Pathophysiology of thrombosis and pregnancy morbidity in the antiphospholipid syndrome

Authors


Olga Amengual, Department of Internal Medicine II, Hokkaido University Graduate School of Medicine, N15W7, Kita-Ku, Sapporo, Japan. Tel.: +81 11 706 5915; fax: +81 11 706 7710; e-mail: olgaam@med.hokudai.ac.jp

Abstract

Eur J Clin Invest 2012; 42 (10): 1126–1135

Abstract

In patients with the antiphospholipid syndrome (APS), the presence of a group of pathogenic autoantibodies called antiphospholipid antibodies causes arteriovenous thrombosis and pregnancy complications. To date, the pathogenicity of the antiphospholipid antibodies has been the focus of analysis. Recently, the antibodies were reported to be capable of direct cell activation, and research on the underlying mechanism is ongoing. The antiphospholipid antibodies bind to the membranes of vascular endothelial cells, monocytes and platelets, provoking tissue factor expression and platelet aggregation. This activation functions as intracellular signalling, independent of the cell type, to activate p38MAPK and the transcription factor NFκB. Currently, there are multiple candidates for the membrane receptors of the antiphospholipid antibodies that are being tested for potential in specific therapy. Recently, APS was reported to have significant comorbidity with complement activation, and it was proposed that this results in placental damage and cell activation and, therefore, could be the primary factor for the onset of pregnancy complications and thrombosis. The detailed mechanism of complement activation remains unknown; however, an inflammation-inducing substance called anaphylatoxin, which appears during the activation process of the classical complement pathway, is thought to be a key molecule. Complement activation occurs in tandem, regardless of the pathology of APS or the type of antiphospholipid antibody, and it is thought that this completely new understanding of the mechanism will contribute greatly to comprehension of the pathology of APS.

Introduction

Antiphospholipid syndrome (APS) is an autoimmune disorder defined by the persistent presence of antiphopholipid antibodies (aPL) in plasma of patients with vascular thrombosis and/or pregnancy morbidity. The clinical features and laboratory manifestations associated with aPL have considerably broadened since the first description of APS in 1983 [1] including thrombocytopenia, haemolytic anaemia, cardiac valve disease, pulmonary hypertension, nephropathy, skin ulcers, livedo reticularis, cognitive dysfunction and atherosclerosis [2].

An international consensus on classification criteria for APS was stated in Sapporo (Sapporo criteria) [3], and they were revised in 2006 in Sydney [4]. Definition of APS is made when at least one of the two clinical criteria (vascular thrombosis or pregnancy morbidity) occurs in a patient whose laboratory tests for aPL are positive (Table 1).

Table 1.   Revised classification criteria for the antiphospholipid syndrome [3]
  1. Antiphospholipid syndrome is present if at least one of the clinical criteria and one of the laboratory criteria are met.

  2. ELISA, enzyme-linked immunosorbent assay.

Clinical Criteria
 Vascular thrombosis
  ≥ 1 clinical episodes of arterial, venous or small vessel thrombosis, in any tissue or organ confirmed by objective validated criteria by imaging or histopathology in the absence of significant evidence of inflammation in the vessel wall
 Pregnancy morbidity
  ≥ 1 unexplained deaths of a morphologically normal foetus at or beyond the 10th week of gestation, or
  ≥ 1 premature births of a morphologically normal neonate before the 34th week of gestation owing to eclampsia, severe pre-eclampsia or placental insufficiency, or
  ≥ 3 unexplained consecutive spontaneous abortions before the 10th week of gestation (maternal anatomic or hormonal abnormalities and paternal and maternal chromosomal causes excluded)
Laboratory Criteria
 Lupus anticoagulant present in plasma, on ≥ 2 occasions at least 12 weeks apart, detected according to the guidelines of the International Society on Thrombosis and Haemostasis
 IgG and/or IgM anticardiolipin antibodies present in medium or high titre in serum or plasma, on ≥ 2 occasions at least 12 weeks apart, measured by a standardised ELISA
 IgG and/or IgM antiβ2 glycoprotein I antibodies present in titre >99th percentile, in serum or plasma, on ≥ 2 occasions at least 12 weeks apart, measured by a standardised ELISA

The relevant antibodies found in APS are directed against specific plasma proteins that possess an affinity for anionic phospholipids, such as β2 glycoprotein I (β2GPI) and prothrombin [5,6]. APL can be categorised into those antibodies detected by solid-phase enzyme-linked immunosorbent assays (ELISA) such as anticardiolipin antibodies, antiβ2GPI antibodies or those that prolong phospholipid-dependent coagulation time, called lupus anticoagulant.

There are two aspects of APS, vascular manifestations and pregnancy complications. Thrombus formation is the key event of vascular manifestations in APS, and many pathogenic mechanisms have been proposed to explain the thrombotic predisposition in this syndrome. However, obstetrical complications in patients with APS cannot be caused solely by thrombosis in the uteroplacental vasculature, and additional pathways have been raised to pregnancy problems in APS [7].

The mechanisms of thrombosis production in patients with APS are not completely clarified. However, the interaction between aPL and cells involved in the regulation of haemostasis is one of the mechanisms responsible of the thrombophilic state in APS. The aPL-cell interaction induces a perturbation in the cells that results in a pro-thrombotic/pro-inflammatory response and subsequently thrombosis.

Complement activation, one of the mechanisms related to obstetric complications in APS has also been involved in the production of thrombosis in patients with aPL. In this manuscript we discuss the aPL-cell interaction and the role of complement in the aPL-associated complications as the major pathogenic mechanism (Table 2).

Table 2.   Antiphopholipid antibodies (aPL)-mediated pathogenic mechanisms
  1. β2GPI, β2 glycoprotein I.

aPL-mediated thrombosis
 Interference with the components of the coagulation cascade
  Protein C pathway
  Protein Z pathway
  Contact activation pathway
  β2GPI-thrombin interaction
 Impairment of fibrinolysis
 Cell interaction
  Induction of proinflammatory phenotype on endothelial cells
  Induction of procoagulant activity on endothelial cells and monocytes
  Release of membrane-bound microparticles
  Pro-coagulant effects on platelets
  Disruption of the annexin V shield
 Complement activation
aPL-mediated foetal loss
 Intraplacental thrombosis
 Inflammation
 Inhibition of syncitium-trophobalst differentiation
 Disruption of the annexin V shield
 Complement activation

Antiphospholipid antibody-associated thrombosis

The association between aPL and thrombotic events is well established. Evidence from animal models of APS indicates that aPL may play a causal role in the vascular abnormalities in both the venous and arterial territories [8,9]. In an animal model of photochemically induced arterial thrombosis, monoclonal antibodies raised against human β2GPI promoted thrombus formation [8]. Ramesh et al. [9] demonstrated that aPL inhibit the activation of nitric oxide and that the resulting decline in nitric oxide production underlies the promotion of leucocyte–endothelial cell (EC) adhesion and arterial thrombosis in mice. Injection of aPL in mice increased thrombus formation, carotid artery tissue factor (TF) activity, as well as peritoneal macrophage TF activity and expression [10]. Furthermore, enhanced thrombus formation was observed in femoral vein of mice treated with aPL [11]. Vega-Ostertarg et al. [12] found that mice injected with aPL have an enlargement in the thrombus size in the postcapillary venular endothelium in the cremaster muscle. Rapid endothelial deposition of fibrinogen and intravascular platelet–leucocyte aggregates were detected by intravascular microscopy on the mesenteric vessels of rats receiving an intraperitoneal injection of bacterial lipopolysaccharide followed by infusion of immunoglobulin G (IgG) purified from patients with APS [13].

Despite the persistent presence of aPL in circulation, thrombotic events in patients with aPL only occur occasionally, suggesting that the presence of aPL is necessary but not sufficient for clot formation in vivo. The ‘two-hit hypothesis’ has been proposed in which aPL (first hit) can only exert their prothrombotic influence in the presence of another thrombophilic condition (second hit). This ‘two-hit hypothesis’ was shown in an animal model of APS in which the injection of aPL in rats only resulted in increased thrombus formation when rats were pretreated with lipopolissacharide, but not when were injected with buffer [13].

Antiphospholipid antibodies and cell interactions

The major antigen structures recognised by aPL in patients with APS are phospholipid–binding proteins, β2GPI and prothrombin, expressed on the membranes of different cell types. The antibody forms a complex with the corresponding antigen, leading to the cell perturbation, the activation of cell signalling pathways, the transcription of procoagulant substances, adhesion molecules and subsequently thrombus formation.

Studies on the pathogenicity of aPL have been carried out mainly on the corresponding target molecules especially on the function of β2GPI and their modifications by aPL. However, to evaluate whether antiβ2GPI antibodies can block the function of β2GPI is difficult as true physiological role of β2GPI in coagulation cascade is not elucidated. Individual with complete β2GPI deficiency does not have any particular phenotype [14]. Thus, the recent trend is to favour the hypothesis that the function of aPL on prothrombotic cells, via β2GPI, is more important than the function of β2GPI.

Membranes of activated platelets with negatively charged phospholipids are an important source of catalytic surface for blood coagulation. Activated factor X and thrombin are generated on activated platelets, and procoagulant microparticles shed by platelet activation. Platelets are prone to agglutinate and aggregate after exposed to aPL [15], and circulating activated platelets are found in patients with APS [16]. β2GPI binds to membranes of activated platelets and inhibit the generation of activated factor X. Antiβ2GPI antibodies interfere with this inhibition [17]. Thus, activated platelets may be a predominant immune target of antiβ2GPI antibodies and direct action of aPL in platelets contribute to APS-related thrombosis.

The endothelium is a predominant target of aPL. Pathogenic aPL binding to β2GPI cause the up-regulation of adhesion molecules [18], TF [19] and endothelin-1 [20] causing a pro-inflammatory and prothrombotic EC phenotype. Prothrombin also binds to ECs, and this binding is enhanced by a human monoclonal IgG antiprothrombin antibody, IS6. IS6 up-regulates expression of TF and E-selectin on ECs [21].

Antiphospholipid antibodies exert also effect in the stimulation of the release of microparticles from ECs [22]. Microparticle production is a hallmark of cell activation, but the role of microparticle in the pathophysiology of thrombosis has not been elucidated. Antiphospholipid antibodies bind to the negatively charged membrane of monocytes and induce TF up-regulation [23,24]. Monocytes are the source of most majority of circulating TF-bearing microparticles [25] and TF up-regulation is a major feature of monocyte activation in the APS [26].

Cell receptors for antiphosphospholipid antibody interactions

The cell activation mediated by aPL might require an interaction between phospholipid-binding plasma protein and a specific cell receptor(s). A number of potential receptors for the binding of β2GPI to cellular membranes have been identified including annexin A2, apolipoprotein E receptor 2 (ApoER2′), low-density-lipoprotein receptor (LDL-R) -related protein, megalin, Toll-like receptor (TLR) 2, TLR 4, the very-LDL-R and P-selectin glycoprotein (GP) ligand-1. β2GPI also directly binds to the platelet adhesive receptor GPIbα and to the platelet factor 4 (PF4) [27–32]. Most of these receptors are expressed on various cell types and whether those different receptors are involved in the pathophysiology of thrombosis is still matter of debate.

Annexin A2 is a receptor for tissue-type plasminogen and its ligand plasminogen. Annexin A2 is a membrane-bound protein found on the surface of ECs and monocytes, and on the brush-border membrane of placental syncytiotrophoblasts [33]. Annexin A2 interacts with the β2GPI-antiβ2GPI antibody complex on the ECs and monocyte surfaces, mediating cell activation [27,28]. The involvement of annexin A2 in aPL-mediated pathogenic effects has been reported in vitro and in vivo models [34]. However, it is unlikely that annexin A2 per se is actually involved in cellular activation because it lacks transmembrane domain. The activation of signalling responses requires the presence of another transmembrane adaptor protein(s) that associates with annexin A2 on the ECs surface [29]. TLR-4 was identified as a potential putative adaptor protein for annexin A2 [28].

Several groups reported that TLR-2 and TLR-4 are involved in aPL-mediated cell activation [30,35,36]. TLR-4 signalling was shown in ECs after the incubation with aPL [29], but a direct interaction between TLR4 and β2GPI remains to be confirmed. Binding of β2GPI to TLR2 on endothelial surface has been reported [37].

Megalin/gp33 is an endocytic receptor that internalises multiple ligands including apolipoprotein E and B100. Megalin was shown to behave as a receptor of β2GPI and β2GPI-phospholipid complex [38]. Pennings et al. [39] demonstrated that dimeric β2GPI can interact with LDL-R family members, including megalin.

Apolipoprotein E receptor 2 is a member of the LDL-R family expressed in many cell types. Studies on platelets suggested ApoER2′ as a receptor of β2GPI [40]. The blockage of the platelet ApoER2′ using a receptor-associated protein abrogated the increased adhesion of platelets to collagen induced by β2GPI-anti-β2GPI antibody complex [41]. Using a recombinant soluble form of LDL-binding domain 1 of ApoER2′, it was shown that the interaction between β2GPI and ApoER2′ mediated the aPL action in endothelium [9]. The importance of ApoER2′ in the induction of prothrombotic state mediated by aPL was confirmed in vivo in a murine model of thrombosis and using ApoER2′ deficient mice [42]. Injection of aPL caused a significant increase in thrombus formation, vascular TF activity and monocyte activation in the murine model of thrombosis, which were significantly reduced in the ApoER2′ deficient mice. Those data support the role of ApoER2′in thrombus formation in APS; however, the role of other potential receptors cannot be excluded as demonstrated by the partial protection from thrombogenic effects of aPL in ApoER2′-deficient mice.

β2 glycoprotein I directly binds to GPIbα subunit of the platelet adhesion receptor GPIb/IX/V in vitro [35,36]. The platelet GPIbα subunit has the von Willebrand factor as the most important ligand, but also serves to localise factor XI and thrombin on the platelet surface. Binding of β2GPI to GPIbα enables antiβ2GPI antibodies, directed against domain I, to activate platelets, resulting in thromboxane production and also to the activation of the phosphoinositol-3 kinase (PI3-kinase)/Akt pathway [31] contributing to the platelet adhesion and aggregation.

The involvement of Fcγ receptor on cellular activation has been investigated in vivo [8] and in vitro studies on platelets [26], monocytes [26] and ECs [28]. Results suggest that this receptor is not strictly necessary for cellular activation.

The direct binding of β2GPI to PF4 derived from platelet granules has been reported [43]. PF4 is a member of the C–X–C chemokine family secreted by activated platelets and has ability to bind to the platelets surfaces. PF4 contributes to the natural dimerisation of β2GPI, leading to the stabilisation of β2GPI binding onto the phospholipid cell surfaces which facilitates the antibody recognition. The β2GPI-PF4 complex is strongly recognised by serum of patients with APS [43]. Moreover, platelets may be activated by β2GPI-antiβ2GPI antibody-PF4 or β2GPI-PF4 complexes. Almost every cell type can be a source of PF4 especially under some stimulation. Both, β2GPI and PF4 are abundant in plasma; thus, the preformed β2GPI-PF4 complexes may prime several pro-coagulants cells culminating in coagulation.

Those potential receptors proposed to be involved in the aPL-mediated cell activation have significantly increased in the last years, and additional studies are needed to clarify their biological and pathological roles.

Signalling pathways of cell activation

The signal transduction mechanisms involved in aPL-mediated cell activation have been the centre of interest for many researchers. How pathogenic aPL recognition of phospholipid-binding proteins on the cell surface elicits a transmembrane signal to modify intracellular events is not completely understood.

The adapter molecule myeloid differentiation protein (MyD)88-dependent signalling pathway and the nuclear factor kappa B (NFkB) have been involved in the ECs activation by aPL [44,45]. Incubation of ECs with antiβ2GPI antibodies resulted in a redistribution of NFkB from the cytoplasm to the nucleus, and this effect was accompanied by an increased expression of TF and leucocyte adhesion molecules [46]. The p38 mitogen-activated protein kinase (MAPK) pathway is an important component of intracellular signalling cascades that initiate various inflammatory responses. It is recognised that the p38 MAPK pathway has a crucial role in mediating the effect of aPL in different cell types [24,47,48]. Activation of p38 MAPK increases activities of cytokines such as tumour necrosis factor (TNF) alpha, IL-1β and macrophage inflammatory cytokine 3β [24,36]. Monocytes stimulated by monoclonal antiβ2GPI antibodies from patients with APS induce phosphorylation of p38 MAPK, a locational shift of NFkB into the nucleus and up-regulation of TF expression. Such activation was not seen in the absence of β2GPI, indicating that the disturbance of monocyte by anti-β2GPI antibodies is started by interaction between the cell and the autoantibody-bound β2GPI [24,44]. The implication of p38 MAPK in cell activation has been also demonstrated in platelets [47] and ECs [48]. Pretreatment of platelets with p38 MAPK-specific inhibitor, SB203580, completely abrogated aPL-mediated platelet aggregation. The induction of TF expression was also reported through the simultaneous activation of NFkB via the MAPK pathway and of the MEK-1/ERK pathway, but an inhibitor of the MEK-1/ERK pathway could not suppress the TF expression, implying the main role of p38 MAPK in those reactions [49].

Purified IgG from APS patients with venous thrombosis, without pregnancy morbidity, caused phosphorylation of NFkB and p38MPK and up-regulation of TF in monocytes. These effects were not seen with IgG fractions from patients with obstetric APS alone, suggesting that aPL from patients with different clinical aspects of APS may trigger different signalling responses [44]. Figure 1 shows the procoagulant cell activation as one of the pathogenic mechanisms of thrombosis mediated by aPL.

Figure 1.

 Pathogenic mechanisms of cell activation mediated by antiphosphopholipid antibodies. Antiphospholipid antibodies interact with monocytes or endothelial cells through binding to phospholipid-binding protein (β2GPI or prothrombin) on cell surface. This interaction might require a specific cell receptor (s) and results in p38MAPK phosphorylation, nuclear translocation of NFkB and up-regulation of procoagulant substances and adhesion molecules, and subsequently thrombus formation. p38 MAPK, p38 mitogen-activated protein kinase; NFkB, nuclear factor kappa B; β2GPI, β2 glycoprotein I; PAI-1, plasminogen activator inhibitor-1; TNFα, tumour necrosis factor alpha; TF, tissue factor.

Recently, two major findings in the antigenic structures recognised by aPL have been reported: first, the structural changes in β2GPI. β2GPI can exist in two different conformations, plasma β2GPI circulates in a circular (closed) conformation, whereas after interaction with antiβ2GPI antibodies undergoes a major conformational change into a fishhook-like (open) structure [50]. Second, the finding that β2GPI can be reduced by thioredoxin 1 (TRX-1). β2GPI treated with TRX-1 generate free thiols within β2GPI, a process that may affect the function of β2GPI, and may have a regulatory role in platelet adhesion [51]. Those novel biochemical findings into the structural changes that can occur within β2GPI and the consequences of these changes for the function of β2GPI might be relevant to our better understanding of the APS, but further studies are necessary to clarify their roles in the pathogenesis of APS.

Complement activation in APS

Several genes involved in SLE susceptibility have recently been identified and confirmed. Of those, particularly IRF5 [52] and STAT4 [53] have been confirmed in several studies to be clearly associated with primary APS. In a study by our institution on gene polymorphisms of STAT4 in Japanese primary patients with APS, the polymorphism rs7574865G/T was related to APS and STAT4 was considered a disease-susceptibility gene for both SLE and APS regardless of race [54]. Of the similarities between SLE and APS, the focus in recent years has turned to complement activation as a common point that has been observed in the mechanism of pathology of both diseases [55].

Hypocomplementemia caused by complement activation is commonly observed in SLE and is thought to be well correlated with disease activity. Many proposed hypotheses have explained the involvement of the hypocomplementemiathis in the disease onset mechanism in SLE. Understanding the mechanism of hypocomplementemia and complement activation in APS is important because it has the potential to be helpful not only in understanding the pathology of APS but also in clarifying the involvement of the mode of complement activation in SLE as well.

Complement system

‘Complement’ is a general term for over 20 serum proteins that function in innate immunity. They would attach to the membrane of organisms and would be activated sequentially in cascade process. They are activated by pattern recognition receptors that have evolved to recognise specific molecular patterns. The complement system has three known activation pathways according to the differences of activation elements: the classical pathway activated by an antigen–antibody complex; the lectin pathway activated by lectin, recognising glycans on microorganisms; and the alternative pathway, in which C3 is bound without recognition molecules. These three pathways culminate in activation of C3, the central step of complement activation. Activation of C3 leads to the production of pro-inflammatory fragment C3a which is called anaphylatoxin and sequentially activates C5. Activation of C5 also produces anaphylatoxin C5a which has especially strong inflammation-inducing effect. C4, a classical pathway protein, also produces C4a on its activation which is another member of anaphylatoxin but has little effect on inducing inflammation. C5b, an another fragment of activated C5, forms a complex with C6, C7, C8 and C9 which is called membrane attack protein (MAC). MAC produces a trans-membrane channel on the membrane of microorganism that disrupts the phospholipid bilayer, leading to cell lysis and death.

Complement activation in APS pregnancy morbidity

Complement activation was first thought to be involved in the pathophysiology of pregnancy complication in APS. Initially, it was believed that thrombosis in placentas was an important mechanism of pregnancy complication in APS. Obstructed blood flow in placenta caused by thrombosis may impair foetal growth because of exchange failure of blood between mother and foetus [56]. Placental thrombosis has been reported, and in vitro studies have confirmed that aPL have the ability to disrupt the annexin A5 anticoagulant shield on trophoblast and EC monolayers [57]. However, other reports were unable to confirm the occurrence of multiple thrombosis in the placentas of patients with APS; in fact, placental thrombosis imaging was either virtually absent or only mild in the analysis of many cases [58].

A recent study pinpointed the trophoblastic basement membrane as a target of aPL, where the aPL were thought to complicate pregnancy by directly inducing localised inflammation. Multiple peritoneal injections of human IgG with aPL activity to pregnant naïve mice enabled embryo implantation and induced considerable placental damage that resulted in foetal resorption and growth retardation. Deposits of human-derived IgG and complement were observed in a pathological study of the placentas of these mouse models, accompanied with local TNF secretion and a temporary but a clear increase in serum TNF levels [59–61].

Obstetrical problems were markedly suppressed in these models when complement was deficient or when antibodies against the complement were administered, such that complement activation was thought to be an important mechanism in pregnancy complications by aPL [62].

The anaphylatoxin C5a is especially important and causes placental damage, having a strong inflammation-inducing effect that causes localised placental inflammation and promotes TF expression in neutrophils infiltrated to the placenta [63,64]. In addition, anti-complement activation is involved in the effect of heparin preventing aPL-related pregnancy complications [65]. However, considering the fact that the mouse models were given large doses (10 mg each) of IgG fractions containing aPL in human serum, it is unclear whether they adequately reflect the behaviour of aPL in vivo in human.

The involvement of complement activation in pregnancy morbidities in patients with APS has been vigorously studied, with a series of reports suggesting a clear involvement in mouse models. In one retrospective study, complement deposition was found in the placental tissues of women positive for aPL [66]. There are some reports discussing the dysfunction of complement-regulating factor. Complement-regulating factor mutations are related to preeclampsia in patients with SLE or APS [67]. Decreased expression of decay-accelerating factor (CD55, a complement regulatory factor) in the endometrium has been confirmed in aPL-positive pregnancies (by endometrial biopsy) [68]. However, there remains a lack of conspicuous evidence that complement activation is directly related to the APS pregnancy morbidity outcome. Preliminary data from recent reports indicate that the histology of placental specimens from patients with APS shows evidence of complement activation compared with control placental specimens; however, complement deposition can be detected both in abortive specimens and in placentas at term without a clear relationship with either pregnancy outcome or therapy [45].

Although large prospective analyses are needed to demonstrate definite conclusions about the involvement of complement in APS-related pregnancy morbidity, the potential role of complement in aPL-mediated clinical manifestations should not be neglected. In addition to causing acute local inflammation, complement components are able to modulate the functions of procoagulant cells (monocyte, ECs) and decidual or trophoblast cells [69].

Complement activation in APS thrombosis

According to the positive results showing the relationship between complement activation and pregnancy morbidity in the APS murine model, an intensive study was carried out that investigated the relationship between thrombosis and complement activation in this model.

Mice deficient in C3 (C3−/−) and C5 (C5−/−) were used to investigate the role of complement activation in APS thrombosis. Each was administered aPL–IgG or control IgG, and thrombosis was induced via standardised pinch injury to the femoral vein [70]. The sizes of the thrombi in C3−/− and C5−/− mice were significantly reduced compared to those of wild-type mice. Additionally, mice treated with monoclonal anti-C5 antibody developed smaller thrombi compared with mice that did not receive the monoclonal antibody [71,72]. Complement activation using C5a production observed in the aPL-administered mice was found to induce TF expression on neutrophils, resulting in modified prothrombin time [71]. These phenomena suggest a possible mechanism by which aPL activation of complement pathway can initiate coagulation. Finally, antiβ2GP1 antibodies were found to initiate thrombus formation, with decreased thrombotic occlusions in C6-deficient rats and in mice treated with anti-C5 antibody [21].

A recent report investigated the significance of complement activation in patients with primary APS [72]. From the analysis of serum complement levels (C3, C4, CH50) and anaphylatoxins (C3a, C4a, C5a) in patients with primary APS, non-SLE connective tissue disease and healthy subjects, it revealed that complement levels were significantly lower in patients with primary APS compared with those with the other groups. Most patients with primary APS showed elevated serum levels of C3a and C4a related to hypocomplementemia. Among the patients with primary APS, no correlation was found between any particular clinical manifestation and hypocomplementemia. Hypocomplementemia is frequently found in patients with primary APS, reflecting complement activation and consumption rather than deficiency as suggested by the correlation between high serum C3a concentrations and low serum C3 levels. This conclusion strengthens the recognition of crosstalk between complement activation and prothrombotic status in APS (Fig. 2).

Figure 2.

 Pathogenic mechanisms of complement activation in antiphospholipid syndrome. Complement classical pathway which is initiated by C1q protein is significantly activated in the serum of patients with antiphospholipid syndrome. Activation of the pathway proceeds as cascade reaction producing anaphylatoxins, the fragments of the complement proteins that amplify the activation of monocytes, platelets or endothelial cells. Activation of these cells and molecules induces expression of TF or adhesion molecules and platelet aggregation. β2GPI, β2 glycoprotein I; TF, tissue factor.

Crosstalk between complement and coagulation pathways

There is lesser evidence of APS thrombosis and complement activation than pregnancy complications, but various aspects of involvement have become known for complement activation and thrombosis. Complement activation is increasingly being recognised as a major contributor of vascular inflammation [73]. Complement deposition has been frequently observed in atherosclerotic lesions [74], and accumulating evidence suggests that complement plays a significant role in ischaemia/reperfusion injury [75]. C3a and C5a enhance leucocyte recruitment and support the host inflammatory response [76]. C5a level elevations have been associated with increased cardiovascular risk in patients with advanced atherosclerosis [77].

Recently, the complement pathway has also been identified as having an effect on the coagulation pathway itself. Both have highly substrate-specific reactions that proceed with the cascading activation of many different serine proteases, but some reactions also cross from one cascade to the other. For example, activated factor XII (FXIIa), an initiator of the intrinsic coagulation pathway, degrades and activates C1, an initiator of the classical complement pathway [78], while thrombin directly degrades C5 in the absence of C3 to produce the anaphylatoxin C5a [79]. In addition, C5a increases the expression of TF [80], and the membrane attack complex degrades prothrombin to thrombin [81]. Thus, the complement and coagulation pathways have a close relationship. This mechanism is very effective at the site of trauma by causing anaphylatoxin production at place of haemostasis; complement activation induces inflammation at the trauma site and effectively prevent microorganism infiltration. It is possible that a complex combination occurs in the involvement of complement activation in APS thrombosis, such as with vascular injury, direct activation of the coagulation pathway or cellular activation.

Conclusion

Ongoing research focused on cell receptors and intracellular signalling pathways involved in the cell activation mediated by aPL substantially advance the understanding of the thrombotic mechanism in APS. Further studies are needed to clarify the biological role of the numerous potential receptors proposed for aPL–cell interaction.

Complement activation seems to be an essential factor for disease manifestation in pregnancy morbidity in patients with aPL from the results of the experiments in in vivo models. Accumulating evidences are offering promising prospects on the involvement of complement activation in thrombosis related to aPL. Although the definite conclusion that complement activation is a part of the process of disease manifestation can only be induced from the result of large prospective studies, there is no doubt that the clarification of the mechanism of complement activation in APS would be a key to a better understanding of pathogenesis of APS.

Acknowledgments

This work was supported by the Japanese Ministry of Health, Labour and Welfare, the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT) and the Japanese Society for the Promotion of the Science (JSPS). Olga Amengual is a postdoctoral researcher granted by JSPS/MEXT (ID 0940106, Project number 21-40106).

Address

Department of Internal Medicine II, Hokkaido University Graduate School of Medicine, Sapporo, Japan (K. Oku, O. Amengual, T. Atsumi).

Ancillary