The inextricable link between atherosclerosis and prototypical inflammatory diseases rheumatoid arthritis and systemic lupus erythematosus

The increased burden of cardiovascular disease in patients with rheumatoid arthritis and systemic lupus erythematosus has recently become the focus of intense investigation. Proatherogenic risk factors and dysregulated inflammation are the main culprits, leading to enhanced atherosclerosis in subgroups of patients with inflammatory diseases. Common molecular pathways shared by atherosclerosis and inflammatory disease may be involved. In this review we map the key determinants of the increased incidence of cardiovascular disease in patients with inflammatory diseases at each step of the atherogenesis.


Introduction
Research during the past 20 years has driven a major shift in how atherosclerosis is conceptualized. Initially branded as a passive accumulation of lipids in the vessel wall, atherosclerosis is now recognised as an 'inflammatory disease' [1]. Striking similarities can be identified between atherosclerosis and prototypical inflammatory diseases ( Figure 1). In parallel, there is growing evidence that cardiovascular disease (CVD) is the leading cause of mortality in patients with chronic inflammatory diseases [2], including rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), Sjögren's disease and systemic sclerosis.
This review first summarizes the impact of CVD on the lives of patients with inflammatory diseases. Second, we map the key molecular determinants of the increased prevalence of CVD in patients with inflammatory diseases at each step of the progression of atherosclerosis (initiation, progression and thrombotic complications). We focus on RA and SLE, for which more evidence is currently available.
(page number not for citation purposes) MI compared with age-matched and sex-matched control individuals [11]. In a Canadian cohort of SLE patients, the relative risk for MI was 10.1, that for death due to CHD was 17, and that for stroke was 7.9 [12]. SLE patients have increased subclinical atherosclerosis compared with the general population, with a greater prevalence of carotid plaques and increased IMT [13,14]. Myocardial single photon emission computed tomography scanning has revealed coronary artery disease in 40% of patients with SLE [15,16], although coronary artery calcification is more prevalent in lupus without underlying CVD [17].

Risk factors for cardiovascular events in patients with rheumatoid arthritis and systemic lupus erythematosus
Clustering of traditional atherogenic risk factors RA and SLE patients have an overall increased frequency of traditional cardiovascular risk factors [3,10,18,19]. Smoking is associated with subclinical atherosclerosis in patients with RA [20]. Hypertension is also an important risk factor for CVD in RA and SLE [21,22].
A particular type of dyslipidaemia is present in patients with RA. This is characterized by low high-density lipoprotein (HDL), raised triglycerides and low levels of low-density lipoprotein (LDL) [23]. Interestingly, such a lipid profile is detectable in those individuals who go on to develop RA at least 10 years before the onset of symptoms [24]. Dyslipo-proteinaemia in SLE is slightly different from that in RA, characterized by elevated levels of very-low-density lipoprotein cholesterol and triglycerides, high or unchanged LDL, low levels of HDL and disturbance in chylomicron metabolism. Hence, low HDL levels are a common trait in RA and SLE. Low HDL in the general population exhibits a strong inverse correlation with CVD events, even in the presence of low LDL or statin therapy. Changes in lipoprotein functions have also been described. For instance, LDL from RA patients has an increased capacity to bind proteoglycans, which is considered an important early step in atherogenesis [25]. High levels of the pro-atherogenic lipid particle lipoprotein-a have been detected in both diseases, as well as elevated levels of circulating oxidized LDL (oxLDL) [26]. Paraoxonase, an enzyme that binds to HDL and prevents oxidation of LDL, was shown to have reduced activity in patients with SLE and primary antiphospholipid syndrome [27].
Patients with RA and SLE have a higher prevalence of metabolic syndrome [28], which includes obesity, dyslipidaemia, hypertension and insulin resistance. Additionally, patients with RA and SLE are less physically active than individuals without the disease, which may also contribute to accelerated atherosclerosis.

Traditional therapies and cardiovascular risk
The effect of chronic use of drugs commonly employed to treat RA and SLE in promoting or suppressing atherogenesis Similarities between the atherosclerotic plaque and rheumatoid arthritis joint. The (a) atherosclerotic plaque has many features in common with (b) rheumatoid arthritic synovium. First, in both diseases, blood-borne mononuclear cells are recruited to sites that are devoid of any significant inflammation in physiological conditions. Second, upregulation of cytokines and matrix-degrading enzymes is central to the pathogenesis of both diseases. Third, both in rheumatoid arthritis and atherosclerosis, immune cells do not target resident cells in the same way that diabetogenic T cells directly destroy pancreatic islets. Instead, immune cells begin complex interactions with the resident cell types, which proliferate, change their properties and phenotype, and contribute to the inflammatory process and tissue destruction.
is extensively reviewed elsewhere [23,29]. Such drugs can affect the progression of atherosclerosis either by acting on the inflammatory process or by altering cardiovascular risk factors. Corticosteroids, for instance, promote insulin resistance, weight gain, fluid retention and hypertension. Nonetheless, within the context of inflammatory diseases their net effect is likely to be complex because of their beneficial effects on the ongoing inflammation. In the COBRA (Combination Therapy in Early Rheumatoid Arthritis) study [30], high-dose steroids used in conjunction with other disease-modifying antirheumatic drugs improved the total cholesterol to HDL ratio, with a linear association with the Disease Activity Score. However, in a different study [31] glucocorticoid exposure was associated with carotid plaque and arterial incompressibility.
The effect of methotrexate is similarly complex. Methotrexate use causes elevation in serum homocysteine levels by folate depletion. This effect could promote atherogenesis as because hyperhomocysteinaemia is an independent risk factor associated with CVD in the general population as well as in patients with SLE [32]. However, in a prospective study involving 1,240 patients with RA [33], therapy with methotrexate was associated with improved survival by reducing CVD mortality by 70%. Folate supplementation is a plausible approach to counteract hyperhomocysteinaemia [34] and is currently a matter of debate. It is worth noting that in a trial of 5,442 women conducted over 7 years [35], daily supplementation with a combination of folate and vitamins B 6 and B 12 was able to reduce hyperhomocysteinaemia, but it was unable to abolish CVD risk.
Nonsteroidal anti-inflammatory drugs are among the most commonly used anti-inflammatory interventions. This group of drugs has a much broader range of actions than simply inhibition of inflammation, including their effect on prostaglandin metabolism, in which atheroprotective mediators (for example, prostacyclin) are also inhibited. In a recent metaanalysis [36] selective cyclo-oxygenase-2 inhibitors and highdose regimens of ibuprofen and diclofenac were found to be associated with a moderate increase in the risk for adverse cardiovascular events.

Inflammatory burden and atherosclerosis
Both SLE and RA patients exhibit an increased prevalence of pro-atherogenic risk factors, which may very well contribute to the burden of CVD events and atherosclerotic disease. However, there is growing evidence that this is only one side of the story.
In a number of studies, even after correction for classical risk factors for CVD, both RA and SLE remain independently associated with clinical and subclinical atherosclerosis [9,18] (Tables 1 and 2). In a cohort of 296 SLE patients followed for a mean of 8.6 years, CVD events were substantially higher than would be predicted on the basis of traditional factors alone, using the Framingham risk stratification method [12]. In a study conducted in the Pima Indian population, the number of swollen joints was predictive of death from CVD [37]. Similarly, the presence of severe extra-articular disease clearly increased the risk for cardiovascular events in patients with RA [2,38]. Duration of the disease has also been associated with the presence of carotid plaques [39]. A recent study conducted in 631 RA patients [40] showed that the rate at which IMT increased was proportional to the duration of disease. Another study [41], in 155 RA patients, also demonstrated that patients with longstanding disease have an increased Framingham risk score compared with patients with early disease. (The Framingham risk score is used to derive an estimated risk for developing coronary heart disease within 10 years and is based on age, sex, total and HDL cholesterol, blood pressure, diabetes and smoking.) High Framingham scores were associated with coronary artery calcification. The increased risk for CVD in RA patients was also shown to be present 2 years before the fulfilment of criteria for RA [7]. In addition, patients with very early inflammatory polyarthritis are at increased risk for premature death from CVD when their serum is positive for rheumatoid factor [42]. These findings suggest that even preclinical inflammation may contribute to increased risk for cardiovascular events.

Disease initiation and endothelial activation
In humans, atherosclerotic lesions occur in the large and medium sized arteries and are characterized by an asymmetrical focal thickening of the innermost layer of the arterial wall, the intima. Atherosclerosis is a multifactorial disease with strong epidemiological associations with a variety of lifestyle-related and genetics-related risk factors (for example, dyslipidaemia, smoking, diabetes mellitus and hypertension). The common pathway, through which all of these risk factors lead to atherosclerosis development, is endothelial activation and subsequent endothelial dysfunction. The sequence of events that leads to early stages of atherogenesis has been mostly characterized under conditions of hypercholesterolaemia. Excess LDLs enter the intima where they are retained by matrix proteoglycans. Here they are subject to oxidative and enzymatic modifications, leading to formation of pro-inflammatory phospholipids that may induce endothelial cell activation [43].
Pro-inflammatory mediators such as tumour necrosis factor (TNF)-α, are strongly linked to endothelial activation and dysfunction [44]. Persistent endothelial activation leads to endothelial dysfunction, characterized by reduced production of nitric oxide and increased expression of adhesion molecules and chemokines [1], resulting in impaired vasorelaxation and recruitment of blood-borne mononuclear cells such as monocytes and T lymphocytes. Endothelial dysfunction has been detected in RA and SLE patients by noninvasive tests such as the brachial flow-mediated dilatation, as extensively reviewed elsewhere [45,46]. Endothelial dysfunction might reflect CVD risk factor burden in these individuals, but it could also indicate systemic inflammation and prolonged exposure of endothelial cells to pro-inflammatory cytokines such as TNF-α (vide infra). Interestingly, in animal models as well as in human clinical trials in RA and SLE patients, high doses of statins can reduce inflammation and improve endothelial function [47]. Statins are effective in lowering LDL-cholesterol, but they also have immunomodulatory properties [48]. Both mechanisms of action might be at play in reducing endothelial dysfunction.

Progression of atherosclerotic disease
Although atherosclerosis is initiated by the exposure to traditional lifestyle risk factors, inflammation is a key driving force at each stage of atherosclerosis development [43], with both innate and adaptive cellular components of inflammation involved in disease progression ( Figure 2).

Myeloid cells, cytokines and atherosclerosis progression
Recruitment of blood-borne mononuclear cells is a key step in atherogenesis. Even in hypercholesterolaemic murine models  of atherosclerosis, genetic deletion or blockade of chemokines (such as C-C chemokine ligands 2 and 5, and fractalkine and its cognate receptors) and adhesion molecules (such as vascular cell adhesion molecule-1) abrogates monocyte recruitment and atherosclerosis in mice [43]. Blood-borne monocytes recruited through the endothelium undergo differentiation into macrophages induced by macrophage colony-stimulating factor. During this process scavenger receptors are upregulated, mediating the uptake of oxLDL, causing the formation of foam cells, which are the hallmark of atherosclerotic lesions. Macrophages in atherosclerotic plaques exhibit features of activation and release a host of pro-inflammatory mediators including TNF-α, which leads to the engagement of the pro-inflammatory cytokine cascade, resulting in IL-1 and IL-6 production. IL-6 in turn stimulates production of large amounts of acute-phase reactants, including C-reactive protein, serum amyloid A and fibrinogen in the liver.
Cytokine upregulation is a general feature of inflammatory diseases and could be a mechanism that links increased atheroma formation in patients with RA and SLE. Several cytokines with relevance to CVD could contribute to the pathogenesis of atherosclerosis in RA and SLE.
TNF-α plays a dominant role in RA. TNF-α induces a proinflammatory cascade, with upregulation of other mediators, namely IL-1, IL-6 and IL-8, prostaglandins and matrix metalloproteinases (MMPs). Upregulation of adhesion molecules by TNF-α on the vascular endothelium leads to cellular recruitment of neutrophils and monocytes. It also stimulates the proliferation of fibroblasts [49]. Introduction of the anti-TNF  and reduces carotid IMT [55] and MI [56].
Recently, macrophage migration inhibitory factor (MIF) has emerged as a potential link between RA, SLE and atherosclerosis development (for review, see [57]). Originally, described as 'a soluble factor released from T cells in delayed hypersensitivity responses inhibiting migration of macrophages', MIF is considered a pleiotropic cytokine with roles in several inflammatory diseases. MIF induces the pro-inflammatory mediators TNF-α, IL-1, IL-6 and MMPs. It can activate T cells, promote angiogenesis and induce proliferation of cells, while inhibiting p53 expression and apoptosis of the same cells [57,58]. MIF can be induced by oxLDL, which is an initiating factor in atherogenesis, and so expression of MIF early on may enhance pro-inflammatory responses and lesion progression [58]. MIF immunoneutralization or genetic deletion reduces disease activity in animal models of arthritis and lupus. In mice with advanced atherosclerosis, MIF blockade led to plaque regression and reduced monocyte and T-cell content in the plaques. Interestingly, glucocorticoids can induce the expression of MIF in multiple cell types [57], which is a function of potential relevance to the development of CVD in those patients receiving chronic treatment. Increased serum levels of MIF have been detected in SLE patients compared with healthy control individuals [57], and correlations were reported between MIF serum levels and SLE-related damage scores and glucocorticoid exposure.
The type I interferon cytokines (interferon [IFN]-α and IFN-β) play a causative role in SLE. Type I IFN-α is produced by dendritic cells in response to viral or bacterial infections, sensed through Toll-like receptor (TLR)-9. Auto-antigens such as self-DNA, exposed by apoptotic cells, may also trigger TLR-9 signalling, resulting in IFN-α production [59]. IFNs play many roles including inhibition of cell proliferation, interfering with viral infections, proliferation of memory T cells and increased antigen presentation activity by dendritic cells [60]. Type I IFNs were recently linked to impairment of vascular repair by circulating endothelial progenitor cells. SLE patients have decreased numbers of endothelial progenitor cells, with increased endothelial cell apoptosis, and decreased pro-angiogenic and repair capacity [59]. IFN-α has anti-proliferative effects on endothelial progenitor cells in vitro. SLE patients have elevated levels of IFN-α, which correlates with disease severity [60] and impaired endothelial function [61]. Interestingly, IFN-α produced by plasmacytoid dendritic cells has been detected in human carotid plaques, where it could sensitize antigen-presenting cells to the TLR-4 ligand lipopolysaccharide, thereby promoting the production of pro-inflammatory molecules [62]. By contributing to endothelial dysfunction/damage and inducing pro-inflammatory responses within the atherosclerotic plaque, IFN-α could promote atherosclerosis in patients with inflammatory diseases.

Lymphocytes and immune dysregulation
T cells infiltrate atherosclerotic lesions, even in early stages of disease. Such infiltrates are predominantly CD45RO-expressing memory CD4 + T cells, most of which bear the T-cell receptor αβ. They appear to be in close contact with major histocompatibility complex (MHC) class II expressing macrophages and dendritic cells. Both human and animal studies have thus far shown a predominantly T-helper (Th)1 pattern in atherosclerosis, with immunopositivity for the classic Th1 cytokine IFN-γ and Th1-stimulating cytokines IL-12 and IL-18. In apoE -/mice lacking IFN-γ or its receptor, the development of atherosclerosis is inhibited [43].
During the past decade, an unusual subpopulation of CD4 + T cells lacking the co-stimulatory receptor CD28 have been suggested to link atherosclerotic syndromes with RA. Such CD4 + T cells exhibit a spectrum of regulatory receptors that are usually only seen on natural killer cells. Such regulatory receptors include perforin, members of the killer immunoglobulin-like receptor family, and the receptor for the chemokine fractalkine. In RA, circulating CD4 + CD28 null cells are increased and correlate with preclinical atherosclerotic disease and endothelial dysfunction. Such cells are also increased in the circulation and in plaques of patients with acute coronary syndrome (ACS) [63] and possess endothelial cell cytotoxic activity [64].
An additional mechanism that could also explain the increased incidence of CVD in patients with inflammatory diseases is the failure of regulatory feedback mechanisms, such as production of anti-inflammatory cytokines IL-10 and transforming growth factor (TGF)-β by regulatory T cells. T-cell-specific genetic deficiency in the TGF-β receptor leads to plaque destabilization, whereas regulatory T cells have the ability to reduce atherosclerotic burden in hyperlipidaemic mice [43]. In RA CD4 + CD25 + regulatory T cells have a defect in their suppressive functions, which was restored during treatment with TNF blockers [65].
The decreased clearance of apoptotic bodies has also been evoked to explain increased inflammation in lupus-like models of disease. The gld and lpr mouse strains have inactivating mutations in Fas ligand and Fas, respectively, and they suffer from lupus-like disease. (Fas ligand and Fas are transmembrane proteins on the cell surface; binding of Fas ligand to its receptor Fas induces apoptosis.) Although these mice typically do not exhibit signs of atherosclerosis, they are susceptible to atherosclerotic lesion formation when they are fed an atherogenic diet [66]. ApoE -/mice crossed with gld mice developed lesions containing a twofold to threefold increase in macrophages and CD3 + lymphocytes as compared with apoE -/mice [67]. A similar increase in lymphocytes and macrophage accumulation within the lesions were observed after bone marrow transfer from lupusprone strains developed on a C57BL/6 background to LDL receptor-deficient mice [68,69], even in the presence of lower levels of cholesterol and triglycerides. Furthermore, an increase in the expression of the MHC class II molecule I-Ab in mice with lupus-like disease compared with control animals was observed, and there were significantly higher antibody titres to oxLDL and cardiolipin [67][68][69].

Thrombotic complications
ACS such as MI occur when thrombosis ensues on an atherosclerotic plaque, suddenly limiting blood flow through the coronary artery. Acute thrombosis is often triggered by sudden plaque rupture or erosion [43], but heterogeneity of pathological substrates for thrombosis has been documented [70]. MI is often caused by plaques that only moderately restrict the vessel lumen (by <50% of the lumen diameter) [71].

Plaque vulnerability
The chance of plaque rupture is determined by the strength and thickness of the fibrous cap, surrounding the inflammatory and necrotic 'core' of the plaque. The cap, made up of smooth muscle cells (SMCs) either recruited from the media or derived from progenitors, forms a protective barrier encasing the core, preventing the prothrombotic materialtissue factor and phospholipids -from contacting the arterial blood flow [43]. 'High-risk' atherosclerotic plaques are characterized by their composition rather than by their impact on the vessel lumen. A high-risk plaque is characterized by a large lipidic/necrotic core with increased number of activated monocytes and lymphocytes. The fibrous cap is also thinner due to SMC apoptosis and proteolytic degradation of the fibrous proteins. On the contrary, a 'stable' plaque is characterized by its thick fibrous cap, containing increased amounts of fibrous proteins and SMCs, and generally a smaller proinflammatory/prothrombotic core; such plaques are less prone to rupture.
There is evidence that plaques obtained from patients who have experienced an ACS have a higher content of macrophages and activated T lymphocytes at sites of coronary thrombosis [72]. Also, they are often found in clusters within the fibrous cap, in particular in the 'shoulder' area of the lesion, the border between normal vessel and atherosclerotic plaque, where plaque rupture is thought to occur most often. A retrospective autopsy study comparing 41 patients with RA with sex-, age-and autopsy date-matched control individuals did not identify significant differences between groups in severity of the coronary stenoses in cross-sectional area or the number of vessels affected. However, RA patients with clinical history of CVD were more likely to have coronary lesions with a 'high-risk' rather than a 'stable' plaque phenotype [73]. This finding is in agreement with a long 'silent' history of coronary artery disease, with increased incidence of MI and sudden death [5,6].
Plaque stability can be influenced by cytokines via a variety of mechanisms. First, cytokines can affect the amount of fibrous proteins produced by SMCs. IFN-γ inhibits collagen production, whereas TGF-β has the opposite effect. Second, cytokines such as TNF-α and IL-1 can induce expression of MMPs in SMCs and macrophages, which degrade the fibrous cap. MIF can induce MMP expression and may also contribute to plaque instability by increasing monocyte recruitment and retention in the plaque. Third, apoptosis of SMCs can be induced by IFN-γ and TNF-α. Finally, cytokines, in particular TNF-α, can induce expression of tissue factor, which is the major initiator of the thrombotic cascade and whose increased expression is a feature of plaques in patients with ACS. Therefore, pro-inflammatory cytokines can have both a localized effect on the plaque architecture and determine occurrence of thrombotic events, and a more systemic effect, in which they affect cholesterol levels, insulin resistance and endothelial function.  [75,76], and studies in larger sample sizes are awaited to clarify the role of this SNP in the aetiology of RA [75].

Genetic associations
Interestingly, polymorphisms in the region of the TNFAIP3 gene were recently linked to RA and SLE [77,78]. TNFAIP3 encodes the de-ubiquitinating enzyme A20, an endogenous inhibitor of the nuclear factor-κB (NF-κB) pathway. NF-κB is a transcription factor that is activated by TNF or IL-1/TLR signalling pathways, which induces transcription of proinflammatory genes. In atherosclerosis, NF-κB is activated at sites of the arterial wall that are prone to lesion development.
In animal studies, ApoE -/-A20 -/mice had increased atherosclerosis, whereas mice with A20 over-expression had decreased atherosclerosis [79]. SNPs in the TNFAIP3 gene region may cause reduced expression or reduced activity of A20 [77], therefore contributing to an uncontrolled inflammatory response and autoimmunity and potentially accelerated atherosclerosis in these patients.

Antiphospholipid antibodies
Antiphospholipid antibodies are the hallmark of the antiphospholipid syndrome, which is characterized by recurrent thrombosis and is associated with SLE. Between 30% and 50% of patients with SLE have antiphospholipid antibodies, and approximately one-third develop antiphospholipid syndrome [23]. Many of these antibodies, such as anticardiolipin (aCL), recognize phospholipid-binding plasma proteins such as β 2 -glycoprotein I (GPI), which is anticoagulant. Nevertheless, there is some new evidence to suggest that they may also play a role in the development of atherosclerotic disease. Antibodies to β 2 -GPI may lead to endothelial activation and/or damage, as well as enhance oxLDL uptake by macrophages, contributing to recruitment of immune cells to the vascular wall. Furthermore, immunization with β 2 -GPI leads to vascular lesions in LDL receptor -/mice, suggesting that immune recognition might also play a role [80].
However, in SLE cohorts no correlation was found between aCL antibodies and indices of atherosclerosis extent and severity such as carotid IMT [81,82]. However, association was found with thrombotic complications such as MI. This paradox can potentially be explained by the ability of aCL to cause thrombosis even in the absence of underlying advanced atherosclerotic plaque [81]. Some patients with RA may have elevated levels of aCL [83]. In contrast with SLE patients, however, RA patients exhibited an increased carotid IMT compared with RA patients without elevated auto-antibodies [13]. However, in one study, only 17 out of 83 patients were found to have increased antibodies to aCL, and larger cohorts are needed to address this observation.

Conclusions
An increased prevalence of pro-atherogenic risk factors in patients with RA and SLE compared with age-matched control individuals may very well contribute to the initiating stages of disease. However, the dysregulated inflammation that characterizes inflammatory diseases may increase susceptibility to traditional risk factors and the likelihood that a chronic inflammatory response will ensue at the arterial sites. Potentially, shared pathogenic pathways are also at play. The dissection of the molecular mechanisms underlying atherosclerosis and classic inflammatory diseases will no doubt help us to understand how inflammation causes CVD in both the general population and in the subgroups of patients with inflammatory diseases.