Patients with systemic lupus erythematosus exhibit accelerated atherosclerosis, a chronic inflammatory disease of the arterial wall. The impact of B cells in atherosclerosis is controversial, with both protective and pathogenic roles described. For example, natural IgM binding conserved oxidized lipid epitopes protect against atherosclerosis, whereas anti-oxidized low-density lipoprotein (oxLDL) IgG likely promotes disease. Because BAFF promotes B cell class-switch recombination and humoral autoimmunity, we hypothesized that excess BAFF would accelerate atherosclerosis. In contrast, BAFF overexpression markedly reduced hypercholesterolemia and atherosclerosis in hyperlipidemic mice. BAFF-mediated atheroprotection required B cells and was associated with increased protective anti-oxLDL IgM. Surprisingly, high–titer anti-oxLDL IgM production and reduced atherosclerosis was dependent on the BAFF family receptor transmembrane activator and CAML interactor. In summary, we identified a novel role for B cell–specific, BAFF-dependent transmembrane activator and CAML interactor signals in atherosclerosis pathogenesis, of particular relevance to the use of BAFF-targeted therapies in systemic lupus erythematosus.
Atherosclerosis is a chronic inflammatory disease of the vascular intima modulated by both the innate and adaptive arms of the immune system (1). Patients with autoimmune disease, particularly systemic lupus erythematosus (SLE) and rheumatoid arthritis, have a markedly increased risk of atherosclerotic cardiovascular disease, suggesting additional immunologic factors impact atherogenesis in the setting of autoimmunity (2). In SLE, elevated cardiovascular risk cannot be fully explained by Framingham risk factors, accumulated organ damage or exposure to immunosuppressive therapy (3, 4), and a large randomized trial of statins in SLE failed to identify any cardioprotective effects (4, 5). These data emphasize the urgent need for an improved understanding of the underlying immune mechanisms affecting atherosclerosis with a goal of targeted therapies able to modulate this life-threatening disease process.
Although B cells clearly impact the pathogenesis of systemic autoimmunity, their role in atherosclerosis has been controversial, with both atheroprotective and atherogenic effects described. For example, initial studies using splenectomized apolipoprotein E-null (Apoe−/−) and B cell-deficient (μMT) low-density lipoprotein receptor-null (Ldlr−/−) mice demonstrated accelerated atherosclerosis in the absence of B cells (6, 7). These atheroprotective roles for B cells have been attributed to the production of protective natural IgM Ab binding phosphorylcholine (PC) on Streptococcus pneumoniae surface capsule, apoptotic cells, and oxidized low-density lipoprotein (oxLDL) (8). Supporting this idea, Ldlr−/− mice unable to secrete IgM (sIgM−/−) develop accelerated atherosclerosis (9), whereas both passive immunization with anti-PC IgM and active S. pneumoniae immunization attenuates atherosclerosis (10, 11). In addition, adoptive transfer of wild type, but not sIgM−/−, peritoneal B1a B cells reduces atherosclerosis in splenectomized Apoe−/− mice (12), consistent with B1 B cells being the predominant source of natural IgM (13).
In contrast to these protective functions, B cells can actively promote atherosclerosis, particularly in autoimmune settings. Deletion of FcγRIIb (an inhibitory Fc-receptor expressed on B cells and other cell types) promotes the development of class-switched IgG2c Ab to oxLDL and leads to enhanced atherogenesis in Ldlr−/− and Apoe−/− mice (14, 15). Further, B cell depletion with anti-CD20 Ab decreases the development of atherosclerosis in both Ldlr−/− and ApoE−/− murine models of atherosclerosis (16, 17). Protection in this setting correlates with a decline in anti-oxLDL IgG and coordinate preservation of anti-oxLDL IgM titers (17), consistent with the resistance of natural Ab–producing B1 B cells to anti-CD20 depletion (18). In addition to these animal studies, consistent findings are noted in patients with SLE, where those with the lowest titers of anti-PC IgM have the highest risk of atherosclerosis (19, 20), whereas lupus patients with a history of cardiovascular disease have elevated IgG anti-oxLDL Ab titers compared with other SLE patients and population controls (21).
Together, these human and animal studies support the paradigm whereby preformed natural IgM Ab protect against the development of atherosclerosis, whereas class-switched autoantibodies against related oxidized lipid epitopes promote atherogenesis. However, the specific B cell-intrinsic signals required for class-switch recombination to pathogenic Ab isotypes in atherosclerosis have not been fully addressed.
BAFF is a crucial B cell survival cytokine that has been closely linked to lupus pathogenesis. In addition to the development of lupus-like disease in BAFF transgenic (BAFF-Tg) mice (22, 23), BAFF levels are elevated in a subset of SLE patients (24) and a BAFF-inhibiting therapeutic Ab, belimumab, demonstrated clinical efficacy in human lupus (25). Because atherosclerotic cardiovascular disease is accelerated in humans with autoimmune disease, we hypothesized that the overexpression of BAFF in hyperlipidemic mice would promote atherogenesis by promoting pathogenic class-switched IgG anti-oxLDL Ab. In contrast to this idea, BAFF overexpression markedly reduced hypercholesterolemia and atherosclerosis in Apoe−/− mice. Despite BAFF family receptor expression on non-B cell lineages, these protective effects of BAFF required the presence of peripheral B cells. Further, BAFF overexpression correlated with a marked increase in atheroprotective anti-oxLDL IgM Ab titers, which was dependent on the B cell surface receptor transmembrane activator and CAML interactor (TACI). Together, our findings uncover a novel role for BAFF-dependent TACI activation in promoting protective B cell functions in atherosclerosis, of relevance to the pathogenesis of atherosclerotic cardiovascular disease in patients with and without systemic autoimmunity.
Materials and Methods
Apoe−/− (26), BAFF-Tg (23), μMT (27), and Taci−/− (28) mice on C57BL/6 background and relevant murine crosses were bred and maintained in the specific pathogen-free animal facility of Seattle Children’s Research Institute (Seattle, WA). To control for potential genetic differences between strains, all studies were performed on littermate controls. In addition, mean percentage C57BL/6J genetic background based on single nucleotide polymorphism analysis of representative experimental animals was 97.2 and 98.4% in Apoe−/− and Apoe−/−.BAFF-Tg mice, respectively (https://www.jax.org/jax-mice-and-services/breeding-and-rederivation-services/genome-scanning). All animal studies were conducted in accordance with Seattle Children’s Research Institute Institutional Animal Care and Use Committee approved protocols.
Induction and quantification of atherosclerosis
Female 6 wk–old mice of indicated genotypes were placed on high-fat, high-cholesterol Western diet (WD) (21% fat, 0.2% cholesterol; Harlan Teklad, TD88137) for 8 or 12 wk. Serum was obtained by cardiac puncture from animals fasted for 6 h prior to sacrifice. The heart and proximal aorta were fixed with 10% formalin, and frozen at −80°C in OCT compound. Serial frozen sections were cut from the aortic sinus, counterstained with Oil Red O, and average lesion area quantified with ImagePro-Plus software (Media Cybernetics), as described (29).
Serum cholesterol analysis
Fasted serum cholesterol was assayed using colorimetric assay kits (1010-430; Stanbio Laboratory). Plasma lipoprotein profiles were analyzed by fast protein liquid chromatography (FPLC) using sera pooled from five to six individual mice, as described (29).
In vivo BAFF inhibition
The p values were calculated using the two-tailed Student t test, and the one-way ANOVA followed by Tukey’s multiple comparison test (GraphPad Software).
Results and Discussion
Excess BAFF limits hyperlipidemia and atherosclerosis in Apoe−/− mice fed a high-fat, high-cholesterol diet
To test the impact of elevated BAFF levels on murine atherosclerosis, we crossed Apoe−/− with transgenic mice overexpressing BAFF in myeloid cells using the human CD68 promoter (BAFF-Tg) (23). Independent cohorts of 6 wk-old Apoe−/− and Apoe−/−.BAFF-Tg female littermate mice were placed on a high-fat, high-cholesterol WD for 8 wk to rapidly induce hyperlipidemia and atherosclerosis. As predicted, WD-fed Apoe−/− mice developed IgM Ab against PC and the oxidized lipid MDA-LDL (8), with titers of these IgM Ab markedly increased by BAFF overexpression (Fig. 1A, 1C). In addition, Apoe−/−.BAFF-Tg mice developed elevated class-switched IgG Ab against PC and MDA-LDL, with the proinflammatory IgG subclasses IgG2b and IgG2c exhibiting the greatest increase (Fig. 1B, 1C). Because IgG2c Ab bind proinflammatory Fc-receptors on myeloid cells (32), we predicted that increased titers of anti-oxLDL IgG2c Ab would promote atherogenesis in WD-fed Apoe−/− mice. Surprisingly, BAFF overexpression resulted in ∼50% reduction in total serum cholesterol levels after 8 wk on WD without affecting overall weight gain (Fig. 1D, 1E). Analysis of FPLC-separated lipoprotein fractions demonstrated a prominent increase in very low density lipoprotein (VLDL) in WD-fed Apoe−/− mice. Notably, BAFF overexpression resulted in a marked decrease in the VLDL peak without significantly affecting high-density lipoprotein levels (Fig. 1F). We next analyzed the impact of increased serum BAFF levels on atherogenesis by quantification of Oil Red O+ area on serial aortic sinus sections from Apoe−/− and Apoe−/−.BAFF-Tg mice after 8 wk on WD. Strikingly, Apoe−/−.BAFF-Tg mice exhibited a ∼80% reduction in aortic root atheroma (Fig. 1G, 1H). Together, these findings highlight the unexpected observation that increased serum BAFF protects against hyperlipidemia and atheroma formation in an established murine model of atherosclerosis.
BAFF-mediated atheroprotection requires B cells
BAFF exerts biologic activity by binding the cell surface receptors BAFF receptor (BAFF-R), TACI and B cell maturation Ag (BCMA), whereas the related cytokine APRIL (a proliferation-inducing ligand) binds TACI and the BCMA (33). Although BAFF primarily affects B cell function, additional immune lineages express BAFF-family receptors, including BAFF-R expression by activated CD4+ T cells and regulatory T cells (34), and TACI expression on monocytes (35) and dendritic cells (36). In addition, BAFF is produced by adipocytes, with levels of adipose-tissue BAFF expression correlating with obesity in murine models (37, 38). Moreover, expression of all three BAFF-family receptors (BAFF-R, TACI, BCMA) has been observed on adipocytes, with surface levels modulated by proinflammatory stimuli (37, 39). Thus, decreased hyperlipidemia and atherosclerosis in Apoe−/−.BAFF-Tg mice may occur independently of B cells.
For this reason, we examined whether BAFF-mediated atheroprotection required B cells via two parallel strategies. First, we generated atherosclerosis-prone mice lacking B cells by crossing Apoe−/− and μMT strains. Importantly, μMT mice have elevated serum BAFF levels because of the loss of surface BAFF binding by B cells [serum BAFF: 13 ± 1.5 ng/ml (wild type) versus 166 ± 38 ng/ml (μMT); p < 0.0001, by two-tailed Student t test]. To test the B cell-independent impact of BAFF on atherosclerosis progression, we treated WD-fed Apoe−/−.μMT mice with anti-BAFF mAb [10F4 clone (31)] or hamster IgG1 isotype control (Fig. 2A). This dosing strategy was associated with therapeutic trough serum 10F4 levels (31) and resulted in complete binding of free serum BAFF (Fig. 2B, 2C). In contrast to our findings in Fig. 1, the atheroprotective effect of BAFF was not observed in the absence of B cells as the extent of aortic root atheroma was equivalent in 10F4- versus isotype-treated Apoe−/−.μMT mice (Fig. 2D, 2E).
As a second model to test whether BAFF affected the progression of atherosclerosis via a B cell-dependent versus an independent mechanism, we crossed Apoe−/−.BAFF-Tg with μMT mice. Notably, increased serum BAFF was not associated with a reduction in serum cholesterol in B cell-deficient Apoe−/−.BAFF-Tg animals after 8 wk on WD (Fig. 2F). In parallel with total cholesterol levels, we observed no difference in serum lipoprotein profiles, based on FPLC analysis of WD-fed Apoe−/−.μMT versus Apoe−/−.μMT.BAFF-Tg mice (Fig. 2G). Finally, the severity of aortic root atherosclerosis was not affected by BAFF overexpression in the absence of B cells (Fig. 2H, 2I). Although μMT mice exhibit altered lymphoid architecture that may impact atherosclerosis progression, these combined data demonstrate that, despite BAFF-family receptor expression on myeloid lineages and adipocytes, excess BAFF limits hypercholesterolemia and decreases atherogenesis via B cell-dependent mechanisms.
TACI is required for BAFF-mediated atheroprotection
Having documented that BAFF acts on B cells to decrease atherosclerosis, we next asked which BAFF-family receptor (BAFF-R, TACI or BCMA) is the relevant BAFF target in atherosclerosis. BAFF-R signals are required for peripheral B cell survival beyond the transitional stage, and activation of this receptor likely explains the prominent B cell hyperplasia in BAFF-Tg mice (40). However, previous work has demonstrated that BAFF-R-deficient Apoe−/− mice develop reduced atherosclerosis, implying that BAFF-R is unlikely to mediate an atheroprotective effect (41). In addition, BCMA promotes plasma cell survival, but basal serum IgM and IgG titers are not altered in BCMA-deficient mice (42, 43). For these reasons, we hypothesized that TACI-dependent signals might be required to limit hypercholesterolemia and atherosclerosis progression in Apoe−/−.BAFF-Tg mice. In support of this idea, Taci−/− mice have lower serum IgM levels, suggesting that a lack of TACI may decrease the formation of atheroprotective IgM against oxidized lipid epitopes. Further, although TACI was initially hypothesized to act as a negative regulator of BAFF-mediated signals, recent independent studies, including one from our laboratory, demonstrate that TACI signals are required for class-switched autoantibody formation in BAFF-Tg mice (30, 44).
To test whether TACI signals are required for BAFF-mediated atheroprotection, we generated Apoe−/−.BAFF-Tg.Taci−/− mice. Notably, the increase in serum PC and anti-MDA-LDL IgM titers in WD-fed Apoe−/−.BAFF-Tg mice was not observed in the setting of Taci deficiency (Fig. 3A). In addition, WD-fed Apoe−/−.BAFF-Tg.Taci−/− mice developed marked hypercholesterolemia equivalent to Apoe−/− animals, based on both FPLC lipoprotein profile analysis and total serum cholesterol levels (Fig. 3B, 3C). Finally, Apoe−/−.BAFF-Tg.Taci−/− mice developed large atheromatous lesions within the aortic root that were equal in size to WD-fed Apoe−/− controls (Fig. 3D, 3E). Although BAFF-mediated activation of BAFF-R and/or BCMA might also impact atherosclerosis progression in this model, these findings demonstrate that TACI is required for the BAFF-mediated decrease in hypercholesterolemia and attenuation of atherosclerosis in BAFF-Tg mice.
Taken together, our findings highlight an unanticipated role for BAFF in lipid metabolism and the progression of atherosclerosis. We show that excess BAFF results in markedly reduced aortic atherosclerosis via a B cell- and TACI-dependent mechanism. Although TACI signals might exert additional impacts on atheroma formation via B cell-independent mechanisms, our combined observations suggest a novel role for TACI-driven B cell activation in atherosclerotic cardiovascular disease. Although B cell Ag-presentation and cytokine production also impact immune responses independently of Ab formation, prior animal and human studies have strongly implicated Ab binding oxidized lipid epitopes in the pathogenesis of atherosclerosis. Consistent with these data, BAFF-mediated atheroprotection correlated with a prominent TACI-dependent increase in natural IgM Ab titers, further emphasizing the importance of IgM Ab against oxidized lipid epitopes in limiting atherosclerosis. In this context, promoting anti-oxLDL Ab by vaccination has been proposed as an immunomodulatory strategy in atherosclerotic cardiovascular disease (45). Our study suggests that activating B cell TACI signals may significantly enhance the therapeutic efficacy of this approach.
Whereas BAFF has been implicated in the pathogenesis of human SLE, lupus patients also exhibit increased cardiovascular disease (2). Therefore, our findings raise the concern that therapeutic BAFF inhibition may exert unanticipated impacts on cardiovascular risk in SLE. Although anti-CD20-mediated B cell depletion reduced atherosclerosis in murine models (16, 17), BAFF inhibition may have distinct impacts on the relative balance of atheroprotective and proatherogenic Ab titers. Further, TACI-Ig (Atacicept), a dual inhibitor of both APRIL and BAFF currently in clinical development for treatment of SLE, is associated with prominent ∼70% reduction in total serum IgM and ∼40% decrease serum IgG titers (46). Given the critical importance of TACI signals in B cell–mediated atheroprotection, inhibition of both known TACI ligands using TACI-Ig may additionally impact atherosclerosis progression compared with isolated BAFF blockade. Alternatively, BAFF and/or APRIL blockade may be associated with a relative preservation of protective IgM Ab relative to class-switched pathogenic Ab. In this regard, the impact of therapeutic BAFF family inhibition on atherosclerosis progression has not yet been assessed in hyperlipidemic murine models. In summary, our study highlights a novel role for TACI-dependent BAFF signals in protection against atherosclerosis progression, findings suggesting that lupus patients undergoing therapeutic BAFF inhibition should be closely monitored for increased cardiovascular events.
This work was supported by the National Institutes of Health under award numbers R01AI071163 (to D.J.R.), DP3DK097672 (to D.J.R.), U54 AI112983 (to D.J.R.) and K08AI112993 (to S.W.J.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Additional support was provided by the Benaroya Family Gift Fund (to D.J.R.), a Rheumatology Research Foundation Scientist Development Award (to S.W.J.), the Seattle Children’s Research Institute Pediatric Early Research Career award (to S.W.J.), and the Arnold Lee Smith Endowed Professorship for Research Faculty Development (to S.W.J.).
Abbreviations used in this article:
B cell maturation Ag
fast protein liquid chromatography
low-density lipoprotein receptor
malondialdehyde-modified low density lipoprotein
B cell deficient
oxidized low density lipoprotein
systemic lupus erythematosus
transmembrane activator and CAML interactor
very low density lipoprotein
The authors have no financial conflicts of interest.