Abstract
The risk of premature death is high in haemodialysis (HD) patients. Antibodies against cardiolipin (anti-CL) are thrombogenic in diseases such as systemic lupus erythematosus (SLE). CL is easily oxidized (Ox) and plays a role in apoptosis. In this work we studied immunoglobulin (Ig)M anti-CL and anti-OxCL in HD-patients. We conducted an observational study with a prospective follow-up examining the relationship between anti-CL, anti-OxCL and mortality risk in a well-characterized cohort of 221 prevalent HD patients [56% men, median age 66 (interquartile range 51–74) years, vintage time 29 (15–58) months] with a mean follow-up period of 41 (20–48 months). According to the receiver operator characteristic (ROC) analysis, anti-OxCL [area under the curve (AUC) 0·62, P < 0·01], but not anti-CL (AUC 0·52, P = 0·2), is associated with mortality. In crude and adjusted Cox analysis, every log increase in anti-OxCL inversely predicted all-cause [adjusted hazard ratios (HR) 0·62 (0·43–0·89)] and CVD-related [adjusted HR 0·56 (0·32–0·98)] mortality. Patients with anti-OxCL levels below median also had increased all-cause and cardiovascular disease (CVD)-related mortality. Although anti-OxCL and anti-phosphorylcholine (PC) were related positively to each other (ρ = 0·57, P < 0·01), patients with one or two of these autoantibody levels below the median were associated with an incrementally increased death risk. Anti-OxCL were co-factor β2-GPI-independent; anti-CL from patients with anti-phospholipid antibody syndrome were β2-GPI-dependent, while sera from HD-patients less so. Sera from healthy donors was not β2-GPI-dependent. Anti-OxCL IgM is β2-glycoprotein 1 (GPI)-independent and a novel biomarker; low levels are associated with death among HD patients (and high levels with decreased risk). Combination with anti-PC increases this association. Putative therapeutic implications warrant further investigation.
Keywords: antibodies, atherosclerosis, cardiolipin, cardiovascular disease, haemodialysis, inflammation
Introduction
Patients with chronic kidney disease (CKD) have poor prognosis with increased mortality. Atherosclerosis and its consequences, cardiovascular disease (CVD) and infectious complications such as pneumonia, are the main causes of death in this underserved population [1]. Persistent low-grade inflammation and the development of protein-energy wasting (PEW) are believed to be important factors that predispose to premature mortality [2].
Atherosclerosis is the major cause of CVD such as stroke and myocardial infarction (MI), which is regarded as an inflammatory disease, where activated immune competent cells producing cytokines are abundant [3]. Currently there are no anti-inflammatory or immune modulatory therapies developed to target CVD (or atherosclerosis). Further, there are few risk markers or risk factors which reflect inflammation, none of which target immune reactions. High-sensitivity C-reactive protein (hsCRP) is of interest in CVD, although volatility at the individual level of this measure may be a limitation [4,5]. Lipoprotein-associated phospholipase A2 (LP-PLA2) is another emerging inflammatory risk marker [6]. We have reported that natural antibodies against phosphorylcholine (PC) of immunoglobulin (IgM) subclass (anti-PC) are independently established risk factors associated negatively with the development of CVD, low levels leading to high risk [7–11]. As there is a need to identify uraemic patients at risk of developing serious infections, it is noteworthy that anti-PC inhibits the lethal effects of pneumococcal infection in mice [12].
Cardiolipin (CL) is a dimeric phospholipid, which is present in bacteria and in inner membranes of mitochondria of eucariotic cells – more so when metabolic activity is high [13]. High content of double bonds renders CL easily susceptible to oxidation [13].
Antibodies against CL (aCL) are well-known risk factors for both venous and arterial thrombosis, especially in patients with rheumatic diseases such as systemic lupus erythematosus (SLE) [14]. The immunogenicity of CL is dependent upon plasma co-factors such as β2-glycoprotein 1 (β2GP1). There is still a debate regarding the exact nature of how co-factors interact with CL so that it is recognized by thrombogenic antibodies [14]. Although little is known about the clinical role played by anti-OxCL, we have reported recently that low levels of these antibodies predict the development of CVD among healthy 60-year-olds [15].
In this study, we report that IgM antibodies against oxidized CL (anti-OxCL) but not anti-CL IgM are associated negatively (and independently of other known factors) with mortality risk among prevalent haemodialysis (HD) patients. The implications of these findings are discussed.
Subjects and methods
Patients and experimental design
This study includes prevalent patients undergoing HD at six dialysis units in Stockholm and Uppsala, both in Sweden. This is a post-hoc analysis from a cross-sectional study with mortality follow-up that aimed originally at investigating the variability of inflammatory markers in HD patients. Patient recruitment took place between October 2003 and March 2004; the protocol has been described previously in more detail [16]. We have previously published the association between anti-PC and mortality in this patient material [17], data that will be used in the analyses of this study. From the 224 patients included in the study (who survived the first 3 months after inclusion) and followed further for assessment of overall and cardiovascular mortality, anti-OxCL and anti-CL levels were determined in 221 (not enough plasma was available in three patients). The Ethics Committee of Karolinska Institutet and Uppsala University Hospital approved the study protocols. Informed consent was obtained from all patients.
A nephrologist reviewed each patient's medical chart and extracted data pertaining to underlying kidney disease, history of CVD, other co-morbid conditions and survival data. The co-morbidity history of each patient was determined at baseline according to the Davies co-morbidity scoring on a seven-point scale that was simplified into a three-risk-category scale [18]. Nutritional status was evaluated using the subjective global assessment (SGA). For the purpose of this study, PEW was defined as an SGA score > 1. Body mass index (BMI) and nutritional status were assessed on a dialysis day.
Survival was determined from the day of examination, with a mean follow-up period of 41 [interquartile range (IQR) 20–48] months, with no loss of follow-up. Causes of death were registered on the basis of each patient's medical chart and classified as CVD or non-CVD. CVD mortality was defined as death due to myocardial ischaemia or infarction, cardiac arrest or unknown sudden death, acute and chronic heart failure, cerebrovascular accidents, cerebral haemorrhage and ruptured aortic aneurysm. Non-CVD death was defined as that not attributable to a CVD origin. Individuals with unknown causes of death were grouped within the non-CVD group.
Laboratory analyses
Blood samples were collected before the HD session after the longest interdialytic period. Plasma and serum were separated and kept frozen at −70°C if not analysed immediately. Serum concentrations of interleukin (IL)-6 were quantified by immunometric assays on an Immulite Analyzer (Siemens Medical Solutions Diagnostics, Los Angeles, CA, USA). hsCRP (nephelometry), albumin and total cholesterol concentrations were analysed using certified methods at the Department of Laboratory Medicine in Karolinska University Hospital or Uppsala Academic Hospital.
Oxidation of CL
CL was purchased as ethanol solution from Sigma-Aldrich (Sigma-Aldrich, St Louis, MO, USA; product C 1649) and was stored at −20°C. CL was oxidized in aqueous solutions containing 1·5 mmol/l tert-butylhydroperoxide and 20 μmol/l CuSO4. Oxidation of CL was confirmed by mass spectrometry (electrospray ionization mass spectrometer; Micromass, Beverly, MA, USA) [15].
Determination of antibodies with enzyme-linked immunosorbent assay (ELISA)
IgM antibodies to OxCL were determined by ELISA. Serum from two donors with anti-OxCL IgM levels above median levels were used as internal standard and tested on every plate. The plateau of antibody binding was reached with an antigen concentration of 10 μg/ml. Immulon 1B plates (Thermo Labsystems, Franklin, MA, USA) were coated with OxCL (10 μg/ml, 50 μl/well) in ethanol. Coated plates were incubated overnight at 4°C. After five washings with phosphate-buffered saline (PBS), the plates were blocked with 2% bovine serum albumin (BSA)–PBS for 2 h at room temperature and washed as described above. Serum samples were diluted (1:50) in 0·2% BSA–PBS and added at 50 μl/well.
Plates were incubated overnight at 4°C and washed as described above. Alkaline phosphatase-conjugated goat anti-human IgM (diluted 1:7000 in the sample buffer) were added at 100 μl/well and incubated at 4°C overnight. After five washings, colour was developed by adding the alkaline phosphatase substrate (PNPP) at 100 μl/well and incubating the plates for 60 min at room temperature in the dark. The plates were read in an ELISA Multiscan Plus Spectrophotometer at 405 nm. All samples were measured in duplicate in a single assay and the intra/interassay coefficient of variation was below 15%.
IgM antibodies to CL where measured by a standard ELISA kit (Orgentec Diagnostika GMbH, Mainz, Germany) according to the manufacturer's instructions, except for the sample dilution. Samples were diluted 10 times less than instructed to allow for measurement of all, not only very high, antibody levels as reported previously [15].
Specificity of antibodies
In order to investigate the specificity of anti-OxCL IgM, competition assays were performed. At a dilution giving 50% of maximal binding to OxCL, sera were preincubated with different concentrations of OxCL or CL coated overnight onto glass tubes. After vortexing, the tubes were incubated at 4°C overnight and centrifuged at 13 000 g for 30 min (4°C). The supernatants were tested for antibody binding to OxCL [15]. The percentage of inhibition was calculated as follows:
 |
Effect of β2-glycoprotein 1 (GPI) on the binding of anti-OxCL or anti-CL with ELISA
The assay was performed essentially as described above for the anti-OxCL or anti-CL ELISA. Plates were coated with OxCL or CL, blocked with 2% BSA–PBS, then incubated with different concentrations of β2GPI or diluent alone with sera from high-titre patients from patients with the anti-phospholipid antibody syndrome (APS) for 1 h at room temperature. Each value shown is the mean ± standard deviation (SD).
Statistical analyses
All variables were expressed as mean ± SD, median (IQR) or percentage of patients, unless indicated otherwise. Statistical significance was set at the level of a P-value < 0·05 with two-sided tests. After normality assessment, comparisons between two groups were then assessed with Student's unpaired t-test and the Mann–Whitney or χ2 tests, as appropriate. Because many of the variables followed a non-normal distribution, Spearman's rank correlation (ρ) was used to determine correlations of anti-OxCL with other variables. In order to evaluate the diagnostic power of anti-OxCL for mortality, the receiver operator characteristic (ROC) curve analyses were performed. Survival analyses were made with the Kaplan–Meier survival curve and the Cox proportional hazard model and using anti-OxCL as a continuous variable, or by simple stratification of anti-OxCL according to the median value. The univariate and multivariate Cox regression analyses are presented as hazard ratios [HR; 95% confidence intervals (CI)]. Cox adjustments were performed on the basis of correction for factors that are known to be important predictors of the outcome in this disease. All statistical analyses were performed using the statistical software jmp 9.2 (SAS Institute, Cary, NC, USA).
Results
β2GPI, a co-factor for anti-CL, could induce increased anti-CL binding to CL in sera from APS patients. However, β2GPI sera from HD patients had a lower effect on binding, while β2GPI had no effect on binding in sera from healthy individuals (Fig. 1a,b). Preincubation with OxCL could inhibit up to 60–70% of IgM anti-OxCL binding to OxCL, while CL (exposed to air overnight) had low capacity to influence anti-OxCL binding.
Fig. 1.
Importance of β2-glycoprotein 1 (GPI) in antibody binding to antigens. Enzyme-linked immunosorbent assay (ELISA) of immunoglobulin (Ig)M anti-cardiolipin (CL) (a) and IgM anti-oxidized (Ox)CL (b) and dependency of β2-GPI in sera from uraemic patients [haemodialysis (HD) patients], patients with the anti-phospholipid antibody syndrome and healthy controls (a), who had been incubated with different concentration of β2-GP1 [for 1 h at room temperature (RT)]. β2-GPI increased anti-CL binding to CL but not anti-OxCL binding to OxCL among patients with anti-phospholipid antibody syndrome. The effect of anti-CL among uraemics was intermediate, while anti-CL in sera from healthy controls was β2GPI-independent. Antibody levels were determined by ELISA, and the optical density (OD) value where no β2GPI was added was set at 1 as a standard. The results are expressed as mean ± standard deviation.
Anti-OxCL IgM presented a non-normal distribution, with a median value of 73·5 (IQR 51·6–105·8, range 7·7–354·6). General characteristics of the patients studied are presented in Table 1, together with a comparison of individuals with low and high anti-OxCL IgM as stratified by the median value. Patients with low levels of anti-OxCL were older and had lower levels of both anti-PC and anti-CL. No differences were observed with regard to other demographic characteristics, co-morbidities, inflammation or nutritional markers. Anti-OxCL did not differ according to the presence/absence of PEW (defined as SGA > 1), diabetes, CVD and/or other co-morbidities reflected in the Davies Comorbidity score. In univariate analysis, anti-OxCL IgM was associated negatively with age and strongly positively with anti-PC and anti-CL IgM levels.
Table 1.
General characteristics of the prevalent haemodialysis patients included in the study stratified by the median anti-oxidized cardiolipin (OxCL) value, as well as univariate correlations with aOxCL levels
| All patients (n = 221) | ≤ Median aOxCL (n = 110) | > Median aOxCL (n = 111) | Spearman's rho | |
|---|---|---|---|---|
| Age (years) | 66 (51–74) | 70 (58–76) | 60 (48–72)** | −0·17†† |
| Males (%) | 56 | 55 | 57 | – |
| Vintage time (months) | 28 (13–57) | 28 (14–57) | 28 (12–51) | −0·01 |
| Davies comorbidity score (%) | 18/57/25 | 16/57/28 | 21/57/23 | – |
| Diabetes (%) | 27 | 44 | 51 | – |
| Cardiovascular disease (%) | 64 | 66 | 63 | – |
| C-reactive protein (mg/l) | 6·7 (2·5–21·0) | 8·2 (2·5–18·0) | 6·1 (2·4–22·5) | −0·03 |
| Interleukin-6 (pg/ml) | 8·7 (5·4–15·5) | 9·1 (5·9–15·1) | 8·2 (4·6–16·5) | −0·05 |
| Protein-energy wasting (%) | 47 | 44 | 51 | – |
| Body mass index (kg/m2) | 24·4 ± 5·2 | 24·8 ± 5·4 | 23·9 ± 5·1 | −0·10 |
| s-Albumin (g/l) | 34·4 ± 4·6 | 34·8 ± 4·2 | 34·0 ± 5·0 | −0·07 |
| Cholesterol (mmol/l) | 4·4 ± 1·1 | 4·4 ± 1·1 | 4·3 ± 1·1 | 0·02 |
| Anti-PC (U/ml) | 41·5 (20·6–82·3) | 25·1 (16·2–51·3) | 58·9 (36·3–95·3)*** | 0·56†† |
| Anti-CL-IgM | 121·6 (74·1–128·2) | 84·7 (55·2–144·4) | 157·0 (115·7–208·0)*** | 0·57†† |
| Anti-OxCL-IgM | 73·5 (51·6–105·8) | 51·6 (35·9–62·5) | 105·8 (87·1–124·8) | – |
Patients were divided according to anti-OxCL-immunoglobulin (Ig)M median value (73·5 units). Data are presented as median (interquartile range), average ± standard deviation or percentage; protein energy wasting is defined as SGA score > 1; Davies comorbidity score depicts the proportion of individuals with low, medium and high comorbidity risk. SGA: subjective global assessment; aOxCL, oxidized autoantibodies against cardiolipin. Significantly different from aOxCL-IgM below median value: **P < 0·001; ***P < 0·0001. Statistically significant univariate correlation with aOxCL-IgM as assessed by Spearman's rank test: †P < 0·05; ††P < 0·01.
Anti-CL IgM also presented a non-normal distribution, with a median value of 55·1 (range 5·7–202·5). Stratification by median value showed that patients with anti-CL IgM below the median were older and presented significantly lower levels of both anti-PC and anti-CL (data not shown). No differences were observed with regard to other demographic characteristics, co-morbidities, inflammation or nutritional markers. Anti-OxCL IgM did not differ according to the presence/absence of PEW (defined as SGA > 1), diabetes, CVD and/or other co-morbidities reflected in the Davies Comorbidity score.
During a median follow-up of 41 (IQR 19–48) months, 109 deaths occurred (49%), 44 of them attributed to cardiovascular causes as follows: myocardial ischaemia and infarction (n = 13), cardiac arrest/sudden death (n = 11), cerebrovascular accident (n = 4), cerebral haemorrhage (n = 3) and other causes of cardiac failure (n = 12). ROC curves versus all-cause mortality generated with serum anti-OxCL IgM showed a significant (P<0·01) good prediction with an area under the curve (AUC) of 0·62. As ROC curves versus anti-CL IgM did not correlate with mortality (AUC 0·53, P = 0·2), no further survival analysis was performed for this antibody.
Kaplan–Meier analysis according to the median anti-OxCL IgM value showed that patients with anti-OxCL IgM below the median had higher mortality (Fig. 2). Hazard ratios for all-cause and cause-specific mortality are presented in Table 2, together with the adjustment for potential confounders. In crude analysis, anti-OxCL IgM was associated inversely with all-cause and CVD-specific mortality (e.g. the higher the anti-OxCL IgM levels the lower the mortality risk). No association was observed with non-CVD mortality (data not shown). These associations remained statistically significant even after adjusting for age, sex, smoking, dialysis vintage and comorbidities. Additional sensitivity adjustment for inflammation (IL-6 levels) and PEW (SGA > 1) did not modify the results (data not shown).
Fig. 2.
(a) Kaplan–Meier for all-cause mortality according to median levels of oxidized autoantibodies against cardiolipin-immunoglobulin M (OxCL-IgM) during a follow-up period of 41 (20–48) months (log-rank χ2 8·30; P = 0·004). (b) Kaplan–Meier test for all-cause mortality according to whether none, one or two of these antibodies are decreased [anti-phosphorylcholine (PC) and OxCL-IgM] based on median values. Total number of patients 202 (log-rank χ2 15·4; P = 0·0004).
Table 2.
Crude and adjusted all-cause mortality hazard ratios according to aOxCL levels (per log increase) in 221 prevalent HD patients
| All-cause mortality | CVD mortality | ||||
|---|---|---|---|---|---|
| Number of deaths | 109 (49%) | 44 (20%) | |||
| Covariates | HR (95% CI) | P | HR (95% CI) | P | |
| Crude | aOxCL-IgM (per log increase) | 0·58 (0·41–0·81) | 0·002 | 0·49 (0·30–0·82) | 0·008 |
| Adjusted | aOxCL-IgM (per log increase) | 0·62 (0·43–0·89) | 0·009 | 0·56 (0·32–0·98) | 0·04 |
| Age (years) | 1·05 (1·03–1·07) | <0·0001 | 1·03 (1·01–1·06) | 0·03 | |
| Male | 1·13 (0·76–1·69) | 0·54 | 1·01 (0·54–1·92) | 0·92 | |
| Smoking (yes) | 1·64 (0·83–3·24) | 0·06 | 2·51 (0·97–6·58) | 0·05 | |
| Vintage (months) | 1·01 (0·99–1·02) | 0·43 | 1·01 (0·99–1·01) | 0·80 | |
| Comorbidities (Davies score) | 2·16 (1·56–3·01) | 0·003 | 2·58 (1·54–4·46) | 0·0002 | |
HR: hazard ratio; aOxCL: oxidized autoantibodies against cardiolipin; Ig: immunoglobulin.
Because we have shown previously in the same patient material that anti-PC values are associated inversely with mortality, we hypothesized that concurrent decreases in both these antibodies predicts outcome. In 202 of the 221 patients in whom both anti-PC and anti-OxCL IgM levels had been measured, we performed a survival analysis depending on whether none, one or two of these antibodies were decreased according to the stratification by median value. Kaplan–Meier analysis (Fig. 2b) shows a different impact on all-cause mortality for this grouping, whereby decreases in both antibodies render the highest death rates (Table 3). These groupings were explored through Cox regressions for both all-cause and CVD-specific mortality. In both crude and adjusted models, patients with decreases in both antibodies showed the highest hazards of dying from all-cause mortality. Regarding CVD mortality, we observed a similar effect size, although perhaps our relatively low number of CVD deaths made this association borderline significant (P = 0·05), and should be considered as a trend.
Table 3.
Crude and adjusted all-cause mortality hazard ratios according to according to whether none, one or two of these antibodies are reduced (anti-PC and aOxCL) based on median values
| All-cause mortality | CVD mortality | ||||
|---|---|---|---|---|---|
| Number of deaths | 109 (49%) | 44 (20%) | |||
| Covariates | HR (95% CI) | P | HR (95% CI) | P | |
| Crude | Zero antibodies reduced (n = 70) | 1·00 | 1·00 | ||
| One antibody reduced (n = 61) | 1·86 (1·08–3·25) | 0·02 | 1·91 (0·79–4·88) | 0·15 | |
| Two antibodies reduced (n = 71) | 2·67 (1·63–4·52) | <0·0001 | 3·13 (1·43–7·53) | 0·003 | |
| Adjusted | Zero antibodies reduced (n = 70) | 1·00 | 1·00 | ||
| One antibody reduced (n = 61) | 1·50 (0·83–2·78) | 0·17 | 1·41 (0·56–3·73) | 0·45 | |
| Two antibodies reduced (n = 71) | 2·04 (1·19–3·61) | 0·008 | 2·23 (0·98–5·53) | 0·05 | |
| Age (years) | 1·05 (1·03–1·07) | <0·0001 | 1·03 (0·99–1·06) | 0·05 | |
| Sex (men) | 1·25 (0·82–1·94) | 0·29 | 1·02 (0·54–1·99) | 0·93 | |
| Smoking (yes) | 1·20 (0·87–1·66) | 0·25 | 2·64 (1·01–2·05) | 0·04 | |
| Vintage (months) | 1·01 (0·99–1·02) | 0·54 | 1·01 (0·99–1·01) | 0·94 | |
| Comorbidities (Davies score) | 4·18 (2·19–8·19) | <0·0001 | 2·50 (1·49–4·34) | 0·0004 | |
CVD: cardiovascular disease; PC: phosphorylcholine; aOxCL: oxidized autoantibodies against cardiolipin; HR: hazard ratio.
Discussion
The chief finding is that low levels of IgM antibodies to OxCL, but not to CL, are associated with increased mortality among HD patients. Further, as the prognostic value of low levels of anti-OxCL was independent of other risk factors, our finding implies that low anti-OxCL levels represent a novel potential mechanism for premature death. Indeed, both the major causes of early death in uraemia, CVD and infections, could be related to CL, OxCL and anti-OxCL. Our findings accord with our recent report that, for the first time, anti-OxCL was demonstrated to be a protection marker for CVD, as studied in a cohort of healthy 60-year-olds [15].
Cardiolipin (CL) is a phospholipid with an unusual dimeric structure, which is synthesized by bacteria and is present exclusively in eucariotes in inner mitochondrial membranes. CL concentration is therefore increased in cells where the metabolic activity is high, as in cardiac cells (thereby the name) [13]. It is interesting to note that mitochondria have an origin as early forms of bacteria [19]. Because the content of double bonds in CL is high, it is easily susceptible to oxidative modification [20]. Previous reports indicate that CL is present in low-density lipoprotein (LDL) [21], an interesting finding which, however, was not confirmed in a recent report [22]. If CL undergoes oxidation in the atherosclerotic plaque, either as a component of LDL or of other compounds, this could promote oxidation of LDL. Oxidized LDL (OxLDL) is abundant in atherosclerotic lesions and has proinflammatory properties, activating T cells, endothelial cells and monocytes/macrophages [23–25]. As atherosclerotic lesions are characterized by activated immune competent cells producing proinflammatory cytokines, OxLDL could be of importance as an underlying factor [3], and it is possible that early oxidation of CL could initiate oxidation of LDL which could generate other inflammatory phospholipids, which could be causative of the oxLDL-induced immune activation [26,27]
Although it is not clear if OxCL is the underlying factor, CL is likely to play an important role in lung tissue damage in pneumonia [28]. Further, as oxidized forms of CL are implicated in tissue damage caused by radiation and other damaging agents [29,30], it is possible that anti-OxCL contributes to the defence against infection-related tissue damage. A major feature of atherosclerotic plaques is the abundance of dead cells in a necrotic core [31]. These are known to promote inflammatory reactions; an important question is why the cells die in the first place, and why they are not cleared effectively. It is therefore of interest that CL undergoes oxidation early during apoptosis [32]. Further studies are needed to determine if anti-OxCL could potentiate phagocytosis and clearance of apoptotic cells and the components thereof. Recent findings that anti-PC, another type of natural antibody, facilitates such phagocytosis of apoptotic cells [33], supports this hypothesis. We have shown recently that there is a negative association between anti-PC and risk of CVD and atherosclerosis development [8]. In addition, using the same cohort we have demonstrated that low anti-PC levels predict premature mortality [17]. Although we found an association between anti-PC and anti-OxCL, PC and oxCL have different antigenic structures. Here we report that having a combination of below median levels of both anti-OxCL and anti-PC synergistically increases the risk of death. This finding implies that the antibodies recognize different epitopes, even though they may not be as distinct as is the case with more specific, recognizing peptide antigens after somatic mutation.
In this study, anti-CL IgM was not associated with increased risk of death among HD patients and, further, there were no associations between very high levels of anti-CL IgM and death (or death by CVD). Thus, the role of anti-CL IgM as a risk factor for CVD in SLE is present only when anti-CL IgM levels are exceedingly high [14,34]. As this study did not detect a role for anti-CL IgM in death or CVD in the uraemic milieu, further studies are needed to determine if IgG or other anti-PL subclasses could play a role.
The extensively oxidized CL used here did not show cross-reactivity with CL as antigen [15], and we have not studied other forms of oxidized or modified CL. The commercial kit used herein for anti-CL determinations is developed to not promote CL oxidation in the experimental settings.
We cannot exclude the possibility that some of the anti-OxCL binding to OxCL is present in immune complexes. However, we think this is beyond the scope of the present paper, which focused on the totality of antibodies binding to OxCL.
Other limitations in the study deserve mention. As the number of patients is somewhat low, larger studies are needed to confirm our findings. The accuracy of death certificate data is limited, because autopsies are performed infrequently. Finally, as studies suggest that more patients with chronic kidney disease die of CVD before they reach the dialysis facility [35], this cohort of prevalent patients constitute a selected group of survivors.
The exact antigen for anti-CL and other anti-phospholipid antibodies has been much discussed. Even though there is no consensus, as some researchers suggest that the lipid moiety, even as oxidized, is a major antigen, (most) others think that β2GPI is the antigen, either by itself or in combination with phospholipids. There is general agreement, however, that β2GPI seems to indeed be of major importance in many, if not most cases, in patients with APS with or without SLE [14]. It has also been much discussed whether anti-CL in the general population, when present at high levels, is as dangerous as when the patient also suffers from SLE [14]. However, recent studies indicate that this is not the case [36]. Our finding that anti-CL in sera from healthy donors is β2GPI-independent could provide an explanation as to why anti-CL is a more clear-cut risk marker among patients with SLE and the APS which is common in SLE.
Air-exposed CL (which leads to a degree of oxidation) is, in fact, is a risk marker in a mouse model of atherosclerosis, and not a protection marker as in humans [20]. Such air-exposed oxidized CL was β2GPI-dependent as a co-factor for optimal recognition by antibodies [37], as opposed to the OxCL studied herein. The degree and type of oxidation of CL could thus play an important role in determining the pathogenicity of the corresponding antibodies.
Taken together, our data suggest that low levels of anti-OxCL IgM serve as a novel biomarker to identify HD patients at increased risk. The possibility that anti-OxCL IgM (potentially in combination with anti-PC) could be used as a novel treatment strategy in dialysis patients certainly deserves further study. In future, immunomodulation to raise levels of anti-OxCL, possibly in combination with other antibodies as IgM anti-PC, through active or passive immunization could certainly be of interest.
Acknowledgments
We would like to thank the patients and personnel involved in the creation of this cohort. We are also indebted to our research staff at KBC (Annika Nilsson, Ann-Kristin Emmot and Ulrika Jenson) and KFC (Björn Anderstam, Monica Ericsson and Ann-Kristin Bragfors-Helin). The MIMICK cohort was supported by an unrestricted grant from Amgen Inc. We also benefited from the Swedish Medical Research Council (VR), Loo and Hans Ostermans and Westmans Foundations, Vinnova, CIDaT, AFA, Torsten Söderbergs Foundation and the 6th Framework Program of the European Union, Priority 1: Life sciences, genomics and biotechnology for health (grant LSHM-CT-2006-037227 CVDIMMUNE).
Disclosure
J. F. is named as inventor on patents and patent applications relating to antibodies against phospholipid epitopes. A. F. is named as co-inventor on a patent application related to antibodies against phospholipid epitopes.
Author contributions
A. F. and J. F. wrote the paper, X. H. and J. S. conducted the experiments and co-wrote the paper, J. J. C. made statistical calculations and co-wrote the paper, O. H., P. B. and P. S. co-wrote the paper and contributed to clinical aspects. All participated in discussions about the findings.
References
- 1.Stenvinkel P. Chronic kidney disease: a public health priority and harbinger of premature cardiovascular disease. J Intern Med. 2010;268:456–467. doi: 10.1111/j.1365-2796.2010.02269.x. [DOI] [PubMed] [Google Scholar]
- 2.Stenvinkel P, Carrero JJ, Axelsson J, Lindholm B, Heimburger O, Massy Z. Emerging biomarkers for evaluating cardiovascular risk in the chronic kidney disease patient: how do new pieces fit into the uremic puzzle? Clin J Am Soc Nephrol. 2008;3:505–521. doi: 10.2215/CJN.03670807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Frostegard J, Ulfgren AK, Nyberg P, et al. Cytokine expression in advanced human atherosclerotic plaques: dominance of pro-inflammatory (Th1) and macrophage-stimulating cytokines. Atherosclerosis. 1999;145:33–43. doi: 10.1016/s0021-9150(99)00011-8. [DOI] [PubMed] [Google Scholar]
- 4.Ridker PM, Stampfer MJ, Rifai N. Novel risk factors for systemic atherosclerosis: a comparison of C-reactive protein, fibrinogen, homocysteine, lipoprotein(a), and standard cholesterol screening as predictors of peripheral arterial disease. JAMA. 2001;285:2481–2485. doi: 10.1001/jama.285.19.2481. [DOI] [PubMed] [Google Scholar]
- 5.Stenvinkel P, Lindholm B. C-reactive protein in end-stage renal disease: are there reasons to measure it? Blood Purif. 2005;23:72–78. doi: 10.1159/000082014. [DOI] [PubMed] [Google Scholar]
- 6.Zalewski A, Nelson JJ, Hegg L, Macphee C. Lp-PLA2: a new kid on the block. Clin Chem. 2006;52:1645–1650. doi: 10.1373/clinchem.2006.070672. [DOI] [PubMed] [Google Scholar]
- 7.Caidahl K, Hartford M, Karlsson T, et al. IgM-phosphorylcholine autoantibodies and outcome in acute coronary syndromes. Int J Cardiol. 2012;167:464–469. doi: 10.1016/j.ijcard.2012.01.018. [DOI] [PubMed] [Google Scholar]
- 8.Frostegard J. Low level natural antibodies against phosphorylcholine: a novel risk marker and potential mechanism in atherosclerosis and cardiovascular disease. Clin Immunol. 2010;134:47–54. doi: 10.1016/j.clim.2009.08.013. [DOI] [PubMed] [Google Scholar]
- 9.Fiskesund R, Stegmayr B, Hallmans G, et al. Low levels of antibodies against phosphorylcholine predict development of stroke in a population-based study from northern Sweden. Stroke. 2010;41:607–612. doi: 10.1161/STROKEAHA.109.558742. [DOI] [PubMed] [Google Scholar]
- 10.de Faire U, Su J, Hua X, et al. Low levels of IgM antibodies to phosphorylcholine predict cardiovascular disease in 60-year old men: effects on uptake of oxidized LDL in macrophages as a potential mechanism. J Autoimmun. 2010;34:73–79. doi: 10.1016/j.jaut.2009.05.003. [DOI] [PubMed] [Google Scholar]
- 11.Sjoberg BG, Su J, Dahlbom I, et al. Low levels of IgM antibodies against phosphorylcholine-A potential risk marker for ischemic stroke in men. Atherosclerosis. 2009;203:528–532. doi: 10.1016/j.atherosclerosis.2008.07.009. [DOI] [PubMed] [Google Scholar]
- 12.Briles DE, Forman C, Hudak S, Claflin JL. Anti-phosphorylcholine antibodies of the T15 idiotype are optimally protective against Streptococcus pneumoniae. J Exp Med. 1982;156:1177–1185. doi: 10.1084/jem.156.4.1177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Schlame M. Cardiolipin synthesis for the assembly of bacterial and mitochondrial membranes. J Lipid Res. 2008;49:1607–1620. doi: 10.1194/jlr.R700018-JLR200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Frostegard J. Atherosclerosis in patients with autoimmune disorders. Arterioscler Thromb Vasc Biol. 2005;25:1776–1785. doi: 10.1161/01.ATV.0000174800.78362.ec. [DOI] [PubMed] [Google Scholar]
- 15.Su J, Hua X, Vikstrom M, et al. Low levels of IgM antibodies to oxidized cardiolipin increase and high levels decrease risk of cardiovascular disease among 60-year olds: a prospective study. BMC Cardiovasc Disord. 2013;13:1. doi: 10.1186/1471-2261-13-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Carrero JJ, Qureshi AR, Axelsson J, et al. Comparison of nutritional and inflammatory markers in dialysis patients with reduced appetite. Am J Clin Nutr. 2007;85:695–701. doi: 10.1093/ajcn/85.3.695. [DOI] [PubMed] [Google Scholar]
- 17.Carrero JJ, Hua X, Stenvinkel P, et al. Low levels of IgM antibodies against phosphorylcholine-A increase mortality risk in patients undergoing haemodialysis. Nephrol Dial Transplant. 2009;24:3454–3460. doi: 10.1093/ndt/gfp309. [DOI] [PubMed] [Google Scholar]
- 18.Davies SJ, Phillips L, Naish PF, Russell GI. Quantifying comorbidity in peritoneal dialysis patients and its relationship to other predictors of survival. Nephrol Dial Transplant. 2002;17:1085–1092. doi: 10.1093/ndt/17.6.1085. [DOI] [PubMed] [Google Scholar]
- 19.Martin W, Hoffmeister M, Rotte C, Henze K. An overview of endosymbiotic models for the origins of eukaryotes, their ATP-producing organelles (mitochondria and hydrogenosomes), and their heterotrophic lifestyle. Biol Chem. 2001;382:1521–1539. doi: 10.1515/BC.2001.187. [DOI] [PubMed] [Google Scholar]
- 20.Horkko S, Miller E, Dudl E, et al. Antiphospholipid antibodies are directed against epitopes of oxidized phospholipids. Recognition of cardiolipin by monoclonal antibodies to epitopes of oxidized low density lipoprotein. J Clin Invest. 1996;98:815–825. doi: 10.1172/JCI118854. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Deguchi H, Fernandez JA, Hackeng TM, Banka CL, Griffin JH. Cardiolipin is a normal component of human plasma lipoproteins. Proc Natl Acad Sci USA. 2000;97:1743–1748. doi: 10.1073/pnas.97.4.1743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Dashti M, Kulik W, Hoek F, Veerman EC, Peppelenbosch MP, Rezaee F. A phospholipidomic analysis of all defined human plasma lipoproteins. Sci Rep. 2012;1:139. doi: 10.1038/srep00139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Frostegard J, Nilsson J, Haegerstrand A, Hamsten A, Wigzell H, Gidlund M. Oxidized low density lipoprotein induces differentiation and adhesion of human monocytes and the monocytic cell line U937. Proc Natl Acad Sci USA. 1990;87:904–908. doi: 10.1073/pnas.87.3.904. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Frostegard J, Wu R, Giscombe R, Holm G, Lefvert AK, Nilsson J. Induction of T-cell activation by oxidized low density lipoprotein. Arterioscler Thromb. 1992;12:461–467. doi: 10.1161/01.atv.12.4.461. [DOI] [PubMed] [Google Scholar]
- 25.Berliner JA, Territo MC, Sevanian A, et al. Minimally modified low density lipoprotein stimulates monocyte endothelial interactions. J Clin Invest. 1990;85:1260–1266. doi: 10.1172/JCI114562. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Frostegard J, Huang YH, Ronnelid J, Schafer-Elinder L. Platelet-activating factor and oxidized LDL induce immune activation by a common mechanism. Arterioscler Thromb Vasc Biol. 1997;17:963–968. doi: 10.1161/01.atv.17.5.963. [DOI] [PubMed] [Google Scholar]
- 27.Huang YH, Schafer-Elinder L, Wu R, Claesson HE, Frostegard J. Lysophosphatidylcholine (LPC) induces proinflammatory cytokines by a platelet-activating factor (PAF) receptor-dependent mechanism. Clin Exp Immunol. 1999;116:326–331. doi: 10.1046/j.1365-2249.1999.00871.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ray NB, Durairaj L, Chen BB, et al. Dynamic regulation of cardiolipin by the lipid pump Atp8b1 determines the severity of lung injury in experimental pneumonia. Nat Med. 2010;16:1120–1127. doi: 10.1038/nm.2213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Tyurina YY, Tyurin VA, Kapralova VI, et al. Oxidative lipidomics of gamma-radiation-induced lung injury: mass spectrometric characterization of cardiolipin and phosphatidylserine peroxidation. Radiat Res. 2011;175:610–621. doi: 10.1667/RR2297.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Tyurina YY, Tyurin VA, Kaynar AM, et al. Oxidative lipidomics of hyperoxic acute lung injury: mass spectrometric characterization of cardiolipin and phosphatidylserine peroxidation. Am J Physiol Lung Cell Mol Physiol. 2010;299:L73–85. doi: 10.1152/ajplung.00035.2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Geng YJ, Libby P. Progression of atheroma: a struggle between death and procreation. Arterioscler Thromb Vasc Biol. 2002;22:1370–1380. doi: 10.1161/01.atv.0000031341.84618.a4. [DOI] [PubMed] [Google Scholar]
- 32.Kagan VE, Tyurin VA, Jiang J, et al. Cytochrome C acts as a cardiolipin oxygenase required for release of proapoptotic factors. Nat Chem Biol. 2005;1:223–232. doi: 10.1038/nchembio727. [DOI] [PubMed] [Google Scholar]
- 33.Chen Y, Khanna S, Goodyear CS, et al. Regulation of dendritic cells and macrophages by an anti-apoptotic cell natural antibody that suppresses TLR responses and inhibits inflammatory arthritis. J Immunol. 2009;183:1346–1359. doi: 10.4049/jimmunol.0900948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Svenungsson E, Jensen-Urstad K, Heimburger M, et al. Risk factors for cardiovascular disease in systemic lupus erythematosus. Circulation. 2001;104:1887–1893. doi: 10.1161/hc4101.097518. [DOI] [PubMed] [Google Scholar]
- 35.Keith DS, Nichols GA, Gullion CM, Brown JB, Smith DH. Longitudinal follow-up and outcomes among a population with chronic kidney disease in a large managed care organization. Arch Intern Med. 2004;164:659–663. doi: 10.1001/archinte.164.6.659. [DOI] [PubMed] [Google Scholar]
- 36.Pengo V, Banzato A, Bison E, Bracco A, Denas G, Ruffatti A. What have we learned about antiphospholipid syndrome from patients and antiphospholipid carrier cohorts? Semin Thromb Hemost. 2012;38:322–327. doi: 10.1055/s-0032-1304719. [DOI] [PubMed] [Google Scholar]
- 37.Horkko S, Miller E, Branch DW, Palinski W, Witztum JL. The epitopes for some antiphospholipid antibodies are adducts of oxidized phospholipid and beta2 glycoprotein 1 (and other proteins) Proc Natl Acad Sci USA. 1997;94:10356–10361. doi: 10.1073/pnas.94.19.10356. [DOI] [PMC free article] [PubMed] [Google Scholar]