Lipid peroxidation is enhanced in patients with systemic lupus erythematosus and is associated with arterial and renal disease manifestations




Cardiovascular disease with premature atherosclerosis is common in patients with systemic lupus erythematosus (SLE). We previously identified elevated levels of oxidized low-density lipoprotein (OxLDL) together with elevated levels of autoantibodies related to OxLDL as risk factors for cardiovascular disease in female patients with SLE. Autoantibodies to OxLDL are common in SLE and cross-react with anticardiolipin antibodies (aCL). We therefore hypothesized that lipid peroxidation is enhanced in patients with SLE in general.


One hundred forty-seven female patients with SLE and 60 age- and sex-matched controls were compared. A monoclonal antibody to oxidized phospholipids, EO6, was used to determine oxidation epitopes on LDL. Anti-OxLDL and autoantibodies to malondialdehyde (MDA)–modified LDL, cardiolipin, and oxidized aCL were determined by chemiluminescence technique.


As determined by binding of EO6, patients with SLE had a higher level of oxidized phospholipids on LDL (P = 0.005) compared with controls. The level of OxLDL (e.g., oxidized phospholipid/apolipoprotein B) was associated with arterial disease (P = 0.006) and renal manifestations (P = 0.04). As reported previously, levels of aCL, autoantibodies to OxLDL, and autoantibodies to MDA-modified LDL were enhanced and were closely correlated in SLE. Anticardiolipin antibodies from these SLE patients recognized mainly oxidized forms of cardiolipin, indicating that antigenic epitopes on cardiolipin are related to lipid peroxidation in patients with SLE.


In general, patients with SLE (particularly those with cardiovascular disease) had more oxidized epitopes on LDL compared with controls. Furthermore, aCL in these patients recognized epitopes generated during lipid peroxidation. Thus, “neo” self antigens on lipoproteins, generated during oxidation, are present in SLE and may be of importance for the development of premature cardiovascular disease and possibly also for other autoimmune phenomena observed in SLE.

Systemic lupus erythematosus (SLE) is an autoimmune systemic disease, and 90% of the cases of SLE occur in women. These patients are at high risk of cardiovascular disease, which often affects women with SLE before menopause (1, 2), the time of life at which women normally are protected from coronary heart disease (3). Premature atherosclerosis was recently demonstrated in patients with SLE (4, 5). In epidemiologic studies, women ages 44–50 years had a 50-fold increased risk of myocardial infarction as compared with controls from the Framingham study (1), and the relative risk for coronary heart disease was 7.5, after adjusting for Framingham risk factors (6). We recently identified both traditional and nontraditional risk factors for cardiovascular disease among female patients with SLE. Briefly, these were markers of inflammation (raised levels of acute-phase reactants and tumor necrosis factor α), dyslipidemia, enhanced low-density lipoprotein (LDL) oxidation, antiphospholipid antibodies (aPL; lupus anticoagulants and antibodies to oxidized LDL [anti-OxLDL]), and high levels of homocysteine (2, 7, 8).

Enhanced titers of antibody to anionic phospholipids, including cardiolipin and lupus anticoagulant, are features of the antiphospholipid syndrome (APS), which is characterized by fetal loss, autoimmune thrombocytopenia, and thrombosis (9). Among patients with SLE, 30–50% have aPL; APS develops in approximately one-third to one-half of these patients (10) and is often referred to as secondary APS. Although this is not included in the definition of APS, some anti-OxLDL can also be regarded as aPL that cross-react with aCL (11, 12).

In APS and SLE, a significant fraction of aCL recognize oxidized phospholipids (OxPL) (12, 13). Furthermore, enhanced lipid peroxidation in patients with APS, as determined by secretion of oxidation products of arachidonic acid in urine, was recently reported (14). An enhanced monocyte expression of tissue factor by oxidative stress in patients with APS, with antioxidant therapy providing beneficial effects, was also reported (15). These findings indicate that lipid peroxidation may be of importance in APS.

According to the modified LDL hypothesis, oxidation of LDL is an important factor in atherogenesis (16). As a result of oxidation, a variety of immunogenic neoepitopes are formed on OxLDL. For example, oxidation of the phosphorylcholine (PC)–containing phospholipids renders them antigenic, and, in fact, such OxPL form ligands on OxLDL recognized by macrophage scavenger receptors, leading to enhanced uptake of OxLDL by macrophages and foam cell formation (17–19). Through this route, macrophages become lipid laden and develop into the characteristic foam cells of the atherosclerotic lesions (18, 20–22). OxLDL is also chemotactic, immune stimulatory, and has toxic properties that promote local inflammatory processes in atherosclerotic lesions (23, 24). Furthermore, OxLDL elicits a humoral immune response, with production of autoantibodies to oxidation-specific epitopes of OxLDL (anti-OxLDL).

EO6 is an IgM monoclonal antibody cloned from apolipoprotein E (Apo E)–deficient mice that was subsequently shown to be equivalent to the classic T15 naturally occurring antibody that specifically binds to the PC moiety of the cell wall polysaccharide of pathogens such as Streptococcus pneumoniae (25). This antibody confers optimal protection to mice against lethal infection with streptococci. EO6 was subsequently shown to recognize the PC headgroup of OxPL, but not the same moiety on native, unoxidized phospholipids. Thus, EO6 binds to OxPL on OxLDL and on apoptotic cells and via this mechanism blocks the uptake of OxLDL and apoptotic cells by macrophage scavenger receptors such as CD36 (19). Therefore, the epitope recognized by EO6 is a key ligand for scavenger receptors (18). In this study, we used EO6 to detect enhanced numbers of OxPL epitopes on Apo B-100 particles (e.g., OxPL/Apo B, a measure of OxLDL that we term OxLDL-EO6) in a large cohort of unselected female patients with SLE. These epitopes were associated with arterial and renal disease manifestations. Furthermore, high-titer aCL were shown to bind primarily to oxidized forms of cardiolipin, further supporting a role for enhanced lipid peroxidation in these patients. The implications of these findings are discussed below.


Study group.

The study group comprised 147 unselected women with SLE (mean ± SD age 47.1 ± 13.1 years) who were being evaluated at the Department of Rheumatology, Karolinska Hospital. All patients fulfilled at least 4 of the American College of Rheumatology (ACR) 1982 revised classification criteria for SLE (26). The investigation included an interview and a physical examination by a rheumatologist. SLE disease activity was determined using the SLE Disease Activity Measure (SLAM) (27), and organ damage was determined with the Systemic Lupus International Collaborating Clinics (SLICC)/ACR Damage Index (28). Laboratory examinations were performed on blood samples that were obtained from patients who had fasted overnight. The control group comprised 60 healthy age-matched women (mean ± SD age 46.6 ± 13.2 years) who were selected from among blood donors. The patients and controls were from the same population. Two patients who were receiving lipid-lowering therapy (statins) were not excluded from the study.

Arterial disease was defined as myocardial infarction (confirmed by electrocardiography and a rise in the level of creatine kinase), angina pectoris (coronary insufficiency confirmed by exercise stress test), intermittent claudication (peripheral atherosclerosis confirmed by angiography), or peripheral arterial thrombosis (confirmed by angiography). Nephritis was confirmed by biopsy in most patients but was also considered to be present if patients had persistent proteinuria (>0.5 gm/day) or if cellular casts were present, as defined by the ACR 1982 revised criteria (26). The study was approved by the local ethics committee of Karolinska Hospital and was conducted in accordance with the Helsinki Declaration.

Routine laboratory tests.

Anti–double-stranded DNA (anti-dsDNA) antibodies were determined by immunofluorescence using Crithidia lucilliae kinetoplast assay. Antinuclear antibodies, Sjögren's syndrome A antibodies, Sjögren's syndrome B antibodies, and rheumatoid factor were determined using routine laboratory techniques at the Department of Immunology, Karolinska Hospital. Levels of total cholesterol, LDL, high-density lipoprotein (HDL), and triglycerides in plasma were determined at the Department of Clinical Chemistry, Karolinska Hospital, using routine techniques.

Lipids and reagents.

Cardiolipin and reduced cardiolipin were obtained from Avanti Polar Lipids (Alabaster, AL). Reduced cardiolipin is a hydrogenated preparation of cardiolipin in which all of the unsaturated fatty acids of cardiolipin are reduced, rendering a cardiolipin preparation that is unable to undergo lipid peroxidation (12). Luminol was obtained from Boehringer Mannheim (Indianapolis, IN).

LDL was isolated from pooled plasma obtained from healthy donors, by sequential preparative ultracentrifugation under conditions to minimize oxidation and proteolysis, and LDL was oxidized by copper or modified by malondialdehyde (MDA), as previously described (29).

Chemiluminescence immunoassay for antibody binding to OxLDL and cardiolipin.

The chemiluminescence immunoassay was performed with modifications as previously described (13). To determine antibody binding to OxLDL, both copper-modified OxLDL and MDA-modified LDL were used. The assays were performed using 96-well, white, round-bottomed MicroFluor plates (Dynatech, Chantilly, CA). Plates were coated overnight at 4°C with 50 μl of copper-modified OxLDL or MDA-modified LDL (5 μg/ml) or reduced cardiolipin (25 μg/ml) in Tris buffered saline (TBS) with 0.27 mM EDTA and 20 μM butylated hydroxytoluene. In the case of cardiolipin, the immunogen was added to microtiter wells in organic solvent and allowed to plate overnight to allow evaporation of solvent and plating of antigen, which is the standard method for plating cardiolipin. We have previously shown that this approach results in significant oxidation of cardiolipin, but not reduced cardiolipin (12, 13). The plates were washed 4 times with TBS buffer.

The plasma samples were diluted 1:250 in TBS containing 2% bovine serum albumin (BSA) and incubated in wells for 1 hour at room temperature. After washing, the amount of antibody bound was measured with alkaline phosphatase–labeled goat anti-human IgG (A-3187; Sigma) diluted 1:45,000 in TBS containing 2% BSA and 1 mM of MgCl2 and ZnCl2, or IgM (A-3437; Sigma) diluted 1:37,000 in the same buffer for 1 hour at room temperature. After 4 washes with TBS, 25 μl of a 50% solution of LumiPhos 530 (Lumigen, Southfield, MI) was added to each well, and the plates were incubated for 90 minutes at room temperature in the dark. Luminescence was determined using a Lucy 1 Luminometer and WINLCOM software (Anthos Labtec Instruments, Salzburg, Austria) as previously described (30). Each determination was performed in triplicate. All samples were measured in a single assay, and the coefficients of variance for low and high standards were 6–10%.

Determination of OxLDL epitopes and LDL-containing immune complexes.

The EO6 epitope concentration on Apo B-100–containing particles was measured by a chemiluminescence modification of a previously described assay (31, 32). This sandwich assay uses an anti–Apo B-100 antibody, MB47, to capture LDL, and a biotin-labeled anti-OxLDL antibody, EO6, to measure the amount of the EO6 epitope present on the LDL, which is an oxidized phospholipid.

The procedure was as follows: 96-well, white microtiter plates (MicroLite 2; Dynatech) were incubated with 50 μl of 10 μg/ml MB47 in 50 mM Tris HCl (pH 7.4) containing 150 mM NaCl, 0.27 mM EDTA, and 0.02% NaN3 (TBS) overnight at 4°C. After washing the plates 3 times with washing buffer (TBS) containing 0.001% apoprotinin, using an automated plate washer (Microplate washer, model 1550; Bio-Rad, Hercules, CA), a 50-fold dilution of plasma in TBS containing 2% BSA (BSA/TBS) was added to the MB47-coated plates (50 μl/well) and incubated for 2 hours at room temperature.

At this dilution of plasma, the amount of Apo B-100 particles saturates the capacity of MB47 binding, and, thus, in each well an equal number of Apo B–containing particles are bound. This has been verified by demonstrating that biotinylated MB24, another Apo B–specific antibody that binds to an Apo B epitope distinct from that of MB47 (30), binds equally to each well.

After incubation, the plates were washed as described above, and 10 μg/ml of biotin-labeled EO6 in BSA–TBS was added to the plates (50 μl/well) and incubated overnight at 4°C. After washing the plates as described above, 10,000-fold diluted avidin-conjugated alkaline phosphatase (NeutrAvidin, alkaline phosphatase–conjugated; Pierce, Rockford, IL) in BSA–TBS containing 1 mM MgCl2, and 1 mM ZnCl2 was added to the plates (50 μl/well) and incubated for 1 hour at room temperature. The plates were then washed 4 times with washing buffer, and 50% LumiPhos 530 (Lumigen) in distilled water was added to the plates (30 μl/well) and incubated for 1.5 hours at room temperature in the dark. The chemiluminescence was read on an MLX microtiter plate luminometer (Dynatech). Data are expressed in relative light units (RLUs), measured over 100 msec. All samples were measured in a single assay, and the intraassay coefficients of variance of low and high standards were 6–10%. Data are expressed as absolute amounts of EO6 bound per well (in RLUs). We refer to this measurement as OxLDL-EO6, which is a measurement of OxPL/Apo B.

To detect LDL immune complexes in sera, 96-well, white, round-bottomed MicroFluor plates (Dynatech) were coated with 10 μg/ml of MB47 overnight at 4°C. The plates were washed 4 times with TBS. Plasma was then added at a 1:50 dilution in TBS containing 2% BSA for 2 hours at room temperature. The level of IgG or IgM bound to the LDL was then determined by addition of alkaline phosphatase–labeled goat anti-human IgG and IgM, as indicated above. As noted above, the amount of LDL bound per well was demonstrated to be equal, and thus the amount of IgG or IgM bound (expressed as RLUs) is normalized to equivalent numbers of LDL particles.

Absorption experiments.

Cardiolipin was added to microtiter wells and its solvent evaporated, leading to exposure to air, which results in rapid and extensive oxidation of the cardiolipin (12, 13). To demonstrate the specificity of the aCL, the aCL were first preincubated with either cardiolipin or reduced cardiolipin to preabsorb antibody populations. Cardiolipin or reduced cardiolipin was dried onto the surface of glass tubes for 3 hours, and serum samples (1:50 dilution in 2% BSA–TBS) were added. After vortexing, the tubes were incubated at 37°C for 60 minutes and centrifuged at 13,000 revolutions per minute for 30 minutes (4°C). The supernatants containing unbound antibodies were then tested for antibody binding to the cardiolipin, using the standard procedure as described. Control experiments were performed, in which similar preincubations were conducted in tubes to which no cardiolipin had been added.

Statistical analysis.

Variables were tested for skewness, and skewed continuous variables were logarithmically transformed. When appropriate, nonparametric tests (Mann-Whitney U test) were used for comparisons between the groups, whereas analysis of variance/Student's t-test was used for normally distributed variables. Simple regression (or, in the case of non-normally distributed variables, Spearman's rank correlation) was used to calculate correlation coefficients between oxidation epitopes, antibody levels, and other parameters. Analysis of covariance with Fisher's protected least significant difference was used to calculate the influence of possible confounding factors. P values less than 0.05 were considered significant.


Basic and clinical characteristics of the patient group.

The mean ± SD age of the patients was 47.1 ± 13.1 years, the mean disease duration was 12.6 years, and 52% of the patients were treated with prednisolone. Among women in the control group, the mean ± SD age was 46.6 ± 13.2 years. The basic characteristics of the study group are shown in Table 1. Twelve patients had previous arterial disease, 5 had myocardial infarction, 7 had angina, 3 had intermittent claudication, and 2 had peripheral arterial thrombosis.

Table 1. Characteristics of the female patients with systemic lupus erythematosus*
CharacteristicMean ± SDMedianRangeNo. of patients% of patients
  • *

    The Systemic Lupus Erythematosus Disease Activity Measure (SLAM) is a measure of disease activity (27), and the Systemic Lupus International Collaborating Clinics/American College of Rheumatology (SLICC/ACR) is a measure of cumulative disease damage (28). The disease manifestations/criteria were defined according to the 1982 revised ACR criteria for classification of systemic lupus erythematosus (SLE) (26). Anti-dsDNA = anti–double-stranded DNA; ANA = antinuclear antibodies; SSA = Sjögren's syndrome A antibodies; SSB = Sjögren's syndrome B antibodies.

Age, years47.1 ± 13.14819–80147
Disease duration, years12.6 ± 9.811.50–54
No. of ACR SLE criteria64–10147
SLAM score60–20147
SLICC/ACR Damage Index10–10147
Dermal manifestations12786
Arterial disease128
Venous thromboses1812
Antiphospholipid syndrome2014
Presently receiving prednisolone7752
Months receiving prednisolone88.0 ± 101.3480–480
Present prednisolone dose, mg5.0 ± 8.71.250–60147
Smokers, yes/previous/no36/41/6725/28/47
Anti-dsDNA positive ever8960
ANA positive ever14699
SSA positive4229
SSB positive2014
Rheumatoid factor positive2620

Oxidation-related epitopes and immune complexes and autoantibodies.

Apo B-100 particles (chiefly LDL in these women without hypertriglyceridemia) isolated from patients with SLE expressed significantly higher levels of the EO6-specific epitopes as compared with controls; that is, they had higher levels of OxLDL-EO6 (OxPL/Apo B) (Table 2). The OxPL/Apo B value correlated with IgG anti-OxLDL among SLE patients, and with IgG anti–MDA-modified LDL among both patients and controls. (Table 3). LDL containing immune complexes of the IgM and IgG types did not differ between patients with SLE and controls (Table 2). The levels of IgG and IgM autoantibodies against OxLDL, MDA-modified LDL, and cardiolipin were significantly higher in the patient group compared with the control group (Table 2).

Table 2. OxLDL-related values in patients with SLE and controls*
VariableIg classSLE patientsControlsP
  • *

    Values are the mean ± SD relative light units/100 msec. Oxidized low-density lipoprotein (OxLDL) was made by copper-catalyzed oxidation of LDL. SLE = systemic lupus erythematosus; OxLDL-EO6 = oxidized phospholipid (OxPL) epitopes on LDL as measured by binding of EO6 to Apo B-100 particles, captured by MB47 (OxPL/Apo B); NS = not significant; MDA = malondialdehyde.

OxLDL-EO6 (OxPL/Apo B) 15,616 ± 1,45811,674 ± 2,2750.005
Immune complexesIgM3,242 ± 2003,298 ± 313NS
Immune complexesIgG7,166 ± 3546,926 ± 519NS
Anti-OxLDLIgM14,242 ± 1,1285,900 ± 1,7610.0002
Anti-OxLDLIgG18,110 ± 1,30811,308 ± 2,040<0.0001
Anti–MDA-modified LDLIgM40,035 ± 1,96424,937 ± 3,0640.0002
Anti–MDA-modified LDLIgG31,346 ± 1,64114,629 ± 2,560<0.0001
Anticardiolipin antibodiesIgM5,270 ± 6522,060 ± 1,0700.0006
Anticardiolipin antibodiesIgG18,626 ± 1,59713,037 ± 2,5900.01
Table 3. Correlations between OxLDL-EO6 (OxPL/Apo B) and antibody levels to OxLDL in SLE patients and controls*
  • *

    Oxidized low-density lipoprotein (OxLDL) was made by copper-catalyzed oxidation of LDL. OxLDL-EO6 = OxLDL as measured by binding of EO6 to Apo B-100 particles, captured by MB47 (EO6/Apo B); OxPL = oxidized phospholipid; SLE = systemic lupus erythematosus; MDA-LDL = malondialdehyde-modified LDL.


Analysis of antibodies against cardiolipin.

Cardiolipin was extensively oxidized when incubated for 3 hours or overnight, as determined by measurement of conjugated dienes (data not shown). Figure 1 shows that cardiolipin that had been oxidized by exposure to air for 3 hours out-competed, binding to cardiolipin (which was, in fact, oxidized as shown previously [13]) as opposed to reduced cardiolipin, which did not inhibit binding to cardiolipin significantly.

Figure 1.

Competition immunoassay of the binding of IgG from 20 anticardiolipin antibody–positive patients with systemic lupus erythematosus (antibody levels >2 SD above the mean in the control group). Each serum sample was diluted 1:50 with 2% bovine serum albumin–Tris buffered saline buffer and preincubated with 25 μg of cardiolipin exposed to air for 3 hours (OxCL) or reduced cardiolipin (CLred). After preincubation, the immune complexes were pelleted by centrifugation and the supernatants tested for IgG binding to cardiolipin using a standard procedure. Results are expressed as the percent inhibition of binding when compared with control incubations carried out under the same conditions in the absence of phospholipid. Each point represents the mean of triplicate determinations. RLU = relative light unit.

LDL oxidation epitopes, antibodies, lipids, and association with clinical manifestations.

Patients with SLE and arterial disease (including myocardial infarction, angina, claudication, or peripheral arterial thrombosis) had significantly higher levels of OxLDL-EO6 (OxPL/Apo B) and anti-OxLDL of the IgM subclass than did those with no history of arterial disease. MDA-modified LDL and anti-OxLDL of the IgG subclass were not associated with arterial disease. Renal manifestations of lupus were associated with OxLDL-EO6 and also with anti-OxLDL of the IgM subclass (Table 4).

Table 4. Association of OxLDL-EO6 (OxPL/Apo B) and anti-OxLDL with clinical manifestations*
 PresentNot presentP
  • *

    Values are the mean ± SD relative light units/100 msec. Oxidized low-density lipoprotein (OxLDL) was made by copper-catalyzed oxidation of LDL. Active disease was defined as a SLAM score ≥7, and a SLICC/ACR score >1 represented the presence of cumulative disease damage. OxLDL-EO6 = OxLDL as measured by binding of EO6 to Apo B-100 particles, captured by MB47 (OxPL/Apo B); NS = not significant; MDA-LDL = malondialdehyde-modified LDL (see Table 1 for other definitions).

Arterial disease   
 OxLDL29,526 ± 32,56314,283 ± 153460.006
 Anti-OxLDL IgM21,163 ± 17,20912,403 ± 125630.03
 Anti-OxLDL IgG18,804 ± 16,43618,639 ± 17064NS
 Anti–MDA-LDL IgM42,265 ± 27,88140,159 ± 24769NS
 Anti–MDA-LDL IgG33,565 ± 19,03730,115 ± 19541NS
Venous thrombosis   
 OxLDL21,222 ± 19,36714,787 ± 17,7180.08
 Anti-OxLDL IgM16,380 ± 13,24112,672 ± 13,080NS
 Anti-OxLDL IgG21,641 ± 24,64918,228 ± 15,6140.07
 Anti–MDA-LDL IgM41,360 ± 25,34640,264 ± 24,898NS
 Anti–MDA-LDL IgG36,653 ± 20,58129,432 ± 19,167NS
 OxLDL21,198 ± 26,55912,731 ± 10,2640.04
 Anti-OxLDL IgM14,854 ± 13,29412,235 ± 12,9940.04
 Anti-OxLDL IgG20,399 ± 19,78517,742 ± 15,242NS
 Anti–MDA-LDL IgM40,398 ± 25,69340,398 ± 24,567NS
 Anti–MDA-LDL IgG32,371 ± 18,30529,257 ± 199,777NS
 OxLDL15,368 ± 16,78618,058 ± 27,294NS
 Anti-OxLDL IgM13,543 ± 13,6349,455 ± 6,091NS
 Anti-OxLDL IgG19,072 ± 17,56114,898 ± 8,802NS
 Anti–MDA-LDL IgM40,839 ± 25,52636,524 ± 18,278NS
 Anti–MDA-LDL IgG31,361 ± 19,89121,123 ± 11,359NS
 OxLDL12,867 ± 13,52318,331 ± 21,2330.02
 Anti-OxLDL IgM13,093 ± 14,18413,158 ± 12,118NS
 Anti-OxLDL IgG16,087 ± 11,57521,036 ± 20,474NS
 Anti–MDA-LDL IgM38,024 ± 21,33142,616 ± 27,727NS
 Anti–MDA-LDL IgG31,075 ± 20,22029,606 ± 18,754NS
Active disease   
 OxLDL18,582 ± 23,80813,182 ± 10,790NS
 Anti-OxLDL IgM13,431 ± 13,87712,956 ± 12,457NS
 Anti-OxLDL IgG20,341 ± 19,72217,165 ± 13,638NS
 Anti–MDA-LDL IgM42,075 ± 25,64939,137 ± 24,129NS
 Anti–MDA-LDL IgG30,805 ± 19,98430,065 ± 19,010NS
Disease damage   
 OxLDL17,196 ± 21,23712,738 ± 8,321NS
 Anti-OxLDL IgM14,309 ± 14,21010,801 ± 10,5110.06
 Anti-OxLDL IgG19,102 ± 17,54717,996 ± 16,446NS
 Anti–MDA-LDL IgM42,210 ± 24,64037,035 ± 26,228NS
 Anti–MDA-LDL IgG31,058 ± 19,77929,194 ± 18,480NS

Blood lipid levels were available for the patients with SLE; the mean ± SD values were 5.3 ± 1.2 mmoles/liter for total cholesterol, 3.1 ± 1.1 mmoles/liter for LDL, 1.5 ± 0.5 mmoles/liter for HDL, and 1.6 ± 0.9 mmoles/liter (median 1.3) for triglycerides. Patients with arterial disease had higher levels of total cholesterol (P = 0.001), LDL (P = 0.005), and triglycerides (P = 0.01) but not HDL or blood pressure compared with those with no arterial disease (data not shown). Patients with nephritis had higher levels of total cholesterol (P = 0.01) and triglycerides (P = 0.02) but not HDL or blood pressure compared with patients without nephritis (data not shown). Smoking was not more common in patients with arterial disease or nephritis than in those without these conditions (data not shown). Because the OxPL/Apo B assay is, by design, independent of Apo B levels (see Patients and Methods), there was no association between OxLDL-EO6 (OxPL/Apo B) and LDL levels (data not shown).

The association between OxLDL-EO6 and arterial disease and between anti-OxLDL IgM and arterial disease remained significant (P = 0.01 and P = 0.02, respectively) when factors that were independently associated with arterial disease (cholesterol, triglycerides, and age) were included as covariates. Similarly, the association between OxLDL-EO6 and nephritis and between anti-OxLDL IgM and nephritis remained significant (P = 0.02 and P = 0.02, respectively) when factors that were independently associated with nephritis (cholesterol and triglycerides) were included as covariates. Other than OxLDL-EO6 measurements, no significant associations with leukopenia were observed.


In this study, we demonstrated that female patients with SLE had more OxLDL-EO6 (i.e., Apo B lipoproteins containing OxPL recognized by monoclonal antibody EO6) compared with controls (19). Arterial and renal manifestations of lupus were associated with higher levels of these oxidized epitopes on LDL.

The epitope recognized by EO6 is the PC headgroup of OxPL, which is present on both OxLDL and minimally modified LDL (e.g., LDL that has undergone early oxidative changes of lipids but, in contrast to OxLDL, changes that are insufficient to cause scavenger receptor recognition) (33). Minimally modified LDL is readily deposited in the artery wall and is of importance for endothelial activation and recruitment of monocytes to the intima. These monocyte–endothelial interactions are inhibited by a platelet-activating factor (PAF) receptor antagonist (34). Recently, it was shown that the same PC epitope is also present on apoptotic cells and pneumococci (18), and that PC is a target not only for scavenger receptors and EO6 antibodies but also for C-reactive protein (35). Thus, the immune response to PC/OxLDL is part of a greater immune network that may have evolved to provide protection against both external invaders such as bacteria and internal debris from dying cells (18).

Our study provides evidence to support the hypothesis that there is enhanced lipid oxidation in SLE. Iuliano et al also showed that patients with SLE had enhanced urinary excretion of isoprostanes, consistent with enhanced lipid peroxidation (14). In keeping with this hypothesis is a recent study that demonstrated depressed activity of the antioxidant enzyme paraoxonase in the circulation of patients with SLE (36). These findings are also supported by studies in hamsters, in which both infection and inflammation induced LDL oxidation as expressed by another LDL epitope, lysophosphatidylcholine (37). We have previously shown that patients with SLE have antibodies to lysophosphatidylcholine (38).

Another finding reported here is that an enhanced content of OxLDL-EO6 (OxPL/Apo B) was associated with arterial disease. The present findings confirm and extend our earlier results, indicating that OxLDL is an important factor contributing to the very high risk of cardiovascular disease in patients with SLE (2).

LDL cholesterol is also a well-known risk factor for cardiovascular disease, and in this study LDL levels were associated with arterial disease. Because OxLDL-EO6, as detected here, is a measure of OxPL/Apo B, it is independent of LDL particle number. Our measure of OxLDL thus represents a risk factor that is independent of LDL levels alone. High levels of LDL and enhanced LDL oxidation may thus be regarded as separate risk factors for cardiovascular disease in patients with SLE. In recent studies, Lp(a) has been demonstrated to be closely related to the EO6 measurement of OxLDL (39). Consistent with this are previous reports indicating that the level of Lp(a) is elevated in SLE (40).

OxLDL was also associated with renal disease in SLE. It is well recognized that renal disease and renal failure are strong risk factors for cardiovascular disease. OxLDL (as determined by another technique) has also been implicated in renal disease in the general population (41), and OxLDL was demonstrated in sclerotic and mesangial regions of biopsy specimens obtained from patients with chronic renal disease (42). Furthermore, expression of macrophage scavenger receptors can be induced in human mesangial cells (43). Whether enhanced LDL oxidation has a role in the development of renal manifestations in SLE or occurs secondary to nephritis cannot be determined from this study.

It is not known where the epitopes recognized by EO6 are generated. In the circulation there are strong antioxidant defenses, and in animal experiments OxLDL is rapidly taken up by the liver (44). It is therefore highly unlikely that any significant degree of oxidation of LDL occurs within the circulation. In general, LDL is believed to become oxidized after entering the intima, where it is retained by a network of proteoglycans, which further facilitates oxidative and enzymatic modifications (45). LDL may also be oxidized in other tissues, including the liver (46). Previous studies have indicated that the oxidation epitopes demonstrated here are more likely to be derived from oxidation of phospholipids in tissues. In turn, this suggests that the oxidized phospholipids found on plasma Apo B-100 particles reflect lipid peroxidation that occurs in atherosclerotic lesions or other sites of inflammation.

We can also confirm earlier observations that autoantibodies related to epitopes of OxLDL and oxidized cardiolipin are enhanced in SLE (11, 38, 47). In this study, arterial and renal disease manifestations were weakly associated with anti-OxLDL of the IgM subclass. We previously observed an association between IgG anti-OxLDL antibodies and arterial disease in a smaller cohort of older patients with SLE (2). In 2 other studies, anti-OxLDL and arterial disease were not associated (48, 49). These differences between studies may be related to both the selection of patients and the methods. Although in this study anti-OxLDL titers were generally increased, the elevated titers were not associated with the disease manifestations studied.

As reported previously, in patients with APS, many aCL recognize OxPL (12, 13), and in the present report we have demonstrated that in SLE plasma with high levels of aCL, antibody binding to cardiolipin was out-competed by oxidized cardiolipin but not by reduced cardiolipin. This clearly shows that many aCL in plasma with high titers of OxPL from patients with SLE recognize oxidized cardiolipin. Because aPL are pathogenic in animal studies (50) and also predispose to arterial and venous thromboses (51, 52), the possibility that potent antioxidant therapy may be valuable in this patient group should be considered (14). Furthermore, the epitope recognized by EO6 promotes monocyte–endothelial interactions, which can be inhibited by PAF antagonists. This is consistent with a recent study in which PAF antagonists ameliorated atherosclerosis in experimental animals (53). Interference with PAF-related inflammatory effects may thus also be of therapeutic potential in these patients.

In conclusion, enhanced lipid peroxidation may play an important role in SLE, especially with respect to the premature atherosclerotic process observed among these patients. If further evidence can be provided to support this hypothesis, potent antioxidant therapy may be of value for patients with SLE.


We are grateful to Göran Karlsson and Gerdur Gröndal for including patients and to Jill Gustafsson, Eva Jemseby, and Åsa Lundberg for their help with management of the patient cohorts and blood sampling.