The canonical complement system in the blood has been extensively reviewed and is a key chapter in all immunology textbooks. The complement system is important in protecting from infection and in clearance of immune complexes as evidenced by evaluation of individuals with deficiencies in various complement components. The reader is pointed to some very comprehensive recent reviews of complement for a more detailed look at the system (1–3).

As illustrated in Figure 1 , complement activation occurs through three major pathways: the classical, lectin and alternative. Each pathway is initiated by a distinct mechanism, but they share the common goals of target recognition and opsonization, promoting the inflammatory response as well as lysis of abnormal cells or pathogens. The classical pathway is typically initiated by antibodies bound to antigens, which recruit the C1 complex (C1q, C1r, C1s—collectively called C1 esterase). This leads to cleavage of C4 and C2, generating the classical pathway C3 convertase, C4b2b, with C4b covalently attached to activating surfaces. Note: There is ongoing debate among complement experts regarding nomenclature. Older literature refers to the larger fragment of C2 as C2a, while newer conventions prefer C2b to maintain consistency across the cascade (4). The lectin pathway is triggered by recognition of specific carbohydrate patterns on pathogens. Mannose-binding lectin (MBL), collectins and ficolins bind these carbohydrates and activate MASPs (MBL-associated serine proteases), leading to the same C3 convertase (C4b2b) as in the classical pathway. The alternative pathway, distinct from the classical pathway, has the capacity to deposit on a pathogen without need for prior contact/exposure. It relies on the slow, continuous hydrolysis of C3 in plasma, a process known as “tickover”, to form C3(H₂O). Upon encountering an activating surface, C3b is generated and binds Factor B, which is then cleaved by Factor D (also known as adipsin), resulting in the formation of the alternative pathway C3 convertase, C3bBb. C3, present in high serum concentrations (1–2 mg/ml), is cleaved by either C4b2b or C3bBb. The resulting C3b fragment binds covalently to pathogens or proteins, promoting opsonization and further amplification of the cascade. Addition of another C3b to each C3 convertase forms the C5 convertases (C4b2b3b and C3bBbC3b), which cleave C5 into C5a and C5b, the final proteolytic step in complement activation. C5a and C3a, small molecular weight anaphylatoxins, bind to their respective G-protein-coupled receptors (C5aR1, C5aR2, and C3aR1), eliciting pro-inflammatory responses. C5b, in turn, initiates assembly of the membrane attack complex (MAC, C5b-9), leading to target cell lysis.

Complement activation is tightly controlled by both membrane-bound and fluid phase regulator proteins to prevent host tissue damage. In the classical pathway, C1 inhibitor (C1INH) and C4 Binding Protein (C4BP) are soluble factors that block early steps by targeting C1 and C4, respectively. Factor H (FH), a key fluid-phase regulator, primarily controls the alternative pathway C3 convertase by promoting the cleavage of C3b (known as cofactor activity) and accelerating convertase decay (known as decay accelerating activity). CD55 (Decay-accelerating factor, DAF) promotes dissociation of the convertases, while CD46 (Membrane cofactor protein, mCP) serves as a cofactor for the Factor I (FI)-mediated proteolytic inactivation of C3b and C4b deposited on cells. Both CD55 and CD46 are widely expressed in human tissues and regulate all three pathways. CR1 (CD35, C3b/C4b receptor) present on human erythrocytes has cofactor activity and decay accelerating activity for both the classical and alternative pathways. CD59 (Protectin), another membrane-bound regulator, prevents MAC formation by inhibiting C9 polymerization. Other regulators like CSMD1 and CSMD3, though less well characterized, may have cofactor activity or effects on MAC, and are also implicated in neurological processes (5–7).

Species-specific complement differences complicate the study of the role of complement in normal pregnancy and preeclampsia. For example, Crry (CR1-related gene/protein Y) in rodents (but not humans) performs functions analogous to both CD46 and CD55 in humans. In mice and rats, CD46 is only expressed in the testes suggesting a specialized role in reproduction in comparison to a more global role and distribution for CD46 on human cells. Thus, use of animal models to investigate a role for CD46 in pregnancy is not readily translatable to human pregnancy. Activators of the alternative pathway also differ between species. For example, rabbit erythrocytes activate the alternative pathway in human serum, but not rabbit (8) and thus are used to assay alternative pathway activation in human serum. Zymosan however activates the alternative pathway in multiple species. Human CD59 can protect against cell lysis by human complement, but not by rat and guinea pig complement (9). Eculizumab is an anti-C5 antibody therapeutically effective in the human but does not react with mouse C5. The mouse monoclonal antibody BB5.1 has been used for proof-of-concept studies in mice to accumulate evidence for the importance of C5. Also, novel anti-C5 monoclonal antibodies capable of inhibiting complement in multiple species have been developed to assist in preclinical studies to evaluate complement involvement (10) and are useful tools for proof-of-concept studies in rat, but with low reactivity in mouse. ELISA assays developed for measurement of human complement components and activation products are not guaranteed to be accurate for measurement of mouse or rat complement, so quantitation of components in pre-clinical studies must be carefully planned and validated, along with effectiveness of inhibitors. These differences pose challenges in translating animal model findings to human pregnancy and diseases such as preeclampsia.

Complement receptors mediate the downstream effects of activation products. CR1 (CD35) binds C3b and C4b, along with their breakdown products, iC3b and iC4b, and plays a major role in clearing immune complexes, especially in primates (11). C3a and C5a, potent inflammatory mediators, signal through G protein coupled receptors C3aR and C5aR1/2. Notably, C3a can be converted to C3a_{des arg} (also called ASP, acylation-stimulating protein) and C5a to C5a_{des arg} via carboxypeptidase cleavage. While C3a binds C3aR and C5a can interact with C5aR1 and C5aR2, both C3a and ASP have been reported to interact with C5aR2 on adipocytes to promote triglyceride synthesis, though this interaction is controversial (12, 13) due to the conflicting findings regarding the exact nature and significance of this interaction. While the initial studies identified ASP as a potential ligand for C5aR2 that stimulates triglyceride synthesis in adipocytes, other studies provided contradicting evidence. Therefore, further research is needed to fully understand the intricate relationship between C3a/ASP, C5aR2, and their roles in adipocyte function and metabolic regulation.

Beyond traditional pathways, complement components can also be activated by extrinsic proteases, a process independent of convertases. Proteases from the coagulation, fibrinolytic, and kallikrein-kinin cascades (e.g., plasmin, thrombin, kallikrein) can directly cleave C3, C5, and Factor B, generating active fragments such as C3b, C3a, C5b, C5a, and Bb. These fragments can then initiate effector responses themselves or go on to assemble convertases. This highlights the complement system’s versatility: both as a tightly coordinated cascade and as a modular set of components that contribute to immune surveillance, inflammation, and tissue homeostasis.

The intracellular complement system, also referred to as the ‘complosome,’ has garnered growing attention for its roles in maintaining cellular homeostasis and modulating local innate immune responses ( Figure 2 ). Although it has been known for decades that complement proteins are synthesized by lymphocytes and other immune cells outside the liver (14), recent studies have revealed previously unrecognized functionality of this system within cells. This was first demonstrated in T cells (15), and has since been extended to the central nervous system (16–18) and the pancreatic islets (19–21) where intracellular complement activation products contribute to both intracellular signaling and autocrine effects in the local microenvironment. Emerging evidence suggests that the tickover product C3(H₂0) can be shuttled from the extracellular space into the cytosol (22), where it is cleaved in lysosomes by Cathepsin L to generate C3a and C3b. Intracellular C3a interacts with the G protein coupled receptor C3aR either within the cell or at the plasma membrane, to evoke effector responses. Meanwhile, C3b covalently binds to membrane surfaces, tagging them for immune recognition. There is also evidence that C5 is cleaved in the lysosome to form C5a and C5b. C5a then engages C5aR1 on lysosomes and mitochondria, influencing key cellular processes such as metabolism, pathogen clearance and regulation of autophagy. The specific intracellular complement components vary across different cell types, and this area remains an active and rapidly expanding field of study. At the fetal-maternal interface, for example, trophoblasts are known to synthesize complement components such as C4, C3 and terminal components (23), while endothelial cells in the spiral artery produce C1q (24). The precise roles of these locally synthesized or intracellularly active complement components in supporting normal pregnancy outcomes are currently under intense investigation.

A favorable pregnancy centers on the placenta, a unique temporary organ formed in the uterus to connect mother to the developing fetus. Many adverse pregnancy outcomes result from defective placentation. Although placental structure varies across species, the critical sites for complement system regulation may be conserved or differ slightly. In humans, rats and mice, the placenta is hemichorial i.e. the fetal epithelium is in direct contact with the maternal blood (29). Rodent placental layers are termed decidual, junctional and labyrinth whereas layers in the human placenta have different terminology (extravillous, extravillous trophoblast column, villous). Again, recent reviews provide excellent background of the details (30). The other key feature of the placenta is that the decidual layer is maternal and the trophoblasts are fetal (both mom and dad). This fetal-maternal interface represents a potential immunological challenge, making complement regulation crucial to prevent maternal rejection of the allogeneic fetus. In rodents, three layers of trophoblasts, including two syncytiotrophoblasts and one cytotrophoblast, separate maternal blood from fetal capillaries, whereas in humans, only a single syncytiotrophoblast layer is present. Trophoblasts invade the maternal decidua to remodel uterine spiral arteries to a low resistance blood vessel. This invasion is deeper in rats and humans than in mice. Given these differences, complement regulation is particularly important at two sites: the point of trophoblast invasion into the spiral arteries and the syncytiotrophoblast layer where maternal blood directly contacts fetal tissue. Smith Jackson’s review offers a comprehensive look at complement’s role in placental development, emphasizing the importance of balancing complement activation with regulation (31).

Pregnancy represents an allogeneic-type mismatch between the fetus and the maternal immune system (32, 33). Since the fetus encounters maternal immunocompetent cells during trophoblast invasion, healthy placentation depends on a collaboration between efficient trophoblast invasion and the mother’s immune system. During trophoblast invasion, many immune cells infiltrate the decidua, allowing them to reach the endometrium and spiral arteries. In healthy pregnancies, non-recognition of these trophoblasts by maternal immune cells promotes maternal-fetal tolerance. The complement system regulates key processes to establish and maintain maternal-fetal tolerance during a healthy pregnancy (34, 35). Placental damage in the human is prevented by fluid phase and membrane bound regulators of complement activation, such as, DAF (CD55), MCP (CD46), CD59 (36), FH, FI, or complement receptor 1 (CR1) (37). These regulators assist in the clearance of placental fragments that enter the maternal circulation as a result of syncytiotrophoblast turn-over. Furthermore, levels of many complement proteins increase during pregnancy. Improper clearance of such components, driven by an inadequately regulated complement system, may cause deposition of debris in tissues and vascular walls, leading to an overly exuberant inflammatory response, resulting in endothelial injury and placental dysfunction (38). Complement dysregulation or overactivation can further activate programmed death of trophoblasts, known as apoptosis, significantly disrupting their migration and placental vascularization. This in turn can lead to placental ischemia and release of inflammatory mediators by the placenta. Due to its potent and rapid nature, complement activation at the fetal maternal interface requires strict control for a successful pregnancy.

In women who develop preeclampsia, early impairments in placental blood flow and oxygenation, often due to poor spiral artery remodeling (39, 40) lead to placental inflammation and apoptosis. These changes lead to systemic activation of leukocytes and endothelial cells over the course of pregnancy (41, 42). Prior to the onset of clinical disease, there is an increase in the release of soluble fms-like tyrosine kinase-1 (sFLT1) by trophoblasts. sFLT1 acts as an anti-angiogenic factor which binds and sequesters placental growth factor (PlGF) and vascular endothelial growth factor (VEGF) (43–45). Elevated sFLT1 levels, along with decreased PlGF and VEGF, are associated with hypertension and proteinuria, and eventual development of preeclampsia (45, 46).

Angiogenic imbalance and excessive complement activation are both associated with development of preeclampsia and abnormal placentation, but whether angiogenic imbalance results in complement activation or vice versa is unresolved. Banadakoppa et al. demonstrated that increased complement activation in trophoblasts correlates with elevated secretion of sFLT1 protein (47). As complement activation increases, intracellular sFLT1 levels decline due to enhanced secretion of sFLT1 protein from trophoblast cells. Notably, this secretion closely correlated with trophoblast release of C5b-9. Supporting this, Collier et al. found strong spatial and quantitative associations between placental C5b-9 and sFLT1, with sFLT1 localization observed near immunofluorescence signals for C4d and C5b-9 on the syncytiotrophoblast membrane (48). In vitro studies with human umbilical vein endothelial cells (HUVECs) (49) demonstrated that addition of PlGF increased the expression and secretion of FH, a key complement regulatory protein. In contrast, sFLT1 decreased FH levels making the HUVECs more susceptible to complement activation. These findings suggest that an imbalance in angiogenic factors may enhance complement activation by downregulating complement regulation via FH.

Studies exploring the relationship between angiogenic imbalance and excessive complement activation have also been conducted in preclinical animal models as understanding of the sequence of events preceding fetal loss, growth restriction and adverse maternal consequences could have important ramifications for therapeutic development. In the rat model of placental ischemia induced hypertension (50), inhibition of complement activation from GD14–19 resulted in attenuation of the maternal hypertension, but had no effect on the reduced VEGF or reduced fetal weight. These data suggest complement activation was not critical for the reduction of VEGF in late pregnancy within this model. The BPH mouse model, which exhibits pre-pregnancy borderline hypertension, obesity, and dyslipidemia, is considered a relevant model of superimposed preeclampsia (51). During pregnancy, BPH dams develop high blood pressure, fetal loss, growth restriction and abnormal placentation, features characteristic of preeclampsia. Work by Sones et al. showed complement activation in this model recruits neutrophils very early in gestation at e8.5 leading to increased tumor necrosis factor (TNF), reduced placental VEGF, abnormal placentation and fetal death by e12.5. Complement inhibition restored VEGF levels, suggesting that complement activation leads to angiogenic imbalance (51) and adverse pregnancy outcomes. However, further analysis suggests a more complex picture. In the BPH model, dysregulated VEGF pathway genes occurred at the start of decidualization (e5.5) and precedes the complement deposition in the placenta seen at the end of decidualization and beginning of placentation (e7.5) suggesting that angiogenic dysregulation may initiate complement activation rather than result from it (52). Similarly, in the abortion prone CBA/J X DBA/2 mouse model characterized by fetal rejection and growth restriction without maternal hypertension (53), blocking of complement activation reduced sFLT1 levels and rescued fetal growth. Taken together, these studies indicate that the relationship between angiogenic imbalance and complement activation may differ depending on the stage of pregnancy and the model. The two processes may operate in parallel or redundantly rather than in a strictly linear fashion. Overall, evidence suggests that in early gestation, angiogenic imbalance likely precedes and contributes to complement activation, impairing placental development. In contrast, during later stages, complement activation may be more directly involved in adverse maternal outcomes such as hypertension, independent of angiogenic imbalance. Continued studies in animal models are needed to further clarify these complex interactions.

Studies by Agostinis in human trophoblasts have demonstrated the importance of C1q in the trophoblast invasion of the spiral arteries and their remodeling for proper placental development. This non-canonical effect of C1q does not require pathway activation. C1q is synthesized by endothelial cells in the decidua as well as by endovascular trophoblasts and macrophages. It facilitates connections between endothelial cells and endovascular trophoblasts, contributing to effective spiral artery remodeling during placentation (24, 54).

Mice deficient in complement components or regulators have revealed extensive information about the role of complement in pregnancy and preeclampsia (55). In the C1q-deficient mouse, trophoblast invasion into decidua is impaired, resulting in hallmark features of preeclampsia, including hypertension and reduced fetal weight (56–58). These effects occur despite the absence of overt pathway activation, further supporting a structural or signaling role for C1q. Mating experiments revealed that only when wild-type (WT) females were mated with C1q-deficient males did preeclampsia-like outcomes manifest, including fetal demise and maternal hypertension. Since the placenta is influenced more by paternal genes, WT females carrying heterozygous (C1q-deficient) placentas developed all major symptoms of preeclampsia. In contrast, C1q-deficient females mated with WT males exhibited normal pregnancies, indicating that the presence of C1q on fetal-derived placental trophoblasts, rather than maternal C1q, is critical for a healthy pregnancy (56). Another key finding in these studies is that the impaired placental development in the absence of C1q leads to reduced placental blood flow and placental ischemia along with secondary mechanisms of tissue injury and inflammation with endothelial damage. The WT females mated with C1q deficient males had greater generation of C5a and other inflammatory mediators later in pregnancy compared to a normal pregnancy suggesting exaggerated canonical complement activation also occurred following the impaired placental development due to C1q deficiency (58). In addition, the WT females mated with C1q deficient males maintained high blood pressure 2 months postpartum when fetal-derived placental trophoblasts lacking C1q were clearly not present. Sutton et al. expanded these findings by investigating the vascular dysfunction postpartum of wild-type (WT) females mated with C1q-deficient males. In this study, vascular dysfunction persisted up to 6 months postpartum, emphasizing a very lasting impact of a compromised placenta on maternal cardiovascular health following pregnancy (59).

Immunohistochemical analyses of placentae from preeclamptic women has also shown that deposition of complement C1q differed significantly between patients with preeclampsia and controls and further demonstrated that early-onset preeclampsia patients had more C1q deposition than late-onset preeclampsia patients (60). This C1q deposition in the syncytiotrophoblast region was thought to reflect complement activation from placental dysfunction rather than the non-canonical function of C1q bridging trophoblasts and endothelial cells in spiral artery remodeling. Also, placental dysfunction is more typically observed in early onset preeclampsia, leading to a greater need for clearance of apoptotic cells through C1q, as well as being associated with more C1q deposition (61).

Additional studies using the CBA/J X DBA/2 mouse model, described as both abortion-prone and a model for preeclampsia, have examined the role of MBL and lectin pathway in pregnancy complications (62, 63). Maternal hypertension does not occur in this model, so its usefulness as a model of preeclampsia is overstated. Affected pregnancies were characterized by high rates of fetal resorptions, intrauterine growth restriction, proteinuria and angiogenic imbalance (64). These mice do show exaggerated blood pressure response to angiotensin infusion and exercise during pregnancy compared to controls. Importantly, pregnancy failure is associated with lectin pathway activation and the terminal complement complex because MBL deficiency and anti C5 treatment in this model prevented pregnancy loss (63).

Locally generated C3 activation products in the reproductive tract promote embryonic survival and growth. In the oviduct, C3 convertase is formed from maternal epithelial or blood-derived C3, Factor B, and Factor D, likely via the alternative pathway, resulting in formation of C3b, which is then cleaved to iC3b by local embryonic regulator Crry (rodents) or MCP (human). This locally produced iC3b acts as a potent embryotrophic factor, increasing size of the blastocyst (65–68) and supporting reproductive success. Embryonic lethality is apparent in the absence of Crry and is also consistent with impaired early pregnancy in C3 deficient mice (69). Fetal and placental weights are lower in C3 deficient mice with increased fetal resorptions and aberrant placental development including smaller spongiotrophoblast area and labyrinth.

In the BPH mouse model of superimposed preeclampsia, placental insufficiency, fetal demise and growth restriction occurs early in the pregnancy with high blood pressure in the dam appearing in later pregnancy. Two novel reagents were used to assess events early in pregnancy and inhibited the C3 convertases at the site of inflammation via targeted binding by Complement Receptor 2 (CR2): CR2-Crry and CR2-FH (51). These reagents rescued the placental morphology, fetal demise and growth restriction in the BPH mouse in early pregnancy indicating the importance of complement through C3, but the effect on maternal high blood pressure in later pregnancy was not assessed. BPH mice are hyperphagic and become obese with increased visceral reproductive white adipose tissue. Sones hypothesized that the increased reproductive white adipose tissue in the BPH mouse before and during pregnancy is a source of complement factors that contribute to the angiogenic imbalance and adverse pregnancy outcomes. They found that levels of C3 and Factor B are increased in reproductive fat in BPH mice compared to C57Bl/6 controls, suggesting the reproductive white adipose tissue may be a source of complement components influencing reproductive outcomes (70). Caloric restriction in the BPH mouse attenuated C3 expression at the maternal fetal interface early in pregnancy and restored the angiogenic imbalance to that of a control mouse. These data suggest the reproductive white adipose tissue may be a source of complement components working in a non-canonical fashion to influence reproductive outcomes but further studies are needed to determine if complement from adipose tissue is mechanistically connected to maternal hypertension and fetal development later in pregnancy (70).

Besides iC3b, the cleavage fragment C3a also plays an important role in embryonic development. The C3a-C3aR axis has been shown to regulate cell attraction during collective migration of neurocrest cells in zebrafish and Xenopus embryos (71, 72). Beyond development, C3a and C5a have also been evaluated as biomarkers of preeclampsia. In 1989, Haeger et al. first reported an increase in both C3a and C5a in women with preeclampsia and HELLP syndrome (73). Later, Lynch et al. showed that plasma C3a levels were elevated in early pregnancy among women who developed hypertensive disorders of pregnancy (848 vs. 757 ng/L, P<0.001) (74). Among women with obesity (BMI >30 kg/m²) and elevated plasma C3a levels before 20 weeks gestation, the odds of preeclampsia were significantly increased (aOR 8.8, 95%CI 3-24, P<0.001) (75). However, He et al. evaluated plasma samples as early as 6–8 weeks of gestation but found no difference in C3a or C5a levels between women who developed preeclampsia and those who did not (76). Further studies over the past decade have been mixed, in part due to variable study design and assay systems. Four studies have reported elevated plasma C3a levels in preeclampsia compared to healthy pregnancies (35, 73, 77, 78) while two found no difference (79, 80). Similarly, five studies showed increased plasma C5a levels in preeclampsia or HELLP compared to healthy controls (73, 77–80), whereas two did not (81, 82). Burwick et al. demonstrated a fivefold increase in urinary C5a levels in women with severe preeclampsia compared to matched healthy controls (80). A follow up study, evaluating the same cohort of patients, showed that urinary C5a levels did not correlate significantly with urinary albumin levels (r=0.37, P=0.08), but did correlate strongly with urinary levels of KIM-1, a marker of proximal tubule injury (r=0.60, P=0.001) (83). Investigators hypothesized that complement activation in severe preeclampsia may propagate tubular injury in the kidney, as seen in mouse models of renal ischemia, where upregulation of KIM-1 in the proximal tubule is mediated by products of complement activation (84, 85). Notably, among the subset of studies analyzing severe preeclampsia specifically, all showed significantly elevated plasma C5a levels compared to healthy controls, suggesting that terminal complement activation is more consistently associated with severe disease. Standardization of methodologies and assay systems will be critical in future complement biomarker studies in preeclampsia to improve comparability and reliability of these results.

In addition to plasma biomarkers, evaluation of amniotic fluid offers insights into local complement activity. Banadakoppa et al. measured C3a, C4a and C5a in amniotic fluid in mid-gestation and found significantly higher C3a levels in women who later developed early-onset preeclampsia (<34 weeks) compared to controls (318.7 ng/mL vs. 254.5 ng/mL, P=0.04), while C4a and C5a levels did not differ significantly (86).

Animal studies further support the role of C3a and/or C5a in the pathophysiology of preeclampsia. Regal et al. used a rat model of placental ischemia-induced hypertension and demonstrated that inhibition of complement activation with soluble CR1, prevented maternal hypertension but did not reverse fetal resorptions or growth restriction (50). Specific inhibition of C3aR or C5aR attenuated the hypertension induced by placental ischemia, implicating C3a and C5a in disease progression (87). Using mesenteric artery myography, the cholinergic vasodilation was impaired following the RUPP surgery, and this was rescued by the C5a antagonist but not the C3a antagonist, indicating that the complement system through C5 cleavage is important for the vascular dysfunction. Around the same time, Burwick and Feinberg (88) were the first to show that C5 blockade with eculizumab could also be used effectively to treat preeclampsia and HELLP syndrome, consistent with the protective effects seen with soluble CR1 and C5a antagonism in animal models.

Angiotensin II is a potent vasoconstrictor and binds to two major receptors: Type 1 (AT1R) and Type 2 (AT2R). Wallakut et al. initially demonstrated increased autoantibodies to the AT1R (AT1-AA) in preeclampsia compared to normal pregnancies (89). These autoantibodies are not unique to preeclampsia but are also increased in a variety of conditions such as refractory hypertension (90), aortic dissection (91), or kidney transplant rejection. AT1-AA activate, rather than block, the AT1R, and enhance angiotensin II signaling. This activation is blocked by AT1R antagonists such as losartan. Since, antigen antibody interactions are known to activate the complement system, it seems reasonable to hypothesize that complement activation in preeclampsia may be due to AT1-AA interaction with the AT1R. In humans, the AT1-AA belong to the IgG3 isotype which are strong activators of the complement system.

Several animal models have been used to investigate the role of complement and AT1-AA in preeclampsia. In one such model female rats expressing the human angiotensinogen gene are mated with males expressing the human renin gene. Human renin from the fetus crosses the placenta to cleave human angiotensinogen that cannot be cleaved by rat renin, resulting in increased angiotensin II, AT1-AA production and hypertension (92). These rats demonstrate albuminuria, IgG deposition and complement component staining (C1q, C3, C3c and C5b-9) in the kidneys, implicating complement activation in the pathophysiology. However, no direct test of complement involvement in placental pathology or resulting hypertension was performed, presenting as a clear gap in knowledge in determining a critical role for complement in this model. In a mouse model, injection of human AT1-AA induced preeclampsia-like symptoms and complement activation (93, 94). These effects were attenuated by losartan and by high doses of C3aR antagonist, SB290157 (95) suggesting the importance of C3a in the preeclampsia-like symptoms induced by injection of the human AT1-AA. However, these studies are limited by the high doses of SB290157 used that may have off-target effect and partial agonist activity. The limitations of the C3a antagonist SB290157 include C3a receptor agonist activity and cross-reactivity with C5aR2 at higher doses (96, 97). Although development of C3a antagonists remains limited and improved pharmacological tools to block the C3a receptor are still lacking (1), future studies could employ C3aR knockout mice to further investigate whether C3a plays a critical role in mediating AT1-AA–induced preeclampsia-like symptoms.

LaMarca’s group employed a similar model in the rat showing that AT1-AA injection during pregnancy results in hypertension, increased sFLT1 and oxidative stress, i.e. all hallmarks of preeclampsia. Both an endothelin A antagonist and an AT1R blockade reduced the blood pressure (98) confirming that AT1-AA interact with AT1R to cause the high blood pressure in pregnancy. AT1-AA were also detected in the rat reduced uterine perfusion pressure (RUPP) model of placental ischemia, where placental ischemia is induced on gestation day 14, resulting in hypertension and detectable AT1-AA by day 19. The hypertensive response following placental ischemia was prevented by an AT1-AA inhibitory peptide (99), implicating AT1-AA as critical. Regal used the same RUPP rat model to indicate that complement activation was critical for placental ischemia induced hypertension and verified that losartan attenuated RUPP-induced hypertension (100). However, losartan did not affect complement activation. Thus, in the rat model of placental ischemia induced hypertension, both AT1-AA and complement activation are critical mechanistic pathways. However, the evidence does not indicate that AT1-AA initiate complement activation to cause hypertension following placental ischemia.

Besides AT1-AA, placental ischemia in the rat may invoke recruitment and deposition of IgM antibodies, another potent trigger of complement activation (100). Ischemia-reperfusion studies across various organ systems have consistently implicated IgM as a key initiator of complement-mediated injury (101–104) with the specific complement pathways activated depending on the affected tissue and context. In preeclampsia, increased IgM deposition has been observed in both the placenta and kidneys of affected pregnancies (105, 106), as well as in the placental ischemia induced model of hypertensive pregnancy in rats (100). Regal et al. used the RUPP rat model and demonstrated deposition of both IgM and C3 in placenta of RUPP rats (87). Thus, placental IgM deposition following placental ischemia could be responsible for the complement activation and resultant hypertension.

The role of autoantibodies against C1q and FH in preeclampsia have also been investigated (107, 108). In a study by Dijkstra et al. (107), maternal serum samples from women with preeclampsia, healthy pregnancies, and nonpregnant women were analyzed across cohorts from the Netherlands, Finland, and Norway. No differences were found in anti-C1q or anti-FH autoantibody levels between preeclamptic and normal pregnancies. However, factor H levels were lower, particularly in early-onset cases, suggesting a role for impaired complement regulation in its development, despite no difference in circulating autoantibodies. Agostinis et al. measured C1q bound-circulating immune-complexes (CIC-C1q) in the serum of pregnant women with preeclampsia compared with healthy controls and detected higher CIC-C1q levels in preeclampsia (P=0.0004). Women with early-onset preeclampsia showed even greater CIC-C1q levels compared to those with late-onset preeclampsia. In contrast to the CIC data, preeclampsia patients had lower levels of C1q and anti-C1q compared with healthy controls (108). Study authors noted that the significant difference in anti-C1q levels among preeclampsia patients in their study, in contrast to the study by Dijkstra et al. (107), could be explained by the finding that differences in anti-C1q levels are most pronounced in the first trimester of pregnancy (108). Low anti-C1q levels in early pregnancy could reflect abnormal placentation and subsequent apoptosis occurring in the first trimester among women who later develop preeclampsia.

To further explore the importance of antibody as an important initiator of complement activation in placental ischemia induced hypertension, B cell depletion experiments were conducted using the rat model of placental ischemia induced hypertension with disparate results. While LaMarca reported that depleting rat B cells with anti-human CD20 (rituximab) reduced RUPP-induced hypertension (109, 110), Regal’s group found no such effect using anti-rat CD20 (111–113) where B cells in the blood and peritoneum were certainly depleted, but circulating IgM antibody was still apparent. Given the short window between placental ischemia and hypertension onset (GD14 to GD19), it is plausible that natural IgM, present independently of prior immunization is more relevant as an initiator of complement activation than the generation of an IgG3 autoantibody within 5 days of the placental ischemic event. In addition, rituximab targets human CD20, but not rat CD20 (113, 114), raising concerns about the specificity of B cell depletion with rituximab in rat models. Further studies are needed to determine whether specific B1 or B2 cell subsets and the antibodies they produce play a critical role in placental ischemia induced hypertension and complement activation. Such insights could better inform the development of targeted therapeutic strategies in humans.

Another strategy to explore the importance of B cells in initiating complement activation through production of antibody is the use of adoptive transfer techniques. Herrock adoptively transferred B2 cells from RUPP rats into normotensive pregnant rats and induced hypertension but did not cause an increase in circulating C3a (109, 115). In a similar model, placental CD19+ B cells from preeclamptic patients adoptively transferred into normal pregnant rats also induced hypertension along with complement activation, but a critical role for the complement activation in the increased blood pressure was not evaluated. In the RUPP rat model, Herrock et al. (116) demonstrated that an antibody to B cell activating factor (BAFF) prevented the placental ischemia induced hypertension and circulating AT1-AA, but did not significantly decrease circulating BAFF or complement activation. A surgical sham control group or isotype control injection for the anti-BAFF was not evident in this study and is a clear limitation that complicates interpretation of the data. Their data suggests that the complement activation is not initiated by the AT1-AA in this model. Since inhibition of complement activation with soluble CR1 attenuated placental ischemia induced hypertension in the rat as well (50) certainly multiple redundant pathways can drive the hypertensive phenotype.

Five studies have demonstrated that plasma C5b-9 levels are increased in preeclampsia compared to healthy controls (35, 78, 80, 82, 117) while two studies found no difference (77, 81). Plasma C5b-9 levels are also elevated in other hypertensive disorders of pregnancy, including gestational hypertension, chronic hypertension, and preeclampsia without severe features (80, 117) suggesting that terminal complement activation (C5a and C5b-9) may be a shared feature across these conditions.

Burwick et al. first showed that urine C5b-9 levels are increased more than 4-fold among women with severe preeclampsia. This finding has been validated across three different clinical studies (80, 82, 117). In contrast to plasma C5b-9 levels, increased urinary C5b-9 levels appear to be more specific to women with severe preeclampsia (80, 117). In one study, urinary excretion of C5b-9 was detected in 96% of cases with severe preeclampsia compared to just 12% of controls with chronic hypertension and 8% of healthy controls (80). In a multicenter observational study, Burwick et al. showed that urine C5b-9 concentrations differentiated preeclampsia with severe features from other hypertensive disorders of pregnancy, with area under the receiver operating characteristic curve (ROC) of 0.74 (95% CI, 0.68–0.80) (117). Notably, 40% of women with severe preeclampsia (42/104) had urine C5b-9 concentration 22 ng/mL or greater compared with 0% (0/137) of those with chronic or gestational hypertension (P<0.001) and 11% (6/57) of those with preeclampsia without severe features (P<0.001). After multivariable logistic regression, including adjustment for total urine protein, women with urinary C5b-9 ≥22 ng/mL had tenfold increased odds of having severe preeclampsia (adjusted OR 10.0, 95% CI 3.5–29, P<0.001).

Under physiological conditions, urinary excretion of C5b-9 is not expected due to its large size (greater than 1,000,000 Daltons) (118). However, C5b-9 may form at the glomerular membrane with shedding into the urine, or C5b-9 may result from complement-mediated inflammation and cellular injury at the proximal tubule (83–85, 119, 120). Burwick et al. showed that urinary excretion of C5b-9 is increased in association with other urinary proteins such as albumin and KIM-1, indicative of glomerular and proximal tubule injury (83). Guseh et al. demonstrated that urinary excretion of C5b-9 is also strongly associated with the anti-angiogenic state in preeclampsia, characterized by plasma elevation of sFLT1 and suppression of PlGF and VEGF levels (121). Pregnant women with detectable levels of C5b-9 in the urine, compared to those without detectable levels, had significantly higher plasma levels of sFLT1 (32,029 versus 4556 pg/mL, P < 0.0001) and significantly lower levels of PlGF (15.6 vs. 226 pg/mL, P < 0.0001) and VEGF (119 vs. 153 pg/mL, P = 0.001).

In addition to systemic effects, C5b-9 deposition contributes directly to placental injury in preeclampsia. Rampersad et al. showed that under hypoxic conditions, placental trophoblast deposition of C5b-9 increases with enhanced apoptotic cell death (122). The surface density of C5b-9 staining in preeclamptic placentas is increased 2-fold compared to placentas from uncomplicated pregnancies. Subsequent studies by Banadakoppa et al. (47) and Collier et al. (48) similarly found increased trophoblast deposition of C5b-9 in placentas from women with preeclampsia and HELLP syndrome, compared to placentas from healthy pregnancy. In the study by Collier et al., the degree of placental C5b-9 staining was similar in preeclampsia and HELLP syndrome (48).

Factor D, also known as adipsin, is a key protein in the alternative pathway, where it facilitates the cleavage of Factor B and formation of the C3 convertase (C3bBb). In addition to its immunologic role, Factor D is produced in large quantities by adipose tissue and contributes to metabolic regulation. Mice deficient in adipsin (Factor D) are fertile with no obvious abnormalities (123). However, when subjected to diet induced obesity, glucose homeostasis is impaired (124). These data suggest that without Factor D, the alternative pathway C3 convertase is not generated, resulting in reduced C3a and subsequent disruption of glucose regulation. In humans, serum levels of adipsin are significantly higher in late pregnancy in preeclampsia (125) and potentially even in early pregnancy (126) compared to normal pregnancy, suggesting it may play an important role in fat metabolism and glucose homeostasis that are altered during gestation and preeclampsia.

Preeclampsia is a condition unique to humans, making it inherently challenging to model in animals. Maternal hypertension is generally considered an essential feature for an animal model, while other characteristics, such as abnormal placental development or dysfunction, proteinuria, angiogenic imbalance, endothelial dysfunction, and fetal growth restriction are desirable to more accurately replicate the human condition. The strengths and limitations of different animal models used in pre-clinical studies have been recently reviewed (127) with no single animal model emerging as superior. In this review, we have focused on animal models of preeclampsia in which the role of complement has been investigated; however, complement involvement remains largely uninvestigated in the majority of available animal models. Collectively, these animal studies of complement in preeclampsia and pregnancy point to the complexity and heterogeneity of preeclampsia pathogenesis. This is paralleled by the human studies where evidence for complement activation differs with severity of the preeclampsia and the clinical scenarios. Consequently, a one-size-fits-all treatment approach is unlikely to be effective, highlighting the need for individualized therapeutic strategies tailored to the underlying mechanisms in each patient.

Human Complement Receptor 1 (CR1; CD35) reacts with C3b and C3b degradation products and is present on human erythrocytes, but not rodent erythrocytes. Immune complexes tagged with C3b degradation products are cleared via the reticuloendothelial system by a CR1 dependent mechanism. A study by Feinberg et al. (128) reported reduced CR1 expression in erythrocytes of preeclamptic patients compared to normal pregnancy, suggesting that the reduced clearance of immune complexes may add to the general inflammation associated with preeclampsia. This reduction in CR1 could be inherited or acquired based on studies in autoimmunity, suggesting that evaluation of genetic variations in CR1 should be investigated in preeclampsia.

Mice with a homozygous deficiency in Crry exhibit embryonic lethality before 10 dpc (days post coitus) (129) whereas the heterozygote embryos survive. The death of homozygous embryos is associated with excessive C3 deposition and neutrophil infiltration. This lethality is rescued in a C3-deficient animal providing evidence that fetal demise is dependent on C3. Further studies have shown that the maternal alternative complement pathway is primarily responsible for this outcome since neither IgM nor C4 deficiency rescues the Crry homozygous knockouts, whereas Factor B deficiency does (130). In addition, C6 deficiency is ineffective in rescue indicating that the Membrane Attack Complex is not critical (131). While much has been learned from these global knockout models, recent work by Banadakoppa has refined our understanding using a placenta-specific approach. By employing intrauterine injection of shRNA (short hairpin RNA that acts as silencing RNA) to dose dependently reduce Crry in the mouse placenta at 10.5 dpc, after the critical window for complement-mediated embryonic lethality in homozygous knockouts, the researchers observed that ~ 30% reduction in placental Crry led to increased C3 deposition in the labyrinth at 17.5 dpc, decreased fetal weight and increased blood pressure in the dam (132) – symptoms indicative of preeclampsia. This data establishes the importance of placental Crry in maintaining normal fetal growth and blood pressure during pregnancy and reinforces the concept that the extent of complement activation modulates the severity of preeclampsia.

In mice, CD55 is not expressed in trophoblasts or early embryos, and mice deficient in the GPI anchored form of DAF are fertile (133). Although DAF is detectable in the mouse placenta by 10.5 dpc (130) this timing does not align with early embryonic loss observed in Crry-knockout mice, suggesting that DAF does not play a critical role in preventing complement-mediated fetal demise. CD46, another complement regulatory protein, is expressed in mouse testes (130) and plays a role in the interaction between oocyte and sperm during fertilization (134, 135). CD46 can also act as a C3b receptor on the surface of the lymphocyte for intracellularly generated C3b (134, 135).

Mice deficient in CD59 are healthy and fertile (136), suggesting that CD59 is not essential for a normal pregnancy. However, in preeclampsia, placental deposition of C5b-9 is increased which may be due to increased complement activation or due to impaired regulation by CD59. Buurma et al. found that placental mRNA expression of complement regulatory proteins CD55 and CD59 were increased 2-fold and 4-fold, respectively, in placentas of women delivered for preeclampsia compared to placentas from uncomplicated deliveries (105). The investigators hypothesized that upregulation of CD55 and CD59 in the face of increased complement activation suggests the presence of a feedback mechanism to maintain trophoblast integrity. An observational case control study of preeclampsia with severe features also found that plasma levels of CD59 were elevated while urinary levels were reduced, relative to normotensive controls (137). Although soluble (plasma) CD59 is less effective than its membrane-bound form, it may still contribute to systemic complement regulation in preeclampsia.

The family of CSMD (CUB and Sushi Multiple Domains) proteins includes CSMD1, 2 and 3 with high expression in the brain. The primary protein sequence of CSMD proteins is homologous with complement control proteins and they have been shown to have the ability to regulate complement activation at the C3 convertase and/or C5b-9 steps in complement activation (5, 6). With mutations in the CSMD1 protein, male and female infertility can occur (138). In addition, in CSMD1 knockout mice neurodevelopmental abnormalities occur, and altered expression of CSMD1 occurs in the peripheral blood of schizophrenic patients (139). CSMD3 knockout progeny have neurodevelopmental abnormalities consistent with autism, as well as fetal growth restriction similar to offspring of preeclamptic pregnancies.