IgG Autoantibodies against β2-Glycoprotein I Complexed with a Lipid Ligand Derived from Oxidized Low-Density Lipoprotein are Associated with Arterial Thrombosis in Antiphospholipid Syndrome

We recently reported [J. Lipid Res. 42 (2001), 697; 43 (2002), 1486; 44 (2003), 716] that β2-glycoprotein I (β2GPI) forms complexes with oxidized LDL (oxLDL) and autoantibodies against these complexes are present in patients with SLE and antiphospholipid syndrome (APS). The relationship of β2GPI/oxLDL complexes and IgG autoantibodies against β2GPI complexed with oxLig-1 (an oxLDL-derived ligand) with clinical manifestations of APS was studied in 150 APS and SLE patients. The β2GPI/oxLDL levels of APS patients were similar to those of SLE patients without APS, but they were significantly higher than healthy individuals. There was no difference in the complex levels among the patients with arterial, venous thrombosis, or pregnancy morbidity. IgG anti-β2GPI/oxLig-1 levels of APS were significantly higher than those of SLE without APS and healthy individuals. Further, antibody levels of APS patients with arterial thrombosis were significantly higher than those patients with venous thrombosis and pregnancy morbidity. Thus, oxidation of LDL leads the complex formation with β2GPI in SLE and APS patients. In contrast, anti-β2GPI/oxLig-1 autoantibodies were generated only in APS and were strongly associated with arterial thrombosis. These results suggest that autoantibodies against β2GPI/oxLDL complexes are etiologically important in the development of atherosclerosis in APS.


INTRODUCTION
High serum levels of antiphospholipid antibodies have been associated with thromboembolic events of both the arterial and venous vasculature, and with pregnancy morbidity (miscarriages and fetal loss). These features are major criteria for the classification of the antiphospholipid syndrome (APS), a clinical entity that may be present in the context of a systemic autoimmune disorder (secondary APS), or in the absence of an underlying disease (primary APS) Gharavi et al., 1987). Antiphospholipid antibodies, anti-cardiolipin antibodies (aCL) or lupus anticoagulants, are a heterogeneous group of autoantibodies with a possible pathogenic role in the development of the clinical manifestations of APS. These antibodies are characterized by their reactivity to negatively charged phospholipids, phospholipid/ protein complexes, and certain proteins presented on suitable surfaces (i.e. activated cell membranes, oxygenated polystyrene) Roubey, 1994).
Several plasma proteins that participate in coagulation and interact with anionic phospholipids have been described as antiphospholipid cofactors, i.e. b 2 -glycoprotein I (b 2 GPI), prothrombin, and annexin V. These protein cofactors have been shown to be relevant antigenic targets for antiphospholipid antibodies (Matsuura et al., 1990;McNeil et al., 1990). b 2 GPI is a 50 kDa single-chain polypeptide composed of 326 amino acid residues, arranged in 5 homologous repeats known as complement control protein domains. In vitro, b 2 GPI binds strongly to anionic molecules, such as negatively charged phospholipids, heparin, and lipoproteins, as well as to activated platelets and apoptotic cell membranes. Further, b 2 GPI has anticoagulant properties, as it has been shown to inhibit the intrinsic coagulation pathway, prothrombinase activity, and ADP-dependent platelet aggregation (Sheng et al., 1998). It has also been reported to interact with several elements in the protein C, protein S anticoagulant system (Merrill et al., 1999). b 2 GPI's fifth domain contains a patch of positively charged amino acids that likely represents the binding region for phospholipids (Bouma et al., 1999;Hoshino et al., 2000).
Venous thromboembolic complications represent the most common clinical finding in APS patients Ginsburg et al., 1992;Bick and Baker, 1999). However, over 25% of the patients enrolled into a European cohort of 1000 APS patients presented an arterial thrombotic event (myocardial infarction, cerebrovascular accident, angina, etc.) as the initial clinical manifestation (Cervera et al., 2002). More recently, the premature (or accelerated) development of atherosclerosis has been recognized in autoimmune patients (Ward, 1999;Aranow and Ginzler, 2000;van Doornum et al., 2002). The traditional risk factors for atherosclerosis failed to account for these changes (Esdaile et al., 2001). Increased levels of autoantibodies against oxidized low-density lipoprotein (oxLDL), phospholipids, and Lp(a), have been proposed as alternative mechanisms as well as certain biochemical and genetic abnormalities (Lockshin et al., 2001). Oxidation of LDL (oxLDL) plays an important pathogenic role in early events leading to atherosclerosis (Berliner and Heinecke, 1996;Steinberg, 1997). oxLDL is a pro-inflammatory chemotactic agent for macrophages and T lymphocytes, which have a central role in atherogenesis (McMurray et al., 1993). In addition, oxLDL has been found in human and rabbit atherosclerotic lesions (Yla-Herttuala et al., 1989), and shown to be an immunogen producing autoantibodies in patients with autoimmune disorders, such as systemic lupus erythematosus (SLE) and APS (Salonen et al., 1992;Vaarala et al., 1993). The participation of the immune system in the development of atherosclerosis is becoming apparent and some antiphospholipid antibodies may also be possible participants (Vaarala, 1996;Romero et al., 1998;Tinahones et al., 1998).
b 2 GPI has also been localized in human atherosclerotic lesions by immunohistochemical staining (George et al., 1999), which suggests a role of b 2 GPI (and antiphospholipid antibodies, i.e. anti-b 2 GPI antibodies) in atherosclerosis. In 1997, we (Hasunuma et al., 1997) reported that Cu 2þ -oxLDL, unlike native LDL, binds to b 2 GPI. In vitro macrophage uptake of oxLDL was slightly decreased in the presence of b 2 GPI, as compared to oxLDL alone. In contrast, the addition of an antiphospholipid antibody, i.e. b 2 GPI-dependent aCL (or antib 2 GPI), together with b 2 GPI, resulted in a significant increase of oxLDL uptake by macrophages. It is wellknown that oxLDL uptake by macrophages is inhibited with polyinosinic acid, a scavenger receptor blocker. However, the increased b 2 GPI and anti-b 2 GPI antibody dependent uptake was not affected by polyinosinic acid and it is most possible that macrophage Fcg receptors were involved in the binding. This mechanism may be relevant to the development of atherosclerosis in patients with APS. The b 2 GPI-specific ligand on the oxLDL particles (oxLig-1, 7-ketocholesteryl-9-carboxynonanoate) respon-sible for the oxLDL interaction with b 2 GPI has been isolated and identified. Increased macrophage uptake of liposomes (as a model of oxLDL) has also been reported when oxLig-1/b 2 GPI/antibody complexes were applied (Kobayashi et al., 2001;Liu et al., 2002). Most recently, we have reported that oxidatively modified LDL interacts in vivo with b 2 GPI, and detected b 2 GPI/oxLDL complexes, autoantibodies against b 2 GPI/oxLig-1 complexes, and IgG immune complexes containing b 2 GPI and oxLDL in serum samples from SLE and APS patients (Kobayahsi et al., 2003).
In the present study, serum levels of b 2 GPI/oxLDL complexes and IgG anti-b 2 GPI/oxLig-1 autoantibodies were measured in patients with APS, and their association with clinical manifestations of APS was assessed.
Our results indicate that oxidation of LDL leads the complex formation with b 2 GPI, and that these complexes commonly appear in the blood stream of patients with APS as well as in SLE patients with or without APS. However, autoantibodies against b 2 GPI/oxLig-1 were only generated in APS patients. Further, these antibodies showed a stronger correlation with arterial thrombosis when compared to venous thrombosis. These results may indicate etiological importance of IgG antib 2 GPI/oxLDL (oxLig-1) autoantibodies in the development of atherosclerosis in APS patients.

Patients
Serum samples from 150 patients were utilized in the study. One hundred samples were obtained from APS patients enrolled in the Registry for the APS (Oklahoma Medical Research Foundation, Oklahoma City, OKwww.slrapls.org). The clinical diagnosis of APS was based on the Sapporo criteria for the classification of APS (Wilson et al., 1999). All patients had a positive lupus anticoagulant and/or IgG b 2 GPI-dependent aCL ELISA result on 2 or more occasions. Twenty-four patients were classified as primary APS and 76 as secondary APS to SLE. Eighty-eight of the APS patients were females and 12 males. The mean age was 44.6 years (range 18 -82 years). A separate population of 50 patients meeting the 1982 ACR criteria for SLE (Tan et al., 1982), with no history of antiphospholipid antibodies, was used as control. In addition, 43 serum samples from healthy blood bank donors were also included in this study as controls.
Three major clinical manifestations for APS were recorded: venous thrombosis, arterial thrombosis and pregnancy morbidity. Venous thrombotic events included deep-vein thrombosis (DVT), pulmonary embolism (PE) and superficial phlebitis confirmed by Doppler ultrasound, venography or ventilation-perfusion scanning. Arterial thrombotic events included myocardial infarction (MI), cerebrovascular accident (CVA) or peripheral arterial thrombosis. Pregnancy morbidity was evaluated separately, including pregnancy loss after 10 weeks of gestation and/or late pregnancy complications as previously defined (Wilson et al., 1999). Fourteen of the APS patients had a history of thrombocytopenia (platelet count , 100,000 mm 3 ). In all cases, thrombocytopenia was present in combination with at least one of the above clinical manifestations, since the Sapporo criteria were used. The clinical characteristics and classification of the APS patients studied are summarized in Table I.
The Registry for the APS has been approved and monitored by the Internal Review Boards (IRB) of the Oklahoma Medical Research Foundation, New York University Medical Center and (previously) Saint Luke's-Roosevelt Hospital Center in New York City. Informed consent was given to all participants according to FDA/ICH guidelines and institutional requirements. The current project was preapproved by the Registry Advisory Board. A material transfer agreement and inter-institutional assurances were initiated in accordance with current regulations.

Monoclonal Antibodies
The following monoclonal antibodies were used to develop and calibrate the ELISA tests for measuring b 2 GPI/oxLDL complex and anti-b 2 GPI/oxLig-1 antibodies: WB-CAL-1 monoclonal antibody reactive to b 2 GPI (IgG2a, k) derived from a NZW x BXSB F1 mouse, a spontaneous model of APS (Hashimoto et al., 1992), and EY2C9 monoclonal anti-b 2 GPI antibody (IgM) established from peripheral blood lymphocytes of APS patients (Ichikawa et al., 1994). Both monoclonal antibodies bind only to b 2 GPI complexed with Cu 2þ -oxLDL and negatively-charged phospholipid, such as CL and phosphatidylserine, but not to monomeric (free) b 2 GPI in solution. 1D2 (Yamasa Corporation, Choshi, Japan) is an IgG murine monoclonal antibody specific for human ApoB-100 and the antibody binding is not affected by the oxidation of LDL.

Purification of Human b 2 GPI
Human b 2 GPI was purified from fresh normal plasma as previously described (Finlayson and Mushinski, 1967) with slight modifications. Briefly, human plasma was first precipitated with 70% perchloric acid, extensively dialyzed against Tris/NaCl buffer (pH 8.0) and concentrated before loading into a heparin column (Amersham Biosciences, Piscataway, NJ). Pooled b 2 GPI fractions were again dialyzed against sodium acetate/NaCl buffer (pH 4.8) and concentrated. This preparation was then loaded into a CM cellulose column (Sigma-Aldrich, St. Louis, MO) and b 2 GPI fractions were pooled, dialyzed against sodium acetate/NaCl buffer, concentrated at approximately 1 mg/ml and stored at 2 708C until use. The reactivity of b 2 GPI was checked by ELISA and the purity was assessed by SDS-PAGE.

LDL Purification and Oxidation
LDL was isolated by ultracentrifugation of fresh normal human plasma in EDTA/KBr solutions as described (Havel et al., 1955). LDL (d ¼ 1.019 -1.063 g/ml) was adjusted to a concentration of 100 mg/ml based on protein concentration. The LDL fraction was oxidized with 5 mM CuSO 4 in 10 mM phosphate buffer containing, 150 mM NaCl, pH 7.4 (PBS) at 378C for 12 h. Oxidation was terminated by the addition of EDTA (at a final concentration of 1 mM), and extensively PBS containing EDTA. The degree of oxidation was measured using the thiobarbituric acid reactive substance (TBARS) procedure (Ohkawa et al., 1979).

ELISA Procedure for b 2 GPI/oxLDL Complexes
In the present study, the ELISA for b 2 GPI/oxLDL complexes was performed in the presence of b 2 GPI to ensure the detection of all possible forms of oxLDL. oxLDL is predominantly present as a complex with b 2 GPI but it may be present as free oxLDL. Monoclonal antibody against complexed b 2 GPI (WB-CAL-1) was coated onto 96-well microtiter plate (Immunlon 2HB, Dynex Technologies Inc., Chantilly, VA) by incubating 50 ml/well of 5 mg/ml of WB-CAL-1 in PBS, pH 7.4, overnight at 2-48C. The plate was blocked with PBS containing 1% non-fat dry milk (nfdm) for 1 h. Fifty microliters of 30 mg/ml of human b 2 GPI in PBS was added to each well, followed by 50 ml of the serum samples diluted at 1:25 in PBS-nfdm, and incubated for 2 h at room temperature. The wells were washed 4 times with PBS containing 0.05% Tween-20 between each step. Biotinylated 1D2 (anti-human ApoB-100) antibody diluted in PBS-nfdm was added to the wells and incubated for 1 h at room temperature, followed by horseradish peroxidase (HRP)streptavidin. Color was developed with tetramethylbenzidine (TMB)/H 2 O 2 and the reaction was stopped with 0.36N sulfuric acid. Optical density was measured at 450 nm. Serum oxLDL concentration (indicated as U/ml) was calculated as a complex with b 2 GPI, against a reference curve built with 2-fold serial dilutions of a known concentration of oxLDL added to wells containing b 2 GPI. The unit value was arbitrarily derived from the concentration of the material used in the reference curve. A normal cut-off value for the assay was established at 23 U/ml by testing 43 samples from healthy blood donors (mean þ 3 standard deviations).

ELISA for IgG Anti-b 2 GPI/oxLig-1 Antibodies
The ELISA procedure used in the study has been previously described by Kobayashi et al. (2001) with slight modification. Fifty microliters of 100 mg/ml of oxLig-1 (7-ketocholesteryl-9-carboxynonanoate) in ethanol was coated onto Immunlon 2HB plates by evaporation.
The plate was blocked with 1% BSA for 1 h at room temperature and washed. Fifty microliters of 30 mg/ml of human b 2 GPI in PBS containing 0.3% BSA was added to the oxLig-1 coated wells to allow complex formation. Fifty microliters of serum or plasma samples diluted 1:100 in PBS containing 0.3% BSA were subsequently added to the wells and incubated for 1 h at room temperature. The wells were washed 4 times with PBS containing 0.05% Tween-20 between steps. Diluted HRP-conjugated antihuman IgG antibody was added to the wells and incubated for 1 h. Color was developed with TMB/H 2 O 2 and the reaction stopped with 0.36 N sulfuric acid. Optical density was measured at 450 nm. To establish the initial performance of the assay and to select a strong reactive sample to be used as control, monoclonal antibody, EY2C9, and HRP-conjugated anti-human IgM antibody were used. Level of IgG anti-b 2 GPI/oxLig-1 antibodies in samples (expressed in U/ml as defined above) was calculated against the curve prepared with a selected serum positive sample. A normal cut-off value for the assay was established at 10 U/ml by testing 43 samples from healthy blood donors (mean þ 3 standard deviations).

ELISA for aCL and Anti-b 2 GPI Antibodies
All APS samples were tested for IgG aCL and anti-b 2 GPI antibodies on commercially available ELISA test kits (Corgenix Inc., Westminster, CO), following the manufacturer's instructions. The aCL ELISA test uses exogenous bovine b 2 GPI thus measuring b 2 GPI-dependent antibodies. The anti-b 2 GPI ELISA test uses purified human b 2 GPI as antigen in the absence of exogenous phospholipids.

Statistical Analysis
Statistical analysis was performed with a SigmaStat program (SPSS Science Inc., Chicago, IL). Student's t test was performed to compare the results between different groups and Chi-square test was used to assess the relationship between antibodies and clinical manifestations. Sensitivity, specificity, positive predictive value (PPV) and odds ratio of anti-b 2 GPI/oxLig-1 antibodies were calculated by 2 £ 2 contingency table analysis. Ninetyfive percent confidence intervals for odds ratios were also calculated. Pearson's product moment correlation was performed to assess the association of individual values between variables. A p value of 0.05 or less was considered as significant.

RESULTS
Serum Levels of b 2 GPI/oxLDL Complexes Figure 1 shows that most APS patients had elevated serum levels of b 2 GPI/oxLDL complexes with a mean level of 96.7^72.3 U/ml, while none of the healthy controls reacted above the cut-off (mean 12.4^3.7 U/ml, p ¼ 5:8 £ 10 29 These results indicate that oxidation of LDL leads the complex formation with b 2 GPI and the complexes commonly appear in APS patients and SLE patients with/or without APS. In addition, b 2 GPI/oxLDL complexes were particularly high in a subgroup with apparent increased vasculopathy as evidence by both arterial and venous thrombotic history.

Serum IgG Anti-b 2 GPI/oxLig-1 Antibodies
Thirty-six percent of the APS patients had elevated levels of IgG anti-b 2 GPI/oxLig-1 antibodies with a mean level of 22.5^64.9 U/ml, significantly higher as compared with SLE patients without APS (9.1^5.1 U/ml, p ¼ 0:02) and to healthy controls (5.7^1.4 U/ml, p ¼ 0:005). There was no difference between primary and secondary APS with regard to the antibody levels. The mean IgG antib 2 GPI/oxLig-1 level of each subgroup was: 23.4^41.9 U/ml for arterial thrombosis ðn ¼ 45Þ with 40% classified as positive, 12.3^16.5 U/ml for venous ðn ¼ 39Þ with 36% positives, and 8.6^6.3 U/ml for pregnancy morbidity ðn ¼ 15Þ with 20% positives (Fig. 2). The mean level of the venous thrombosis ð p ¼ 0:05Þ and the pregnancy morbidity ð p ¼ 0:01Þ subgroups were statistically lower as compared with that of arterial thrombosis subgroup. These results indicate significantly higher serum levels of IgG antib 2 GPI/oxLig-1 antibodies in primary and secondary APS patients as compared with SLE patients without APS and healthy controls. In addition, APS patients with a history of arterial thrombosis had significantly higher antibody levels, as compared with patients with venous thrombosis or pregnancy morbidity.
Relationship of IgG Anti-b 2 GPI/oxLig-1 Antibodies with aCL and Anti-b 2 GPI Antibodies Due to the prominent presence of b 2 GPI in the antigenic mixture used to detect IgG anti-b 2 GPI/oxLig-1 antibodies, the relationship of these antibodies with b 2 GPI-dependent antiphospholipid antibodies was evaluated. FIGURE 2 Serum levels of IgG anti-b 2 GPI/oxLig-1 antibodies measured by ELISA in healthy controls, SLE without clinical or serologic manifestations of APS (diseased controls), and 100 APS patients classified into groups according to their history of arterial thrombosis, venous thrombosis or pregnancy morbidity. The cut-off (horizontal broken line) was established at 10 U/ml (mean þ 3 standard deviations from 43 healthy subjects). The horizontal solid lines indicate the mean IgG anti-b 2 GPI /oxLig-1 antibody level of each group.
IgG anti-b 2 GPI/oxLig-1 versus anti-b 2 GPI antibodies also showed a little dislocating distribution pattern. This pattern may suggest the presence of distinct populations of antibodies, some are much reactive for b 2 GPI directly and others are to b 2 GPI/oxLig-1. Twelve (27%) of the APS patients in the arterial thrombosis subgroup had antibodies reacting to both b 2 GPI and b 2 GPI/oxLig-1, while only 4 (10%) in the venous thrombosis and none in the pregnancy morbidity groups had this dual reactivity.

Comparative Clinical Performance
The clinical performance (relative sensitivity and positive predictive value-PPV) of IgG anti-b 2 GPI/oxLig-1 antibodies for the history of thrombosis (arterial and venous) and pregnancy morbidity in APS patients was evaluated by 2 £ 2 contingency table analysis. Table II shows that IgG anti-b 2 GPI/oxLig-1 antibodies were 38.6% sensitive for total thrombosis (arterial and venous combined) with a PPV of 94% ð p ¼ 0:001Þ: The specificity of this antibody for total thrombosis was 93.7%. The PPV for arterial thrombosis was 90% and for venous thrombosis 88% (p ¼ 0:002 and 0.005, respectively). The relative sensitivity for pregnancy morbidity was 20% with a PPV of 60% ð p ¼ 0:309Þ: These results indicate that IgG anti-b 2 GPI/oxLig-1 antibodies are found predominantly in those autoimmune patients who have a history of vasculopathy, with a stronger association for arterial than venous thrombosis in patients with APS.

DISCUSSION
The cholesterol that accumulates in macrophage-derived foam cells of atherosclerotic lesions is derived from circulating lipoproteins, mainly LDL, but LDL must be modified before it can induce foam cell formation (Ross, 1999). Oxidation of LDL is an effective mechanism that modifies LDL, increasing its macrophage uptake via scavenger receptors and intracellular accumulation. Several studies have demonstrated that atherosclerosis is an inflammatory disease, involving the dysregulation of cholesterol homeostasis by aberrant interactions between lipid-modulating elements and mediators of inflammation (Steinberg, 2002). Although the initiating inflammatory factor(s) remain unknown, likely candidates include oxLDL, immunological injury, homocysteine and infectious agents. An active role of antibodies in this process has been proposed (Virella et al., 2002) as recent prospective studies have indicated that b 2 GPI-dependent aCL or anti-b 2 GPI antibodies are associated with MI and stroke in men (Vaarala, 1998;Brey et al., 2001).
Our results indicate that oxidation of LDL is a common occurrence in APS and SLE patients without APS, and has demonstrated the presence of circulating b 2 GPI/oxLDL complexes in these patients (Fig. 1). Although it can be hypothesized that this might be related to chronic inflammation of the vasculature that occurs in autoimmune patients, the mechanism(s) for the increased oxidation of LDL found here are not known. b 2 GPI binds to oxLDL, not to native LDL, possibly promoting its clearance from circulation (Hasunuma et al., 1997) and preventing thrombus formation. Circulating b 2 GPI/oxLDL complexes have been implicated as atherogenic autoantigens, and their presence may represent a risk factor or an indirect but significant contributor for thrombosis and atherosclerosis (Kobayahsi et al., 2003) in an autoimmune background. As numerous interacting inflammatory, oxidative and coagulation factors are thought to contribute to the development of atherosclerosis, the oxidative modification of LDL may play a role in the initiation, progression and terminal events in these vascular lesions (Ross, 1999).
The high-density lipoprotein (HDL)-associated enzyme paraoxonase (PON) has anti-oxidant activity that protects LDL from oxidation (Durrington et al., 2001). Decreased PON activity has been reported in patients with high levels of aCL (Lambert et al., 2000). Furthermore, IgG antib 2 GPI antibodies have been associated with reduced PON activity in SLE and primary APS patients (Delgado-Alves et al., 2002). PON activity is also known to increase with lipid-lowering drugs (Belogh et al., 2001;Senti et al., 2001), and in one study, cholesterol-lowering statins prevented the in vitro endothelial cell activation normally induced by anti-b 2 GPI antibodies (Meroni et al., 2001). Antioxidant treatment for 4 -6 weeks has been observed to decrease the titer of circulating aCL antibodies in SLE and APS patients (Ferro et al., 2002). Vascular injury as seen in autoimmune patients may affect PON activity or any other anti-oxidant mechanism, triggering LDL oxidative changes. Taken together, these findings provide additional support to the hypothesis that oxidative stress plays an important role in antiphospholipid antibody production and development of thrombosis in APS.
The mean level of IgG anti-b 2 GPI/oxLig-1 antibodies was highest in APS patients with arterial thrombosis (Fig. 2). The coexistence of these autoantibodies with b 2 GPI/oxLDL complexes, suggest that these two elements interact perhaps forming circulating immune complexes. This observation along with the increased macrophage uptake of b 2 GPI/oxLDL complexes in the presence of anti-b 2 GPI/oxLig-1 antibodies, provides a possible explanation for the accelerated development of atherosclerosis in autoimmune patients. Two groups (Zhao et al., 2001;Kobayahsi et al., 2003) using similar assay systems have recently shown increased serum levels of oxLDL and antibodies to oxLDL in APS patients with history of arterial thrombotic events. It is possible that APS patients also present immune complexes (b 2 GPI/oxLDL/ antibody). The ELISA system used in this study seems to detect only free (unbound) antibodies to b 2 GPI/oxLDL (oxLig-1) complexes. Although preliminary, our results suggest that IgG anti-b 2 GPI/oxLDL (oxLig-1) antibodies may represent a distinct subset of antiphospholipid antibodies and that they may coexist with other antibodies. IgG anti-b 2 GPI/oxLDL (oxLig-1) antibodies appear to be a useful serologic marker with high specificity for APS and might possibly have a pathogenic role in atherosclerotic risk in autoimmune patients.