Pre-eclampsia (PE) is a serious clinical condition that occurs in 3-5% of pregnant women in the second half of pregnancy after 20 weeks and more frequently after 34 weeks of gestation and is characterized by hypertension, proteinuria, liver and cerebral involvement (1). The disorder is responsible for maternal morbidity and mortality and adverse pregnancy outcomes, including preterm delivery, intrauterine growth retardation and fetal death (2). The diverse clinical presentations of PE argue for a multifactorial nature of this syndrome whose pathogenesis has not yet been clearly defined despite the efforts made over recent years to identify the possible causes.

As the clinical signs disappear with the termination of pregnancy after placental expulsion, the placenta has been recognized to play a major role in the development of PE. This newly formed organ at the feto-maternal interface undergoes structural and functional changes that start with defective remodeling of decidual spiral arteries, particularly evident in the early-onset PE (3, 4), and endothelial dysfunction (5, 6) resulting in high resistance vessels, reduced tissue perfusion and oxidative stress (7).

These changes stimulate an inflammatory response in the second stage of PE involving both cells and soluble molecules of the innate immune system (8, 9).

Complement (C) is a critical component of innate immunity and has been implicated in the development of PE. The system plays an important role in host defense and homeostasis following activation via the classical, the lectin and the alternative pathways, but it may also cause tissue damage under conditions of unrestricted activation (10–12). While the alternative pathway functions as an amplification loop for the classical and lectin pathways, the activation of latter two pathways requires the intervention of recognition molecules to initiate the triggering process. C1q serves this function for the classical pathway and acts in association with the serine proteases C1r and C1s. Conversely, the lectin pathway is triggered by a set of different recognition molecules belonging to the collectin family (mannose-binding lectin aka MBL, collectin-10 aka CL-10 and collectin-11 aka CL-11) and the ficolin family (ficolin-1, ficolin-2, and ficolin-3). All the recognition molecules of the lectin pathway are found associated with a set of serine proteases named MASPs (MASP-1, MASP-2 and MASP-3) and regulatory molecules named MAPs (MAP-1 and MAP-2) (13). Activation of the classical and lectin pathways leads to the enzymatic cleavage of C4, C3 and C5, and eventually results in the release of potent anaphylatoxins, C3a and C5a, and the assembly of the C5b-9 complex. Evidence has been collected over the years suggesting an involvement of C in placental alterations of PE (14, 15). Increased levels of C3a, C5a and the soluble sC5b-9 complex have been reported in the circulation of preeclamptic patients (16, 17) and in the urine of patients with severe eclampsia suggesting their contribution to C-mediated renal damage (18). The finding of high levels of Bb, a split product of the C factor B of the alternative pathway, in the early phase of gestation in women who later developed PE led Lynch and colleagues to propose Bb as an early biomarker for the development of PE (19). The observation, however, was not confirmed in a prospective study performed in Caucasian patients (16). These conflicting results may be due to racial differences in the patients studied since the levels of Bb were found to be significantly higher in African-American PE patients (20).

Immunohistochemical analysis of PE placenta revealed increased deposition of C components and C activation products in chorionic villi compared to control placentae (21–25). Diffuse C4 staining has been found to be associated with placental pathologic changes and fetal growth restriction, and has been proposed as biomarker for adverse pregnancy outcomes in PE patients (23, 25). However, the pathway of C activation and the mechanisms involved in C-mediated tissue damage have not been fully elucidated. More direct evidence for C involvement in PE has been obtained in an animal model where CBA/J female mice mated to DBA/2 males have high resorption rate and fetal growth retardation (26). The pregnant mice manifest features common to PE, including albuminuria, endotheliosis and increased sensitivity to angiotensin II that correlates with adverse pregnancy outcomes (27). The fetal loss in the CBA/J female mice can be prevented by administering either an inhibitor of the C3 convertase, Crry-Ig, or a neutralizing anti-C5 monoclonal antibody or a C5a receptor antagonist (28). The interesting observation by Singh et al. that pregnant C1q-deficient mice present the classical manifestations seen in PE patients, including hypertension, albuminuria, and increased levels of soluble VEGF receptor 1 (sFlt-1), suggests that C1q may exert a surprising protective role in pregnancy (29). In this respect, it is important to note that C1q is widely distributed in maternal decidua in normal pregnancy and is locally synthesized and secreted by macrophages as well as endothelial cells and extravillous trophoblasts (30, 31). Considering the controversial findings in the literature herein we aimed to investigate the contribution of C1q expressed by perivascular trophoblasts to the physiologic process of decidual vascular remodeling and to explore the impact that a defect in the local expression of C1q may have on the vascular changes. In addition, we sought to determine whether locally expressed C1q is involved in the activation of the classical pathway at placental site, and whether C activation may also be triggered via the lectin pathway as observed in an animal model of PE (32).

Previous studies have documented C deposits on chorionic villi and in particular on villous trophoblasts of PE placentae, suggesting the contribution of C to the development of placental alterations in PE (21–25). We have now extended the C analysis to the decidual tissue based on the understanding that the pathologic process starts in the decidua involving primarily the vessels and subsequently affects the villi. The data presented in this study show that C is activated in PE placenta through the lectin pathway and that the role of C1q is to promote vascular remodeling in decidua rather than to activate the C classical pathway.

Examination of the placentae from healthy controls revealed a negligible presence of C components, most likely due to an inflammatory-like process that develops in placental decidua as a result of tissue remodeling associated with embryo implantation (8, 9). The only exception to this general rule is the diffuse deposition of C1q on decidual vessels and stroma in all placental samples examined. This is consistent with our previous findings that C1q is synthesized and secreted by endothelial cells and extravillous trophoblasts in the decidua in normal pregnancy (30, 31). The increased expression of C1q observed in decidua and on the surface of villi in the PE patients may be attributed to the binding of C1q to endothelial cells and trophoblasts as a result of cellular changes that occur in PE.

Different from C1q the other C components and the C activation products were focally distributed on decidual vessels and villous trophoblasts and were most likely deposited in areas of histopathologic alterations of the placental tissue. The localization of C4d and C3d in PE placenta is a clear indication of C activation that may reach completion with the deposition of terminal C complex revealed by a monoclonal antibody directed against a neoantigen of C9 expressed in C5b-9 complex (39). Linear deposits of C4d on syncytiotrophoblasts were reported to be associated with lower gestational age and fetal growth restriction in PE patients suggesting the contribution of C activation to the development of these clinical manifestations (23, 25). However, it is unclear whether C activation is a secondary event triggered by cells undergoing apoptosis or is directly responsible for trophoblast damage. It should be pointed out that syncytrophoblasts are relatively resistant to C-mediated killing, which requires the presence of cytotoxic antibodies and a marked decrease in the expression of the C regulators CD46 and CD59 (40). C dependent cytotoxic antibodies are infrequently found in PE patients except for those with antiphospholipid syndrome associated with PE (41). In addition, C regulators have been reported to be normally expressed both at molecular and protein levels in the placenta of PE patients (23, 24). However, we cannot exclude a sublytic effect of C on syncytiotrophoblasts resulting in the release of microparticles that circulate in maternal blood in normal pregnancy and at increased level in PE patients (42).

Most likely C activation plays a more critical role in decidua promoting inflammation and stimulating a hypercoagulable state that are both features of maternal abnormalities implicated in the second stage of PE pathogenesis (43). In this context C5a and C5b-9 have a significant impact on vascular endothelium by inducing release of proinflammatory cytokines, enhancing the expression of tissue factor and increasing vascular permeability (44). Whatever the mechanism involved in C-mediated placental damage in human preeclampsia, C activation has been shown to play an essential role in a mouse model of PE causing adverse pregnancy outcomes that can be prevented by the administration of C inhibitors (28, 32, 45).

The co-localization of C1q and C4 on syncytiotrophoblasts of PE placentae in the absence of MBL led Buurma and colleagues (23) to suggest that C is activated via the classical pathway possibly triggered by the binding of antibodies, although the association between C4d and immune complex deposits was not statistically significant. Our failure to detect C1r and C1s deposition at sites of C1q and C4d deposition in this study rules out the direct activation of the classical pathway and favors the activation of the lectin pathway. Consistent with this is our previous report that pregnancy loss in a mouse model of PE is mediated by MBL-dependent activation of the lectin pathway (32). However, we could not detect MBL in human PE placentae indicating that the results obtained in mice do not fully replicate the observations in PE patients as suggested by the different molecular organization of human and mouse MBL. Although mice and higher primates possess two MBL genes, MBL1 and MBL2, only one MBL gene (MBL2) give rise to protein in man, while MBL1 is expressed as pseudogene (MBL1P1) (46). By contrast, the two forms of MBL in mice are encoded by two distinct functional genes, known as mbl-a and mbl-c, and, consequently, the protein concentration of mouse MBL may be 10-fold higher than that of human MBL. Another important point to consider is that ficolin-3, the most abundant and most potent lectin pathway activator in man, is not present in mice, making a direct comparison of the lectin pathway activity between mice and humans difficult (47). Thus, the absence of MBL does not exclude activation of the lectin pathway by other initiators including ficolins and collectins as suggested by the strong staining of decidual vessels and villous syncytiotrophoblasts for ficolin-3, MASP-1, and MASP-2. The finding of ficolin-3 deposited on villous syncytiotrophoblasts of PE placentae is in agreement with the observations by Wang and colleagues (48), who reported binding of ficolins to cells undergoing apoptosis and suggested that they may contribute to promote local inflammatory responses. While we are not excluding this possibility, our data showing deposition of MASPs and other C components on syncytiotrophoblasts indicate that ficolin-3 most probably acts as trigger of the lectin pathway. Ficolin-3-mediated C activation is likely to plays an important role in decidual vascular endothelium by contributing to endothelial activation and dysfunction as documented by the increased levels of endothelial-derived factors observed in PE patients, including inflammatory cytokines (6) and soluble adhesion molecules (49). Based on our results, we would speculate that the ficolin-3 gene, FCN3, may be turned on locally, particularly in the vascular bed, in response to different stimuli (50) and contributes to the pathophysiologic mechanisms causing PE in humans, a process that cannot be replicated in murine models (51). The subarachnoid hemorrhage in humans is another example of human pathologic condition associated with ficolin-3-dependent C activation (52).

We have previously provided evidence for an alternative role of C1q in normal pregnancy by showing impaired remodeling of the decidual vessel in C1q deficient pregnant mice (31), a recognized animal model of PE (27), and reduced levels of C1q mRNA transcripts in PE placental tissue (53). The finding that the unremodeled decidual vessels more frequently observed in PE patients are surrounded by trophoblasts that do not stain for C1q indicates that the vascular changes are associated with the expression of C1q in these cells. This is consistent with the observation that the defective remodeling of blood vessels in C1q deficient pregnant mice can be reversed by mating these animals with C1q sufficient mice, resulting in an increased number of remodeled vessels and the perivascular appearance of C1q-expressing trophoblasts. The EVT that invade the decidua in heterozygous and normal animals share the ability to synthesize C1q. The decreased and restricted expression of C1q to perivascular trophoblasts in decidual EVT from PE patients could be explained by the special microenvironment organized around the remodeled vessels that helps C1q-containing EVT to exert disruptive activity on the vascular wall (54). The mechanisms by which C1q modulates EVT effector functions remain to be defined.

In conclusion, we have shown that C components and C activation products are deposited on placental tissue obtained from PE patients but are absent in the control placentae except for C1q that is constitutively expressed at the placental level in physiological pregnancy. Our data support the notion that in human PE C is locally activated through the lectin pathway triggered by ficolin-3 whilst MBL does not appear to be involved. Equally, C1q is not implicated in the activation of the classical C pathway and should be considered a biomarker of EVTs that contribute to decidual vascular remodeling.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The study was reviewed and approved by the Regional Ethical Committee of FVG (CEUR), Udine, Italy (CEUR-2020-Os-156; Prot. 0022668/P/GEN/ARCS). The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by The UK Home Office.

FT, BB, and AM designed the study, analyzed the results, and wrote the manuscript. PG, MB, and RB provided resources, expertise and critically reviewed the manuscript. AG, VC, and GM performed the immunohistochemical staining. GR and FF organized patient recruitment. AM and CA prepared figures and tables and edited the manuscript. All authors contributed to the article and approved the submitted version.

This research was supported by grants from the Institute for Maternal and Child Health, IRCCS Burlo Garofolo, Trieste, Italy (RC20/16, RC23/18, 24/19 to G.R. and RC09/21 to CA).

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.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.