B cell depletion reduces the development of atherosclerosis in mice

To directly assess the role of B cells in atherosclerosis, we examined lesion development in mice with or without B cell depletion. We first used Apoe^−/− mice fed a high fat Western diet, a model previously shown to be associated with significant B cell activation and previously used to demonstrate the protective role of B cells in atherosclerosis (Caligiuri et al., 2002). To deplete B cells, mice were treated every 3 wk with a previously validated mouse monoclonal CD20 antibody (Uchida et al., 2004a,b) for either 6 or 12 wk. Control mice received a control mAb. As expected (Uchida et al., 2004a; Hamaguchi et al., 2005), treatment with CD20 mAb led to sustained and profound reduction of the number of mature B cells in the blood (Fig. 1 a), spleen (Fig. 1 b), peritoneum, and bone marrow (Fig. S1). B220^high IgM⁺ cells were severely depleted (92–100%) at all studied sites. Spleen B220^low IgM⁺ cells also showed a marked reduction (∼80%). However, as previously observed (Uchida et al., 2004a), immature bone marrow B220^low IgM⁺ cells (Fig. S1) were less sensitive to CD20 mAb–mediated depletion. Treatment with CD20 mAb for 6 wk did not affect plasma cholesterol levels (6.4 ± 0.9 vs. 6.3 ± 0.8 g/liter in control and CD20 mAb–treated groups, respectively; P = 0.88) but unexpectedly led to a significant reduction, not acceleration, of atherosclerotic lesion development (Fig. 1 c). We subsequently analyzed the experiments of Apoe^−/− mice treated for 12 wk under a high fat diet and still found a significant reduction of atherosclerosis at two different vascular sites (Fig. 1 d and Fig. S2) despite similar plasma cholesterol levels (18.7 ± 1.1 vs. 17.9 ± 1 g/liter in control IgG and anti-CD20–treated groups, respectively; P = 0.68). To rule out the possibility that the atheroprotective effect of CD20 mAb treatment was caused by the use of a mouse model that generates excessive inflammation in response to a very high lipid overload, we examined the effect of B cell depletion in Apoe^−/− mice fed a chow diet. Treatment of these mice with CD20 antibody for 12 wk also resulted in significant reduction of lesion development (Fig. 1 e) despite similar plasma cholesterol levels (5.5 ± 0.6 vs. 5.7 ± 0.8 g/liter in control IgG and CD20 mAb–treated groups, respectively; P = 0.96). The elevated plasma cholesterol levels in Apoe^−/− mice are mostly of the very low LDL subtype, whereas elevated LDL is the major atherosclerosis risk factor in humans. Thus, we examined the effects of B cell depletion in the Ldlr^−/− mouse model. Again, treatment of Lldr^−/− mice with CD20 mAb led to marked B cell depletion (Fig. S3) and to significant reduction of atherosclerosis (Fig. 1 f). Overall, these studies provide solid evidence for an unsuspected proatherogenic role of B cells in three mouse models of atherosclerosis.

We next addressed the potential mechanisms responsible for atheroprotection after B cell depletion. We found that treatment with CD20 depleting antibody resulted in a profound reduction of IgG type anti-oxLDL antibodies both at 6 and 12 wk of treatment (Fig. 2), which was consistent with the profound depletion of B220^high cells in blood, spleen, and bone marrow (Fig. 1 and Fig. S1). Reduction of anti-oxLDL IgG antibodies might have limited the potentially deleterious consequences of immune complex formation on atherosclerosis (Hernández-Vargas et al., 2006). However, in other studies and particularly after splenectomy, profound reduction in anti-oxLDL IgG levels was observed in association with acceleration, not reduction, of atherosclerosis (Caligiuri et al., 2002). Thus, in the absence of studies directly addressing the role of IgG type anti-oxLDL antibodies on atherosclerosis, changes in anti-oxLDL IgG levels after CD20 mAb treatment could not be held responsible for lesion reduction. Levels of IgM type antibodies against either copper-oxidized or malondialdehyde-modified LDL were also reduced after 6 or 12 wk of CD20-targeted therapy (Fig. 2). IgM type antibodies are endowed with atheroprotective properties (Lewis et al., 2009), and their reduction after CD20 mAb therapy could not account for atheroprotection. It is interesting to note, however, that IgM type anti-oxLDL and T15id⁺ IgM antibodies showed a much lower reduction compared with IgG type antibodies (Fig. 2), which is consistent with the lower efficiency of CD20 mAb on peritoneal B1 cell depletion (Fig. S4; Hamaguchi et al., 2005). This might have preserved an atheroprotective pathway. IgM type antibodies dominate the humoral response to oxLDL in Apoe^−/− mice (Palinski et al., 1994, 1996) and are increased even at a young age (before the initiation of CD20 mAb treatment in this study), which may also explain, at least in part, the persistence of a significant IgM level after CD20 immunotherapy, a treatment that does not dramatically affect preexisting antibody titers (DiLillo et al., 2008).

We next examined atherosclerotic lesion composition to gain more insight into the mechanisms of atheroprotection. In addition to the significant reduction of macrophage accumulation (Fig. S5), CD20 mAb treatment was associated with a marked reduction of T lymphocyte accumulation within the lesions (Fig. 3 a), suggesting a role for B cells in driving T cell–dependent lesion inflammation. As expected at this stage of lesion formation, very few B cells were detected within the plaques or within the adventitial layer (Fig. S6), suggesting that modulation of lesion T cell accumulation by CD20 mAb therapy most likely occurred as a consequence of systemic modulation of T cell function. To address this hypothesis, we examined T cell activation and proliferation. Interestingly, we consistently found marked reductions in CD69 and CD44^high expression on spleen-derived CD4⁺ T cells of mice treated with CD20 antibody compared with controls at both 6 wk (not depicted) and 12 wk (Fig. 3, b and c) of a high fat diet, indicating reduced T cell activation. B cell depletion also led to significant reduction of in vivo BrdU staining of effector CD4⁺CD25⁻ T cells, suggesting reduced proliferation (Fig. 3 d). This is consistent with the lower proliferative potential of CD4⁺ cells recovered from CD20-treated mice and co-cultured with purified CD11c⁺ cells in the presence of CD3 stimulation (Fig. S7 a). In addition, replacement of control CD11c⁺ cells with DCs isolated from CD20-depleted mice led to a reduction of (control) T cell proliferation in vitro (Fig. S7 a). This is also consistent with the marked reduction of CD40 expression on CD11c⁺ DCs recovered from CD20 mAb–treated mice (Fig. 3 e). Moreover, we show that incubation of DCs with activated B cells leads to enhanced CD40, CD80, and MHC-II expression on DCs (Fig. S7 b). Thus, a major consequence of B cell depletion using CD20 antibody is a marked reduction of DC and T cell activation in vivo, which could potentially account for its atheroprotective effect.

T cell–derived cytokines significantly alter lesion development (Tedgui and Mallat, 2006). Therefore, we examined the consequences of B cell depletion on cytokine production by purified T cells. We found a marked reduction of proatherogenic IFN-γ by purified T cells recovered from CD20 mAb–treated mice compared with controls (Fig. 4 a). Of note, this was associated with a deviation of the immune response toward a significant increase of T cell–derived (Fig. 4, b and c) and circulating IL-17A production in CD20 mAb–treated animals compared with controls (26 ± 4 vs. 0 pg/ml, respectively). Replacement of control CD11c⁺ cells with DCs isolated from CD20-depleted mice significantly shifted (control) T cell cytokine production toward a higher IL-17/IFN-γ ratio (Fig. S7 a), indicating a contribution of DCs to the observed cytokine switch. Recent studies in our laboratory identified an unexpected regulatory and protective role for IL-17A in atherosclerosis under a context of reduced IFN-γ production (Taleb et al., 2009). In addition, IL-17A has been shown to modulate Th1 polarization (O’Connor et al., 2009). Thus, the T cell cytokine profile induced by B cell depletion is compatible with the observed atheroprotective effects.

To examine whether CD20 mAb–induced changes in T cell cytokine profile (reduced IFN-γ and increased IL-17) could be responsible for CD20 mAb–dependent atheroprotection, CD20 mAb was administered to Apoe^−/− mice (on a high fat diet for 6 wk) in the presence of control or anti–IL-17A neutralizing antibody. IL-17 neutralization led to increased IFN-γ production in the spleen (not depicted) and atherosclerotic aortas, and completely abrogated the atheroprotective effects of CD20 mAb therapy (Fig. 5, a and b) despite similar circulating cholesterol levels, similar efficiency of B cell depletion (Fig. S8), and no significant changes in anti-oxLDL antibodies levels (Fig. S8). We reproduced these results in a second separate experiment (Fig. 5 c). Collectively, these results identify a hitherto unsuspected role for B cells in driving the development of atherosclerosis through modulation of T cell activation and cytokine production.

Our results may seem in disagreement with previous studies showing that both μMT deficiency (Major et al., 2002) and splenectomy (Caligiuri et al., 2002) accelerate atherosclerosis in mice. However, these studies did not directly address the role of mature B cell depletion on atherosclerosis in immunocompetent mice. Several other concomitant immune cell dysfunctions (Moulin et al., 2000; Ngo et al., 2001; João et al., 2004; AbuAttieh et al., 2007) might have contributed to enhanced lesion development in μMT-deficient animals. Furthermore, the reported limitation of atherosclerosis acceleration in splenectomized mice after reconstitution with purified B cells could have been confounded by the reduction of plasma cholesterol levels in B cell–reconstituted mice (Caligiuri et al., 2002), and could not be selectively attributed to B cells because T cell reconstitution also resulted in atheroprotection (Caligiuri et al., 2002). Finally, it should be noted that although B cell depletion significantly limited lesion development in our present studies, the roles of specific subtypes of B cells in driving or controlling atherosclerosis merit further investigation. More particularly, it will be important to address the respective roles of regulatory (Yanaba et al., 2008) versus nonregulatory B cells in these processes (Bouaziz et al., 2007). Finally, it could be argued that skewing T cell responses from Th1 to IL-17 production in the absence of mature B cells would not be helpful in other autoimmune/inflammatory diseases known to be mediated in part by Th17 cells. However, recent studies have provided data indicating that not all Th17 cells are endowed with the same pathogenic potential (McGeachy et al., 2007). These issues merit further investigation.

In conclusion, we provide strong evidence that mature B cell depletion reduces the development of atherosclerosis in mice. Our results challenge the paradigm that overall B cell function is atheroprotective and show that a major B cell role in atherosclerosis is to drive T cell activation toward enhanced proatherogenic Th1 immune response and limited production of atheroprotective IL-17 (Fig. S9). Although limited vascular B cell infiltration is detectable in the early stages of atherosclerosis, B cell accumulation substantially increases with time. It localizes within and around advanced atherosclerotic coronary lesions and atherosclerotic abdominal aortic aneurysms (Moos et al., 2005) of mice and humans, and is even prominent in vascular inflammation associated with other immune-mediated diseases (Aubry et al., 2004). Inhibition of excessive B cell activation either through depletion or immune modulation might substantially limit vascular inflammation and atherosclerotic lesion development.

All mice were on a C57BL/6 background. Apoe^−/− mice were 10-wk-old males maintained on a chow diet for 12 wk or put on a Western diet (20% fat, 0.15% cholesterol, 0% cholate) for either 6 or 12 wk. Ldlr^−/− mice were 10-wk-old males put on a Western diet for either 6 or 12 wk. At 10 wk old, mice were treated i.p. with a previously validated mouse CD20 mAb (Uchida et al., 2004a,b) or control IgG (200 µg every 3 wk) for either 6 or 12 wk. In some experiments, mice received an i.p. injection of either purified neutralizing anti–IL-17A–specific antibody (200 µg/mouse, twice per week; Uyttenhove and Van Snick, 2006; Uyttenhove et al., 2007; Taleb et al., 2009) or control IgG for 6 wk. Experiments were conducted according to the French veterinary guidelines and those formulated by the European Community for experimental animal use (L358-86/609EEC), and were approved by the Institut National de la Santé et de la Recherche Médicale.

Quantification of lesion size and composition was performed as previously described (Taleb et al., 2007).

CD11c⁺ and CD4⁺ cells were purified and processed for cell proliferation assays and cytokine production as previously described in detail (Taleb et al., 2007). For cell proliferation experiments, CD4⁺ cells were cultured in round-bottom 96-well microplates. Cells were stimulated with 1 µg/ml of purified soluble CD3-specific antibody (BD) in the presence of 10⁴ antigen-presenting cells purified on CD11c-coated magnetic beads (Miltenyi Biotec). For cytokine measurements, CD4⁺ T cells were cultured at 10⁵ cells/well for 48 h on anti-CD3–coated microplates (5 µg/ml) in the presence of 2 × 10⁴ CD11c⁺ cells. IL-17 and IFN-γ production in the supernatants was measured using specific ELISAs (BD and R&D Systems).

Allophycocyanin (APC)-conjugated anti-CD3ε (145-2C11), FITC- or PE-Cy7–conjugated anti-CD4 (RM4-5), APC-conjugated anti-CD25 (PC61.5), PE-conjugated anti-CD69 (H1.2F3), APC-conjugated anti-IgM (II/41), FITC-conjugated anti-CD86 (GL1), PE-conjugated anti-CD80 (16-10A1), APC-conjugated anti-CD40 (1C10), PE-Cy7–conjugated anti-CD11c (N418), PE-Cy7–conjugated anti-CD11b (M1/70), and PE-conjugated anti-CD45R (B220; RA3-6B2) were from eBioscience. FITC-conjugated anti-CD5 (53-7.3), biotin-conjugated anti-CD44 followed by APC-conjugated streptavidin, APC-Cy7–conjugated anti-CD45R (B220; RA3-6B2), APC-conjugated anti–IFN-γ (XMG1.2), and PE-conjugated anti–IL-17A (TC11-18H10) were from BD. For blood staining, erythrocytes were lysed using FACS lysing solution (BD). For intracellular cytokine staining, lymphocytes were stimulated in vitro with leukocyte activation cocktail (BD) according to the manufacturer’s instructions for 4 h. Surface staining was performed before permeabilization using an intracellular staining kit (eBioscience). For staining on aortic cells, the aorta was removed and flushed intensively with 2 mM PBS-EDTA. The aorta was cut into pieces and digested using a cocktail of enzymes for 30 min at 37°C (450 U/ml collagenase I, 375 U/ml collagenase XI, 180 U/ml hyaluronidase I, and 180 U/ml DNase I; all from Sigma-Aldrich). Forward scatter and side scatter were used to gate live cells excluding red blood cells, debris, and cell aggregates in total splenocyte, lymph node, bone marrow, aorta, and peritoneum populations. Cells were analyzed using a flow cytometer (FACSCanto II or LSR II; BD).

BrdU labeling was performed as previously described (Fisson et al., 2003). Mini osmotic pumps (ALZET1007D; Charles River), delivering 1.2 mg/day BrdU (Sigma-Aldrich) for 7 d, were transplanted to mice subcutaneously under isoflurane anesthesia 1 wk before sacrifice. Lymph node cells and splenocytes were stained with PE-Cy7–conjugated anti-CD4 (RM4-5) and APC-conjugated anti-CD25 (PC61.5) antibodies. BrdU detection was performed using an FITC BrdU flow kit (BD) according to the manufacturer’s instructions. Cells were analyzed using a FACSCanto II or LSR II flow cytometer.

Bone marrow–derived DCs were cultured from whole bone marrow using 20 ng/ml of GM-CSF (Miltenyi Biotec)–conditioned medium for 10 d. DCs were plated in 24 wells at 4 × 10⁵ cells/well. B cells were isolated from C57BL/6 spleens using a B cell isolation kit (Miltenyi Biotec) according to the manufacturer’s protocol. B cells were incubated overnight in RPMI 1640 with 10 µg/ml anti–mouse IgM (Jackson ImmunoResearch Laboratories, Inc.) and 2.5 µg/ml anti–mouse CD40 (clone HM40-3; BioLegend). After washing, B cells were added on DC cultures at 2 × 10⁶ cells per well and co-cultured for 48 h. DCs were collected for FACS analysis.

Quantitative real-time PCR was performed on an ABI Prism 7700 (Applied Biosystems) in triplicate. Cycle threshold for GAPDH (primers: GAPDH R, 5′-CGTCCCGTAGACAAAATGGTGAA-3′; GAPDH L, 5′-GCCGTGAGTGGAGTCATACTGGAACA-3′) was used to normalize gene expression. The following primer sequences were used: IL-10 (IL-10 R, 5′-GGCGCTGTCATCGATTTCT-3′; IL-10 L, 5′-ATTCATTCATGGCCTTGTAGACAC-3′), TGF-β (TGF-β R, 5′-AGCAACAATTCCTGGCGTTACCTT-3′; TGF-β L, 5′-TGGGCTCGTGGATCCACTTC-3′), and IFN-γ (IFN-γ R, 5′-GCCACGGCACAGTCATTGAAA-3′; IFN-γ L, 5′-TCGCCTTGCTGTTGCTGAAGA-3′).

Specific antibody titers to given antigens in plasma were determined by chemiluminescent ELISA, as previously described (Friguet et al., 1985; Binder et al., 2003; Chou et al., 2009).

Values are expressed as means ± SEM. Differences between values were examined using nonparametric Mann-Whitney or Kruskal-Wallis tests, and were considered significant at P < 0.05.

Fig. S1 shows CD20 mAb–induced B cell depletion in bone marrow and peritoneum. Fig. S2 shows that CD20 mAb treatment reduces the development of atherosclerosis in the thoracic aorta. Fig. S3 shows the efficiency of B cell depletion in Ldlr^−/− mice. Fig. S4 shows the relative efficiency of CD20 mAb treatment on peritoneal B2, B1a, and B1b cell depletion in Apoe^−/− mice. Fig. S5 shows that CD20 mAb treatment reduces macrophage infiltration within atherosclerotic lesions. Fig. S6 shows that very few B cells are detected with atherosclerotic arteries of Apoe^−/− mice fed a Western diet for 12 wk. Fig. S7 shows modulation of DC and T cell functions through B cell activation or depletion. Fig. S8 shows that IL-17A neutralization does not affect plasma cholesterol levels, efficiency of B cell depletion, or production of anti-oxLDL antibodies. Fig. S9 shows the proposed mechanisms for the atheroprotective effect of CD20 mAb.

We are indebted to D. Macé, C. Lahoute, and M.O. Kozma for their valuable technical support.

This work was supported by the Institut National de la Santé et de la Recherche Médicale, the transatlantic Leducq Immunoregulatory Network (C.J. Binder, A. Tedgui, and Z. Mallat), the European Union Seventh Framework Program TOLERAGE (Z. Mallat), the Fund of Scientific Medical Research, the Belgian Federal Service for Scientific, Technical, and Cultural Affairs (C. Uyttenhove and J. Van Snick), and the National Institutes of Health (grants CA105001, AI56363, and AI057157 to T.F. Tedder).

The authors have no conflicting financial interests to disclose.