 0.085 | 0.934 | 10 | 0.810 | 0.437 | 10 | 0.511 | 0.621 | 10 |
C4bp is an inhibitor of the classical pathway occurring usually physiologically in complex with the anticoagulant protein S. Overall, C4bp was found deposited particularly in small syncytial bodies, which appeared as brightly staining particles attached or sometimes shed from the syncytium (see Figures
C-reactive protein and Complement Receptor type 1 (CR1; CD35) were tested with four samples representing one early-onset PE, two late-onset PE and one control specimens. Both staining were negative. CRP and CD35 were subsequently omitted from the protocol.
In general, C3 (C3b, iC3b) detected by an antibody against C3c was abundantly present in the placenta. It was found in the STB layer and in the villous stroma (Figures
The intensity of C3 deposition appeared weaker in the PE group and even more so in the late-onset group when compared with normal pregnancy (Figures
The samples were stained separately also for C3d because it is the C3 activation product that remains covalently bound in areas of C3b deposition thus reflecting a longer period of C activation. C3d deposition was also present in the placenta (Figures
dg | C3b/iC3b |
C1q |
C4bp |
C4 |
|||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Control |
Control |
Control |
Control |
||||||||||
df | df | df | df | ||||||||||
Sum | PE | −1.417 | −0.172 | 20 | −1.695 | 0.106 | 20 | −0.775 | 0.448 | 19 | 0.848 | 0.407 | 19 |
Mean | PE | −1.366 | 0.187 | 20 | −1.493 | 0.151 | 20 | −0.602 | 0.554 | 19 | −0.664 | 0.515 | 19 |
Sum | PE | 0.458 | 0.652 | 20 | 0.045 | 0.964 | 20 | −0.411 | 0.685 | 20 | |||
Mean | PE | 0.407 | 0.689 | 20 | −0.398 |
0.695 | 18.961 | −0.330 | 0.745 | 20 |
Factor H is a key regulator of the C amplification cascade in the alternative pathway. There was significant variation in the FH staining patterns in different placentae (Figures
In certain PE placentae, only the STB layer of some villi and fibrin clusters had become strongly stained for FH, while the rest of the tissue remained negative (Figure
Component 1 | Component 2 | Pearson’s |
||
---|---|---|---|---|
Early-onset PE | ||||
DAF | C4 | 0.824 | 0.023 | |
C3b/iC3b | s-Eng | 0.829 | 0.021 | |
Late-onset PE | C1q | FH | 0.021 | |
C4 | C9 | 0.921 | 0.026 | |
Control | DAF | MCP | 0.634 | 0.049 |
C4bp | MCP | −0.659 | 0.038 | |
FH | C9 | 0.671 | 0.034 |
The terminal complex of C activation was assessed by staining for tissue associated C9. It could represent deposited polymeric MAC or tissue-bound SC5b-9 complexes. Positive staining for C9 was found in distinct regions of the placentae (Figures
The membrane-bound regulators of complement MCP, DAF, and CD59 were typically observed in a dual pattern, where both the apical and the basal layer of the STB stained positive. Membrane regulators were mainly observed in the STB layer and to a lesser extent in the villar endothelium. DAF showed the weakest stainings of the membrane regulators. Unlike the soluble regulators, membrane-bound regulators were not observed in the syncytial bodies (Figure
We found s-endoglin to be significantly more abundant in PE pregnancies than in healthy controls (Table
In PE, villi in 83% (10/12) of the placentae had distinct and circumferential endoglin deposition on the apical sides of the STB, sometimes penetrating through to the basal layer. Villi with negative staining in the STB were scarce (Figures
Differences between women with and without PE were seen in the classical pathway of C activation and in the binding of protective C regulators to the placental structures. The critical differences were mostly observed at the STB layer and in lesions representing injured tissue structures. C1q deposition on the STB was least abundant in the late-onset PE group. Interestingly, in our small study sample we observed that
Abundant evidence suggests that immunological mechanisms are involved in the various steps of PE pathogenesis. These include an incomplete spiral artery remodeling that causes poor placental development and creates turbulent and constrained blood flow to the villi (
The overall deposition of C1q was stronger in the early-onset than in the late-onset PE, which may reflect the difference in the etiopathogenesis between these two patient groups. Placental dysfunction is typically observed in the early-onset disease. We observed a higher frequency of necrotic large villi and an increased number of fibrinoid necrotic areas in the early-onset PE placentas (
Our results support the theory, that C1q has an important role in the maintenance of immune tolerance by clearing apoptotic and self-antigens. C1q has an important ability to recognize altered or exposed structures of self thereby leading to their efficient clearance by phagocytes without lysis and inflammation (
High-intensity areas of C1q deposition were negatively correlated with the soluble regulator FH in the late-onset PE group. This could be a reflection of a C1q/FH balance. Our results suggest that C1q could also bind to structures exposed by a turbulent blood flow during PE. These structures could be within the connective tissue, in vascular endothelia or on the trophoblastic cells. It was recently shown that FH can bind independently from and even compete with C1q for binding to apoptotic surfaces and other targets (
We found an increasing frequency of
C4bp is a major soluble protein that binds to C4b and regulates the classical pathway C3 convertase. We observed C4bp binding typically and intensely to apoptotic fragments and structures including shed and knotted syncytium. Because of this, it was not possible to quantitatively differentiate if fluorescence intensity levels in relation to disease status were due to the different levels of physical damage of the placenta or to truly different levels of C4bp (
Factor H is the most important regulator of the alternative pathway. Generally, FH was observed in abundance in most of the placentae. The absence of STB staining for FH in the PE cases, a trend which was also apparent with other C components, is likely due to the loss of surface negative charge and/or disturbed structure of the STB in the preeclamptic placentae. In clusters of C3b deposition, e.g., in areas of STB damage, FH would also bind. Considering that FH is a soluble molecule, it seems atypical that in 83% of cases and 40% controls, intense FH deposition was observed in the stromas of the tissue. This suggests an increased requirement of the tissue for protection from C attack in majority of the placentae from preeclamptic pregnancies. Interestingly, FH is known to be produced extrahepatically in certain tissues, including the placenta (
C3 is the most central component of the C system. We found C3 abundantly in the placenta. In the high-intensity area analysis, C3b/iC3b deposition correlated with C1q and negatively with FH deposition in the early-onset PE patients. This suggests that in the most difficult cases of PE where, the alternative pathway has become activated and, as a result, C3 deposits were observed, regulation by FH had failed to protect the placenta. This relationship was missing in the late-onset PE and control groups. There were no clear differences in the intensity or distribution of C3d between placentas from PE patients and healthy controls. The kinetics of C3d deposition differs from that of C3b and iC3b. C3d is usually found as a remnant from prolonged complement activation, especially in basement membranes.
C3d was localized most clearly to the sub-syncytium basement membrane. There was no difference in the intensity of C3d expression between the control and PE groups. This is in contrast to Sinha et al., who observed that more C3d deposition occurred in the trophoblast basement membrane in PE than in normal placentae, and that the positivity was more marked in the severe PE group than in patients with mild PE (
While MCP is an important inhibitor of C activation in the healthy maternal–fetal interface, C4bp apparently has a different role in binding mainly to apoptotic structures and damaged STB. This was reflected by a negative correlation between C4bp and MCP intensity in the controls. The correlation between MCP and DAF (CD55), which was observed in the high-intensity areas of the control placentae showed that these regulators synergize each other in the regulation of C activation in the third trimester placenta. They are probably both needed for the control of C activation on the syncytium. Membrane-bound regulators of complement DAF, MCP, and CD59 have been observed in the healthy first trimester and term placentae, where they serve to protect the developing placenta from C attack (
Increased endoglin staining was observed in PE samples. Our results corroborate the findings of Sitras et al. and Nishizawa et al. who both found increased levels of endoglin in PE, but could not differentiate late- and early-onset patient groups with regard to endoglin expression (
Here, we have described for the first time the expression of a large panel of C system components in the placentae of PE pregnancies. It is apparent that complement is involved in multiple ways in both normal pregnancy as well as in PE.
A notable difference between two soluble C regulators was observed. C4bp was found to bind directly to apoptotic syncytial structures while fetal FH apparently provides an overall, broad scale protection to the placental tissue.
Correlation of MAC deposition with the classical pathway activating components in the patient groups and negative correlation in the controls may be indicative of C regulation breakdown and/or uncontrolled classical pathway activation in PE.
In our small cohort partial
Anna Inkeri Lokki and Jenni Heikkinen-Eloranta did the laboratory work for the project, analyzed the data, and drafted the manuscript. Jenni Heikkinen-Eloranta described the patient material with help from Hannele Laivuori and Terhi Saisto. Hannele Laivuori is the head of the FINNPEC board and together with Terhi Saisto and Jenni Heikkinen-Eloranta she collected the samples and clinical data. Hanna Jarva and Seppo Meri are experts in the complement system and participated in planning the laboratory methodology (Hanna Jarva) and providing the materials and laboratory space (Seppo Meri). Marja-Liisa Lokki is a specialist in the field of MHC genetics and planned, analyzed, and helped to interpret the C4 genetic portion. Hannele Laivuori and Seppo Meri came up with the study question planned the project in collaboration with Anna Inkeri Lokki and Jenni Heikkinen-Eloranta.
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.
We wish to thank Dr. Riikka Tulamo for her advice in the IFL methodology, and Susanna Mehtälä and Dr. Marcel Messing for technical support in IFL analyses. We also acknowledge Heikki Lokki of Department of computer sciences, University of Helsinki for statistical consultation on the high-intensity quantification. Furthermore, the board of investigators, study nurses, and women of the FINNPEC cohort are acknowledged, their participation made this study possible. This research was funded by Academy of Finland (#137529 Jenni Heikkinen-Eloranta, #121196 and #134957 Hannele Laivuori), Jane and Aatos Erkko Foundation, Päivikki and Sakari Sohlberg Foundation, Finnish Medical Foundation, University of Helsinki and State subsidy for Health Research (EVO) (Hannele Laivuori) Sigrid Jusélius Foundation, the Stockmann Foundation and State subsidy for Health Research (EVO TYH2012237) (Seppo Meri). Study conducted at: Haartman Institute, Department of Bacteriology and Immunology, University of Helsinki, Helsinki, Finland.