ApoA-1 (apolipoprotein A-1) is the main component of HDL (high-density lipoprotein) and stabilizes PON-1 (paraoxonase-1), which prevents lipid peroxidation and oxLDL (oxidized low-density lipoprotein) formation. Autoantibodies against apoA-1 [anti-(apoA-1) IgG] have been found in antiphospholipid syndrome and systemic lupus erythematosous, two diseases with an increased risk of thrombotic events, as well as in ACS (acute coronary syndrome). OxLDL levels are also elevated in these diseases. Whether anti-(apoA-1) IgGs exist in other prothrombotic conditions, such as APE (acute pulmonary embolism) and stroke, has not been studied and their potential association with oxLDL and PON-1 activity is not known. In the present study, we determined prospectively the prevalence of anti-(apoA-1) IgG in patients with ACS (n=127), APE (n=58) and stroke (n=34), and, when present, we tested their association with oxLDL levels. The prevalance of anti-(apoA-1) IgG was 11% in the ACS group, 2% in the control group and 0% in the APE and stroke groups. The ACS group had significantly higher median anti-(apoA-1) IgG titres than the other groups of patients. Patients with ACS positive for anti-(apoA-1) IgG had significantly higher median oxLDL values than those who tested negative (226.5 compared with 47.7 units/l; P<0.00001) and controls. The Spearman ranked test revealed a significant correlation between anti-(apoA-1) IgG titres and serum oxLDL levels (r=0.28, P<0.05). No association was found between PON-1 activity and oxLDL or anti-(apoA-1) IgG levels. In conclusion, anti-(apoA-1) IgG levels are positive in ACS, but not in stroke or APE. In ACS, their presence is associated with higher levels of oxLDL and is directly proportional to the serum concentration of oxLDL. These results emphasize the role of humoral autoimmunity as a mediator of inflammation and coronary atherogenesis.
- apolipoprotein A-1
- coronary disease
- oxidized low-density lipoprotein
ApoA-1 (apolipoprotein A-1) is the most abundant protein (70%) of HDL (high-density lipoprotein), whose concentration is known to be inversely correlated with cardiovascular risk [1,2]. HDL-associated apoA-1 plays a crucial role in cholesterol homoeostasis by regulating reverse cholesterol transport and delivering it to the liver . HDL-associated apoA-1 also has anti-inflammatory properties [4,5] and has antioxidant properties by binding and stabilizing PON-1 (paraoxonase-1) enzyme activity [6,7]. PON-1 prevents the formation of lipid peroxidation products, such as oxLDL [oxidized LDL (low-density lipoprotein)], that play a major role in endothelial cell dysfunction, monocyte chemotaxis, foam cell formation and plate rupture. OxLDL is therefore a key molecule in all stages of atherogenesis .
There are an increasing number of studies showing that ‘non-traditional’ factors, such as autoantibodies, can act as cardiovascular risk factors [9–13]. The prototypical example of such a causal association are the antiphospholipid antibodies that are known to be proatherogenic and to account for the increased cardiovascular risk encountered in APS (antiphospholipid syndrome) and SLE (systemic lupus erythematosous) [14–18]. In APS and SLE, antiphospholipid antibodies have been shown to be inversely correlated with PON-1 activity , and low PON-1 activity in plasma, either acquired or innate, has been reported to be a risk factor for myocardial infarction and stroke [20–25].
Furthermore, the identification of antibodies against apoA-1 not only in SLE and APS, but also in ACS (acute coronary syndrome) emphasizes the potential involvement of humoral autoimmunity in atherogenesis and/or atherothrombosis, even in the absence of manifest autoimmune disease .
As stroke shares common cardiovascular risk factors with ACS, and because arterial and venous thrombotic complications are common in patients suffering from APS and SLE [25,27,28], we hypothesized that anti-(apoA-1) autoantibodies are involved in these various vascular diseases. Therefore we designed a prospective observational study to determine whether anti-(apoA-1) IgGs are present in ACS, to determine whether these autoantibodies were also present in patients suffering from stroke and APE (acute pulmonary embolism), and, when present, whether they could account for the low PON-1 activity and the high oxLDLs levels reported in myocardial infarction [29,30].
MATERIALS AND METHODS
Sample collection from the control and patient groups
Four groups of subjects were compared in the present study. The local Ethics Committee approved this protocol, and all subjects provided informed consent. Serum samples from all subjects were collected within 24 h of admission, aliquoted and frozen at −80 °C until use. Results were collected from patients' files in the emergency room.
Sera were obtained from 140 healthy blood donors [70 male and 70 female; median (range) age, 50 (18–70) years], who were recruited during the same period of time as the patient groups.
Recruitment of patients took place from 1 November 2005 until 30 June 2007 at the University Hospital of Geneva, which is the only public primary care hospital in this canton.
Serum was collected from 127 patients [93 males and 34 females; mean±S.D. age, 64±14.5 years; median (range) age, 65 (26–86) years] presenting to the emergency room or Intensive Care Unit of the University Hospital of Geneva with acute chest syndrome and subsequently proven significant angiographic coronary stenosis.
Exclusion criteria were the presence of any known autoimmune disease with the exception of diabetes mellitus, and the inability to give informed consent for any reason, including oro-tracheal intubation. Among the 168 screened patients, 41 did not meet the inclusion criteria, leaving 127 who were eligible for the study. Conventional LVEF (left ventricular ejection fraction) evaluation by echocardiography was performed within 5 days of admission by experienced cardiologists.
Samples from 34 ischaemic stroke patients [17 males and 17 females; mean±S.D. age, 68±14.3 years; median (range) age, 72 (36–89) years] were obtained randomly from the prospective SMS (Stroke Markers Study) cohort (n=79) recruited at the University Hospital of Geneva. Samples were obtained between 0 and 72 h after stroke onset in the presence of symptoms and signs suggestive of an acute or subacute stroke, and confirmed by brain MRI (magnetic resonance imaging) or CT (computer tomography).
Patients were excluded if the timing of stroke onset was not definite, if there was intracranial haemorrhage, traumatic brain injury, metastatic cancer, liver cirrhosis, chronic or acute renal failure, recent myocardial infarction (within 3 months) or any psychiatric condition. A total of 287 out of the 366 screened patients did not meet the inclusion criteria or they declined to be included in this study. From the 79 patients remaining, 34 samples were received.
Serum samples were collected from 58 patients [24 males and 34 females; mean±S.D. age, 66±16 years; median (range) age, 72 (30–93) years] presenting to the emergency room with dyspnoea and/or with chest-pain-associated shortness of breath and subsequent proven pulmonary embolism by spiral chest CT.
In the present study, patients were excluded if creatinine clearance (according to the Cockroft–Gault formula) was <30 ml/min, <18 years of age, pregnant or taking anticoagulants. A total of 350 patients were screened and, of those, 77 had proven APE by spiral chest CT. Three of the patients were excluded because they had an autoimmune disease and 16 patients declined to participate in the study, leaving 58 patients who were eligible for the study.
Determination of human antibodies to apoA-1 by ELISA
Anti-(apoA-1) IgGs were measured as described previously [26,31], with minor modifications. Maxisorb plates (Nunc) were coated with purified delipidated apoA-1 (20 μg/ml; 50 μl/well) for 1 h at 37 °C. After three washes with PBS/2% (w/v) BSA (100 μl/well), all wells were blocked for 1 h with PBS/2% (w/v) BSA at 37 °C. Samples were diluted 1:50 in PBS/2% (w/v) BSA and incubated for 60 min. Samples at the same dilution were also added to a non-coated well to assess individual non-specific binding. After six further washes, 50 μl/well of AP (alkaline-phosphatase)-conjugated anti-(human IgG) (Sigma–Aldrich), diluted 1:1000 in PBS/2% (w/v) BSA, was incubated for 1 h at 37 °C. After six more washes (150 μl/well) with PBS/2% (w/v) BSA, the phosphatase substrate p-nitrophenyl phosphate (50 μl/well; 1 mg/ml; Sigma–Aldrich) dissolved in 4.8% diethanolamine (pH 9.8) was added. Each sample was tested in duplicate and A405 was determined after 20 min of incubation at 37 °C (VERSA Max; Molecular Devices). The corresponding non-specific binding was subtracted from the mean absorbance for each sample.
Controls for anti-(apoA-1) IgG
We selected three different samples, the first representing a negative control, the second being close to the cut-off value, and the third being twice the cut-off value. On each plate, these samples were used as controls. As neither reference material nor calibrators are currently available for this test, we expressed results as A405.
Intra-assay and inter-assay variations of the anti-(apoA-1) IgG ELISA
Repeatability and reproducibility were determined at two levels. At a high level (A405=1.2, i.e. twice the cut-off value), the intra- and inter-assay coefficients of variation were 10% (n=10) and 17% (n=10) respectively. At the cut-off level, the intra- and inter-assay coefficients of variation were 16% (n=10) and 12% (n=8) respectively.
IgG specificity for apoA-1 assessment
Specificity of these autoantibodies to apoA-1 was confirmed by conventional saturation tests. Briefly, 50 or 100 μg/ml of apoA-1 was added to serum with anti-(apoA-1) IgG before starting the procedure described above. Tests were performed in quadruplicate. Spiking with 50 and 100 μg/ml of apoA-1 resulted in a dose–response inhibition of 60 and 87% respectively. No significant difference in non-specific absorbance values was detected between the ACS group and the stroke or APE groups of patients [mean absorbance±S.D., 0.13±0.06 and 0.10±0.09 (P=0.09) or 0.12±0.02 (P=0.4) respectively]. Moreover, there was no difference in the mean non-specific absorbance between patients with ACS who tested positive for anti-(apoA1) IgG and those who were negative for anti-(apoA1) IgG (mean absorbance±S.D., 0.138±0.028 and 0.115±0.08 respectively; P=0.3).
Sera from 140 healthy donors were used to assess a reference range of anti-(apoA-1) IgG values. The upper reference range was set at the 97.5th percentile of the distribution curve, i.e A405=0.64 (Figure 1). Values ranged from 0.15 to 0.71. A serum sample was therefore considered positive if the absorbance value was above 0.64.
Classical autoantibody measurement
ANA (anti-nuclear antibody), ANCA (anti-neutrophil cytoplasmic antibody) and RF (rheumatoid factor) measurements were performed in the Laboratory of Clinical Immunology and Allergy, University Hospital of Geneva, Geneva, Switzerland, using routine indirect immunofluorescence and ELISA techniques. IgG and IgM autoantibodies to anti-β2GPI (anti-β2 glycoprotein I) were measured by ELISA, as described previously .
cTnI (cardiac troponin I) concentrations were determined with the Unicell DXI 800™ (Beckman Coulter). CK (creatine kinase), CRP (C-reactive protein), creatinine, total cholesterol, triacylglycerols (triglycerides) and HDL (all in mmol/l) were determined using a Synchron LX20 pro™ (Beckman Coulter) auto-analyser. PON-1 activity was assessed as described previously .
OxLDL levels were determined using a commercially available ELISA kit (Mercodia), and samples were run in duplicates. The intra-assay variation coefficient for a sample at a value of 70 units/l was 3%.
Analyses were performed using Statistica™ software (StatSoft). Fischer's exact test, χ2 test combined with the Yates correction test, Mann–Whitney U test and Student's t test were used when appropriate to compare the different groups. Spearman rank correlations were performed to establish correlations between the variables. A two-sided P value was considered significant if <0.05.
Characteristic of the study subjects
The clinical characteristics of the patients with ACS, APE and stroke are shown in Table 1. The prevalence of traditional cardiovascular risk factors, such as diabetes, dyslipidaemia, pre-existing coronary heart disease, higher male/female ratio and smoking, was not significantly higher in the ACS group than in APE group. With the exception of atrial fibrillation, smoking status and a higher female/male ratio, the stroke group did not differ significantly from the ACS group (Table 1).
Prevalence of anti-(apoA-1) IgG
On admission, the prevalence of anti-(apoA-1) IgG was 11% in the ACS group, 0% in the stroke group, 0% in the APE group and 2% in the reference group (Figure 1 and Table 2). As shown in Table 2, statistically significant differences were observed between the ACS and control groups (P=0.0044), between the ACS and stroke groups (P=0.042), and between the ACS and APE groups (P=0.0057). No significant difference was found between the control, APE and stroke groups (Table 2).
Moreover, patients with ACS had significantly higher median titres of anti-(apoA-1) IgG than the control (P=0.04), stroke (P=0.005) and APE (P<0.00005) groups (Table 2).
Association of anti-(apoA-1) IgG with circulating oxLDL levels
Patients with ACS had a median oxLDL value that was significantly higher than that obtained in 40 controls (52.1 compared with 39.3 units/l respectively; P=0.006). Patients with ACS positive for anti-(apoA-1) IgGs had significantly higher LDL (P=0.04) and oxLDL (P<0.00001) levels than those who tested negative (Table 3 and Figure 2). Linear regression showed a significant correlation between the titre of anti-(apoA-1) IgG and oxLDL levels (r=0.44, P<0.00001), which remained significant after a log-transformation of oxLDL values required owing to their non-parametric distribution (r=0.38, P<0.000006). This correlation was still significant using Spearman's rank test (r=0.28, P<0.05; Figure 3). In patients with ACS, there was a significant correlation between LDL values and oxLDL (linear regression after logarithmic transformation of oxLDL: r=0.24, P=0.0007; Spearman's rank test: r=0.18, P<0.05), but not with the other lipid parameters or CRP.
Using the oxLDL/LDL ratio to take into account LDL in the reported association between anti-(apoA-1) IgG and oxLDL, the correlation between anti-(apoA-1) IgG and oxLDL/LDL was still significant (linear regression: r=0.3, P=0.00006; Spearman's rank test: r=0.22, P<0.05), and patients with ACS positive for anti-(apoA-1) IgG had a significantly higher oxLDL/LDL ratio than patients with ACS who tested negative for these autoantibodies (64 compared with 18.2 respectively; P<0.00001).
There were no significant associations between oxLDL and traditional cardiovascular risk factors, such as gender, age, diabetes, hypertension, dyslipidaemia, familial history, obesity and smoking status (results not shown).
The results in Table 3 show that no significant differences were observed for hypertension, diabetes, dyslipidaemia, smoking status, positive familial history, BMI (body mass index), previous coronary heart disease and lipid levels between patients with ACS positive or negative for anti-(apoA-1) IgG.
There was no significant difference, however, in PON-1 activity in plasma between patients with ACS who were positive for anti-(apoA-1) IgG and those who were negative (P=0.88; Table 3). These plasma PON-1 values were within the normal range , and no correlation was found between anti-(apoA-1) IgG titres and PON-1 activity.
All of the patients who tested positive for anti-(apoA-1) IgG were negative for anti-β2GPI and -RF antibodies. Three were positive for anti-cardiolipin antibodies, and three others were slightly positive for ANAs on the Hep-2 cell line, but were negative for anti-nucleosome- and anti-nucleoprotein-specific antibodies. All of the patients were negative for ANCA-specific anti-MPO (myeloperoxidase) and anti-PR3 (proteinase 3) antibodies when tested by ELISA.
The present prospective study confirms that anti-(apoA-1) autoantibodies are present in a significant subset of patients with ACS without any clinical features of autoimmune diseases. On the other hand, anti-(apoA-1) autoantibodies were not associated with other arterial or venous thromboembolic events, such as stroke or APE. Moreover, the present study is the first to support a strong and significant association between anti-(apoA-1) IgG titres and circulating oxLDL levels. The latter are known to reflect the burden of oxidative-induced endothelial damage and foam cell formation [29,30] and to be related to more complicated atherosclerotic lesions in CAD (coronary artery disease) . Because of the striking differences observed in median oxLDL values in patients with ACS positive for anti-(apoA-1) IgG compared with those who were negative and the correlation observed between oxLDL levels and anti-(apoA-1) IgG titres, it is tempting to hypothesize that increased oxLDL levels observed in the ACS group could be partly due to the presence of anti-(apoA-1) IgG, even if the underlying mechanisms are currently unknown. Indeed, anti-(apoA-1) autoantibodies are still not proven to be biologically active and to inhibit apoA-1 function, and it is not clear whether and how these autoantibodies are involved in atherogenesis. Anti-(apoA-1) IgGs could simply be a marker reflecting a non-specific activation of the immune system in an inflammatory context, but without any role in atherogenesis. On the other hand, autoantibodies could compromise apoA-1 function, thereby dampening cholesterol transport or increasing oxidation product formation, particularly oxLDL. They could interfere with scavenger-receptor-mediated uptake of oxLDL, as anti-(apoA-1) polyclonal antibodies have been shown in vitro to impede scavenger receptor B-1 function in human endothelial cells .
However, at variance with what has been shown regarding anti-HDL and anti-β2GP1 autoantibodies , we did not find any difference in PON-1 activity between patients with ACS positive and negative for anti-(apoA-1) IgG. This suggests that the anti-(apoA-1) autoantibodies detected are different from anti-HDL autoantibodies and that, if pathogenic, anti-(apoA-1) autoantibodies are not likely to interfere with the protective antioxidant functions of PON-1, but act through different mechanisms. For example, LCAT (lecithin-cholesterol acyltransferase) is also able to hydrolyse phospholipids and thereby reduce oxLDL levels , and has been shown to be decreased in a model of mice lacking apoA-1 .
The prevalence of anti-(apoA-1) IgG in ACS reported in the present study was lower than observed previously (11 compared with 23%) . This difference is most probably explained by the small number of patients enrolled in our previous study , where the prevalence 95% confidence interval was 11–40%, just encompassing the actual calculated prevalence percentage.
In SLE, anti-(apoA-1) IgG has been reported to be associated with the activity of the disease  and to correlate significantly with monocyte TNFα (tumour necrosis factor α) production . In ACS, these autoantibodies have been reported to strongly correlate with SAA (serum amyloid A) protein levels , an acute-phase protein which promotes atherogenesis by different mechanisms, such as monocyte and polymorphonuclear leucocyte chemotaxis , induction of matrix metalloproteinase synthesis  and conversion of anti-inflammatory into pro-inflammatory HDLs [41,42]. Our present results, combined with the findings mentioned above, strengthens further the link between anti-(apoA-1) IgG and major proatherogenic factors, and lends weight to the hypothesis that humoral autoimmunity is involved in atherogenesis, even in the absence of autoimmune disease. As stroke and ACS have common cardiovascular risk factors, it was surprising not to find any anti-(apoA-1)-IgG-positive patients in our stroke cohort, even if patients suffering from myocardial infarction within 3 months were excluded. This observation deserves further investigation in larger cohorts, but could partly be explained by the fact that cardioembolism (unrelated to vessel wall inflammation pathophysiology) was a potentially frequent aetiology of stroke in our cohort, as reflected by the significantly higher prevalence of atrial fibrillation in this group compared with the ACS group (35 compared with 7% respectively; P=0.0001).
It is noteworthy that three of the 14 (21%) patients positive for anti-(apoA-1) IgG were also positive for anti-cardiolipin antibodies, but not for anti-β2GP1 antibodies. Even if none of these patients were known to have APS or any other autoimmune disease, we cannot exclude the concomitant presence of anti-cardiolipin antibodies in these samples. Another explanation for this phenomenon is a possible cross-reactivity between anti-(apoA-1) IgG and anti-cardiolipin antibodies, as has been shown in a few patients with SLE and APS , even if the signal of anti-(apoA-1) IgG was inhibited in a dose-dependent manner by apoA-1 in our assay.
The main limitation of the present study is related to the fact that it was designed as a prospective observational study and not as a case-control study. Thus controls used in the present study were not matched for sex, age and other cardiovascular risk factors, raising the possibility that some demographic factors could have acted as confounders, blunting the observations reported. However, because we included two other groups of patients (stroke and APE) that were of the same median age as patients in the ACS group but negative for anti-(apoA-1) IgG, we assume that age is unlikely to have biased our results. Moreover, when patients with ACS were divided into two subgroups according to their positivity for anti-(apoA-1) IgG, we did not observe prevalence differences for any of the traditional cardiovascular risk factors shown in Table 3. For this reason, even if differences in gender ratio, hypertension, diabetes, dyslipidaemia and smoking status were observed between the ACS and APE groups, but not when the ACS group was compared with stroke group (except for smoking), we believe that the presence of anti-(apoA-1) autoantibodies is not linked with the presence of any of the traditional coronary risk factors mentioned.
In conclusion, anti-(apoA-1) IgG are present in ACS, but apparently not in other prothrombotic conditions commonly associated with APS and SLE, such as stroke and APE. In ACS, their presence is associated with higher levels of LDL and oxLDL, and they are significantly and positively correlated with serum oxLDL levels, a major marker and mediator of atherosclerosis. Although the causal mechanisms of this association are currently not known, these results highlight the potential role of anti-(apoA-1) IgG in coronary atherogenesis, and emphasize the role of humoral autoimmunity as a mediator of atherosclerotic processes. Determining how these autoantibodies are implicated in atherogenesis is still speculative at the moment, but opens innovative research perspectives.
We are indebted to the staff of the Clinical Chemistry and Clinical Immunology and Allergy Laboratories of Geneva University Hospital for their kind and skilful assistance. We also thank Dr Maryse Bouchard for her kind grammatical assistance prior to submission. This work was supported by a grant for Research and Development from The University Hospital of Geneva (PRD 05-09-II).
Abbreviations: ACS, acute coronary syndrome; ANA, anti-nuclear antibody; ANCA, anti-neutrophil cytoplasmic antibody; APE, acute pulmonary embolism; apoA-1, apolipoprotein A-1; APS, antiphospholipid syndrome; BMI, body mass index; CAD, coronary artery disease; CK, creatine kinase; CRP, C-reactive protein; CT, computer tomography; cTnI, cardiac troponin I; β2GPI, β2 glycoprotein I; HDL, high-density lipoprotein; LDL, low-density lipoprotein; LVEF, left ventricular ejection fraction; oxLDL, oxidized LDL; PON-1, paraoxonase-1; RF, rheumatoid factor; SLE, systemic lupus erythematosous
- © The Authors Journal compilation © 2008 Biochemical Society