The main finding in this study is that IgM antibodies against oxidized cardiolipin are negatively associated with risk for development of CVD. Individuals (men and women) with low levels of anti-OxCL (below lowest quartile) had an almost 2-fold increased risk, independent of other risk factors. Further, high levels of IgM anti-OxCL were protective: values above 86th percentile were associated with a more than 50% reduction of risk of CVD. The associations were especially striking for men; when analysed separately, the risk of development of MI was almost 2.5 in the lowest quartile and even though stroke cases were relatively few, it is interesting to note that the risk of stroke was 12-fold for men with anti-OxCL in the lowest quartile (though confidence intervals were high). In addition, men who within 5–7 years developed CVD had significantly lower median values at baseline than men who did not. There were also significant associations for women, where a two-fold risk for MI was noted below the highest quartile of anti-OxCL levels.
Further, we report that IgG anti-OxCL were not associated with CVD when divided in quartiles, however, very high antibody levels were indeed strong protection markers, with a decreased risk of CVD (OR of 0.23 above 95th percentile).
CL is abundant in the environment of mitochondrial oxidative burst, and is easily oxidized due to its unsaturated double bonds
[16, 17]. Further, it was demonstrated that cardiolipin is the only phospholipid in mitochondria undergoing early oxidation during apoptosis, by a cardiolipin-specific peroxidase activity of cardiolipin-bound cytochrome c
[18, 19]. Oxidized cardiolipin is needed for release of proapoptotic factors and thus may play a role in induction of apoptosis, eg through effects on release of cytochrome c, since cytochrome c anchorage in mitochondria depends on CL, and CL loss and/or oxidation can induce detachment and release, leading to apoptosis
[18–20]. Though the OxCL studied herein is extensively oxidized it is thus still possible that similar forms of OxCL are present in vivo. The non-mutually exclusive possibility that enzymes such as phospholipase A2 may play an important role in CL-modification (as in LDL-modification) is also interesting and deserves further studies.
Our findings are comparable with our recent observations on another phospholipid-related epitope, phosphorylcholine (PC). We have reported that antibodies to PC (anti-PC) are negatively associated with CVD and atherosclerosis, with the most striking observations at low levels, which are associated to increased risk of CVD
[5–8, 21] while high levels are associated with a favourable development with less atherosclerosis
. It should be noted that PC is not present in CL (and thus not in OxCL) and further, cross reactivity between these antigens was relatively low (unpublished observation). In contrast to anti-PC, anti-OxCL was a protective factor for CVD at high levels, since high levels were associated with a 50% decreased risk.
In an interesting report it was demonstrated that CL is quickly oxidized when coated on ELISA-plates for determination of anti-CL and many anti-CL may recognize this variant of OxCL which could contribute to the antiphospholipid antibody syndrome
. Further, this form of CL, oxidized by exposure to air, had ß2GPI as a co-factor
. Interestingly, ß2GPI does not appear to play a role in antibody binding to OxCL. However, we cannot exclude that other protein-OxCL complexes could play a role, but if so, the antigen is not recognized by atherotrhombogenic antibodies.
The nature of the antigen recognized by anti-CL is still debated, almost three decades after their discovery
, but there are many studies indicating that the plasma co-factor ß2GPI increases the binding to CL and plays a role in the antigenicity and also thrombogenicity.
The oxCL we used did not show a high cross-reactivity with CL as antigen. The commercial kit used herein for determination of antibodies against CL is developed to detect native CL. In contrast to IgM anti-OxCL, IgM anti-CL was not significantly associated with CVD either negatively or positively in the whole study group, though among men, median level of anti-CL IgM was lower among cases. However, there were trendwise associations between high IgM anti-CL-levels and increased risk of CVD among men and almost threefold but non-significant risks for women, when the highest percentiles were studied. Further, IgG anti-CL (which in general is believed to be more prothrombotic than IgM
), were not associated with CVD, not even trendwise. Whether some of the anti-OxCL binding to OxCL are present in immune complexes is an interesting possibility, which is beyond the scope of the present paper, which focused on the totality of oxCL binding antibodies.
The role of anti-CL in the general population is not clear, though such antibodies may play a role in subgroups as in survivors of MI (below 45 years of age)
. In one study on recurrent coronary events among patients with a history of MI, high IgG anti-CL, but also, interestingly, low IgM anti-CL were associated with increased risk
. In our study of 60-year olds with no history of MI or CVD, we did not determine such associations and determinations also of anti-OxCL in patients with a history of MI and/or CVD are clearly of interest. However, it appears likely that anti-CL mainly plays a role as a risk factor for CVD (including both arterial and venous thrombosis) in systemic rheumatic diseases like SLE where such antibodies are strong risk factors for CVD
[11, 27], and possibly among individuals with a history of CVD.
To identify mechanisms which could explain the protective role of anti-OxCL, we determined interactions with oxLDL, a major factor in atherosclerosis. OxLDL could inhibit antibody binding to oxCL but not to CL, which indicates cross reactivity. In an interesting previous report, it was demonstrated that aOxLDL cross react with anti-CL
. The most likely explanation to this difference is that CL is to some extend oxidized during the incubation used in that publication
. The finding that oxLDL cross reacts with oxCL could have several implications. Since CL is present in LDL
, our finding most likely indicates that oxCL is a component of oxLDL. We recently demonstrated that oxCL but not CL is pro-inflammatory (unpublished observation) and one possibility is therefore that anti-OxCL is anti-inflammatory.
We demonstrated herein is that oxCL but not CL inhibits uptake of oxLDL in macrophages. OxLDL is taken up through specific scavenger receptors, which are not down-regulated when exposed to increasing amounts of oxLDL (as opposed to the uptake of LDL). Inhibition of the scavenger function is generally believed to be atheroprotective, preventing foam cell formation in the vascular wall which is a key process in development of atherosclerosis. In line with this, mice defective in scavenger receptor function develop less atherosclerosis
. Binding and uptake of oxLDL through oxCL could thus promote atherogenesis.
It should be noted, however, that recent research indicates that different scavenger receptors may play different roles and the role of scavenger receptors may vary depending on disease stage and type
It has previously been demonstrated that murine monoclonal IgM recognizing forms of ox-CL can discriminate apoptotic cells from healthy cells
. Further, natural IgM binding to apoptotic cells increase apoptotic cell phagocytosis and also have direct anti-inflammatory properties
. It is possible that natural IgM including anti-PC and aOxCL could counter atherosclerosis development by increasing phagocytosis of dead and dying cells in the lesions.
It is interesting to note that the associations here were independent on traditional risk factors, and also hsCRP. Further, presence of hypertension in combination with low anti-OxCL increased the risk.