Maria Laura Bertolaccini
Introduction
In clinical practice, anticardiolipin antibody (aCL) detected by ELISA and lupus anti- coagulant (LA) detected by clotting assays have been standardized for the diagnosis of the antiphospholipid syndrome (APS). However, it is now established that antiphos- pholipid antibodies (aPL) are a large and heterogeneous family of immunoglobulins, which, despite their name, do not seem to bind phospholipids, but are directed to plasma proteins with affinity for anionic surfaces (i.e., phospholipids).
Amongst these phospholipid binding proteins, the best studied is β
2-glycopro- tein I ( β
2-GPI), which bears the cryptic epitope(s) for aCL binding. These epitopes are exposed when β
2-GPI binds to negatively charged phospholipids such as cardi- olipin, or irradiated plastic plates [1]. Several studies have highlighted the significance of anti– β
2-GPI antibodies (anti– β
2-GPI) as an alternative ELISA with higher specificity than the conventional aCL ELISA [2–4].
Prothrombin, another phospholipid binding protein, was first proposed as a pos- sible co-factor for LA by Loeliger in 1959 [5]. In subsequent years, the interest regarding this protein has increased and several groups have investigated the significance of antiprothrombin antibodies.
Prothrombin
Prothrombin is a single chain glycoprotein, synthetized in the liver, recognized very early as the prime contributor to the blood coagulation process. This protein is found in plasma at a concentration of around 2.5 µmol/L. Its gene spans 21 kilobase pairs [6] on chromosome 11. Mature prothrombin contains 579 amino acid residues with a molecular mass of 72 kD, including 3 carbohydrate chains and 10 γ-car- boxyglutamic acid residues [7].
The tenase complex, entailing factor Xa and factor V, calcium and phospho- lipids as co-factors, physiologically activates prothrombin. Once negatively charged, phospholipids bind prothrombin and tenase converts prothrombin to thrombin, which triggers fibrinogen polymerization into fibrin [8]. In addition, thrombin binds thrombomodulin on the surface of endothelial cells and activates protein C, which then exerts its anticoagulant activity by digesting factor V and depriving in this way the tenase complex from its most important co-factor. Due
319
to this negative feedback pathway prothrombin/thrombin behaves as an “indi- rect” anticoagulant.
The prothrombin molecule consists of 3 functional domains: Gla, kringle, and catalytic. During its liver byosinthesis, prothrombin undergoes γ carboxylation (10 glutamic acid residues in proximity of its amino terminus). These γ carboxyglu- tamic residues, known as Gla domains and located on fragment 1 of the prothrom- bin molecule, are essential for the calcium dependency of phospholipid binding to prothrombin, necessary for the conversion of prothrombin to biologically active α- thrombin. Two kringle domains follow this region and are involved in pro(throm- bin) binding to fibrin [6]. Tenase selectively hydrolyses 2 peptide bonds on the catalytic domain of the prothrombin molecule. Cleavage at Arg273–Thr274 results in the liberation of prothrombin fragment 1+2 (residues 1–273) and prothrombin 2 (residues 274–581); further cleavage at Arg322–Ile323 results in the formation of α- thrombin. The latter, one of the most potent enzymes known, not only converts fibrinogen to fibrin but acts upon factors V, VIII, XIII, protein C, platelets, and endothelial cells [9]. The schematic representation of the prothrombin molecule is shown in Figure 26.1.
Figure 26.1. Schematic representation of the prothrombin molecule. Cleavage at Arg273–Thr274 results in the
liberation of prothrombin fragment 1+2 (residues 1–273) and prethrombin 2 (residues 274–581); further cleav-
age at Arg322–Ile323 results in the formation of α-thrombin.
Antiprothrombin Antibodies
History
In 1959, Loeligler [5] was the first to describe the “LA co-factor phenomenon” when he described that the addition of normal plasma to that of a patient with LA increased the degree of coagulation inhibition. A low plasma level of prothrombin was also found; suggesting that the co-factor associated with this phenomenon was most likely to be prothrombin. Later, Rapaport et al [10] described the case of a child with LA who underwent recurrent bleeding episodes. Further investigation showed a severe prothrombin deficiency, a prolonged prothrombin time, and a pro- longed activated partial thromboplastin time (aPTT). In 1983, Bajaj et al [11] were the first to provide evidence of the presence of non-neutralizing prothrombin- binding antibodies in 2 patients with LA and hypoprothrombinemia. These anti- bodies bound prothrombin without inhibiting its conversion to thrombin. The authors postulated that hypo-prothrombinemia resulted from the binding of anti- bodies to prothrombin and rapid clearance of these complexes from the circulation.
In 1984, Edson et al [12] demonstrated prothrombin–antiprothrombin complexes in the plasma of patients with LA but without severe hypo-prothrombinemia in pro- thrombin crossed immunoelectrophoresis experiments. These findings were later confirmed by Fleck et al [13], who found antiprothrombin antibodies in 72% of the LA-positive patients and showed that these antibodies had LA activity. In 1991, Bevers et al [14] highlighted the importance of antiprothrombin antibodies in causing LA activity. After incubation with cardiolipin containing liposomes, the LA activity remained in 11/16 patients with LA. These 11 samples demonstrated LA activity in a phospholipid-bound prothrombin dependent fashion. Subsequently, Oosting et al [15] showed that 4/22 LA inhibited endothelial cell mediated prothrombinase activity and the IgG fraction containing LA activity bound to phospholipid–prothrombin complex. In 1995, Arvieux et al [16] showed that antiprothrombin antibodies could be detected by a standard ELISA using prothrombin coated onto irradiated plates.
Immunological Characteristics
Antiprothrombin antibodies are commonly detected by ELISA using irradiated plates
[16] or in complex with phosphatidylserine [17, 18]. The mode of presentation of
prothrombin in solid phase seems to influence its recognition by antiprothrombin
antibodies. In fact, antiprothrombin antibodies cannot bind when prothrombin is
immobilized onto non-irradiated plates [16, 17], but does so if prothrombin is immo-
bilized on a suitable anionic surface, adsorbed on γ-irradiated plates, or exposed to
immobilized anionic phospholipids. An analogy between the behavior of these anti-
bodies and anti– β
2-GPI has been suggested. Antiprothrombin antibodies might be
directed against cryptic or neoepitopes (antigens) exposed when prothrombin binds
to anionic phospholipids and/or may be low affinity antibodies binding bivalently to
immobilized prothrombin. Wu et al [19] observed that human prothrombin under-
goes a conformational change upon binding to phosphatidylserine containing sur-
faces in the presence of calcium. On the other hand, Galli et al [17] demonstrated
that antiprothrombin antibodies could be of low affinity and suggested that pro-
thrombin complexed with phosphatidylserine could allow clustering and better ori-
entation of the antigen, offering optimal conditions for antibody recognition.
A high percentage of antiprothrombin antibodies have species specificity for the human protein [14], but a minority react with bovine prothrombin [20]. The epitope(s) recognized by these antibodies are being investigated. Rao et al [21]
demonstrated binding of aPT to prethrombin 1 and fragment 1, as well as the whole prothrombin molecule, when using purified IgG preparations from LA-positive patients. One of these antibodies reacted with immobilized thrombin. Malia et al [22]
also demonstrated binding of antiprothrombin antibodies to fragment 1+2. Recently, Akimoto et al [23] showed that 61.5% of aPT bound to fragment 1 and 38.4% to prethrombin (fragment 2 + α-thrombin). Overall, these data suggests that the domi- nant epitopes are likely to be located near the phospholipid binding site of the pro- thrombin molecule, although they may have a heterogeneous distribution.
It has been reported that most LA depend on the presence of phospholipid bound prothrombin, as well as phospholipid bound β
2-GPI, and the anticoagulant properties of aPT have been studied by several groups. Perpimkul et al [24] showed LA activity due to antiprothrombin antibodies in 9/10 samples from LA-positive patients. Galli et al [17] and Horbach et al [25] reported the existence of 2 types of circulating anti- prothrombin antibodies which may be distinguished on the basis of their effect in coagulation assays: (1) functional, which cause LA activity, and (2) non-functional, which do not contribute to the LA activity, probably caused by a different epitope specificity of antiprothrombin antibodies [26]. Recently, a human monoclonal antiprothrombin antibody with LA activity has been raised and a semi-quantitative LA assay has been established in an attempt to help in the standardization of this test [27].
Clinical Significance
Prothrombin appears to be a common antigenic target of aPL [28]. However, its clinical significance is far from clear. Most of the studies available in the literature investigated the clinical significance of aPT detected by ELISA using prothrombin coated on irradiated plates and positive correlations with the clinical manifestations of the APS have been reported [29–31]. Little data are available on antibodies directed to phosphatidylserine–prothrombin complex (aPS-PT).
Antiprothrombin Antibodies and Thrombotic Events
Petri et al [29] reported that aPT have potential predictive value for thrombosis in a cohort of patients with systemic lupus erythematosus (SLE). Subsequent studies failed to show aPT as a risk factor for thromboembolic events, undoubtedly reflecting the heterogeneity and lack of appropriate standardization for the detec- tion of these antibodies. Pengo et al [32] found no correlation between the presence of aPT and thrombosis in 22 APS patients with thrombosis. Galli et al [33] showed aPT in 58% of APS patients; however they found no correlation with thrombotic events. At variance, Puurunen et al [30], Horbach et al [25], and Muñoz-Rodriguez et al [34] reported a positive correlation between aPT and thrombosis in SLE.
Puurunen et al [30] found a positive correlation between the presence of aPT and
deep vein thrombosis (DVT) in a SLE population. Horbach et al [25] investigated
the clinical significance of aPT in 175 patients with SLE and found that both IgG and
IgM aPT were more frequent in patients with a history of venous thrombosis. We
also reported a correlation between the presence of aPT and the occurrence of vas-
cular events when we studied 207 patients with SLE [31]. These findings were
expanded by Muñoz-Rodriguez et al [34], who found an association between the
presence of aPT and thrombotic events in a cohort of 177 patients with various autoimmune diseases.
A recent systematic review of the literature [35] showed that no clear associations with thrombosis are found for antiprothrombin antibodies, irrespective of isotype, site, and type of event and the presence of SLE. This may be explained by different detection methods and by the presence of antibodies directed to different epitopes resulting in different functional properties [17, 25, 26].
Vaarala et al [36] showed a predictive value of 2.5-fold increase in the risk of myocardial infarction or cardiac death in middle-aged men with high levels of aPT.
A nested case-control study estimated aPT to increase the thrombotic risk in men with DVT or pulmonary embolism [37]. Subsequent studies confirmed the associa- tion between aPT and venous thrombosis in patients with and without LA and/or aCL [38, 39]. Recently, Zanon et al [40] reported, after multivariate analysis, that aPT were likely risk factors for recurrent thromboembolism in their population of patients with acute venous thromboembolism.
There are only few papers on the significance of antiprothrombin antibodies when using the phosphatidylserine–prothrombin complex as antigen. In 1997, Galli et al [17] suggested that the prevalence of antiprothrombin antibodies increased from 58%
when using prothrombin coated on irradiated plates (aPT) to 90% when prothrombin was coated with phosphatidylserine (aPS-PT). Funke et al [41] reported that aPS-PT conferred an odds ratio (OR) of 2.8 for venous thrombosis and of 4.1 for arterial thrombosis in patients with SLE. Atsumi et al [18] supported these data by showing that the presence of aPS-PT conferred an OR of 3.6 for APS in 265 Japanese patients with systemic autoimmune diseases.
Table 26.1. Antiprothrombin antibodies and thrombosis: data reported in the literature.*bl/TT*bg
Author Study group N Correlation
Petri et al [29] SLE 100 Yes
Vaarala et al [36] Myocardial Infarction 106 Yes
Pengo et al [32] aPL 22 No
Puurunen et al [30] SLE 139 Yes (DVT)
Horbach et al [25] SLE 175 Yes (venous)
Galli et al [17] aPL 59 No
Swadzba et al [42] SLE + LLD 127 No
Forastiero et al [38] aPL 233 Yes (venous)
Palosuo et al [37] DVT + PE 265 Yes
Bertolaccini et al [31] SLE 207 Yes
Martinuzzo et al [43] Pulmonary hypertension 54 No
Guerin et al [44] Autoimmune diseases 265 No
Funke et al [41] SLE 97 Yes
Sorice et al [45] SLE 38 Yes (APS)
Lakos et al [46] Autoimmune diseases 70 Yes
Muñoz-Rodriguez et al [34] APS 177 Yes
Atsumi et al [18] Autoimmune diseases 265 Yes
Galli et al [47] aPL 72 No
Pasquier et al [48] DVT 241 Yes
Nojima et al [49] SLE 124 Yes
Previtali et al [50] Thrombosis 79 No
Simmelink et al [51] LA positive 46 Yes
Zanon et al [40] VTE 236 Yes
N = number of patients; SLE = systemic lupus erythematosus; aPL = antiphospholipid antibodies; LLD = lupus-like disease; DVT = deep venous thrombosis; PE = pulmonary embolism; APS = antiphospholipid syndrome; LA = lupus anticoagulant; VTE = venous thromboembolism.