|Home | About | Journals | Submit | Contact Us | Français|
Explore aorta B-cell immunity in aged apolipoprotein E-deficient (ApoE−/−) mice.
Transcript maps, fluorescence-activated cell sorting, immunofluorescence analyses, cell transfers, and Ig-ELISPOT (enzyme-linked immunospot) assays showed multilayered atherosclerosis B-cell responses in artery tertiary lymphoid organs (ATLOs). Aging-associated aorta B-cell–related transcriptomes were identified, and transcript atlases revealed highly territorialized B-cell responses in ATLOs versus atherosclerotic lesions: ATLOs showed upregulation of bona fide B-cell genes, including Cd19, Ms4a1 (Cd20), Cd79a/b, and Ighm although intima plaques preferentially expressed molecules involved in non–B effector responses toward B-cell–derived mediators, that is, Fcgr3 (Cd16), Fcer1g (Cd23), and the C1q family. ATLOs promoted B-cell recruitment. ATLO B-2 B cells included naive, transitional, follicular, germinal center, switched IgG1+, IgA+, and IgE+ memory cells, plasmablasts, and long-lived plasma cells. ATLOs recruited large numbers of B-1 cells whose subtypes were skewed toward interleukin-10+ B-1b cells versus interleukin-10− B-1a cells. ATLO B-1 cells and plasma cells constitutively produced IgM and IgG and a fraction of plasma cells expressed interleukin-10. Moreover, ApoE−/− mice showed increased germinal center B cells in renal lymph nodes, IgM-producing plasma cells in the bone marrow, and higher IgM and anti–MDA-LDL (malondialdehyde-modified low-density lipoprotein) IgG serum titers.
ATLOs orchestrate dichotomic, territorialized, and multilayered B-cell responses in the diseased aorta; germinal center reactions indicate generation of autoimmune B cells within the diseased arterial wall during aging.
Beyond their ability to produce antibodies,1 B cells produce proinflammatory or anti-inflammatory cytokines,2,3 present antigen to T cells,4 and regulate B- and T-cell responses.5 Mature naive bone marrow (BM)–derived B-2 cells home into secondary lymphoid organs (SLOs) where they undergo somatic hypermutation and affinity maturation in germinal centers (GCs). Antigen-experienced B-2 cells either become short-lived plasma cells (PCs) residing in SLOs or they develop into long-lived PCs that largely home to the BM.6–8 By contrast, the majority of B-1 cells are located in the peritoneal cavity (PerC) and pleural cavities where they form a pool of quiescent innate B cells. On migration to inflammatory tissues, B-1 cells become activated and self-renew to carry out T-cell–independent protective immune responses.9–12 Recent reports showed differential effects of B-cell subsets in atherosclerosis13–24 with antiatherogenic effects of B-1 cells and proatherogenic effects of B-2 cells.25–27 In addition to SLOs and the BM, B-cell responses may be organized in artery tertiary lymphoid organs (ATLOs) in apolipoprotein E-deficient (ApoE−/−) mice.28,29 Here, we report on local aorta as opposed to systemic B-cell responses during aging.
Materials and Methods are available in the online-only Data Supplement.
MIAME (minimum information about a microarray experiment)-compliant microarrays were prepared as described30,31; data were deposited in the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/) and the gene ontology (http://www.geneontology.org/) data banks (accession GSE40156).30,32 To determine if B-cell–related gene expression changes with aging, microarrays of aortas, SLOs, and blood from wild-type (WT) and ApoE−/− mice were compared. B-cell–related genes were altered in WT aortas during aging (Table I in the online-only Data Supplement). However, there were much more pronounced changes in ApoE−/− when compared with WT aortas. Expression kinetics of some of these genes correlated with the kinetics of ATLO formation32,33 (Figure (Figure1;1; Table I in the online-only Data Supplement). B-cell transcriptomes contained genes that were expressed exclusively by B cells and a majority of genes that respond to B-cell–derived molecules yielding a complex B-cell immunity–related gene map (Figure (Figure1;1; Table I in the online-only Data Supplement). Examples of the magnitude of B-cell immunity–related transcripts in ApoE−/− aortas include a 135-fold increase of Ighm (IgM constant region), a 29-fold increase in Ptpn6 (protein tyrosine phosphatase, nonreceptor type 6; SHP1) regulating the IgM repertoire, a 23-fold increase in the immunosuppressive Lilrb3 (leukocyte immunoglobulin-like receptor, subfamily B with transmembrane and immunoreceptor tyrosine-based inhibitory motif domains), Fcer1g (Fc receptor, IgE, high-affinity I, γ-polypeptide), and Cd28 (CD28 antigen) expression that promotes PC survival (Figure (Figure1;1; Table I in the online-only Data Supplement). In contrast, spleen- and blood-transcript maps were considerably smaller, and the extent of differential expression between WT and ApoE−/− mice was much less pronounced (Figure I in the online-only Data Supplement). The majority of B-cell–associated genes in the spleen and blood were downregulated during aging in both WT and ApoE−/− mice: Ptprc (B220; Cd45; protein tyrosine phosphatase, receptor type, C) involved in cell fate decisions of the B-cell receptor; Aicda (activation-induced cytidine deaminase) regulating somatic hypermutation and Ig class switching; Sykb (spleen tyrosine kinase) participating in B-memory cell survival; Vav3 (Vav3 oncogene) mediating B-cell receptor responses; Tcf3 (transcription factor 3) controlling B-cell ontogeny; Foxp1 (forkhead box p1) impacting B-cell survival; and Malt1 (Malt1 paracaspase) participating in B-cell malignancies. In summary, the spleen and blood gene maps suggested that age-associated changes largely mirrored B-cell senescence rather than genotype/hyperlipidemia-dependent changes (Figure I and Table I in the online-only Data Supplement).
Laser capture microdissection aorta-derived tissues were obtained together with renal lymph nodes (RLNs) and spleen.30,31 B-cell–related genes were expressed at higher levels in ATLOs when compared with aorta adventitia segments from WT or ApoE−/− mice without plaques (Figure (Figure2A;2A; Table I in the online-only Data Supplement). In the adventitia cluster, genes associated with B-cell survival, proliferation, differentiation, and activation, such as immunoglobulin genes (ighm), TACI (tnfrsf13b), B-cell activating factor receptor (tnfrsf13c), CD40 antigen (cd40), histocompatibility 2, class II antigen A, β-1 (h2-ab1), complement components (c1qb), and Myd88 (myd88) were robustly expressed in adventitial regions adjacent to plaques compared with adventitia in regions with no plaques (Figure (Figure2A;2A; Table I in the online-only Data Supplement). Moreover, the adventitia adjacent to plaques contained transcripts coding for Igj chain (immunoglobulin joining chain; Igj) involved in somatic hypermutation and memory B-cell development; CD79a (immunoglobulin-associated α; Ly54) involved in B-cell receptor signaling; and Ms4a1 (CD20) controlling T-cell–dependent humoral immunity (Figure IIA in the online-only Data Supplement). The plaque–ATLO cluster markedly expressed Cd19 (CD19 antigen) in ATLOs involved in B-cell maturation, Cd20, Igj chain, Igm, and Cd79a/b (Figure (Figure2B;2B; Figure IIB in the online-only Data Supplement). In addition, the plaque–ATLO B-cell cluster30,31 showed functional separation in B-cell–related genes in ATLOs versus plaques: bona fide B-cell genes displayed strong expression in ATLOs versus low expression in plaques. For example, Ighm, cd19, ms4a1 (cd20), Igj, and cd79a/b were expressed manifold higher in ATLOs when compared with plaques, which expressed genes that respond to B-cell products (Figure (Figure2A;2A; Figure IIB and Table I in the online-only Data Supplement). In contrast, the transcript atlas showed almost identical levels of B-cell–related genes in WT versus ApoE−/− spleens, RLNs, and blood (Figure I in the online-only Data Supplement; Figure Figure2C2C and 2D). It is also noticeable that the LN–ATLO cluster shows a comparably higher expression in ATLOs versus LNs of innate immune response genes, such as fcgr1, fcgr2b, fcgr3, c4b, and the c1q family, indicating ongoing inflammation in ATLOs (Figure (Figure2C2C and 2D).
B cells present in the aorta of aged ApoE−/− mice predominantly reside in ATLOs, whereas they cannot be observed in plaques of young WT or ApoE−/− mice.30,32,33 Fluorescence-activated cell sorting (FACS) analyses of B cells revealed the magnitude of differences in ATLOs and WT adventitia; and B220 immunostaining confirmed that B cells are located in ATLOs and in the adjacent draining LNs but none in WT adventitia or plaques (Figure (Figure3A3A and 3B). Considerable numbers of T-/B-cell clusters referred to as fat-associated lymphoid clusters were observed in paraaortic adipose tissue of aged ApoE−/− mice and numerous small paraaortic LNs containing B cells lined the tissue adjacent to the adventitia (not shown). There were no differences in the frequency of B cells in SLOs or blood of WT versus ApoE−/− mice (Figure (Figure3C).3C). To obtain evidence for an ongoing GC reaction in ATLOs, CD19, IgM, and IgD antisera together with FACS gating for 4 different populations from CD19+ B cells were used (Figure (Figure3D).3D). IgM+/IgD−, IgM+/IgD+, IgM−/IgD−, and IgM−/IgD+ B cells were identified in abdominal but not thoracic aorta segments: IgM+/IgD− cells represent either immature or transitional B cells (also referred to as T-1 cells) representing the earliest B-cell stage present outside the BM or these cells may represent B-1 B cells34; IgM+/IgD+ and IgM−/IgD+ cells represent mature B-cell stages.35,36 Among mature IgD+ cells, IgM−/IgD+ are mature follicular B-2 cells.37 IgM−/IgD− cells represent either switched Ig+ B cells, GC B cells that have transiently lost Ig expression when undergoing hypermutation of their Ig genes or GC-derived memory B cells.34,38 None of the subsets were found in the abdominal aorta of WT mice (Figure (Figure3D).3D). WT and ApoE−/− SLOs and blood revealed equivalent numbers of these subsets with the exception of an increase in transitional IgM+/IgD− B cells in RLNs of ApoE−/− versus WT mice (Figure (Figure3E).3E). We determined the percentages of IgM+/IgD+ or switched Ig+ B cells in SLOs, blood, WT aortas, and ATLOs. SLO and blood IgM+/IgD+ and switched Ig+ B cells were similar in WT and ApoE−/− SLOs (Figure (Figure3F3F and 3G). Although undetectable in WT adventitia, the percentage of IgM+/IgD+ B cells in ATLOs approached that in SLOs (Figure (Figure3F).3F). However, the percentage of switched Ig+ B cells in ATLOs exceeded those in SLOs or blood (Figure (Figure3G).3G). We determined the number of B-1 cells in the PerC and of plasmablasts and PCs in the abdominal aorta, spleen, and RLNs of ApoE−/− mice (Figure III in the online-only Data Supplement). No change in B-1 B cell subtypes was observed in the PerC of WT versus ApoE−/− mice (Figure IIIA in the online-only Data Supplement). Moreover, aged ApoE−/− abdominal aortas, spleens, and RLNs contained plasmablasts and PCs; some of which expressed interleukin (IL)-10 (Figure IIIB in the online-only Data Supplement).
Naive B cells in SLOs enter GCs to undergo a GC reaction involving somatic hypermutation and affinity maturation of their B-cell receptors. ATLO GC B cells were identified by FACS (IgD−/PNA+/GL-7+): they were undetectable in WT aortas but ranged at ≈9% of all IgD− B cells in ATLOs (Figure (Figure4A4A and 4B). Their number was similar in WT and ApoE−/− spleen and blood although they were more abundant in ApoE−/− when compared with WT RLNs (Figure (Figure4B).4B). We sought evidence for isotype-switching using FACS analyses. Surprisingly, we observed significant numbers of CD19+/IgD−/IgG1+, CD19+/IgD−/IgA+, and CD19+/IgD−/IgE+ B cells in ATLOs (Figure (Figure4C4C and 4D). Although class switching is not restricted to GCs, the presence of GCs and cells that class switched to T-dependent Ig subclasses, such as IgG1, suggests that these cells resemble memory B cells. Intriguingly, the percentage of IgD− B cells that class switched to IgG1 was significantly greater than those in the spleen, RLNs, BM, or blood (Figure (Figure4D).4D). In contrast, there were equivalent percentages of IgG1+ B cells in the spleen, BM, and blood of WT versus ApoE−/− mice. Consistent with rare ATLO formation in the thoracic aorta,32 no GC B cells or class switched B cells were observed there (not shown). These data provide evidence for a disease-specific antigen-dependent B-2 B cell maturation pathway in ATLOs.
Long-lived PCs are major constituents of humoral memory. Long-lived PCs preferentially home to the BM, whereas short-lived PCs remain within SLOs. Nothing is known about PCs in atherosclerosis. As long-lived PCs survive for long periods of time in the BM,39 we determined the composition of ATLO PC subtypes.8 Both long-lived and short-lived PCs were observed in ATLOs (Figure (Figure44E).40,41 Moreover, survival factors for long-lived PCs, including CXCL12, B-cell activating factor,39 and others, are markedly expressed in ATLOs30,32 (Table I in the online-only Data Supplement).
To determine B-cell recruitment by ATLOs, we adoptively transferred Ly5.1 B-2 cells to aged Ly5.2 WT or ApoE−/− mice. After 36 hours, B-2 cells had migrated predominantly to the abdominal aorta of ApoE−/− mice (Figure (Figure5A5A and 5B) although none were recruited to WT aortas. Comparably low but similar numbers of B-2 cells were recruited into the PerCs of WT and ApoE−/− mice. There was no difference in B-cell recruitment into the spleen, RLNs, and BM of WT versus ApoE−/− mice (Figure (Figure5C).5C). Similar data were obtained with B-1 cells (not shown).
B-1 cells are predominantly located in body cavities.42,43 Recent studies showed that B-1a cells reside in the aorta perivascular tissue of young ApoE−/− mice.22 To determine if B-1 cells are located in the aged aorta adventitia, we performed FACS analyses. A high percentage of all B cells, that is, ≈21%, in ATLOs were B-1 cells (Figure (Figure5D),5D), and their relative contribution to all B cells exceeded that in the spleen and RLNs by a large margin (Figure (Figure5D).5D). The reason for B-1 B-cell accumulation is most likely the high expression of CXCL13 in ATLOS. Numbers of total B-1 cells in ATLOs are comparable with that of IgM+/IgD− cells, indicating that most IgM+/IgD− cells found in this compartment are B-1 cells. The abdominal aorta harbored considerably higher numbers of B-1 cells when compared with the thoracic aorta (Figure (Figure5E).5E). The B-1 subtype composition was aberrant as we observed a high number of B-1b versus B-1a cells, which dramatically differs from that relation in the PerC (Figure (Figure55E).9 There was no significant difference in total B cells, B-2, B-1a, and B-1b cells in the PerC of aged WT and ApoE−/− mice (Figure (Figure5F5F and 5G).
In view of skewing of ATLO B-1 cells toward the B-1b subtype (Figure (Figure5D5D and 5E) and a recent report showing that B-1b cells protect against atherosclerosis,21 we searched for mechanisms of immunosuppression within the arterial wall. IL-10–producing B-1a rather than B-1b or B-2 cells were found in the PerC (Figure IIIA in the online-only Data Supplement). However, we observed that the majority (≈72%) of abdominal aorta B-1b cells produced IL-10 though a minor component of B-1a cells and a significant but low proportion of IL-10+ cells in the thoracic aorta (not shown). No or comparably low numbers of B-1a cells or B-2 cells expressed IL-10 (Figure (Figure6A).6A). Moreover, the frequency of IL-10+ B cells in ATLOs is higher than those of their counterparts in the spleen and RLNs of WT or ApoE−/− mice (Figure (Figure6B).6B). Following a report that a subset of PCs secretes IL-10,44 we assessed IL-10 expression in PCs. A significant proportion of ATLO CD138+/CD19+ plasmablasts were IL-10+ PCs (Figure IIIB in the online-only Data Supplement). Similar PCs have been shown to suppress immune responses in disease models.45 We further assessed the phenotype of B-1 cells in the abdominal aorta. ATLO B-1 but to a much lesser extent B-2 cells expressed PD-L1, FasL, and transforming growth factor-β, indicating that these cells exert immunosuppressive functions (Figure (Figure66A).
ELISPOT (enzyme-linked immunospot) experiments were performed. There were no constitutively IgM- and IgG-secreting cells in either the thoracic or abdominal aorta of WT mice (Figure (Figure7A7A and 7B). Few IgM- and IgG-secreting cells were observed in the thoracic aorta of ApoE−/− mice (Figure (Figure7A7A and 7B). However, ATLOs contained abundant IgM- and IgG-secreting cells amounting to ≤80-fold increase of IgM-secreting B cells and a 24-fold increase in IgG-secreting B cells in the abdominal aorta (Figure (Figure7A7A and 7B). Blood contains few (<10 cells per 0.5 mL of blood) IgM- or IgG-secreting cells (data not shown). In the spleen and RLNs, there was no difference in Ig-secreting cells between WT and ApoE−/− mice (Figure (Figure7C).7C). However, IgM-secreting cells were higher in ApoE−/− BM when compared with WT BM raising the possibility of a systemic PC response in ApoE−/− mice. To examine a systemic B-cell response, we determined serum titers of IgM, IgG, and IgE, as well as anti–malondialdehyde-modified low-density lipoprotein (MDA-LDL) IgM and IgG. Aged ApoE−/− mice had significantly higher levels of total IgM but not IgG or IgE levels when compared with aged WT mice (Figure (Figure7D).7D). Although anti–MDA-LDL IgM levels were not different, anti–MDA-LDL IgG levels were significantly higher in ApoE−/− versus WT mice (Figure (Figure77E).
These data identify ATLOs as the principal lymphoid tissue that orchestrates atherosclerosis B-cell immunity during aging of ApoE−/− mice. Atherosclerosis ATLO B-cell responses are specific, robust, highly territorialized, multilayered, and include a comprehensive adaptive B-2 and a substantial aberrant innate B-1 cell component: ATLOs but not WT adventitia harbor an unusual set of class-switched IgG1+, IgA+, and IgE+ B cells, a significant number of IL-10+/PD-L1+/FasL+/TGFβ+ B-1b cells, and both short-lived and long-lived PCs, including a fraction of IL-10+ PCs. This body of data—together with our previous observation that B cells are major constituents of ATLO antigen-presenting cells30—reveal a yet unrecognized scenario of aorta atherosclerosis-specific B-cell immunity, which includes B effector cells, PCs, and several immunosuppressive B-cell subtypes (Figure IV in the online-only Data Supplement).
ATLO B-2 B-cell subtypes include transitional, follicular, GC, and IgG1+, IgA+, and IgE+ B cells—the latter representing class-switched B cells and PCs. These data are the first to suggest that (auto)antigen-dependent hypermutation, proliferation, affinity maturation, Ig class switching, memory cell generation, and differentiation into long-lived PCs may be carried out in the arterial wall. It is becoming evident that ATLOs provide a new paradigm of atherosclerosis-specific B-cell immunity and possibly autoimmunity: ATLO B-cell responses occur in aged animals, whereas aortas of young ApoE−/− or young and aged WT mice do not show a significant aorta B-cell compartment.30,46–48 It should be pointed out, however, that this study falls short of proving antigen-specific ATLO-dependent autoimmune B-2 B-cell generation. In this regard, the observation of a considerable number of PCs in ATLOs deserves special attention: PCs may arise from B-1 cells, from B-2 cells via T-cell–independent mechanisms, or from B-2 cells via T-cell–dependent mechanisms.49 Further studies on the origin of aorta PCs seem warranted as the role of PCs in atherosclerosis remains unknown.
Our data demonstrate that local B-cell immune subsets can be distinguished from those in SLOs, the PerC, and the BM: their aberrant nature manifests itself by the presence of large numbers of IL-10+ B-1b cells, of short-lived and long-lived PCs, and of IL-10+ PCs. Possibly, our aged mice will allow to isolate B cells from ATLOs and SLOs to compare their B-cell receptor repertoire. Moreover, the accumulation of IgA+ and IgE+ B cells in the diseased aorta indicates links of atherosclerosis B-cell immunity to innate inflammatory leukocytes in plaques. IgA, IgE, and IgG act through either activating or inhibitory Fc receptors on virtually all innate immune cells, including macrophages.50 The expression of divergent Fc receptors raises the possibility that Fc receptors may be involved in the dichotomic control of inflammation within diseased arteries: Fcer1g (Cd23) is a high-affinity IgE receptor that is upregulated during aging, and Fcgr1 (Cd64), Fcgr2b (Cd32), and Fcgr3 (Cd16) are prominently expressed in ATLOs.
ATLOs contain multiple B-cell subtypes, including IgM+/IgD−, IgM+/IgD+, IgM-/IgD−, and IgM−/IgD+ B cells. ATLO IgM+/IgD− and IgM+/CD43+ B cells may be B-1 cells. In addition, the presence of class-switched memory B cells suggests that some ATLO IgM+/IgD− B cells may represent IgM+ memory B cells that have not undergone class switching. IgM+ memory B cells considerably contribute to the total population of all memory B cells.51 Whether the population of IgM+/IgD− B cells within ATLOs also includes a fraction of immature or transitional B cells that represent the earliest B-cell stages that are found outside the BM is a possibility that deserves attention. Under physiological conditions, immature B cells immigrate from the BM and specifically home to splenic follicles to undergo differentiation into a transitional B-cell stage and finally either become mature B-2 or marginal zone B cells.52 This final B-cell maturation is accompanied by a shift of the B-cell receptor repertoire that includes counterselection against autoreactive cells that occurs in discrete and tightly controlled steps.53 Hence, it is tempting to speculate that immature B cells home to ATLOs to undergo differentiation into mature B cells in the absence of the proper control mechanisms acting in the spleen: this could allow for the generation of autoreactive atherosclerosis-specific B cells.
We thank Dr Gompf (Leibniz Institute for Aging Research, Jena) for fluorescence-activated cell sorting.
This work was funded by the German Research Council: HA 1083/15-4 to A.J.R. Habenicht; YI 133/2-1 to C. Yin; and MO 3054/1-1 to S.K. Mohanta; the German Centre for Cardiovascular Research (MHA VD1.2), SFB 1123/A1 and Z3, the European Research Council (AdG 249929) to C. Weber, by The British Heart Foundation: PG/12/81/29897 to P. Maffia and RE/13/5/30177; and the European Commission Marie Skłodowska-Curie Individual Fellowship 661369 to G. Grassia.
*These authors contributed equally to this article.
This manuscript was sent to Kathryn Moore, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
The online-only Data Supplement is available with this article at http://atvb.ahajournals.org/lookup/suppl/doi:10.1161/ATVBAHA.115.306983/-/DC1.