Article MDA-LDL vaccination induces athero-protective germinal-center-derived antibody responses Graphical abstract Authors Inmaculada Martos-Folgado, Alberto del Monte-Monge, Cristina Lorenzo, ..., Jose L. Martı́n-Ventura, Hedda Wardemann, Almudena R. Ramiro Correspondence aramiro@cnic.es In brief Martos-Folgado et al. show that immunization with MDA-LDL elicits a GC- derived antibody immune response. Furthermore, the authors demonstrate that GC-derived antibodies drive MDA- LDL-induced athero-protection. These results uncover the mechanism underlying MDA-LDL-driven athero- protection and support the development of MDA-LDL-based vaccines for the prevention or treatment of atherosclerosis. Highlights d MDA-LDL immunization triggers a GC response d Pro-atherogenic mice deficient for GC-derived PCs show accelerated atherosclerosis d MDA-LDL-driven athero-protection is impaired in mice lacking GC-derived PCs Martos-Folgado et al., 2022, Cell Reports 41, 111468 October 11, 2022 ª 2022 The Authors. https://doi.org/10.1016/j.celrep.2022.111468 ll ll OPEN ACCESS Article MDA-LDL vaccination induces athero-protective germinal-center-derived antibody responses Inmaculada Martos-Folgado,1,2 Alberto del Monte-Monge,1 Cristina Lorenzo,1 Christian E. Busse,3 Pilar Delgado,1,9 Sonia M. Mur,1 Laura Cobos-Figueroa,1,4,10 Joan C. Escolà-Gil,5,6 Jose L. Martı́n-Ventura,7,8 Hedda Wardemann,3 and Almudena R. Ramiro1,11,* 1B Lymphocyte Biology Lab, Centro Nacional de Investigaciones Cardiovasculares (CNIC), Madrid, Spain 2Medical Department, Pfizer S.L.U., Madrid, Spain 3Division of B Cell Immunology, German Cancer Research Center, Heidelberg, Germany 4Grupo de Manufactura Avanzada con Láser del Centro Láser (UPM), Madrid, Spain 5CIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM), Madrid Spain 6Institut de Recerca de l’Hospital de La Santa Creu I Sant Pau, Barcelona, Spain 7CIBER de Enfermedades Cardiovasculares (CIBERCV), Madrid, Spain 8Vascular Pathology Lab, IIS-Fundación Jiménez Dı́az-Universidad Autónoma, Madrid, Spain 9Present address: Centro de Biologı́a Molecular Severo Ochoa, Consejo Superior de Investigaciones Cientı́ficas, Universidad Autónoma de Madrid, Madrid, Spain 10Present address: Unidad de Presentación y Regulación Inmunes, Instituto de Salud Carlos III, Madrid, Spain 11Lead contact *Correspondence: aramiro@cnic.es https://doi.org/10.1016/j.celrep.2022.111468 SUMMARY Atherosclerosis is a chronic inflammatory disease of the arteries that can lead to thrombosis, infarction, and stroke and is the leading cause of mortality worldwide. Immunization of pro-atherogenic mice with malondial- dehyde-modified low-density lipoprotein (MDA-LDL) neo-antigen is athero-protective. However, the immune response to MDA-LDL and the mechanisms responsible for this athero-protection are not completely under- stood. Here, we find that immunization of mice with MDA-LDL elicits memory B cells, plasma cells, and switched anti-MDA-LDL antibodies as well as clonal expansion and affinity maturation, indicating that MDA-LDL triggers a bona fide germinal center antibody response. Further, Prdm1fl/fl Aicda-Cre+/ki Ldlr/ pro-atherogenic chimeras, which lack germinal center-derived plasma cells, show accelerated atheroscle- rosis. Finally, we show that MDA-LDL immunization is not athero-protective in mice lacking germinal-cen- ter-derived plasma cells. Our findings give further support to the development of MDA-LDL-based vaccines for the prevention or treatment of atherosclerosis. INTRODUCTION cells and induce the expression of adhesion molecules and che- mokines that recruit monocytes to the intima layer and promote Cardiovascular disease (CVD) remains the leading cause of mor- their differentiation into macrophages. At the intima, macro- tality in the world, with most CVD deaths resulting from myocar- phages internalize OxLDL and eventually give rise to lipid-laden dial infarction and stroke. The main cause underlying thrombosis foam cells, which contribute to plaque growth and the generation and CV events is atherosclerosis, a chronic disorder of large ar- of a necrotic core (Bobryshev et al., 2016; Brophy et al., 2017; teries caused by hyperlipidemia that leads to atheroma plaque Libby, 2021). formation (Hansson and Hermansson, 2011; Hansson et al., Adaptive immunity contributes in various and intertwined ways 2006; Lahoute et al., 2011; Libby, 2021; Skalen et al., 2002). to atherosclerosis modulation. Different subsets of T cells have Thus, there is a critical need to expand and improve our thera- been ascribed different roles in atherosclerosis, which generally peutic and prognostic avenues to atherosclerosis (Libby, 2021; seem to stem from their pro- or anti-inflammatory properties (Ait- Nilsson and Hansson, 2020). Oufella et al., 2006; Clement et al., 2015; Gupta et al., 1997; Nus Atherosclerosis initiates with the infiltration and retention of et al., 2017; Saigusa et al., 2020). In the case of B cells, detailed cholesterol-rich low-density lipoprotein (LDL) particles in the studies on the contribution of different B cell subsets to athero- subendothelial space of the artery wall (Skalen et al., 2002; Ta- sclerosis have revealed a complex landscape. B1 cells, a minor bas et al., 2007). Retained LDL is subject to oxidation, which subset of B cells of fetal origin that produce natural immunoglob- leads to diverse oxidized forms of LDL (OxLDL), including malon- ulin M (IgM) antibodies, are generally believed to be athero-pro- dialdehyde-modified LDL (MDA-LDL) (Hansson and Hermans- tective (Kyaw et al., 2011; Rosenfeld et al., 2015; Shaw et al., son, 2011; Lusis, 2000). OxLDL particles activate endothelial 2000). Marginal zone B cells are splenic non-circulating B cells Cell Reports 41, 111468, October 11, 2022 ª 2022 The Authors. 1 This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ll OPEN ACCESS Article that mediate an innate-like rapid antibody response to blood- LDL is not completely understood, and several mechanisms borne antigens and have athero-protective capabilities (Nus have been proposed to explain MDA-LDL-mediated athero-pro- et al., 2017). Follicular B cells are essential for long-lasting, tection (Ait-Oufella et al., 2006; Binder et al., 2004; Hjerpe et al., T-dependent antibody responses and have been assigned 2010; Zhou et al., 2005). both pro-atherogenic and athero-protective roles (Ait-Oufella Here, we find that athero-protective MDA-LDL immunization et al., 2010; Caligiuri et al., 2002; Grasset et al., 2015; Kyaw triggers an antibody immune response that entails GC formation et al., 2010, 2013; Major et al., 2002; Sage et al., 2012). and the generation of MBCs and PCs. A high-throughput single- Upon antigen encounter and in the presence of T cell help, cell analysis of the immunoglobulin repertoire in MDA-LDL- follicular B cells became activated and can engage into the immunized mice revealed clonal expansion and accumulation germinal center (GC) reaction and give rise to long-lived memory of SHM, featuring a GC response. Using Prdm1fl/fl Aicda-Cre+/ki B cells (MBCs) and high-affinity plasma cells (PCs) with alternate Ldlr/ chimeras, we further show that depletion of GC-derived isotypes. Key to these antibody responses are the class switch PCs aggravates atherosclerosis, indicating a global athero-pro- recombination (CSR) and the somatic hypermutation (SHM) re- tective role of these cells. Finally, we found that athero-protection actions. CSR is a recombination reaction that exchanges the pri- by MDA-LDL immunization is impaired in Prdm1fl/fl Aicda-Cre+/ki mary IgM constant region by a downstream constant region, Ldlr/ chimeras. Together, our results indicate that MDA-LDL generating IgG, IgA, or IgE isotypes. SHM introduces changes immunization triggers a bona fide GC reaction and the generation in the antigen-recognizing, variable region of antibody genes of GC-derived PCs, which contribute to athero-protection. thus giving rise to clonal variants. Coupled to affinity maturation, SHM allows the generation of higher affinity antibodies. Thus, RESULTS GCs are essential for high-affinity humoral responses and under- lie the protective mechanism of many vaccination strategies. Athero-protective MDA-LDL immunization induces GC Disruption of T-B cell interactions, abrogation of the GC reaction, reactions or broad depletion of PCs resulted in smaller plaques (Tay et al., To gain insights into the impact of MDA-LDL immunization on 2018), but they showed more instability (Centa et al., 2019). On atherosclerosis, we immunized Ldlr/ mice with MDA-LDL in the other hand, pro-atherogenic mice where antibody secretion complete Freund’s adjuvant (CFA) or with CFA alone as was abolished had larger plaques (Sage et al., 2017). These re- described before (Figure 1A) (Freigang et al., 1998). We did not sults reveal a complex role of the antibody response in athero- detect significant differences in weight gain or plasma lipids be- sclerosis. Of note, the role of antibody responses after specific tween groups (Figure S1A; Table S1). Quantification of athero- challenge with atherosclerosis antigens was not addressed in sclerotic plaques in aortic roots showed a significant decrease these studies. in atherosclerosis extension in mice immunized with MDA- Indeed, very few antigens have been associated with athero- LDL + CFA compared with those immunized with CFA alone sclerosis. The atherosclerosis adaptive immune response is (Figures 1B–1D), as previously described (Freigang et al., 1998; thought to be triggered by self- and neo-antigens generated dur- George et al., 1998; Palinski et al., 1995; Turunen et al., 2015). ing disease development, the paradigm of which are oxidation- We did not detect changes in the proportions of GC B cells or to- specific epitopes present in OxLDL (Nilsson and Hansson, tal PCs (Figures S1B–S1G). However, the proportion of MBCs 2020). Heat shock protein (Hsp) 60, a mitochondrial chaperone, was significantly increased in MDA-LDL + CFA-immunized has also been associated with atherosclerosis (Almanzar et al., mice compared with CFA controls (Figure 1E). Likewise, the pro- 2012; George et al., 1999; Xu et al., 1992, 1993; Zhu et al., portion of IgG1+ PCs was much larger in MDA-LDL + CFA mice 2001). In addition, the aldehyde dehydrogenase 4 family member (Figure 1F). Indeed, IgM, IgG1, IgG2b, and IgG2c MDA-LDL-spe- A1 (ALDH4A1) mitochondrial protein has been recently identified cific antibodies were greatly increased in serum from MDA-LDL- as an auto-antigen in atherosclerosis (Lorenzo et al., 2021). immunized mice (Figure 1G). Thus, MDA-LDL immunization of Circulating ALDH4A1 is increased in mice and humans with Ldlr/ mice is athero-protective and promotes a B cell response atherosclerosis, and anti-ALDH4A1 antibody is athero-protec- that generates MBCs, switched PCs, and MDA-LDL-specific tive (Lorenzo et al., 2021). However, LDL forms have been by switched antibodies. far the most studied atherosclerosis antigen and, as of today, To dissect the B cell immune response to MDA-LDL in the are considered the best antigenic target for the design of athero- absence of atherosclerosis-dependent variables, we performed sclerosis vaccines (Nilsson and Hansson, 2020). T cells, B cells, immunization experiments in Aicda-Cre+/ki (Robbiani et al., 2008; and anti-OxLDL antibodies have been found in atheroma pla- Rommel et al., 2013) R26tdTom+/ki mice (Madisen et al., 2010), ques and in plasma from atherogenic mice, rabbits, and humans where B cells that have been activated for activation-induced (Hansson et al., 1989; Hollander et al., 1979; Jonasson et al., deaminase (AID) expression become irreversibly tdTom+ (Fig- 1986; Karvonen et al., 2003; Palinski et al., 1989, 1994; Tsimikas ure S2). AID is the enzyme that initiates SHM and CSR; thus, its et al., 2007). expression is a good proxy to identify cells that have had GC expe- MDA forms covalent adducts with lysine residues of the apoB- rience (Albright et al., 2019; Centa et al., 2019; Gitlin et al., 2016; 100 protein and generates MDA-LDL, the predominant form of Mayer et al., 2020). Aicda-Cre+/ki R26tdTom+/ki mice were immu- OxLDL (Hartley et al., 2019; Yla-Herttuala et al., 1989). MDA- nized with MDA-LDL + CFA or with CFA alone as shown in Fig- LDL immunization reduces atherosclerosis (Fredrikson et al., ure 2A. We detected an increase in tdTom+GL7+ GC B cells and 2003; George et al., 1998; Palinski et al., 1995; Zeng et al., tdTom+PDL-2+ MBCs in lymph nodes and spleen of MDA-LDL + 2018; Zhou et al., 2001), but the immune response to MDA- CFA-immunized mice (Figures 2B–2D; GC absolute numbers 2 Cell Reports 41, 111468, October 11, 2022 ll Article OPEN ACCESS Figure 1. MDA-LDL is athero-protective and promotes a specific humoral response (A) Experimental design. Ldlr/ mice were immunized subcutaneously (s.c.) with MDA-LDL in CFA or with CFA alone (day 0). Four booster immunizations were done given intraperitoneally (i.p.) with MDA-LDL in IFA or IFA alone every two weeks (weeks 2, 4, 6, 8), followed by three additional immunizations once a month four weeks later (week 12, 16, 20). Mice were fed with 1% cholesterol HFD and with 1.25% cholesterol HFD from weeks 8–10 and 10–23, respectively. (B) Representative Oil red O/Hematoxylin staining and (C) quantification of the proportion of atheroma plaque in serial 80 mm-spaced aortic sinus cryosections from MDA-LDL + CFA and CFA immunized mice. (D) Area under the curve (AUC) plot of atheroma plaque quantification. The scale bars in images represents 200 mm. Representative flow cytometry plots and quantification of (E) splenic MBCs (B220+PD-L2+) and (F) IgG1+ bone marrow PCs (B220CD138+IgG1+) from MDA-LDL + CFA and CFA mice. Population percentages in FACS graphs indicate the frequency of live cells. (G) Quantification by ELISA of MDA-LDL specific IgM, IgG1, IgG2b and IgG2c levels in serum from LDLR/ mice 23 weeks after primary immunization with MDA- LDL + CFA or CFA. Data analyzed by unpaired t and two-way ANOVA statistical tests. All experiments were performed with 21–28-week-old males. Each dot in the graphs represents a biological replicate (individual mouse). *p % 0.05, **p < 0.01 and ****p < 0.0001. Cell Reports 41, 111468, October 11, 2022 3 ll OPEN ACCESS Article Figure 2. MDA-LDL immunization triggers a GC B cell response in non-atherogenic mice (A) Experimental design. Mice were immunized SC with MDA-LDL in CFA or with CFA alone, followed by two booster immunizations IP with MDA-LDL in incomplete FA (IFA) or IFA alone 2 and 4 weeks later. Fluorescence-activated cell sorting (FACS) analyses were performed 2, 4, 6, and 8 weeks after the primary immunization. (B–D) Representative FACS plots of GC B cells (B220+GL7+tdTom+; top panel) and MBCs (B220+PD-L2+tdTom+; bottom panel) 2 and 6 weeks after primary immunization, respectively (B), and quantification of these populations in (C) lymph nodes (LNs) and (D) spleen (SP). (legend continued on next page) 4 Cell Reports 41, 111468, October 11, 2022 ll Article OPEN ACCESS shown in Figures S2B and S2C). GC B cells were also detected in (Figures 4F and S4A–S4M), and MDA-LDL + CFA-immunized C57-BL6 mice after a single MDA-LDL immunization (Figure S2D). mice had a larger proportion of GC IgG1+ cells (Figures S4N– In addition, PCs accumulated over time in the bone marrow and S4P). Together, these data indicate that MDA-LDL immunization were significantly increased 2 weeks after the third immunization triggers an immune response that involves a GC reaction. (Figures 2E and 2F). This response was accompanied by an in- To focus on Igs relevant to the MDA-LDL response, we identi- crease in both unswitched and switched anti-MDA-LDL antibody fied those expressed by clonally expanded B cells. Thus, we titers in MDA-LDL-immunized mice, as measured by enzyme- defined ‘‘B cell clones’’ as clusters of three or more Igs sharing linked immunosorbent assay (ELISA) (Figures 3A–3D). the same V and J genes and the same CDR3 length in both To further analyze the specificity of the antibody response to IgH and IgL. We identified 25 B cell clones, 12 of which were MDA-modified epitopes, we measured the antibody titers to exclusively comprised by Igs from the CFA group (1–12), 12 by BSA oxidized with MDA (MDA-BSA) in plasma from MDA-LDL + Igs from the MDA-LDL + CFA group (13–24) (from now on, exclu- CFA-, CFA-, and PBS-immunized mice. We found that MDA- sive clones), and 1 clone containing Igs from both groups (25) LDL immunization led to a dramatic increase in both IgM and (mixed clone) (Table S3; Figure 4G). IgG anti-MDA-BSA antibody titers (Figure 3E). In addition, compe- We found that exclusive CFA clones were disproportionally tition immunoassays with plasma from MDA-LDL-immunized comprised by Igs expressed by GC B cells, while MDA-LDL + mice showed that binding of plasma IgG1 antibodies to MDA- CFA clones were composed by Igs from GC, PC, and MBC pop- LDL was partially inhibited by pre-incubation with MDA-BSA and ulations (Figure 4H) and showed a greater enrichment in PCs LDL particles, while MDA-LDL completely inhibited the binding (Figure 4I). Finally, while in CFA mice we did not observe differ- of antibodies to MDA-LDL, indicating high specificity of the anti- ences in mutation frequency between expanded and non- MDA-LDL; in this case, IgG1 switched antibodies (Figure 3F). expanded Igs, in MDA-LDL + CFA mice, expanded Igs had a Together, these results indicate that MDA-LDL immunization higher mutational load than non-expanded Igs (Figure 4J). This triggers an immune GC response that gives rise to antibodies result shows that B cells expanded in response to MDA-LDL ex- specific for MDA-containing epitopes. press heavily mutated Igs, indicating that they have undergone affinity maturation. SHM and selection in MDA-LDL-immunized mice To gain molecular insights into the antibody response to MDA- GC-derived PCs are athero-protective LDL, we analyzed the immunoglobulin repertoire by single- To address the role of GC-derived antibodies in atherosclerosis, cell sequencing of Ig genes in Aicda-Cre+/ki R26tdTom+/ki mice we transferred bone marrow cells from Prdm1fl/fl Aicda-Cre+/ki immunized with MDA-LDL in CFA or CFA alone. We performed CD45.2 mice into lethally irradiated Ldlr/ mice expressing single-cell sorting and sequencing of GC B cells (B220+ GL7+ the CD45.1 allotype (from now on, Prdm1fl/fl Aicda-Cre+/ki Ldlr/ tdTom+) and MBCs (B220+ GL7 tdTom+) from lymph nodes chimeras). Prdm1 codes for Blimp1, the master transcriptional and GC-derived PCs (B220 CD138+ tdTom+) from bone marrow regulator of PCs (Shaffer et al., 2002); thus, Prdm1fl/fl Aicda- (Figure 4A and S3A). We obtained 597 heavy chain (IgH) and light Cre+/ki Ldlr/ chimeras will be devoid of PCs and antibody chain (IgL) paired sequences from three mice immunized with secretion derived of cells that have expressed AID. Ldlr/ MDA-LDL and two mice immunized with CFA alone (Table S2; CD45.1+ mice transferred with CD45.2+ Prdm1+/+ Aicda-Cre+/ki Figures 4B and S3B). We found minor differences in the variable bone marrow (from now on, Prdm1+/+ Aicda-Cre+/ki Ldlr/ chi- (V) and joining (J) gene J usage between the MDA-LDL + CFA meras) were used as controls (Figure 5A). Prdm1+/+ Aicda-Cre+/ki and CFA repertoires (Figures S4A–S4F), with the VH2 family be- Ldlr/ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras had compa- ing more represented in MDA-LDL + CFA mice, and no differ- rable proportions of CD45.2+ donor-derived cells and of residual ences in CDR3 length (Figures S4G and S4H). Conversely, VH5 CD45.1+ host-derived cells (Figures S5A and S5B). Following and VH10 families were under-represented in the MDA-LDL + bone marrow reconstitution, chimeras were fed with a high-fat CFA repertoire (Figure S4A) diet (HFD) for 9 weeks. We found no significant differences in Regarding SHM, we did not observe significant differences in weight gain (Figure S5C) or plasma lipid profile (Table S4) be- the proportion of mutated Igs or in their overall mutation fre- tween the two groups. quency between the CFA and MDA-LDL + CFA groups Expectedly, the proportion of CD138+ PCs was reduced in the (Figures 4C and 4D). However, while GC, MBC, and PC subsets spleens of Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras compared showed similar SHM frequencies in the CFA group, in MDA-LDL- with control mice (Figure 5B). This reduction was more severe immunized mice, Igs from GC B cells were the most mutated, fol- in IgG2b isotype-switched PCs (Figure 5C), suggesting that the lowed by the Igs from PCs, and with Igs from MBCs harboring CD138+IgM+ PCs remaining in Prdm1fl/fl Aicda-Cre+/ki Ldlr/ the fewest mutations (Figure 4E). Moreover, only IgG1+ cells, chimeras probably represent PCs derived from extrafollicular B but no other isotypes, showed a significant increase of SHM cells. Likewise, plasma IgM and IgG antibody titers were in MDA-LDL-immunized mice compared with control mice reduced in Prdm1-deficient chimeras (Figures 5D and 5E). These (E and F) Representative FACS plots of bone marrow (BM) tdTom+ PCs (B220CD138+tdTom+) 4 weeks after the primary immunization (E) and quantification (F). GC and MBC FACS analysis was performed on 6- to 8-week-old male (n = 27) and female (n = 21) mice. PC FACS analysis was performed on 6- to 8-week-old male (n = 16) and female (n = 8) mice. Each dot in the graphs represents a biological replicate (individual mouse). Population percentages in FACS graphs indicate the frequency of live cells. Data analyzed by two-way ANOVA statistical test. *p % 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Cell Reports 41, 111468, October 11, 2022 5 ll OPEN ACCESS Article Figure 3. Anti-MDA-LDL antibodies recognize epitopes only presented in the MDA-modified LDL Plasma antibody levels in Aicda-Cre+/ki R26tdTom+/ki mice immunized with CFA alone or MDA-LDL + CFA as determined by ELISA. (A–D) Quantification of (A) IgM, (B) IgG1, (C) IgG2b, and (D) IgG2c anti-MDA-LDL antibodies 2, 4, 6, and 8 weeks after primary immunization. (E) Quantification of IgM and IgG anti-MDA-BSA antibodies 6 weeks after primary immunization. (F) Competition immunoassay to measure anti-MDA-LDL IgG1 antibodies. Plasma IgG1 binding to plated MDA-LDL was quantified in the presence (B) or absence (B0) of increasing concentrations (0, 25, 50, 100, 150, 200, and 300 mg/mL) of the indicated competitors (BSA, MDA-BSA, LDL, and MDA-LDL). Results are expressed as ratios of IgG1 binding to MDA-LDL in the presence (B) or absence (B0) of the competitor. Each dot in the graphs represents a biological replicate (individual mouse). Data analyzed by one-way and two-way ANOVA statistical tests. *p % 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. results indicate that a large proportion of PCs, including those que composition and stability as measured by staining of collagen derived from GC B cells, are depleted in our Prdm1fl/fl Aicda- content and necrotic core analysis (Figures S5D–S5G), as well as Cre+/ki Ldlr/ model. macrophage and smooth muscle cell (SMC) immunofluorescence To assess the contribution of GC-derived PC on atherosclerosis, (Figures S5H–S5J). We conclude that GC-derived PCs and/or an- we evaluated atherosclerosis progression in Prdm1fl/fl Aicda- tibodies protect from atherosclerosis development. Cre+/ki Ldlr/ and Prdm1+/+ Aicda-Cre+/ki Ldlr/ chimeras. Atherosclerosis quantification in aortic sinus revealed a significant GC-derived antibodies contribute to MDA-LDL increase in atherosclerosis in Prdm1-deficient chimeras immunization-mediated athero-protection compared with in controls (Figures 5F–5H). Atherosclerosis accel- To assess the role of GC-derived antibodies in MDA- eration did not associate with obvious differences in atheroma pla- LDL-mediated athero-protection, we immunized Prdm1-deficient 6 Cell Reports 41, 111468, October 11, 2022 ll Article OPEN ACCESS (legend on next page) Cell Reports 41, 111468, October 11, 2022 7 ll OPEN ACCESS Article pro-atherogenic chimeras with MDA-LDL. Briefly, we transferred DISCUSSION bone marrow from Prdm1fl/fl Aicda-Cre+/ki or Prdm1+/+ Aicda- Cre+/ki CD45.2 mice into Ldlr/ CD45.1 mice, and reconstituted Here, we have shown that MDA-LDL immunization of proathero- chimeras were immunized with MDA-LDL + CFA, CFA, or PBS genic Ldlr/ mice triggers an Ab immune response that entails and fed with an HFD as indicated in Figure 6A. Chimeras from the generation of GCs, switched Abs, MBCs, and PCs. Our ex- all 6 groups showed similar weight gain (Figure S6). Plasma levels periments in Aicda-Cre+/ki R26tdTom+/ki mice have allowed us of total cholesterol, free cholesterol, LDL cholesterol, and triglyc- to trace activated B cells after MDA-LDL immunization and eride pro-atherogenic indexes (Talayero and Sacks, 2011), but not have confirmed this GC response in a non-atherogenic back- high-density lipoprotein (HDL) cholesterol, tended to be reduced ground, thus eliminating the confounding effects that the athero- in CFA- and MDA-LDL + CFA-immunized chimeras (Figure S7). sclerosis immune response might have. Interestingly, we de- Reduction in pro-atherogenic lipid levels was observed both in tected Abs specific of the MDA moiety in the context of the LDL Prdm1-proficient and -deficient chimeras, indicating that modula- particle. Thus, although native LDL by itself can raise an immune tion of the plasma lipid profile seems to be independent of GC- response (Gistera et al., 2018), our data support that MDA-modi- derived antibodies (Abs). fied LDL is indeed immunogenic, in agreement with previous data Expectedly, we found that the PC population was reduced (Freigang et al., 1998). Of note, our results were obtained by in the spleen of MDA-LDL + CFA-immunized Prdm1fl/fl Aicda- comparing MDA-LDL in CFA with mice immunized with CFA Cre+/ki Ldlr/ compared with immunized control chimeras alone, which is also known to promote athero-protection (Hansen (Figure 6B). Plasma titers of IgG Abs were also reduced in et al., 2001). Interestingly, MDA-LDL promotes an increase in the MDA-LDL-immunized Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras load of SHM, particularly in expanded clones of B cells, strongly (Figure 6C), while total IgM titers were not significantly changed suggesting that affinity maturation is in place in MDA-LDL-immu- (Figure 6D). Likewise, Prdm1-deficient pro-atherogenic chi- nized mice, and, therefore, that MDA-LDL, but not CFA alone, meras showed a sharp decrease in anti-MDA-LDL IgG1, triggers a bona fide GC response, much like other antigenic trig- IgG2b, and IgG2c Ab titers (Figures 6E–6G) and a mild reduc- gers used for vaccination strategies (Gómez-Escolar et al., 2021; tion of anti-MDA-LDL IgM Ab titers (Figure 6H). These results Weisel et al., 2016). In this regard, it is worth mentioning that the indicate that PCs derived from GCs play a critical role in the Goal of Oxidized LDL and Activated Macrophage Inhibition by Ab response triggered by MDA-LDL immunization in athero- Exposure to a Recombinant Antibody (GLACIER) trial was sclerotic mice. approved to assess the athero-protective properties of passive To test the role of GC-derived Abs in atherosclerosis, we then immunization with the monoclonal anti-MDA MLDL1278A Ab. measured atherosclerotic extent in aortic sinus of Prdm1+/+ The results did not reveal athero-protective effects, at least in Aicda-Cre+/ki Ldlr/ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chi- the cohort of subjects with stable CVD analyzed in the trial meras immunized with MDA-LDL in CFA or CFA alone (experi- (Lehrer-Graiwer et al., 2015; Nilsson and Hansson, 2020). How- ment depicted in Figure 6A; Figures 7A and 7B). We found that ever, it is tempting to suggest that alternative Abs can prove valu- both Prdm1+/+ Aicda-Cre+/ki Ldlr/ and Prdm1fl/fl Aicda-Cre+/ki able candidates in future clinical trials. Therefore, evaluation of Ldlr/ chimeras treated with CFA developed atherosclerotic athero-protection after passive transfer of Abs from expanded plaques of similar size (Figure 7C), indicating that CFA athero- clones identified in our study, either individually or in combina- protection is independent of GC-derived Abs. In contrast, upon tion, can provide proof of concept for such strategies. MDA-LDL immunization, Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chi- In our study, we have addressed the general contribution of meras had larger aortic plaque lesions than did Prdm1-proficient GC-derived Abs using Prdm1fl/fl Aicda-Cre+/ki mice. In this model, chimeras (Figure 7D). Therefore, these results indicate that PCs cells that have activated AID expression are depleted of Prdm1. and Abs derived from GC-experienced B cells drive MDA-LDL- We found that Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras showed mediated athero-protection. increased atherosclerosis, indicating an athero-protective effect Figure 4. Single-cell antibody sequencing from single cells of MDA-LDL-immunized mice (A) GC B cells and MBCs from LNs and GC-derived PCs (PC tdTom+) from BM of CFA- (n = 2) and MDA-LDL + CFA-immunized (n = 3) mice (3 times every 2 weeks) were isolated 2 weeks after last boost by flow cytometric single-cell sorting. IgH and IgL genes were amplified and sequenced. Paired Igh and Igk/Igl gene sequences from a total of 597 Igs were obtained and analyzed. (B) Population distribution of all Igs sequenced within each group. (C) Proportion of mutated (SHM) and germ line (GL) Igs in CFA and MDA-LDL + CFA repertoires. (D–F) Number of IgH somatic mutations (nucleotide exchanges) within mutated Igs in (D) all cells, (E) by cell population, and (F) in IgM and IgG1 B cells from each group. (G) Circos plots illustrate the molecular features of expanded Igs. Igs from clonally expanded cell clones are represented as boxes filled in green (CFA exclusive clones), red (MDA-LDL + CFA exclusive clones), and both colors (shared clones). Clone number is shown in the outer layer. Cell population origin, CSR, and number of Ig mutations (SHM) (nucleotide exchanges) are indicated for each B cell. (H) Population distribution of Igs present in exclusive clones from each group. (I) Ratio of expansion (expanded cells/all cells) of each cell population. (J) Number of somatic mutations (nucleotide exchanges) per mutated non-expanded Ig or expanded Ig in each group. Single-cell antibody sequencing experiment was performed using samples from 2 and 3 mice immunized with CFA or MDA-LDL + CFA, respectively. Data analyzed by non-parametric Mann-Whitney statistical test and chi-square test (A–F) and by non-parametric Kruskal-Wallis statistical, Fisher’s exact, and chi-square tests (G–J). *p % 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. 8 Cell Reports 41, 111468, October 11, 2022 ll Article OPEN ACCESS Figure 5. Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras show increased atherosclerosis (A) Experimental design. BM from Prdm1+/+ Aicda-Cre+/ki or Prdm1fl/fl Aicda-Cre+/ki mice was isolated and intravenously (i.v.) transferred to previously irradiated Ldlr/ mice. Four weeks after BM transfer, mice were fed with a 0.2% cholesterol HFD for 9 weeks. (B and C) Quantification of (B) PC percentage (B220/lowCD138+) and (C) IgM-positive (B220/lowCD138+IgM+) and IgG2b-positive (B220/lowCD138+IgG2b+) PCs in SP. (D and E) Quantification of total (D) IgM and (E) IgG antibody titers in plasma at day 0 and after 4 and 9 weeks of HFD as measured by ELISA. (F and G) Representative oil red O/hematoxylin stains (F) and quantification (G) of the proportion of atheroma plaque in serial 80 mm-spaced aortic sinus cry- osections from Prdm1+/+ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras. (H) Area under the curve (AUC) plot of atherosclerosis quantification. The scale bars in images represent 200 mm. The experiment was performed with 18 male and 6 female 11-week-old receptors. Each dot in the graphs represents a biological replicate (individual mouse). Data analyzed by unpaired t statistical test and two-way ANOVA statistical test. *p % 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. of GC-derived PCs and/or Abs. Two Prdm1 models have been PCs in Prdm1fl/fl Aicda-Cre+/ki mice. Therefore, it is possible previously reported in the context of atherosclerosis with diverse that Ab-secreting cells depleted in the Cd19-Cre or Cd23-Cre results, which may be explained by the different experimental models—but not in Prdm1fl/fl Aicda-Cre+/ki mice—have pro- strategies for Prdm1 deletion. In one study, Prdm1 deletion was atherogenic properties. Interestingly, abolishing Ab secretion driven by a Cd23-Cre, thus driving effective depletion of Blimp1 by deletion of Xbp-1 with mb1-Cre, which is expressed from early from all follicular-derived B cells (Tay et al., 2018). In a second B cell precursors, resulted in increased atherosclerosis, as did study, Prdm1 deletion was driven by the Cd19-Cre allele, leading depletion of T cells with anti-CD4 Abs (Sage et al., 2017). These to Blimp1 deletion in the vast majority of B cells (Centa et al., results suggest that the endogenous T cell-dependent response 2019). Although the onset of AID expression shortly precedes B can be athero-protective (Sage et al., 2017), in agreement with cell entry into the GC, our approach allows a more restricted our results. In addition, the differential impact on atherosclerosis depletion of GC-derived PCs than previous reports. This is further observed in these models may be the consequence of a subtle indicated by the persistence of a considerable fraction of IgM+ functional balance between Abs with different antigen Cell Reports 41, 111468, October 11, 2022 9 ll OPEN ACCESS Article Figure 6. Low MDA-LDL-specific antibody responses in the absence of GC-derived PCs (A) Experimental design. BM from Prdm1+/+ Aicda-Cre+/ki or Prdm1fl/fl Aicda-Cre+/ki mice was i.v. transferred to into irradiated Ldlr/recipients. Four weeks later, mice were immunized with MDA-LDL + CFA, CFA alone, or PBS and fed an HFD with 1.25% cholesterol for 15 weeks. Twenty-three weeks after the primary immunization, mice were sacrificed for analysis. (B) Quantification of SP PCs (B220/lowCD138+) from MDA-LDL + CFA-immunized Prdm1+/+ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras by FACS. (C and D) ELISA-based quantification of total (C) IgG and (D) IgM antibody levels in plasma from Prdm1+/+ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras immunized with MDA-LDL + CFA. (E–H) MDA-LDL-specific (E) IgG1, (F) IgG2b, (G) IgG2c, and (H) IgM antibody levels in Prdm1+/+ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras immunized with MDA-LDL + CFA measured by ELISA. All experiments were performed with 8- to 9-week-old male receptors. Each dot in the graphs represents a biological replicate (individual mouse). Data analyzed by unpaired t statistical test. *p % 0.05, **p < 0.01, and ****p < 0.0001. 10 Cell Reports 41, 111468, October 11, 2022 ll Article OPEN ACCESS Figure 7. GC-derived PCs and antibodies are required for MDA-LDL immunization-induced athero-protection (A and B) Representative oil red-O-stained aortic sinus cryosections of (A) Prdm1+/+ Aicda-Cre+/ki Ldlr/ chimeras and (B) Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras immunized with PBS, CFA, or MDA-LDL + CFA. (C and D) AUC values of the atherosclerosis quantifications in five serial 80 mm-spaced aortic sinus cryosections from Prdm1+/+ Aicda-Cre+/ki Ldlr/ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras immunized with (C) CFA or (D) MDA-LDL + CFA. Atherosclerosis quantification of CFA- and MDA-LDL + CFA-immunized mice was normalized to the mean AUC value of their respective PBS-injected Prdm1+/+ Aicda-Cre+/ki Ldlr/ and Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras. All experiments were performed with 8- to 9-week-old male receptors. Each dot in the graphs represents a biological replicate (individual mouse). Scale bars in images represent 200 mm. Data analyzed by unpaired t statistical test. *p < 0.05. specificities, some of which can have athero-protective proper- the immune response to MDA-LDL and its impact on the athero- ties, while others can aggravate the disease. For instance, anti- sclerosis process has hindered further progress of this potential ALDH4A1 Abs are athero-protective (Lorenzo et al., 2021), while therapeutic avenue. Here, we have shown that immunization with anti-Hsp60 Abs aggravate atherosclerosis (Foteinos et al., 2005). MDA-LDL promotes a bona fide GC response and generates In this regard, we have assessed the role of MDA-LDL-specific MBCs, PCs, and highly mutated anti-MDA-LDL Abs, which are PCs and/or Abs on atherosclerosis development by MDA-LDL important drivers of MDA-LDL-derived athero-protection. Our immunization of pro-atherogenic mice under deletion of Blimp1 data thus indicate that MDA-LDL immunization can warrant a in GC-derived PCs (Prdm1fl/fl Aicda-Cre+/ki Ldlr/ chimeras). high-affinity, long-lived Ab immune response, possibly comparable Depletion of GC-derived PCs dramatically decreased the titers to the response elicited by antigens used in vaccination protocols. of IgG Abs in MDA-LDL-immunized mice, and MDA-LDL-spe- cific IgG Abs were barely detectable. In contrast, titers of total Limitations of the study IgM Abs in MDA-LDL-immunized Prdm1fl/fl Aicda-Cre+/ki Ldlr/ A limitation of our study is the Aicda-Cre+/ki mouse model used to chimeras remain unchanged compared with PC-proficient mice, track GC B cells and to deplete Prmd1. As mentioned above, the and anti-MDA-LDL-specific IgM Abs were only mildly reduced. onset of AID expression shortly precedes the actual entering of the Therefore, our approach allows a precise study of Ab contribu- B cell into the follicle, and thus a minor fraction of non-GC B cells tion to MDA-LDL protection because it successfully abolishes could be labeled by our tdTom-based tracing system or be Prdm1 the generation of GC-derived, switched Abs while retaining a depleted. Nevertheless, this model provides a very good approx- substantial proportion of IgM Abs, presumably derived from ex- imation for GC studies because it is very efficient at depleting/la- trafollicular B cell activation. Our results show that MDA-LDL im- beling GC B cells while having only a minor effect on non-GC cells. munization failed to be athero-protective Prdm1-deficient chi- meras. Thus, our study indicates that MDA-LDL immunization STAR+METHODS gives rise to athero-protective Abs of GC origin. Athero-protection by MDA-LDL immunization was first Detailed methods are provided in the online version of this paper described more than two decades ago (Palinski et al., 1995), open- and include the following: ing new therapeutic perspectives based on thorough immunomo- dulation strategies and most specifically via vaccination strategies d KEY RESOURCES TABLE using MDA-LDL. However, our limited knowledge on the nature of d RESOURCE AVAILABILITY Cell Reports 41, 111468, October 11, 2022 11 ll OPEN ACCESS Article B Lead contact REFERENCES B Materials availability B Data and code availability Ait-Oufella, H., Herbin, O., Bouaziz, J.D., Binder, C.J., Uyttenhove, C., Laur- ans, L., Taleb, S., Van Vré, E., Esposito, B., Vilar, J., et al. (2010). B cell deple- d EXPERIMENTAL MODELS AND SUBJECT DETAILS tion reduces the development of atherosclerosis in mice. J. Exp. Med. 207, B Mice 1579–1587. https://doi.org/10.1084/jem.20100155. d METHOD DETAILS Ait-Oufella, H., Salomon, B.L., Potteaux, S., Robertson, A.K.L., Gourdy, P., B Antigen preparation and immunization protocols Zoll, J., Merval, R., Esposito, B., Cohen, J.L., Fisson, S., et al. (2006). Natural B Flow cytometry regulatory T cells control the development of atherosclerosis in mice. Nat. B Bone marrow transfer Med. 12, 178–180. https://doi.org/10.1038/nm1343. B Single-cell sorting and sequencing of immunoglobulin Albright, A.R., Kabat, J., Li, M., Raso, F., Reboldi, A., and Muppidi, J.R. (2019). transcripts TGFbeta signaling in germinal center B cells promotes the transition from light B Enzyme-linked immunosorbent assay (ELISA) for zone to dark zone. J. Exp. Med. 216, 2531–2545. https://doi.org/10.1084/jem. 20181868. immunoglobulin detection Almanzar, G., Öllinger, R., Leuenberger, J., Onestingel, E., Rantner, B., Zehm, B Quantification of atherosclerosis progression S., Cardini, B., van der Zee, R., Grundtman, C., and Wick, G. (2012). Autoreac- B Necrotic core analysis tive HSP60 epitope-specific T-cells in early human atherosclerotic lesions. B Immunofluorescence J. Autoimmun. 39, 441–450. https://doi.org/10.1016/j.jaut.2012.07.006. d QUANTIFICATION AND STATISTICAL ANALYSIS Binder, C.J., Hartvigsen, K., Chang, M.K., Miller, M., Broide, D., Palinski, W., Curtiss, L.K., Corr, M., and Witztum, J.L. (2004). IL-5 links adaptive and natural SUPPLEMENTAL INFORMATION immunity specific for epitopes of oxidized LDL and protects from atheroscle- rosis. J. Clin. Invest. 114, 427–437. https://doi.org/10.1172/JCI20479. Supplemental information can be found online at https://doi.org/10.1016/j. Bobryshev, Y.V., Ivanova, E.A., Chistiakov, D.A., Nikiforov, N.G., and Orekhov, celrep.2022.111468. A.N. (2016). Macrophages and their role in atherosclerosis: pathophysiology and transcriptome analysis. BioMed Res. Int. 2016, 9582430. https://doi.org/ ACKNOWLEDGMENTS 10.1155/2016/9582430. Brophy, M.L., Dong, Y., Wu, H., Rahman, H.N.A., Song, K., and Chen, H. We thank all members of the B Cell Biology Laboratory for useful discussions, A. (2017). Eating the Dead to keep atherosclerosis at bay. Front. Cardiovasc. Rodriguez-Ronchel for the design of the graphical abstract, V.G. de Yébenes for Med. 4, 2. https://doi.org/10.3389/fcvm.2017.00002. critical reading of the manuscript, and V. Labrador for help with microscopy and image analysis. I.M.-F. and C.L. were fellows of the research training program Busse, C.E., Czogiel, I., Braun, P., Arndt, P.F., and Wardemann, H. (2014). Sin- funded by Ministerio de Economı́a y Competitividad (SVP-2014-068216 and gle-cell based high-throughput sequencing of full-length immunoglobulin SVP-2014-068289, respectively), A.d.M.M. is funded by ‘‘la Caixa’’ Foundation heavy and light chain genes. Eur. J. Immunol. 44, 597–603. https://doi.org/ HR17–00247, and A.R.R. was supported by Centro Nacional de Investigaciones 10.1002/eji.201343917. Cardiovasculares (CNIC). The project leading to these results has received fund- Caligiuri, G., Nicoletti, A., Poirier, B., and Hansson, G.K. (2002). Protective im- ing from la Caixa Banking Foundation under the project code HR17-00247 and munity against atherosclerosis carried by B cells of hypercholesterolemic from SAF2016-75511-R and PID2019-106773RB-I00/AEI/10.13039/501100 mice. J. Clin. Invest. 109, 745–753. https://doi.org/10.1172/JCI7272. 011033 grants to A.R.R. (Plan Estatal de Investigación Cientı́fica y Técnica y Cedó, L., Metso, J., Santos, D., Garcı́a-León, A., Plana, N., Sabate-Soler, S., de Innovación 2013–2016 Programa Estatal de I+D+i Orientada a los Retos de Rotllan, N., Rivas-Urbina, A., Méndez-Lara, K.A., Tondo, M., et al. (2020). la Sociedad Retos Investigación: Proyectos I + D + i 2016, Ministerio de Econo- LDL receptor regulates the reverse transport of macrophage-derived unesteri- mı́a, Industria y Competitividad) and co-funding by Fondo Europeo de Desarrollo fied cholesterol via concerted action of the HDL-LDL Axis: insight from mouse Regional (FEDER). CIBERCV (J.L.M.-V.) and CIBERDEM (J.C.E.-G.) are Instituto models. Circ. Res. 127, 778–792. https://doi.org/10.1161/CIRCRESAHA.119. de Salud Carlos III projects. The CNIC is supported by the Instituto de Salud Car- 316424. los III (ISCIII), the Ministerio de Ciencia e Innovación (MCIN), and the Pro CNIC Centa, M., Jin, H., Hofste, L., Hellberg, S., Busch, A., Baumgartner, R., Verzaal, Foundation and is a Severo Ochoa institute (CEX2020-001041-S grant funded N.J., Lind Enoksson, S., Perisic Matic, L., Boddul, S.V., et al. (2019). Germinal by MCIN/AEI/10.13039/501100011033). center-derived antibodies promote atherosclerosis plaque size and stability. Circulation 139, 2466–2482. https://doi.org/10.1161/CIRCULATIONAHA.118. AUTHOR CONTRIBUTIONS 038534. Clement, M., Guedj, K., Andreata, F., Morvan, M., Bey, L., Khallou-Laschet, J., I.M.-F. designed, performed, and analyzed experiments, prepared figures, and Gaston, A.T., Delbosc, S., Alsac, J.M., Bruneval, P., et al. (2015). Control of the wrote the manuscript; A.d.M.M. performed and analyzed experiments and T follicular helper-germinal center B-cell axis by CD8(+) regulatory T cells limits prepared figures; C.L. performed and analyzed experiments; C.E.B. per- atherosclerosis and tertiary lymphoid organ development. Circulation 131, formed and analyzed experiments; P.D. performed and analyzed experiments; 560–570. https://doi.org/10.1161/CIRCULATIONAHA.114.010988. L.C.-F. performed experiments; S.M.M. performed experiments; J.C.E.-G. provided reagents; J.L.M.-V. designed experiments; H.W. designed experi- Foteinos, G., Afzal, A.R., Mandal, K., Jahangiri, M., and Xu, Q. (2005). Anti-heat ments and wrote the manuscript; and A.R.R. conceived the project, designed shock protein 60 autoantibodies induce atherosclerosis in apolipoprotein and analyzed experiments, and wrote the manuscript. E-deficient mice via endothelial damage. Circulation 112, 1206–1213. https://doi.org/10.1161/CIRCULATIONAHA.105.547414. DECLARATION OF INTERESTS Fredrikson, G.N., Soderberg, I., Lindholm, M., Dimayuga, P., Chyu, K.Y., Shah, P.K., and Nilsson, J. (2003). Inhibition of atherosclerosis in apoE-null mice by The authors declare no competing interests. immunization with apoB-100 peptide sequences. Arterioscler. Thromb. Vasc. Biol. 23, 879–884. https://doi.org/10.1161/01.ATV.0000067937.93716. Received: March 31, 2022 Freigang, S., Hörkkö, S., Miller, E., Witztum, J.L., and Palinski, W. (1998). Im- Revised: August 13, 2022 munization of LDL receptor-deficient mice with homologous malondialde- Accepted: September 19, 2022 hyde-modified and native LDL reduces progression of atherosclerosis by Published: October 11, 2022 mechanisms other than induction of high titers of antibodies to oxidative 12 Cell Reports 41, 111468, October 11, 2022 ll Article OPEN ACCESS neoepitopes. Arterioscler. Thromb. Vasc. Biol. 18, 1972–1982. https://doi.org/ Ishibashi, S., Brown, M.S., Goldstein, J.L., Gerard, R.D., Hammer, R.E., and 10.1161/01.atv.18.12.1972. Herz, J. (1993). Hypercholesterolemia in low density lipoprotein receptor George, J., Afek, A., Gilburd, B., Levkovitz, H., Shaish, A., Goldberg, I., Kopo- knockout mice and its reversal by adenovirus-mediated gene delivery. J Clin lovic, Y., Wick, G., Shoenfeld, Y., and Harats, D. (1998). Hyperimmunization of Invest 92, 883–893. https://doi.org/10.1172/JCI116663. apo-E-deficient mice with homologous malondialdehyde low-density lipopro- Jonasson, L., Holm, J., Skalli, O., Bondjers, G., and Hansson, G.K. (1986). tein suppresses early atherogenesis. Atherosclerosis 138, 147–152. https:// Regional accumulations of T cells, macrophages, and smooth muscle cells doi.org/10.1016/s0021-9150(98)00015-x. in the human atherosclerotic plaque. Arteriosclerosis 6, 131–138. https://doi. George, J., Shoenfeld, Y., Afek, A., Gilburd, B., Keren, P., Shaish, A., Kopo- org/10.1161/01.atv.6.2.131. lovic, J., Wick, G., and Harats, D. (1999). Enhanced fatty streak formation in Karvonen, J., Päivänsalo, M., Kesäniemi, Y.A., and Hörkkö, S. (2003). Immuno- C57BL/6J mice by immunization with heat shock protein-65. Arterioscler. globulin M type of autoantibodies to oxidized low-density lipoprotein has an Thromb. Vasc. Biol. 19, 505–510. https://doi.org/10.1161/01.atv.19.3.505. inverse relation to carotid artery atherosclerosis. Circulation 108, 2107– Gisterå, A., Klement, M.L., Polyzos, K.A., Mailer, R.K.W., Duhlin, A., Karlsson, 2112. https://doi.org/10.1161/01.CIR.0000092891.55157.A7. M.C.I., Ketelhuth, D.F.J., and Hansson, G.K. (2018). Low-density lipoprotein- Kyaw, T., Cui, P., Tay, C., Kanellakis, P., Hosseini, H., Liu, E., Rolink, A.G., reactive T cells regulate plasma cholesterol levels and development of athero- Tipping, P., Bobik, A., and Toh, B.H. (2013). BAFF receptor mAb treatment sclerosis in humanized hypercholesterolemic mice. Circulation 138, 2513– ameliorates development and progression of atherosclerosis in hyperlipidemic 2526. https://doi.org/10.1161/CIRCULATIONAHA.118.034076. ApoE(-/-) mice. PLoS One 8, e60430. https://doi.org/10.1371/journal.pone. Gitlin, A.D., von Boehmer, L., Gazumyan, A., Shulman, Z., Oliveira, T.Y., and 0060430. Nussenzweig, M.C. (2016). Independent roles of switching and hypermutation Kyaw, T., Tay, C., Khan, A., Dumouchel, V., Cao, A., To, K., Kehry, M., Dunn, in the development and persistence of B lymphocyte memory. Immunity 44, R., Agrotis, A., Tipping, P., et al. (2010). Conventional B2 B cell depletion ame- 769–781. https://doi.org/10.1016/j.immuni.2016.01.011. liorates whereas its adoptive transfer aggravates atherosclerosis. J. Immunol. Gómez-Escolar, C., Serrano-Navarro, A., Benguria, A., Dopazo, A., Sánchez- 185, 4410–4419. https://doi.org/10.4049/jimmunol.1000033. Cabo, F., and Ramiro, A.R. (2021). Single cell clonal analysis identifies an AID- Kyaw, T., Tay, C., Krishnamurthi, S., Kanellakis, P., Agrotis, A., Tipping, P., Bo- dependent pathway of plasma cell differentiation. Preprint at bioRxiv. https:// bik, A., and Toh, B.H. (2011). B1a B lymphocytes are atheroprotective by doi.org/10.1101/2021.06.03.446763. secreting natural IgM that increases IgM deposits and reduces necrotic cores Grasset, E.K., Duhlin, A., Agardh, H.E., Ovchinnikova, O., Hägglöf, T., Forsell, in atherosclerotic lesions. Circ. Res. 109, 830–840. https://doi.org/10.1161/ M.N., Paulsson-Berne, G., Hansson, G.K., Ketelhuth, D.F.J., and Karlsson, CIRCRESAHA.111.248542. M.C.I. (2015). Sterile inflammation in the spleen during atherosclerosis pro- Lahoute, C., Herbin, O., Mallat, Z., and Tedgui, A. (2011). Adaptive immunity in vides oxidation-specific epitopes that induce a protective B-cell response. atherosclerosis: mechanisms and future therapeutic targets. Nat. Rev. Cardiol. Proc. Natl. Acad. Sci. USA 112, E2030–E2038. https://doi.org/10.1073/pnas. 8, 348–358. https://doi.org/10.1038/nrcardio.2011.62. 1421227112. Lehrer-Graiwer, J., Singh, P., Abdelbaky, A., Vucic, E., Korsgren, M., Baruch, Gupta, S., Pablo, A.M., Jiang, X.c., Wang, N., Tall, A.R., and Schindler, C. A., Fredrickson, J., van Bruggen, N., Tang, M.T., Frendeus, B., et al. (2015). (1997). IFN-gamma potentiates atherosclerosis in ApoE knock-out mice. FDG-PET imaging for oxidized LDL in stable atherosclerotic disease: a phase J. Clin. Invest. 99, 2752–2761. https://doi.org/10.1172/JCI119465. II study of safety, tolerability, and anti-inflammatory activity. JACC. Cardio- Hansen, P.R., Chew, M., Zhou, J., Daugherty, A., Heegaard, N., Jensen, P., vasc. Imaging 8, 493–494. https://doi.org/10.1016/j.jcmg.2014.06.021. Mouritsen, S., and Falk, E. (2001). Freunds adjuvant alone is antiatherogenic Libby, P. (2021). The changing landscape of atherosclerosis. Nature 592, in apoE-deficient mice and specific immunization against TNFalpha confers 524–533. https://doi.org/10.1038/s41586-021-03392-8. no additional benefit. Atherosclerosis 158, 87–94. https://doi.org/10.1016/ Lorenzo, C., Delgado, P., Busse, C.E., Sanz-Bravo, A., Martos-Folgado, I., s0021-9150(01)00418-x. Bonzon-Kulichenko, E., Ferrarini, A., Gonzalez-Valdes, I.B., Mur, S.M., Rol- Hansson, G.K., and Hermansson, A. (2011). The immune system in atheroscle- dán-Montero, R., et al. (2021). ALDH4A1 is an atherosclerosis auto-antigen rosis. Nat. Immunol. 12, 204–212. https://doi.org/10.1038/ni.2001. targeted by protective antibodies. Nature 589, 287–292. https://doi.org/10. Hansson, G.K., Holm, J., and Jonasson, L. (1989). Detection of activated T 1038/s41586-020-2993-2. lymphocytes in the human atherosclerotic plaque. Am. J. Pathol. 135, Lusis, A.J. (2000). Nature 407, 233–241. https://doi.org/10.1038/35025203. 169–175. Madisen, L., Zwingman, T.A., Sunkin, S.M., Oh, S.W., Zariwala, H.A., Gu, H., Hansson, G.K., Robertson, A.K.L., and Söderberg-Nauclér, C. (2006). Inflam- Ng, L.L., Palmiter, R.D., Hawrylycz, M.J., Jones, A.R., et al. (2010). A robust mation and atherosclerosis. Annu. Rev. Pathol. 1, 297–329. https://doi.org/10. and high-throughput Cre reporting and characterization system for the whole 1146/annurev.pathol.1.110304.100100. mouse brain. Nat. Neurosci. 13, 133–140. https://doi.org/10.1038/nn.2467. Hartley, A., Haskard, D., and Khamis, R. (2019). Oxidized LDL and anti- Major, A.S., Fazio, S., and Linton, M.F. (2002). B-lymphocyte deficiency in- oxidized LDL antibodies in atherosclerosis - novel insights and future direc- creases atherosclerosis in LDL receptor-null mice. Arterioscler. Thromb. tions in diagnosis and therapy