Systemic lupus erythematosus (SLE) is an autoimmune disease characterized by systemic inflammation due to pathogenic autoantibodies and pro-inflammatory immune signaling (1). SLE is chronic and may lead to significant deterioration in quality of life by way of many possible complications, the most ominous of which is end-stage organ failure, involving particularly the kidneys (2, 3). Many studies to date have focused on the renal manifestations of SLE, including lupus nephritis and end-stage renal disease, but there still remains much to be elucidated about the link between SLE and cardiovascular events (4). Cardiovascular disease (CVD) is the leading cause of death in the United States and Chung et. al postulate that up to 45% of SLE patients will have cardiovascular complications throughout their lifetimes, owing to systemic inflammation and heightened incidence of metabolic syndrome as compared to the general population (5–7). The incidence of emergent cardiovascular events may be up to twice as high in SLE patients when compared to those without SLE (8, 9). One longitudinal study of female patients found that those with SLE had a 4 fold increase in major adverse cardiovascular events when compared to healthy controls (10). Many patients are unaware that they have manifestations of CVD until they suffer an acute cardiovascular event, and these events are rapidly increasing in frequency such that they may overtake renal lupus as the most fatal complication associated with SLE (11). The profundity of the risk of cardiovascular disease in lupus patients is well-established in literature (8, 9, 12). Several studies have worked with cohorts of SLE patients to assess the longitudinal impact of lupus upon cardiac health, and have, in agreement with other studies, found that this population is much more at risk for developing cardiovascular complications than the general population. It is also well established that SLE may accelerate the formation of atherosclerotic plaques by promoting hypertension and hyperlipidemia (3, 13–15). While this relationship is clearly established in the literature, the mechanistic basis for this link at a molecular level is unclear. The objective of this study is to identify molecules whose baseline serum levels are significantly different in those SLE patients who will go on to develop atherosclerotic plaques vs. those who do not.

Several omics-based screens have been performed previously in order to explore the metabolic implications of systemic lupus erythematosus, including Wu et al., Perl et al., and Ouyang et al., but these have not examined cardiovascular disease (16–18). Omics-based screens have also been performed to identify biomarkers of cardiovascular disease in the general population, and suggest that a decrease in tricarboxylic acid cycle metabolites may be indicative of ischemic heart disease (19). Of particular interest, Leptin, pro-inflammatory high-density lipoprotein, and TNF related weak inducer of apoptosis (TWEAK) have been reported as potential biomarkers of atherosclerosis in SLE in previous studies (10, 20–22).

Information on the precise ways that SLE leads to cardiovascular disease may lead to the development of predictive tools to assess patients' risk of CVD early on when appreciable modifications in cardiovascular risk factors can still be instituted. In the present study, we have used a metabolomic screen to analyze baseline sera of patients with SLE to identify metabolites that are significantly different among patients who experience carotid plaque progression during follow-up compared to those who do not. Samples were analyzed using a metabolomic screen which used liquid chromatography-mass spectrometry (LC/MS) to detect levels of more than 3,000 metabolites. This analysis has illuminated the metabolic processes underlying pathogenesis of SLE complications, and has provided new insights on potential biomarkers for CVD in SLE. These findings may enable physicians to better identify SLE patients at risk for plaque progression, discuss and institute preventative measures and evaluate patients for long term therapy with agents such as statins and cholesterol lowering medications.

Literature has suggested for some time that lupus patients are more likely to suffer from cardiovascular complications than their healthy counterparts. This phenomenon was first studied longitudinally by Urowitz et. al., where it was found that atherosclerotic plaques and greater incidence of myocardial infarction were associated with later age of death in patients with active SLE (29). A study by Manzi et al. found that female SLE patients experienced earlier and more severe onset of cardiovascular disease than their healthy counterparts and were 50 times more likely to suffer a myocardial infarction (14). While many patients with SLE are now able to manage their symptoms in the short term, it is necessary to unravel mechanisms behind long-term implications of lupus such as renal and cardiovascular disease and begin searching for new approaches for early detection of such complications. To this end, it has been suggested that oxidized low-density lipoproteins, which have long been implicated in atherogenesis in the general population, may be used to gauge risk of atherosclerosis in patients with autoimmune disease (20). Our comprehensive metabolic screen aims to build upon the findings of such studies.

The present study represents the very first comprehensive OMICs study in SLE patients with plaque progression during longitudinal follow up. It is also the first attempt to comprehensively identify prognostic metabolites in SLE^PP. Whereas some metabolites were significantly reduced in SLE patients, others were elevated, at baseline, before the emergence of atherosclerotic plaques. Most notable among the downregulated metabolites was the decrease in multiple alpha keto acids in SLE^PP sera, at baseline (Figure 4B). Alpha keto acid derivative byproducts of amino acid breakdown may originate from the liver or may be produced in the mitochondria during the tricarboxylic acid cycle. Both glucogenic and ketogenic amino acids can give rise to the derivatives which were decreased in SLE^PP patients (Figure 4A). This is in line with previous metabolomic findings which aimed to illuminate metabolic imbalances contributing to SLE in general (16).

In the present screen, alpha ketoglutarate and other alpha keto acids were found to be reduced in SLE patients compared to healthy subjects, pointing to potential upregulation of gluconeogenesis to feed an overactive Citric Acid Cycle in SLE (see Figure 5A), which may also explain the altered amino-acid and protein metabolism in SLE by Reactome analysis (Figure 1A), and the reported reductions in circulating amino acids in SLE (16). Increased consumption of alpha keto acids is also evidenced by the increased production of the downstream metabolite, isovalerate (Figures 5A,B). Isovalerate enhances chemotaxis and phagocytosis of granulocytes and monocytes (30, 31). Interestingly, isovalerate has previously been studied as a novel biomarker for diabetic nephropathy and endometrial cancer (32, 33). While the concentrations of alpha keto acids were found to be reduced in SLE serum, these molecules were further reduced in SLE^PP.

Alpha-keto-acids promote differentiation of naïve CD4+ T cells to proinflammatory TH1 cells instead of immunosuppressive Treg cells (34). BCAA, which give way to alpha-keto-acids, are essential for lymphocyte function and altered BCAA metabolites cause impaired T cell activation (35, 36). The alpha ketoacids which were most significantly decreased are synthesized via reactions catalyzed by branched chain amino acid aminotransferase (BCAT) and are further catabolized via an irreversible attachment to coenzyme A which is catalyzed by branched chain alpha-keto acid dehydrogenase (BCKD). It may be postulated that the change in concentration of alpha keto acids in these groups may be due to an increase in BCKD or a decrease in BCAT activities, respectively. Indeed, evidence for such changes have been demonstrated in heart failure (37). Others have also postulated that a change in branched-chain amino acid catabolism is associated with cardiovascular disease, but it is unclear as to whether this is a cause or effect of the disease (38). Clearly, these postulates need to be experimentally verified in lupus as well.

In contrast, metabolites in the n6-PUFA/arachidonic acid lipoxygenase pathway were largely found to be elevated in SLE^PP sera in our metabolic screen. This is also in line with the literature, where the arachidonic acid pathway is a key player in the development of cardiovascular disease, and pro-inflammatory metabolites of n6-PUFA have been implicated in cardiovascular disease (39–41). Most notable is the increase in LTB4 concentration in SLE^PP sera, resonating well with previous reports in the cardiovascular disease literature (42, 43). Indeed, the role of LTB4 and related leukotrienes in inflammation and cardiovascular disease has been extensively documented (43–46). LTB4 causes activation of neutrophils, chemotaxis and adhesion of leukocytes to vascular endothelium and stimulates release of ROS and dsDNA from neutrophils (47–49). LTB4 also enhances the phagocytic activity of neutrophils and T-cell migration (50, 51). Together, these immune events may promote atheroma formation. In conjunction with our findings, this suggests that LTB4 may not only be prognostic of CVD in SLE, it is also likely to be an active driver of inflammation leading to atheroma formation (42, 47, 52, 53). Other metabolites in this pathway including 9-HODE and 13-HODE are also proinflammatory and have been implicated in atherogenesis (52–55). Wang et al. reported that the serum levels of 19 metabolites in conjunction with the use of traditional CVD risk factors was able to better predict which patients would go on to develop CVD than the use of risk factors alone (56). Their selected metabolites included 13-HODE and 9-HODE, which we also found significantly elevated in SLE patients who exhibited plaque progression.

Indeed, we had previously reported that LTB4, 5-HETE, 9-HODE and 13-HODE, all of which are metabolites in the same metabolic pathway (Figure 3B) are significantly elevated in SLE patients, associated with higher disease activity (16). 5-HETE is chemotactic for neutrophils and causes aggregation and degranulation of neutrophils (57–60). 12-HETE is involved in the recruitment and migration of neutrophils (61). High levels of 9/13 HODE have been shown to correlate with high SLE disease activity (16). 9/13-HODE are ligands of PPARg that enhances monocyte maturation and maintain macrophages in chronic inflammatory state (62). One can certainly envision this to promote atheroma formation. In the present study, we find that these same metabolites are highly upregulated only in SLE patients who later exhibited plaque progression, but not in patients who do not (Figure 3A). Taking together our previous metabolomic findings (16), the present metabolomic findings and the rich body of literature in CVD, it is clear that baseline levels of LTB4, 5-HETE, 9-HODE and 13-HODE are strong predictors and pathogenic drivers of future plaque progression and CVD. These findings also suggest that pharmacological inhibitors of this pathway, including LTB4 inhibitors, warrant evaluation in lupus patients bearing these serum biomarkers.

Plasmalogens emerge as another class of metabolites that may hold biomarker potential in SLE^PP. In particular, the ethanolamine plasmalogens 1-(1-enyl-palmitoyl)-2-linoleoyl-GPE, 1-(1-enyl-stearoyl)-2-oleoyl-GPE, 1-stearoyl-2-arachidonoyl-GPE, and 1-(1-enyl-stearoyl)-2-arachidonoyl-GPE as well as the choline plasmalogen 1-palmitoyl-2-arachidonoyl-GPC were reduced in SLE^PP and SLE sera when compared to controls. Many of these molecules have been implicated in early colorectal and breast cancers and are being evaluated for their potential to detect these malignancies early (63, 64). Reduced plasmalogen levels impair phagocytosis by macrophages (65). Plasmalogen levels are reduced in SLE patients and are correlated with increased oxidative stress (66). Nguma et al., have suggested that ethanolamine plasmalogens may even serve a protective role against oxidative stress and have studied the utility of these molecules in the treatment of colitis in murine models (67). Levels of several choline plasmalogens were found by Nishimukai et al. to have a strong inverse correlation to progression of atherosclerosis and have even been suggested as possible biomarkers of CVD (68). Plasmalogens have also been proposed as potential biomarkers of SLE, as they have been found to be reduced in these patients (69). While our findings are consistent with the earlier reports, within our cohort these molecules are even further reduced in lupus patients who go on to develop atherosclerotic plaques. Further study will be necessary to assess the utility of plasmalogens as biomarkers for both SLE and SLE^PP. Two plasmalogens, 1-(1-enyl-palmitoyl)-GPE and 1-(1-enyl-oleoyl)-GPE, were elevated in the SLE^PP group. In a 2018 study of patients with metabolic syndrome, Candi et al. found the same two ethanolamine plasmalogens to be elevated in the diseased population when compared to healthy controls (70). They had suggested that production of certain plasmalogens is elevated in response to the disease process, in order to combat inflammation (70, 71). Our findings in the present study are possibly explained by a similar mechanism, though experimental verification is warranted.

Several aspects of this study could be improved. As we have discussed previously, it would be important to control for the impact of medications, diet, genetics, and the gut microbiome in future evaluations of cardiovascular disease in SLE (71). In addition, the sample sizes of all groups studied need to be significantly increased. The patients used for this study tended to be older, and as such our findings may not be applicable to younger patients, though this needs to be experimentally verified. The clinical utility of these biomarkers as prognostic indicators may be limited to SLE patients with mild disease, as the mean SLEDAI score for both the SLE and SLE^PP groups was 2.0. Moreover, future studies must use larger cohorts to ascertain whether the reported results were confounded by variables such as tobacco use, which was most prevalent in the SLE group, or hypertension, which was present in the SLE and “HC” groups but not the SLE^PP group.

Taken together with the large body of literature in the cardiovascular field that describe the same metabolites detailed here, the current observations offer resounding support for using baseline LTB4, 5-HETE, 9/13-HODE as well as altered alpha keto-acids and plasmalogens as prognostic biomarkers of long-term atherosclerosis and cardiovascular disease in patients with lupus, and for carefully evaluating the need for early institution of preventive measures in these high-risk subjects. Future studies are warranted to validate these results in larger patient cohorts.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by IRB boards of UCLA, Los Angeles, CA and University of Houston, Houston, TX. The patients/participants provided their written informed consent to participate in this study.

MM and CM designed the study. MM provided patient samples. SB, KV, and HD performed the experiments. SB, KV, HD, AT, MM, and CM analyzed the data and wrote the manuscript. All authors have read the manuscript and concur with the conclusions.

This work was supported by NIH R01 AR074096.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.