Blank et al. [21, 22] hypothesized that molecular mimicry between infectious pathogens and β₂GPI may serve as the origin of aPL and APS . Their theory was based on two lines of evidence: the striking association between APS and infectious agents and a strong amino acid similarity between β₂GPI-derived peptides and various common pathogens (Table 3.2). Using a hexapeptide phage display library, they [21] identified several synthetic peptides, which exhibited high homology with proteins from the membrane particles of different bacteria and viruses, as target epitopes for monoclonal aβ₂GPI derived from APS patients. For example, the sequence LKTPRV showed homology to eight different bacteria, such as Pseudomonas aeruginosa, and five kinds of viruses, such as human cytomegalovirus, while the sequence TLRVYK had homology to other bacteria and viruses. Furthermore, by neutralizing pathogenic aβ₂GPI, these peptides inhibited both endothelial cell activation in vitro and induction of experimental APS in vivo.

To demonstrate that molecular mimicry can trigger experimental APS, Blank and coworkers [7] immunized naïve mice with microbial pathogens that share structural homology with the TLRVYK hexapeptide. All immunized mice developed anticardiolipin antibodies (aCL) and aβ₂GPI; the highest levels of aβ₂GPI were found in mice immunized with Haemophilus influenzae, Neisseria gonorrhoeae, Candida albicans, or tetanus toxoid, while the lowest aβ₂GPI levels were found in mice immunized with Klebsiella pneumoniae. Passive transfer of anti-TLRVYK antibodies (from immunized mice) into naïve mice resulted in thrombocytopenia, lupus anticoagulant (LA) activity, and increased fetal loss, i.e., experimental APS similar to mice injected with a pathogenic monoclonal aβ₂GPI [7]. These findings demonstrate that bacteria with protein sequences homologous to β₂GPI can induce aβ₂GPI and APS manifestations [7], and provide evidence for a role for molecular mimicry in experimental APS.

Gharavi et al. [23] induced circulating aβ₂GPI into naïve mice by immunizing with synthetic peptides derived from bacterial or viral proteins that show sequence similarity with a 15 amino acid peptide (GDKV) in the phospholipid-binding domain (Domain V) of β₂GPI. The synthetic peptides included regions from the proteins of human adenovirus, cytomegalovirus, and Bacillus subtilis. Mice immunized with the peptides produced high levels of aCL and aβ₂GPI, suggesting that viral and bacterial proteins may function like β₂GPI and produce aPL through molecular mimicry of β₂GPI. β₂-glycoprotein I-derived synthetic peptides from regions other than Domain V, such as peptide NTLKTPRV from Domain I [21], also share sequence similarities with common bacterial antigens and interact specifically with aβ₂GPI in mice, decreasing its thrombogenic potential [24]. Finally, the relationship between pathogens and β₂GPI extends beyond sequence homology. β₂-glycoprotein I binds to bacterial lipopolysaccharide (LPS), which is recognized by toll-like receptor 4 (TLR4) [25, 26], and functions as an in vivo scavenger of LPS [25]. The peptide sequence in Domain V responsible for LPS binding is conserved in all mammals [25]. An association between TLR4 gene polymorphisms and APS has been reported [27].

The association of infection with aPL and APS has been most thoroughly investigated in two infectious diseases, leprosy and syphilis. Studies in these infections highlight the need for astute differential diagnosis and careful characterization of the observed aPL in determining the pathogenic role of infection-induced aPL.

Some studies have reported much lower rates of aPL: 3% aβ₂GPI and 37% aCL positivity in a cohort of 35 multibacillary leprosy patients from Egypt [36]. Because these patients had leprosy for 15.2 ± 9.2 years, it is possible that their leprosy and any associated reactions (including aPL) would have resolved. No APS or thromboembolic phenomena were reported in these patients. As antibody levels and immune complex levels decrease as bacterial burden decreases, it is important to consider the stage of the disease when assessing aPL and APS. To evaluate patients prior to anti-leprosy therapy, Baeza et al. [37] studied 30 untreated multibacillary LL patients, of whom 23 (77%) were positive for IgG aCL and 23 (77%) were positive for IgM aCL. Additionally, 25 (83%) of LL sera bound to non-bilayer lipid arrangements containing mycolic acid. Levels of aCL IgG and IgM correlated with antibody reactivity to non-bilayer phospholipid (r = 0.77 and r = 0.69, respectively, p < 0.0001). The authors hypothesize that antibodies to non-bilayer phospholipid may disrupt cellular membranes, leading to the release of potentially immunogenic cellular components, such as aCL.

Deep vein thrombosis (DVT) is an emerging risk in patients with multibacillary leprosy who receive thalidomide and corticosteroids for ENL [42–46]. The typical sequence of events preceding DVT is treatment with multi-antibiotic therapy for M. leprae, development of ENL, initiation of treatment with a corticosteroid, addition of thalidomide with corticosteroid taper, and development of DVT during the cross-taper of thalidomide and corticosteroid. In one such case, aβ₂GPI and aCL IgG were slightly elevated when the patient developed DVT after 2 months of thalidomide treatment. The antibody levels returned to normal 8 weeks after the DVT [42].

Lucio’s phenomenon , characterized by recurrent ulcerative lesions affecting mainly the lower extremities, is a severe and potentially fatal immune reaction that occurs in patients with LL. In one case report from Ecuador of Lucio’s phenomenon with APS, aCL IgM was positive, and dermal vessels were occluded by thrombi [47]. Other reports found aCL positivity with vasculitis, not thrombosis [48, 49]. Some reports lack aPL testing but suggest evidence of thrombosis [50]. Levy et al. [51] found that aCL were β₂GPI dependent in only two of 33 (6%) individuals with leprosy, both of whom had Lucio’s phenomenon. Forastiero et al. [52] compared the thrombogenic properties of IgM antibodies isolated from patients with leprosy or APS (and high levels of aCL, aβ₂GPI, and LA) in a murine model of thrombosis. They found that IgM aPL from leprosy patients did not have thrombogenic and pro-inflammatory effects in vivo, when compared to aPL from APS patients, and present this as data supporting their hypothesis that thrombosis risk may relate to aPL type.

A study using mutant β₂GPI showed that APS-derived aPL bound better to Domain V-deleted than to Domain I-deleted β₂GPI, whereas leprosy-derived aPL bound to both mutant forms [53]. Furthermore, an anti-Domain I monoclonal antibody inhibited binding of APS-derived, but not leprosy-derived, aPL to β₂GPI [53]. This difference in binding specificity may explain, at least in part, the different thrombogenic potentials of APS- versus leprosy-derived aPL.

Binding affinity and antibody isotype also play a role in aPL/β₂GPI interactions. Metzger et al. [63] found that sera from syphilis patients bind to β₂GPI under low salt conditions, but not higher (300 mM) salt conditions, in which APS sera still show strong binding, indicating that the binding affinity of aβ₂GPI in patients with syphilis is lower than that in APS patients. Together, the differences in aPL derived from syphilis versus APS patients support the low incidence of aPL-associated thrombosis in syphilis patients.

The chronic triggers that sustain aPL in APS patients are elusive. Emerging data suggest that microbiota, commensal organisms that colonize human hosts, likely contribute to APS pathogenesis [64]. “Commensalism” is a symbiotic relationship between two organisms of different species in which one derives some benefit, while the other is unaffected. The microbiota colonizes every niche of the human body, including the oral mucosa, gastrointestinal tract, sinobronchopulmonary (respiratory) tract, skin, and urogenital tract [65]. The barrier organ with the largest diversity of microbiota is the gut, which will be the focus of this section; other niches may also harbor triggers of aPL.

Gut commensals influence many aspects of innate and adaptive immunity [66]. Most notably, commensals induce key murine CD4 helper T cell subsets implicated in autoimmunity [67]. Segmented filamentous bacteria, species that colonize the murine small intestine by firmly attaching to the epithelium, are capable of inducing Th17 cells that specifically recognize segmented filamentous bacteria antigens [68]. Colonization with human colonic Clostridia species (Cluster IX and XIVa) leads to differentiation of regulatory T cells (Treg) in gnotobiotic mice (germ-free animals colonized with a defined set of microbes) [69]. Dysregulation in both Treg and Th17 cells is a well-established cellular mediator of autoimmunity. Follicular helper T cells (Tfh), a CD4 T cell subset that supports B cell antibody production, are key promotors of humoral immunity and autoantibody formation [70]. Murine Tfh are dependent on the gut microbiota [71], and B cell development occurs not only in the bone marrow but also in the gastrointestinal tract of mice [72].

The gut microbiome , extensively characterized in health and in several immune-mediated diseases [65, 73], is currently under investigation in APS patients [74]. While causal relationships between the microbiome and autoimmune diseases are difficult to establish in human studies, progress has been made in animal models. Several murine models of autoimmune disease are modulated by gut microbiota [67]. These microbiota-related effects are best demonstrated through germ-free rederivation of such models. The most dramatic examples are the loss of spondyloarthropathy and colitis when HLA-B27 transgenic animals are kept germ-free [75]; in contrast, non-obese diabetic mice (prone to type 1 diabetes) are increasingly susceptible to autoimmune destruction of the pancreatic islet cells with increasing hygiene [76]; germ-free animals develop diabetes with an almost 100% incidence [77]. Environmental factors other than cleanliness, for instance, specific dietary components, can also impact the gut microbiota and therefore immune function and autoimmunity (Fig. 3.1) [67, 78–80]. Dietary effects on microbiota observed in other autoimmune diseases [81] may have a similar impact on APS.

Vieira et al. explored the gut microbiota in the (NZWxBXSB) F1 mouse, a spontaneous model of SLE and APS [81, 82]. Mice treated with an oral broad-spectrum antibiotic cocktail (vancomycin, metronidazole, neomycin, and ampicillin), or with vancomycin or ampicillin alone, did not develop serum aβ₂GPI IgG or die from thromboemboli [82]. Bacterial load (16S ribosomal DNA) monitored in the feces of the mice was profoundly suppressed by broad-spectrum antibiotic treatment. These data suggest that gut commensals may play a role in the (NZWxBXSB) F1 model of APS.

Apoptotic cells provide a potential natural target and immunogen for aPL as well as for most autoantibodies found in SLE [83–86]. The apoptotic cell surface contains anionic phospholipid (phosphatidylserine) not present on the surface of viable cells [87–89], enabling the interaction of phospholipid-binding proteins such as β₂GPI [90] and prothrombin [91], both important target antigens of aPL (Fig. 3.2). Interaction of β₂GPI with apoptotic cells generates epitopes that are immunogenic in normal mice [92]. More recent findings indicate that β₂GPI is highly immunogenic when presented in the context of innate immune activation, such as that induced by bacterial LPS [93]. In fact, mice immunized with β₂GPI not only develop high levels of aβ₂GPI and aCL but also multiple SLE-related autoantibodies and lupus-like glomerulonephritis [93]. Salem et al. [94] showed that epitope spread in the autoantibody response from β₂GPI to multiple autoantigens is associated with a strong T cell response to β₂GPI, independent of the particular β₂GPI T cell epitope specificity. The authors propose that B cells specific for apoptotic cell-associated surface autoantigens take up apoptotic cells via antigen-specific surface IgG, leading to surface MHC class II-associated presentation of multiple apoptotic cell-derived peptides, including those from β₂GPI. In this way, a β₂GPI-specific T cell can provide help to a B cell that internalizes an apoptotic cell with surface-bound β₂GPI [93, 94].

Recently, a form of cell death called NETosis has been implicated in several autoimmune diseases, including APS and SLE [95]. Neutrophil extracellular traps (NETs) are responsible for a form of cell death that is distinct from apoptosis or necrosis. Yalavarthi et al. [96] reported that freshly isolated neutrophils from patients with primary APS have higher levels of spontaneous NET release than do those from healthy control subjects. Furthermore, exposure of neutrophils from healthy controls to APS patient serum or IgG, or monoclonal aβ₂GPI IgG, stimulates NETosis. Neutrophil extracellular traps may be immunogenic, as they present captured microorganisms and autoantigens in an inflammatory milieu that stimulates an immune response [97]. However, the role of NETosis in aPL production remains unclear.