Introduction

For an immunologist, pregnancy remains an enigma. As half the genes are derived from the father, the fetus and the placenta must be considered a ‘semi-allograft’. Becoming pregnant is not that different from receiving a transplant from a genetically different individual. Thus, mothers need to adapt to imprinted antigens from the father in order to maintain a successful pregnancy [1, 2]. Although intimate contact with the mother’s uterine tissue ensures optimal nourishment and protection of the fetus throughout its early development, it makes the fetus and placenta potential targets for her immune system. The mismatched ‘organ transplant’ would be quickly rejected without substantial immune suppression [2]. Failure of the placenta to escape maternal immune response can cause pregnancy complications such as miscarriages and preeclampsia [3]. In the majority of cases, the placenta and the fetus avoid the attack of the maternal immune system. Why are the placenta and fetus not rejected in the uterus? Many mechanisms protect the fetus from attack by the maternal immune system. These include the lack of expression of polymorphic classical class I human leukocyte antigen molecules (HLA-A and B) in invasive fetal trophoblasts [4, 5], tryptophan catabolism by the enzyme IDO that suppresses T cell activity [6], the existence of particular NK cells in the uterus that are proangiogenic instead of cytotoxic [7, 8], and a tight regulation of complement system activation [9, 10].

Complement split product C3b is continually deposited on all surfaces in the body that are in contact with plasma. In the absence of natural protective factors, amplification occurs with the deposition of more C3b and other components and generation of anaphylatoxins C3a and C5a that attract phagocytic cells—macrophages and neutrophils—and initiate tissue destruction. Normal host cells are protected from the harmful effects of complement by cell surface complement regulatory proteins [11]. In humans, decay-accelerating factor (DAF) and membrane cofactor protein (MCP) prevent complement activation by inhibiting C3 and C5 convertases [12]. In other species, including mice, structural and/or functional homologs of these proteins can be found. The mouse protein complement receptor 1-related gene/protein y (Crry/p65, Crry) is a self-protecting complement regulatory proteins. Crry demonstrates cofactor activity for factor I-mediated cleavage of both mouse C3b and C4b. In addition, Crry also exhibits decay-accelerating activity for the classical pathway C3 convertase. Mouse Crry thus uses the specific mechanisms of both human MCP and DAF to inhibit complement activation [13]. Complement inhibition at the fetomaternal interface is crucial for the maintenance of pregnancy (Fig. 1) [14].

Fig. 1
figure 1

Appropriate complement inhibition is an absolute requirement for a pregnancy to be successful. Complement inhibitory proteins (MCP and DAF in humans and Crry in rodents) are expressed in the placenta, throughout gestation and control complement activation protecting pregnancies. Figure modified from [16]

Full size image

We have learned much about the crucial contribution of complement inhibitors to a successful pregnancy from mouse studies in the Molina Lab. The important role of Crry in protecting pregnancies in mice was discovered by serendipity while investigating whether the murine gene, complement receptor-related gene Y (Crry), and its product Crry, contributed to the protection of the body from damage inflicted by inflammation [15]. To elucidate Crry function in vivo, Molina’s group created Crry knockout mice. By mating Crry +/− with Crry +/−, the investigators expected that around 25% of the offspring from these mice would be null for Crry and thus have no surface inhibition of C3 or C4 activation. Instead they found no homozygous Crry deficiency among the offspring of these mice, all of the animals that inherited two mutant copies of Crry died as embryos at ~ 10 days post-conception. Although Crry is not the only membrane complement regulator in mice, it appeared that its absence was sufficient to leave the embryo unprotected against attack from maternal complement [16, 17]. Normal mice express early in pregnancy large amounts of Crry on trophoblast, highlighting the importance of complement inhibition at the feto-maternal interface [17]. That Crry −/− mice could be rescued on a factor B (fB)- or C3-deficient maternal background confirms that Crry−/− embryos die in utero due to an uncontrolled complement attack [18]. These mouse studies emphasize the important role of complement inhibition in the maintenance of pregnancy.

The presence of complement inhibitors in the human placenta is well documented. Most of the studies aimed to investigate the presence of complement regulatory proteins in human placentas were performed either in term placenta obtained after normal birth or early placenta obtained at abortion. In the term placenta from uncomplicated pregnancies, complement inhibitor C4b binding protein (C4BP) was detected in the syncytiotrophoblast, outer layer of the placenta that facilitates exchange of material between the mother and the embryo, and factor H in the tissue stroma [19]. Gene and protein expression of CD46, CD55, and CD59 was demonstrated in the placental extravillous trophoblasts in all three trimesters during normal pregnancies [9, 20, 21]. CD46 expression was demonstrated in the villous cytotrophoblasts, but CD55 and CD59 were inconsistently reported. Interestingly, the placental cells most in contact with the maternal tissue, the syncytiotrophoblast and extravillous trophoblasts, express all three complement inhibitors to provide protection against maternal complement activation.

Despite the presence of complement regulators, trophoblasts also synthesize complement components. Trophoblast cell lines and primary trophoblasts isolated from women secrete C3 and C4 protein and have mRNA for components C6, C7, C8, and C9 [22], providing immune protection against infection. Positive staining for C1q, C9, and C3d was also detected in normal placenta [19, 20]. Implantation, placentation, and the first trimester of pregnancy require a strong inflammatory response while protecting the mother and embryo against infection [23]. In this line, higher plasma anaphylatoxins C3a, C4a, and C5a concentrations were found during normal pregnancy compared to non-pregnant women [24,22,23,26]. If complement activation is involved in the pathogenesis of pregnancy complications, levels of complement split products would be expected to be higher than in normal pregnancy and could provide biomarkers to predict pregnancy outcomes. Complement components as biomarkers will be discussed for each pregnancy complication: preterm birth, preeclampsia, and recurrent miscarriages.

Since uncontrolled complement activation can lead to adverse pregnancy outcomes, recent studies suggested that polymorphisms of complement inhibitor genes might be related to abnormal pregnancy outcomes. Polymorphisms of the human CD46 gene in the Caucasian population were observed at increased frequency in females that had unexplained recurrent spontaneous abortions [27,28,29,30,31]. Provocatively, fathers of pregnancies that ended in miscarriage also showed CD46 polymorphisms. Knowing that paternally expressed genes predominate in the placenta, it is tempting to speculate that in pregnancies fathered by an individual carrying a CD46 polymorphism, abnormal placental expression of CD46 might result in excessive complement activation in the placenta and fetal demise. Alterations in the maternal genes encoding for C4b-binding protein were also associated with miscarriages [28]. In contrast, in a study performed in Iran, no genotype differences in CD46 were observed between fertile women and women that experienced recurrent miscarriages [29]. Impaired capacity to limit complement activation was also associated with late pregnancy complications such as preeclampsia. Mutations in CD46, complement factor I, and complement factor H were observed in women with SLE that developed preeclampsia [30]. On the other hand, a study performed in Finland did not find any association between CD46 SNPs and preeclampsia in women with no autoimmune diseases [31]. The ethnic background and the presence of autoimmunity might be important variables in these studies showing discrepant results. Since the placenta is mainly of paternal/fetal origin, paternal gene expression of complement inhibitors should be investigated and correlated with pregnancy outcomes.

Complement activation is associated with pregnancy complications

Preterm birth

Every year, an estimated 15 million babies worldwide are born preterm (before 37 completed weeks of gestation), and this number is rising. The rate of preterm birth (PTB) ranges from 5 to 18% of babies born in different countries. PTB complications are the leading cause of death among children under 5 years of age, responsible for nearly 1 million deaths in 2015 [32]. The lack of preventive measures to diminish the incidence of PTB and its sequelae is rooted in the lack of knowledge about the normal process of parturition. While term and preterm parturition share a cascade of downstream mediators that participate in cervical ripening and uterine contractions, PTB is a syndrome attributable to multiple pathologic processes. Identification of triggers of PTB might lead to therapeutic strategies to stop the cascade of events leading to premature parturition.

Studies in animals identified a role for complement activation in the pathogenesis of PTB. The association between complement activation and preterm delivery was recognized in mice in which preterm delivery was induced by administration of low-dose endotoxin or the progesterone antagonist RU486 [33]. Administration of both endotoxin and RU486 increased cervical deposition of C3, infiltration of macrophages, and plasma levels of anaphylatoxins compared to gestational age-matched controls. Further, a significant increase in matrix metalloproteinase 9 (MMP-9) activity, collagen degradation, and tissue distensibility were observed in the cervix of mice that received low-dose endotoxin intravaginally [33]. In this study, a significant correlation between C5a and the increased release of MMP-9 by macrophages was observed. Notably, mice deficient in C5a receptors (C5aR) did not show increased MMP-9 activity and cervical remodeling or premature delivery in response to low-endotoxin or RU486, suggesting that the complement system might be a good therapeutic target to prevent PTB [33]. Interestingly, progesterone prevented C5a-induced activation of macrophages by causing internalization of C5aR and reduced PTB in mice [33]. Both complement inhibition and progesterone, which modulates C5aR expression by macrophages, seem to be beneficial therapeutic approaches to prevent premature birth. In this line, progesterone has been approved by the FDA for the prevention of PTB in women with a history of prior spontaneous preterm birth [34]. However, since its approval, contradictory results have been reported [35, 36]. While several studies demonstrated the efficacy of progesterone in preventing PTB [35, 37], a recent trial concluded that vaginal progesterone administration was not associated with reduced risk of PTB or associated neonatal adverse outcomes (death, brain injury, neonatal bronchopulmonary dysplasia, or cognitive score at 2 years of age) [37].

In a follow-up study to investigate the different mechanisms/triggers leading to preterm and term delivery, the role of complement was further elucidated. Contrarily to PTD, complement activation was not required for the physiological process that leads to term delivery in mice. Neither complement deposition in cervical tissue nor increased complement split products in plasma were observed in mice that delivered at term [38]. The fact that mice deficient in different complement components do not have extended gestations confirms that complement does not play a critical role in parturition at term. While both PTD and term delivery showed increased MMPs release, collagen degradation, and cervical remodeling, these common downstream mediators seem to be triggered by different initiators. While activation of complement, in particular C5a, might be the initial step leading to macrophage activation and release of MMPs leading to increased cervical distensibility and preterm birth, MMP release at term is not mediated by complement activation products and originates from cervical fibroblasts and columnar epithelial cells rather than inflammatory cells [38].

That intraamniotic infection (chorioamnionitis) causes premature membrane rupture, and preterm birth is another piece of evidence supporting the role of complement activation in triggering preterm birth. Interestingly, clinical studies showed a correlation between elevated levels of complement activation fragments Bb and C3a and spontaneous PTB [39, 40]. Even more, it has been suggested that certain complement activation proteins might be biomarkers for PTB; for example, it has been demonstrated that plasma levels of complement factors B and H at 10–15 weeks of gestation are predictive of the development of PTB later in pregnancy [41].

Fetal brain injury associated with preterm birth

Of infants that were born before term, 25 to 50% experience long-term cognitive, behavioral, attentional, or socialization deficits [42, 43]. MRI studies in preterm infants showed decreased volume in the cerebral cortex, suggesting a role for the cortex in these long-term complications [44]. During pregnancy, the developing brain is particularly susceptible to inflammatory insults, often due to maternal intrauterine inflammation/infection. There is growing association between maternal inflammation and fetal brain injury [45, 46]. In this line, a new role for C5a in abnormal fetal brain development in inflammation-induced PTB was demonstrated in mice [47]. C5a not only plays a crucial role in the cervical ripening and myometrial contractions that lead to PTB but also causes damage to the developing fetal brain cortex in the mouse model [47]. Disruption of cortical dendritic and axonal cytoarchitecture characterized by decreased staining for microtubule-associated protein 2 and neurofilament 200 was observed in the brains of fetuses born preterm compared to age-matched controls. Fetuses deficient in C5aR (−/−) did not show cortical brain damage after PTB, suggesting that C5a-C5aR interaction is required for abnormal cortical development in prematurity. Treatment with an anti-C5 antibody that prevents generation of C5a also prevented cortical fetal brain injury in PTB mice [47]. In support of these in vivo studies, C5a also showed a detrimental effect on fetal cortical neuron development and survival in vitro [47]. Isolated cortical neurons from day 16 fetuses extend their neurites and establish synapses becoming mature neurons after 10 days in culture. While long axons were observed in cultured control cortical neurons, the length of axons in C5a-exposed neurons was considerably reduced. The blockade of C5a with C5aR antagonist peptide (C5aR-AP) increased the number of dendrites and the length of axons compared to neurons incubated with only C5a [47]. Increased levels of glutamate were measured in the supernatant of cortical neurons incubated with C5a and addition of the N-methyl-d-asparate antagonist, MK-801 (dizocilpine maleate) prevented C5a and glutamate-induced diminished axonal length, suggesting that C5a neurotoxicity might be mediated by glutamate excitotoxicity. Excitation resulting from stimulation of the ionotropic glutamate receptors is known to cause neuronal apoptosis [48]; indeed, signs of neuronal death/apoptosis and increased release of lactate dehydrogenase were observed in cortical neuron cultures exposed to C5a. This neuronal cell death was prevented by C5aR-AP, confirming the neurotoxic effects of C5a/C5aR [47].

Identification of a biomarker that can non-invasively detect pregnancies at risk of PTB and fetal brain injury during PTB would be of enormous clinical benefit in determining optimal timing of delivery and implementation of neuroprotective strategies. The described animal studies suggest that measuring complement activation might be useful to detect/predict fetuses at risk. Using a novel non-invasive MRI technique, complement activation was identified as a biomarker of abnormal fetal brain development in PTB [49]. During complement activation, C3 activation fragments (C3b/iC3b/C3d) are covalently attached to the injured tissue. Ultrasmall superparamagnetic iron oxide (USPIO, diameter: 5–40 nm) particles can be imaged by magnetic resonance imaging (MRI) because they shorten the T2 and T2* relaxation time; antibodies against C3 activation products conjugated to USPIOs were used in vivo to detect complement activation/inflammation in the fetal brain and predict pregnancy/fetal outcomes in PTB [49]. USPIO-conjugated anti-C3 antibodies crossed the placenta and the fetal blood brain barrier (BBB). Within the fetal sacs, most of the USPIO-conjugated anti-C3 antibodies were found in the placenta and fetal brain (Fig. 2). Increased signal was detected in fetal brains in PTB mice indicating increased C3 deposition [49]. C3 deposition was associated with abnormal cortical cytoarchitecture, increased complement split product C5a levels, and increased neurodegeneration [49].

Fig. 2
figure 2

Significant reduction in T2*-relaxation time, indicative of increased C3 deposition in the fetal brain in PTB mice. A significant reduction in T2* time was observed in vivo in the fetal brains in utero in PTB mice treated with USPIO-antiC3 compared to age-matched control and PTB mice that received USPIOs or control Ab-USPIOs (a). *Statistically significant compared to control mice injected with USPIO-antiC3 (p < 0.05), to preterm mice that received USPIO (p < 0.05) and SPIO-Ctrl antibody (p < 0.05) and preterm untreated mice (p < 0.05). b T2* maps of maternal abdomen in a PTB-mouse and age-matched control injected with USPIO-antiC3 antibodies. The dashed ovals show the amniotic sacs containing the fetuses. Microphotographs represent one of 4–5 similar experiments

Full size image

Figure 3 summarizes the postulated mechanisms responsible for complement induced cervical remodeling, myometrial contraction, and fetal brain injury in a mouse model of preterm birth.

Fig. 3
figure 3

Role of C5a in cervical remodeling, myometrial contraction, and fetal brain damage in preterm birth induced by intravaginal administration of LPS. a C5a attracts and activates macrophages to the cervix. Activated macrophages release metalloproteinases that digest collagen fibers. Cervical remodeling leads to increase distensibility of thr cervix leading to PTB. b C5a has uterotonic properties and induces myometrial contractions through interaction with C5aR. c C5a affects fetal brain development. Increased neurodegeneration and abnormal architecture are observed in the cortex in fetuses during PTB

Full size image

A recent translational study reinforces the hypothesis that C5a is a crucial mediator in fetal brain injury in PTB. In this clinical study, C5a was found in the cerebrospinal fluid of newborn human infants and its levels were elevated in those born preterm [50]. Interestingly, this difference was not explained by systemic infection. This study demonstrates increased complement activation in the brain of neonates born preterm and strengthens the suggestion that complement might be a potential therapeutic target for preterm fetal brain injury.

Preeclampsia

Preeclampsia (PE) affects ~ 5% of pregnancies and is traditionally diagnosed by the combined presentation of high blood pressure and proteinuria. Maternal organ dysfunction such as renal insufficiency, liver involvement, and neurological and hematological complications can also be observed during the onset of PE. PE is associated with uteroplacental dysfunction, diminished placental blood flow, and fetal growth restriction. PE can be lethal, and this disorder is the main cause of maternal and child mortality worldwide.

Even though PE is predominantly a human disease, with only very rare reports in other species, efforts to develop animal models which mimic the human pathology have yielded inflammatory models that are useful tools to investigate the pathogenic mechanisms of PE and test therapies for prevention and treatment. Animal models of PE have helped establish a role for complement activation in the pathogenesis of this serious pregnancy complication.

The crossing of CBA/J × DBA/2 animals has long been used as a mouse model of spontaneous abortion and PE [51,52,53]. Along with placental dysfunction, these matings resulted in maternal disease characterized by proteinuria, elevated blood urea nitrogen, and glomerular fibrin deposition and endotheliosis. Like the human disease, pregnancy complications in CBA/J × DBA/2 were associated with a dysregulation in angiogenic factors. Adverse pregnancy outcomes in this model resulted from placental insufficiency and functional deficiency of free vascular endothelial growth factor (VEGF) and elevated levels of soluble VEGF receptor 1 (sFlt-1). Interestingly, Tedesco and Chaouat [54] demonstrated that complement activation in this model was activated primarily through the lectin pathway. Analysis of the implantation sites collected from DBA/2-mated CBA/J mice revealed MBL-A deposition as early as 3.5 days of pregnancy. In addition, MBL-A deficiency prevented pregnancy loss in this abortion-prone mating combination [54]. Inhibition of complement, using Crry-Ig, anti-C5 mAb, or antagonists of C5aR, in vivo reverted the angiogenic imbalance, prevented growth restriction, and rescued pregnancies [53]. In addition, targeted inhibition of complement activation by administration of a chimeric soluble CR2-Crry molecule prevented oxidative stress and placental dysfunction, as well as proteinuria and renal pathologic features of preeclampsia in this model [55]. These findings point to blockade of complement activation as a valuable approach to prevent pregnancy-related complications, including PE and intrauterine growth restriction.

Complement activation has also been linked to pathogenesis in other animal models of PE. In the antibody transfer model of PE, pregnant mice receive autoantibodies that bind and activate the major angiotensin receptor AT1R; a role for complement split product C3a in the development of PE has been shown [56]. Autoantibodies to AT1R activate C3aR resulting in hypertension and proteinuria. Blockade of C3aR attenuated sFlt-1 production of soluble VEGF receptor 1 (sVEGFR-1, also known as sFlt-1; a potent anti-angiogenic molecule), placental dysfunction, and intrauterine growth restriction in this model [56].

In humans, complement activation has also been associated with PE. Complement split products, notably C4d, were identified in focal or diffuse staining patterns in the placentae of women with PE [20]. In addition, mRNA expression for CD55 and CD59 was significantly elevated in the placenta, suggesting a compensatory mechanism to control increased complement activation [20]. Other human studies demonstrated increased C4d deposition in the syncytiotrophoblast in cases of PE with fetal growth restriction [57]. Interestingly, C4d deposition was also observed in placentas from first trimester miscarriages, suggesting that C4d can be a biomarker for at-risk placentas.

Regarding circulating complement split products, an increased C3a/C3 ratio and increased sC5b-9 were observed in preeclamptic pregnancies compared to normal pregnancies [58]. Other studies showed increased C5a plasma levels in pregnancies affected by preeclampsia [59, 60].

Recurrent miscarriages

Animal models of recurrent miscarriages and intrauterine growth restriction (IUGR) also demonstrated that excessive complement activation, particularly C5a generation, is associated with impaired angiogenesis and adverse pregnancy outcomes [52]. Increased C5a levels were associated with deficiency of free vascular endothelial growth factor (VEGF), increased levels of soluble receptor for VEGF 1 (VEGFR-1/sFlt-1), presence of inflammatory infiltrates in the placenta and defective placental development in the CBA/J × DBA/2 model of fetal loss. Inhibition of complement activation in vivo has been shown to prevent angiogenic factor imbalance leading to improved pregnancy outcome [53]. In humans, it has been estimated that up to 20% of early pregnancy losses that are not mediated by autoantibodies are associated with hypocomplementemia [61].

Obstetric antiphospholipid syndrome

The antiphospholipid syndrome (APS), characterized by the presence of antiphospholipid (aPL) antibodies, is frequently associated with serious pregnancy complications: recurrent fetal loss, preterm birth, preeclampsia, and IUGR. Thrombosis is also frequently observed. Both pregnancy complications and thrombosis are tightly correlated with excessive levels of complement activation [62,63,64]. Murine models of OAPS in which pregnant mice are injected with human IgG containing aPL antibodies or mouse aPL monoclonal antibodies were fundamental to uncovering the triggering of the classical pathway by aPL antibodies and to establish C5a and neutrophils as key mediators of fetal injury [63, 65]. Using this model, the crosstalk between the complement system and the coagulation system was underscored, identifying C5a as a stimulus for the expression of the coagulation factor tissue factor (TF) on neutrophils. Through a mechanism that involves TF and protease activated receptor (PAR) expression, C5a activates neutrophils at the fetomaternal interface causing placental injury and leading to fetal death in OAPS [66, 67], Fig. 4. Pravastatin, by downregulating TF and PAR-2, rescued pregnancy in the mouse model of OAPS.

Fig. 4
figure 4

C5a-induced tissue factor plays a crucial role in placental and fetal injury induced by aPL antibodies. aPL antibodies bind to the placenta and activate the complement cascade. C5a attracts and activates neutrophils to the fetomaternal interface. C5a activates neutrophils through the expression of tissue factor and protease activated receptor-2 (PAR-2). Increased neutrophil activity results in increased oxidative damage of the placenta and fetal loss [66, 67]

Full size image

Due to their large size, antibodies are not expected to pass the blood-brain barrier (BBB) under physiological conditions. Provocatively, during pregnancy, aPL antibodies cross the placenta and the BBB reaching the developing brain [68]. The BBB becomes more permeable during inflammation. In particular, C5a regulates BBB integrity in inflammatory settings where it affects both endothelial and astroglial cells [69]. In addition, the immaturity of the fetal BBB, characterized by increased permeability, increases the susceptibility of the fetal brain to maternal autoantibodies and other proinflammatory and toxic insults.

Using in vivo SPECT/CT, it was demonstrated that aPL antibodies labeled with Indium-111 are rapidly cleared from circulation in the pregnant mouse and large amounts of the radiolabeled antibody are entrapped in the placentas and fetal brains within the fetal sacs [68]. This was the first study to show in vivo in real time the passage of aPL antibodies to the fetus targeting the fetal brain [68]. Binding of aPL antibodies to the fetal brain was associated with complement deposition, measured using USPIO-labeled anti C3 antibodies [49]. Interestingly, aPL antibodies bind to the fetal brain and activate the complement cascade leading to abnormal fetal brain architecture and abnormal behavior in the offspring [49]. Increased anxiety was observed in the offspring of OAPS mice, suggesting that in utero exposure to aPL antibodies causes abnormal cortical development and that C3 activation might be a footprint for adverse fetal outcomes. In conclusion, the pathogenic effects of aPL antibodies on the placenta and the developing brain seem to be mediated by complement activation, suggesting that complement activation products might be useful biomarkers of pregnancy outcomes and targets for therapy. Interestingly, a recent case report described a pregnant patient treated with the C5-inhibitor eculizumab to prevent APS-related complications [70]. No new thrombosis or catastrophic APS developed during the last week of pregnancy or postpartum after C5 inhibition [70]. Another case report showed efficacy of C5 inhibition with eculizumab in the treatment of PE [71]. These case reports suggest that targeting C5 activation might be a safe treatment option for OAPS and PE, prolonging pregnancy and improving outcomes. While C5 inhibition holds promise in treating pregnancy complications in APS, and complement split products are increased in non-pregnant patients with APS [72], plasma levels of complement C3 in pregnant patients with APS were not predictive of adverse pregnancy outcomes [73].

Not all complement components have detrimental effects on pregnancy—C1q is required for normal placentation

Despite the clear association between dysregulated complement activation and poor pregnancy outcomes, the presence of certain complement proteins is strictly required to assure normal placentation and pregnancy development. The important role of complement component C1q in normal placentation has been reported [74]. C1q is widely distributed in human decidual endothelial and stromal cells and is actively synthesized by migrating extravillous trophoblasts which require C1q for proper adhesion and migration [74]. Impaired placental labyrinth development, decidual vessel remodeling, and increased fetal death were observed in C1q-deficient mice [74, 75]. The contribution of C1q from fetal-derived trophoblasts in assuring good pregnancy outcome was demonstrated by mating C1q sufficient females with C1q-deficient males. C1q sufficient females showed increased fetal death and abnormal placentas only when impregnated by a male deficient in C1q, indicating that trophoblast-derived C1q is crucial for normal placentation and fetal survival. In addition, when compared to normal mice, C1q-deficient mice showed diminished decidual collagenase activity—required for trophoblast migration and placental angiogenesis—and thicker deciduas, providing additional evidence for defective trophoblast invasion of the maternal tissue in the absence of C1q [75]. It has been suggested that such defective invasion of trophoblasts into the maternal decidua is linked to PE. In line with this concept, pregnant C1q-deficient mice developed PE and recapitulated the key features of the human disease: hypertension, albuminuria, endotheliosis, endothelial dysfunction, decreased placental VEGF, and elevated levels of sFlt-1, previously correlated with increased rates of fetal death [75]. Increased sensitivity to angiotensin II and decreased vasorelaxant response to acetylcholine was observed in aortic rings isolated from pregnant C1q-deficient mice, indicating the presence of endothelial dysfunction in these mice. These studies highlight the crucial role of C1q in normal placentation and good pregnancy outcomes.

Conclusion

The hemiallogeneic fetus would be quickly rejected without numerous mechanisms that work in concert to protect the fetus from immunological recognition and rejection. The presence of complement inhibitors at the maternal interface is an important mechanism that help pregnancy survival. Pregnancy complications such as recurrent miscarriages, preterm birth, intrauterine growth restriction, and preeclampsia have been associated with excessive complement activation in animal models and in humans, suggesting that complement split products might be used as biomarkers to predict pregnancy outcomes and blockage of the complement cascade might result in improved pregnancy outcomes. Randomized controlled trials should be organized to test this possibility in women.