Atherosclerosis can be characterized as an inflammatory disease, where activated monocytes/macrophages and T cells are abundant in lesions and inflammation, as determined by C-reactive protein (CRP), is a risk factor [1–3]. Substantial amounts of cytokines, mainly proinflammatory, are produced by cells in the lesions [4, 5]. Although the inflammatory nature of atherosclerosis has been known for decades, the underlying mechanisms are still only partly known and the exact nature of antigens or compounds causing immune activation are not well characterized.
Atherosclerosis is not, as was often previously assumed, an irreversible process, causing disease by slow development of narrowing of the arterial lumen. Instead, the inflammation typical of atherosclerostic lesions, like other inflammatory and autoimmune diseases such as rheumatoid arthritis, can be ameliorated and also regress . Furthermore, cardiovascular disease (CVD) is related more to plaque rupture and atherothrombosis than to a passive thickening of the vessel wall [7, 8]. Atherosclerotic lesions prone to rupture are characterized by increased expression of proinflammatory cytokines as tumour necrosis factor-α (TNF-α) and γ-interferon (IFN-γ) and thus an activated immune response .
Immune mechanisms in atherosclerosis could therefore be of great importance at all stages of disease development, also late, when plaque rupture and atherothrombosis are a prominent manifestation of atherosclerosis.
Background – SLE and CVD
Systemic lupus erythematosus is often considered the prototypic autoimmune disease, where 90% of the cases are found in women. SLE is characterized by a plethora of disease manifestations, including nephritis, arthritis, pleuritis, pericarditis and vasculitis. A typical feature of SLE is the protean production of autoantibodies of different specificities and possibly related to this, disturbed apoptosis . Before treatment with immunosuppression was implemented, SLE was an often fatal disease. In the 1970s it became clear that not only acute disease flares but also later complications caused by CVD was an important clinical problem in SLE and in an important paper by Urowitz et al.  this bimodal mortality pattern of SLE was described.
These early studies have been largely confirmed, demonstrating an association between SLE and CVD and it is now clearly established that SLE patients are at high risk of CVD [12–17]. Of note, SLE women before menopause can be affected, when women are normally protected from coronary heart disease . The risk in some SLE patients is indeed strikingly high: in epidemiological studies, women aged 44–50 had a 50-fold increased risk of myocardial infarction when compared with controls from the Framingham study and the relative risk for coronary heart disease was 7.5, after adjusting for Framingham risk factors [12, 19]. Taken together, besides being an important clinical problem, atherosclerosis and CVD in systemic lupus erythematosus (SLE) patients could shed light on the role played by immune reactions in human atherosclerosis, not least in women because SLE mainly afflicts women.
SLE and atherosclerosis
Although the risk of CVD is raised in SLE, it is still not clear how atherosclerosis is related to this risk. Autopsy and angiographic studies demonstrated that the prevalence of atherosclerotic lesions was high in SLE, and the role of the then relatively recent introduction of corticosteroids in SLE was discussed as a possible underlying cause [20, 21].
It should also be noted that SLE-related CVD and atherosclerosis may differ from other conditions such as diabetes and hypertension, where it is often assumed that these conditions confer an increased risk in general. It could be the case that CVD – like many other disease manifestations of SLE – only affects a subgroup of SLE patients. Clearly, controlled, prospective studies are needed to establish whether CVD is a general feature of the disease or more a complication affecting only a subgroup of patients.
In order to study both risk factors and mechanisms of SLE-related CVD, we designed a nested case–control study from a cohort of above 200 SLE patients, were women with SLE and a history of CVD (SLE cases: defined by previous myocardial infarction, stroke, angina or claudication), were compared with age-matched women with no history of CVD and population control women. By use of B-mode ultrasound we measured the intima media thickness (IMT) of the carotid artery as a surrogate measurement of atherosclerosis. We demonstrated that both IMT, and presence of atherosclerotic plaques (defined as IMT above 1 mm) was higher amongst SLE cases when compared with controls. These findings argue against the possibility that thrombosis or other types of vascular dysfunction in the absence of atherosclerosis is a common cause of CVD in SLE, although it cannot be excluded in some patients .
We also noted that IMT of SLE controls was not increased when compared with population controls, although the presence of plaques was nonsignificantly more common amongst SLE controls than population controls . Roman et al.  and Wolak et al. subsequently reported in a larger case–control study of nonhospitalized SLE patients without signs of renal failure, that the presence of plaques was much more common amongst SLE patients than controls, a finding that was recently confirmed . Another striking finding in the study by Roman et al.  was that SLE patients actually had decreased atherosclerosis, if determined as IMT, e.g. a general intimal thickening and not only prevalence of atherosclerotic plaques. This interesting study also presented other unexpected findings: presence of atherosclerotic plaques was less common in patients with certain autoatibodies [against cardiolipin (CL), Smith antigen and RNP] and corticosteroid treatment was associated with lower prevalence of plaques. Blood pressure was lower than in the control group (which to some extent was recruited from a hypertension study). Patient selection could thus play a role, even more because this SLE cohort was not randomly selected .
One interpretation of these findings is that atherosclerosis in SLE patients is characterized by an increased risk of localized plaque but not increased atherosclerosis in general, as determined by IMT measurements, at least not in patients with less severe disease.
In another recent case–control study, coronary artery calcification, another atherosclerosis-related measurement, was more frequent in patients with lupus without manifest CVD than in control subjects . This study thus suggests that early-onset atherosclerotic disease still is characteristic of SLE.
These findings emphasize the need for a controlled prospective study to determine the more diffuse and generalized role of atherosclerosis, when compared with localized plaques or atherothrombotic factors or even general arterial dysfunction without signs of macroscopic disease, in the development of CVD in SLE. In such a study, it also appears important with randomly selected SLE patients and population-based controls. Other methods such as intravascular ultrasound or vascular stiffness determined by pulse-wave velocity could also add information about atherosclerosis in SLE.
Risk factors for CVD in SLE
It may be a question of discussion which measurements of atherosclerosis are optimal in SLE and reflective of risk and disease. However, it is of importance to establish which are the risk factors (and underlying mechanisms) for CVD. We have recently demonstrated that both traditional and nontraditional risk factors for CVD may explain the high risk of CVD in SLE (Fig. 1). Briefly, these risk factors were markers of inflammation (raised levels of acute phase reactants and TNF-α), dyslipidaemia (raised Tgs and low HDL), enhanced LDL oxidation, anti-phospholipid antibodies [aPL; lupus anticoagulants (LAC) and antibodies to oxidized LDL (aOxLDL)] and high levels of homocysteine [27, 28]. Many SLE patients with a history of CVD are on treatment with blood pressure-lowering drugs nowadays, and in line with this, previous studies indicate that hypertension is an important risk factor for CVD in SLE .
Figure 1. A combination of traditional and nontraditional risk factors contribute to the high risk of cardiovascular disease (CVD) in systemic lupus erythematosus (SLE) patients.
Download figure to PowerPoint
Similar results, where CVD in SLE is associated with traditional and nontraditional risk factors, have been reported in other studies, with comparable, albeit not identical, findings. One study in a multiethnic US cohort reported that another traditional risk factor, smoking, was important . Doria et al.  also described a similar combination of traditional and nontraditional risk factors, and emphasized the cumulative prednisolone dose as one such factor. In line with these findings is another study where Framingham risk factors did not fully account for CVD in SLE .
Yet another recent study indicates that smoking, and surprisingly, antidepressants, were risk factors for atherosclerotic plaques. Furthermore, vascular stiffness determined by pulse-wave velocity was more related to nontraditional risk factors such as immune dysregulation . In line with these studies are observations indicating that endothelial function as determined by flow-mediated dilatation of the brachial artery is impaired in SLE [32, 33].
A combination of traditional and nontraditional risk factors may thus account for the increased risk of CVD in SLE. Endothelial dysfunction and/or atherosclerosis in combination with prothrombotic factors are the underlying mechanisms which most likely cannot be separated but instead act in concert.
Dyslipidaemia characterized by decreased HDL, raised triglycerides, unchanged or only slightly elevated LDL and raised Lp(a) is typical of SLE. Especially in active disease, this lipid pattern has been described as the ‘lupus pattern of dyslipoproteinaemia’ but the underlying mechanisms are not well known [23, 34, 35]. Previous studies demonstrated that children with active SLE have elevated TGs . Furthermore, TGs are elevated in patients with inactive disease but are even higher amongst patients with active SLE  in patients with newly diagnosed SLE without steroid treatment.
We recently reported a strong correlation between elevated TGs and low levels of HDL on the one hand, and SLE disease activity and an association between cumulative disease damage and high cholesterol and TG levels on the other. These findings indicate that blood lipids may be novel disease markers in SLE [23, 27, 28]. Another finding was that elevated circulating levels of TNF-α are strongly associated with high TGs and low HDL in SLE patients [23, 27, 28]. During infections and a systemic inflammation, TNF-α levels are raised and occur together with similar lipoprotein derangement as seen in SLE . In clinical studies, TNF-α given to humans resulted in a rapid increase in circulating TG and VLDL levels . In rats, TNF-α administration induces a rapid and sustained increase in circulating TG-rich VLDL particles, an effect that may be caused by de novo hepatic synthesis of VLDL . One of the first effects attributed to TNF-α was its ability to inhibit lipoprotein lipase (LPL) , the major enzyme which degrades VLDL particles in the circulation. LPL activity has been studied in SLE and has been shown to be reduced by approximately 50% when compared with healthy individuals . In essence, induction of de novo hepatic lipogenesis and inhibition of LPL are two known mechanisms through which TNF-α can induce lipoprotein changes of the kind seen in SLE. It is therefore possible that activity of the TNF-α system contributes to the pattern of high TGs and low HDL seen in SLE. Whether other proinflammatory compounds including resistin and HMGB-1 could also have an effect on the ‘lupus pattern of dyslipoproteinaemia’ remains to be shown.
High-density lipoprotein may, in principle, protect against atherosclerosis not only through reverse cholesterol transport, but also by functioning as an anti-oxidant and anti-inflammatory agent, decreasing endothelial adhesivity [42, 43]. The HDL-associated apolipoprotein A-1 decreases TNF-α production through inhibiting contact-mediated activation of monocytes by binding to stimulated T cells . Thus, low HDL levels seem to be a consequence of active SLE but they may also indirectly contribute to enhanced inflammation in SLE.
When LDL undergoes oxidation, a variety of immunogenic neoepitopes are formed on OxLDL, which renders them antigenic and also promotes recognition by macrophage scavenger receptors, leading to enhanced uptake of OxLDL. One such antigen that is normally cryptic is phosphorylcholine (PC)-containing phospholipids [3, 45, 46]. Through this route macrophages become lipid laden and develop into the characteristic foam cells of the atherosclerotic lesions [3, 47–49]. OxLDL is also chemotactic, immune-stimulatory but additionally possesses inhibitory and even toxic properties [50, 51]. Furthermore, OxLDL elicits a humoral immune response with production of autoantibodies to oxidation-specific epitopes of OxLDL (aOxLDL).
Iuliano et al.  demonstrated that SLE patients have enhanced urinary excretion of isoprostanes, consistent with enhanced lipid peroxidation, and other studies have reported that increased oxidative stress and lipid peroxidation is raised in SLE . In line with this is a report demonstrating depressed activity of the antioxidant enzyme paraoxonase in SLE patients . We demonstrated that LDL oxidation, as determined by a monoclonal antibody, EO6 (recognizing exposed PC) on LDL, is raised in SLE-related CVD [23, 55]. These findings are also supported by studies in hamsters where both infection and inflammation induced LDL oxidation as expressed by another LDL epitope lysophosphatidylcholine (LPC) . We have previously shown that LPC is involved in the antigenicity of oxLDL . Subsequently, antibodies against LPC, an important mediator of many oxLDL-related effects, turn out to recognize PC and participate in the clearance of apoptotic cells and possibly also of oxLDL . Interestingly, a considerable proportion of exposed oxidized phospholipid epitopes appear to be present on Lp(a) .
Autoimmunity-related nontraditional risk factors including aPL and aOxLDL are implicated in CVD both in SLE and in the general population. Enhanced antibody titres to anionic phospholipids (aPL) including CL and LAC, characterize the antiphospholipid syndrome (APS), in addition to fetal loss, autoimmune thrombocytopenia and/or thrombosis . Amongst patients with SLE, 30–50% have aPL and approximately one-third to half of these develop APS , often referred to as secondary APS. Suggested mechanisms causing thrombosis, both venous and arterial, include interference with coagulation mechanism or a direct activating or even damaging effect on EC . Previous studies have indicated that aPL are also associated with CVD in the general population [61, 62].
We recently reported that inhibition of binding of anti-thrombotic plasmaprotein Annexin V to endothelium caused by aPL could represent a novel mechanism for CVD in SLE patients and possibly also in the general population (Fig. 2). We also discovered that Annexin V is abundant in atherosclerotic plaques, especially at sites prone to plaque rupture, and we suggest that although Annexin V may promote plaque growth in advanced disease, it may also stabilize plaques and inhibit plaque rupture .
Figure 2. Potential mechanism of atherothrombosis in systemic lupus erythematosus (SLE). aPL interfere with binding to endothelium of anti-thrombotic Annexin V.
Download figure to PowerPoint
Antibodies against endothelial cells (aEC) are implicated in SLE and are associated with disease activity and vasculitis . Furthermore, they promote endothelial activation by acting directly on endothelial cells . aEC has also been described in CVD in the general population [66, 67]. However, we could not demonstrate that aEC were raised in SLE-related CVD in our study . The capacity of aEC to promote atherogenesis or atherothrombosis in more active SLE patients remains to be evaluated.
Enhanced lipid peroxidation in patients with APS, as determined by secretion of oxidation products of arachidonic acid in urine, was recently reported . These findings indicate that lipid peroxidation may be of importance in APS, and not only in traditional atherosclerosis.
In APS and SLE a significant fraction of aCL recognizes oxidized phospholipids [69, 70]. The role played by aOxLDL appears to be more complex when compared with aPL but more directly related to atherosclerosis. Although not included in the definition of APS, some aOxLDL can also be regarded as aPL and even cross react with aCL [70, 71]. Recent evidence indicates that aOxLDL may also have atheroprotective properties. Active immunization with oxLDL leading to raised aOxLDL levels causes decreased atherosclerosis development in experimental animals . In humans, a growing body of evidence indicates that aOxLDL are low at an early stage of CVD development in nonautoimmune disease and in healthy individuals [72–74] but raised at later stages and in more advanced disease [3, 23, 75, 76]. One intriguing possibility is that natural antibodies against oxLDL-epitopes like PC, mainly of IgM subclass and belonging to the B-1 lineage of B cells, function as protection factors, by promoting clearance of oxidized lipids apoptotic cells and PC-containing bacteria . As discussed, autoantibodies against Smith antigen and RNP and aCL have been described to be negatively associated with atherosclerosis. Clearly, some autoantibodies could have atheroprotective roles and further studies are necessary to determine these differences and underlying mechanisms in detail.
The role of antibodies against heat shock proteins (aHSP) and other HSP measurements in CVD in general is likely to be complex. An atherogenic role played by aHSP65 is supported by animal experiments and clinical studies. One explanation for this finding could be that HSPs are conserved molecules from an evolutionary point of view and aHSP could cross react with HSPs present both on activated endothelial cells and in bacteria, including mycobacteria [77–79]. Whilst antibodies against HSP 60/65 are risk factors [77, 79–81], HSP70 appears to be a protection factor for atherosclerosis . However, we did not detect an association between HSP-related measurements in SLE-related CVD .
Inflammatory factors and cytokines
The presence of activated, immune-competent cells is typical of atherosclerotic lesions  and systemic inflammation, as reflected by a raised serum concentration of CRP, is associated with an enhanced risk of CVD . Accordingly, we recently demonstrated that both ESR and acute-phase reactants including α1antitrypsin, orosomucoid and CRP are associated with CVD in SLE: the potential role of TNF has been discussed above.
Interleukin-10 (IL-10) is a T helper 2 cytokine with inhibitory effects on pro-inflammatory T helper 1 cells, endothelial cells, granulocytes and monocytes/macrophages. However, IL-10 can also be immune stimulatory, by promoting antibody production and B-cell activation. Several authors have reported that IL-10 is raised in SLE patients and has been hypothesized as being one of the causative factors in SLE by induction of autoantibodies (e.g. against DNA) and of apoptosis [87–89]. In the light of this, it is surprising that findings from animal models indicate that IL-10 is an antiatherogenic cytokine [90, 91].
The production of IL-10 is known to be influenced by a functional polymorphism (−1087) and the A-1087 IL-10 allele is associated with a lower capacity for IL-10 production [92–95]. We recently demonstrated that the A allele frequency of the −1087IL-10 gene was positively associated with CVD in SLE . Furthermore, the IL-10 AA genotype is associated with a reduced ratio of atheroprotective-to-atherogenic cytokines in SLE/CVD patients. Both IL-10 and TNF-levels were raised in SLE/CVD patients, but lower IL-10 : TNF ratio was observed in those with the −1087IL10 AA genotype. The A-1087IL-10 allele may thus be one underlying factor contributing to the high risk of CVD in SLE. Developing SLE in spite of low capacity for IL-10 production could thus increase the risk of CVD in SLE. Whether IL-10 could protect some SLE patients from atherosclerosis remains to be elucidated.
Also, other cytokines including anti-inflammatory TGF-β or chemokines may be important in SLE-related CVD but this possibility has not yet been thoroughly studied.
Other inflammatory factors that may be of importance in SLE-related CVD include PAF-acetylhydrolase (PAF-AH) and secretory phospholipase A2 (PLA2) [97–99]. Both play a major role in the degradation of PAF and PAF-like lipids and in the generation of LPC. PAF, PAF-like lipids and LPC are both pro-inflammatory components of oxLDL that could induce and promote the inflammatory reaction in the vessel wall [99–104]. The potential involvement of PAF in atherogenesis has been implicated by the observation that inhibiting PAF activity decreases the development of atherosclerosis in animal models  and antibodies against PAF are associated with early CVD, atherosclerosis and thrombosis independent of other aPL [106, 107].
An increasing body of evidence indicates that PAF-AH may be either a risk factor or a marker for CVD, including stroke and coronary artery disease [108–110]. Our data support this and are compatible with the possibility that PAF-AH promotes atherogenesis by increasing LPC concentration in the arterial wall. Although PLA2 levels were not raised in the circulation of SLE cases in our study, this negative finding does not necessarily rule out the possibility that sPLA2 plays a role in the artery wall in SLE.
Derangements in both apoptotic mechanisms and the complement system are typical in SLE and may well play a role in CVD development but little is known about these aspects from a functional or genetic point of view.
Treatment in SLE and CVD
Ever since their introduction several decades ago, the role of corticosteroids on vessels has been much discussed. Clearly, randomized prospective studies, albeit relatively difficult to accomplish, are needed if this issue is to be completely clarified.
As atherosclerosis is an inflammatory disease, corticosteroids could be expected to have an antiatherogenic effect, a possibility supported by one experimental study using an animal model . By contrast, corticosteroids have well-known metabolic effects that could, in theory, be proatherogenic, and advanced atherosclerosis has been described especially in patients treated with steroids . It should be noted that treatment is implemented due to higher inflammatory reactivity and a history of steroid treatment could simply reflect a higher inflammatory activity which per se could be an important risk factor [23, 86, 112].
Relatively little is known about the role of other treatments in SLE in relation to CVD. We did not find any association between other medications commonly used in SLE and CVD . However, Roman et al.  described a negative association between immunosuppression and atherosclerosis.