Generation of functional murine CD11c⁺ age-associated B cells in the absence of B cell T-bet expression

Abstract

Age-associated B cells (ABC), a novel subset of activated B cells defined by CD11b and CD11c expression, have been linked with both protective anti-viral responses and the pathogenesis of systemic autoimmunity. Expression of the T_H1 lineage transcription factor T-bet has been identified as a defining feature of ABC biology, with B cell-intrinsic expression of this transcription factor proposed to be required for ABC formation. In contrast to this model, we report that Tbx21 (encoding T-bet)-deficient B cells upregulate CD11b and CD11c surface expression in vitro in response to integrated TLR and cytokine signals. Moreover, B cell-intrinsic T-bet deletion in a murine lupus model exerted no impact of ABC generation in vivo, with Tbx21^−/− ABCs exhibiting an identical surface phenotype to wild-type (WT) ABCs. Importantly, WT and Tbx21^−/− ABCs sorted from autoimmune mice produced equivalent amounts of IgM and IgG ex vivo following TLR stimulation, indicating that T-bet-deficient ABCs are likely functional in vivo. In summary, our data contradict the established literature by demonstrating that T-bet expression is not uniformly required for ABC generation.

Introduction

B cells promote the pathogenesis of systemic autoimmunity via distinct effector functions, including pathogenic autoantibody production, cytokine secretion and antigen presentation to cognate CD4⁺ T cells [1]. Given these distinct roles in autoimmune pathogenesis, the functional characterization of B cell subsets driving disease development is of particular importance. In 2011, independent groups identified a new B cell subpopulation in autoimmunity, characterized by either the lack of surface CD21 and CD23 [2], or by the expression of integrins CD11b and CD11c [3]. Since this subpopulation was first identified in aged female mice, B cells expressing this surface phenotype have been termed “age-associated B cells (ABCs)”. Importantly, several lines of evidence directly implicated ABCs in the pathogenesis of autoimmunity. For example, ABCs are expanded in distinct murine lupus models, including the NZB/WF1 [3], Mer^−/− [3], MRL-Fas^lpr [4], BSXB [5, 6], and WAS chimera models [7]. In addition, CD11c⁺ B cells are increased in the peripheral blood in patients with autoimmunity [3], and accumulate in older female mice and humans, mirroring the age and sex bias of human autoimmune prevalence [2, 3]. Finally, sorted ABCs produce anti-nuclear antibodies ex vivo [3], and inducible Cre-mediated deletion of CD11c⁺ B cells reduced autoantibody titers in Mer^−/− mice [8], implicating ABCs as a source for pathogenic autoantibodies. Notably, the accumulation of CD21^lo and/or CD11c⁺ B cells (variably termed atypical memory B cells, tissue-like memory B cells or CD11c⁺ B cells) has also been noted in diverse human chronic infections and autoimmune states [9]. Although the surface markers used to identify these cells differ between studies, these data suggest that the expansion of phenotypically similar CD21^loCD11c⁺ B cells is a characteristic of chronic immune activation.

Given links between ABCs and the development of autoimmunity, defining the transcriptional signals required for ABC generation is of particular importance. In this regard, the transcription factor T-bet has emerged as a defining characteristic of ABCs, with prominent T-bet expression observed in ABCs generated in vitro [10], as well as in ABCs developing in aged female mice [3], during viral infection [10, 11] or in the setting of autoimmunity [8]. Moreover, T-bet over-expression using retroviral vectors promoted CD11b and CD11c expression by naïve B cells [11], and B cell-intrinsic Tbx21 deletion reduced CD11c⁺ B cells and protected against autoimmunity in B6.Sle1,2,3 and Mer^−/− lupusprone mice [12]. Based on these combined data, independent groups have proposed a model wherein T-bet expression is both necessary and sufficient for ABC formation [13–15].

In contrast to this hypothesis, we report that a B cell-intrinsic requirement for T-bet is not a uniform feature of ABC biology. First, we demonstrate that CD11b⁺CD11c⁺ ABCs can be generated after stimulation of Tbx21^−/− B cells with TLR and cytokine signals known to promote ABC generation in vitro. Second, using a chimeric model of B cell-driven SLE, we show that B cell T-bet expression is redundant for ABC formation in vivo. Although the functional roles for this B cell subset remain incompletely defined, ABCs have been proposed to serve as antigen-presenting cells (APCs) during autoimmunity [16], and to be a source for class-switched autoantibodies [3]. Notably, Tbx21^−/− ABCs exhibited a near identical surface phenotype to wild-type (WT) controls, including prominent upregulation of MHC class II and co-stimulatory molecules CD80 and CD86 required for efficient CD4⁺ T cell activation. In addition, both WT and T-bet-deficient ABCs produced class-switched autoantibodies ex vivo following TLR stimulation. Taken together, these data contradict the prevailing model for ABC development, by demonstrating that functional ABCs can be generated in the absence of T-bet.

Results

T-bet independent generation of CD11b⁺CD11c⁺ ABCs in vitro

Prior studies have demonstrated that ABCs develop in response to integrated BCR, TLR, cytokine and co-stimulatory signals. To model these events in vitro, we stimulated resting splenic B cells with combinations of anti-IgM, the TLR7 agonist R848, cytokines IFN-γ and IL-21, and anti-CD40 agonist antibodies. As predicted, after 48 hours in culture, stimulated WT B cells expressed both CD11b and CD11c, with this ABC surface phenotype most robustly induced by combined anti-IgM, R848, IL-21 and anti-CD40 stimulation (Fig. 1A and B; Supplementary Fig. 1) [10]. Importantly, in parallel with CD11b/CD11c upregulation, these integrated signals also promoted B cell T-bet expression with a clear correlation noted between surface CD11c and intranuclear T-bet staining (Fig. 1C). Based on these findings, we anticipated that lack of B cell T-bet would prevent adoption of an ABC surface phenotype by stimulated B cells. In contrast with this hypothesis, a clear population of CD11b⁺CD11c⁺ ABCs developed among in vitro-stimulated Tbx21^−/− B cells (Fig. 1A and B), although the proportion of B cells with CD11b and CD11c expression was modestly reduced in T-bet-deficient vs. -sufficient B cells. While these in vitro stimulation conditions do not fully recapitulate the cellular signals required for ABC generation in vivo, these data indicate that adoption of an ABC surface phenotype can occur independently of T-bet.

Figure 1.

In vitro generation of CD11b⁺CD11c⁺ ABCs independently of T-bet. (A) Percentage of CD11b⁺CD11c⁺ B cells after in vitro stimulation of WT and Tbx21^−/− CD43-microbead purified splenic B cells with indicated combinations of anti-IgM, R848, IFN-γ, IL-21 and anti-CD40. Error bars indicate mean plus standard error of the mean (S.E.M.) **p<0.01; ***p<0.001; ****p<0.0001; by one-way ANOVA, followed by Tukey’s multiple comparison test. (B) Representative flow cytometry plots showing CD11b and CD11c expression by WT and Tbx21^−/− B cells, stimulated as indicated. Number equals percentage within gate. (A, B) Data shown are from one in vitro B cell stimulation, representative of two independent experiments. (C) Left panel: Gating of CD11b^loCD11c^lo, CD11b^midCD11c^mid, CD11b^hiCD11c^hi subsets in WT B cells stimulated with anti-IgM, R848, IFN-γ, and IL-21. Right panels: Histogram showing intranuclear T-bet expression in CD11b^loCD11c^lo (gray), CD11b^midCD11c^mid (dotted line), CD11b^hiCD11c^hi (solid line) subsets. Data shown are from one in vitro B cell stimulation, representative of four independent experiments. (A–C) For each independent in vitro assay, one WT and one Tbx21^−/− mouse was sacrificed for purification of splenic B cells.

Generation of ABC in vivo in the absence of B cell T-bet expression

Based on these in vitro findings, we next tested whether functional ABCs can develop in an autoimmune setting in the absence of B cell-intrinsic T-bet expression. To do this, we utilized a chimeric model of murine lupus developed by our group, in which B cells, but not other hematopoietic lineages, lack expression of Wiskott–Aldrich syndrome (WAS) protein [7, 17–19]. In this model, Was-deficient B cells initiate spontaneous humoral autoimmunity characterized by immune activation, class-switched anti-nuclear antibody (ANA) formation and the development of immune-complex glomerulonephritis. An important advantage of this “WAS chimera model” is that it allows efficient testing of B cell-intrinsic signals impacting autoimmunity. Importantly, diseased WAS chimeras are also characterized by the prominent expansion of ABCs [7], suggesting that this experimental approach could be used to interrogate B cell signals underlying ABC generation.

The age-related accumulation of a unique B cell subpopulation, now referred to as ABCs, was independently described by two groups based upon distinct flow cytometry gating strategies. Hao et al. [2] defined ABCs as CD19⁺AA4.1⁻CD43⁻CD21⁻CD23⁻ B cells in aged mice, while Rubstov et al. [3] reported expansion of CD19⁺CD11b⁺CD11c⁺ B cells in older female mice and autoimmune strains. Given these distinct, but likely overlapping, surface phenotypes, we first compared the surface phenotype of ABCs defined by each gating strategy. In WAS chimeras, the majority of CD11b⁺CD11c⁺ B cells lacked CD21 and CD23 expression (Fig. 2A). In contrast, defining ABCs by CD21/CD23 expression resulted in a more heterogeneous population of B cells, with only 21.6 ± 13.7% (Mean ± 2 SD) of AA4.1⁻CD21⁻CD23⁻ B cells, also expressing CD11b and CD11c (Fig. 2B). Moreover, CD11b⁺CD11c⁺ ABCs exhibited increased granularity (based on side scatter (SSC)), and greater CD80, CD86 and MHC Class II expression compared with follicular mature (FM) B cells, whereas the degree of surface activation marker expression was lower in CD21/CD23-defined ABCs (Fig. 2C). Given this broader heterogeneity of CD21/CD23-defined ABCs, we limited the remainder of our analyses to ABCs as defined by CD11b and CD11c expression.

Figure 2.

Surface phenotype of CD21⁻CD23⁻ vs. CD11b⁺CD11c⁺ ABCs from autoimmune WAS chimeras. (A) Representative flow cytometry plots demonstrating CD21 and CD23 expression (right) on splenic ABCs defined by CD19, CD11b, and CD11c expression (left). (B) FACS plot showing CD11b and CD11c expression (right) on CD43⁻AA4.1⁻CD19⁺CD21⁻CD23⁻-gated splenic ABCs (left). (A, B) Number equals percentage within gate. (C) Representative histograms showing cell size (by side scatter (SSC)), MHCII, CD80, CD86 and FAS expression on splenic CD21⁻CD23⁻ (dashed line) vs. CD11b⁺CD11c⁺ (solid line) ABCs. Gray histogram indicates FM B cells (CD21^intCD23⁺). (A–C) Data shown are from one experiment (n = 8 mice), representative of two independent experiments with eight mice per experiment.

Previous studies have demonstrated that integrated signals downstream of BCR, TLR7 and IFN-γ promote T-bet⁺CD11c⁺ ABC formation during murine viral infection [11]. In addition, a subset of adoptively transferred naïve B cells developed into T-bet⁺ ABCs in response to MHC Class II- and CD40 ligand-dependent co-stimulatory signals from cognate T cells [20]. Thus, we first examined whether B cell-intrinsic deletion of MHC Class II (MHC-II; MhcII) or the IFN-γ receptor (IFN-γR; Ifngr) similarly prevents ABC formation in the WAS chimera model. Consistent with prior studies, loss of either MHC-II or IFN-γR expression on B cells abrogated the accumulation of splenic CD11b⁺CD11c⁺ ABCs in aged WAS chimeras (Fig. 3A; Supporting Information Fig. 2). Moreover, splenic ABCs, whether defined as CD21⁻CD23⁻ or CD11b⁺CD11c⁺, exhibited increased intra-nuclear T-bet staining, confirming a correlation between B cell T-bet expression and the ABC surface phenotype in the WAS model (Fig. 3B). Together with a previously reported role for TLR7 in ABC formation [7], these findings indicate that similar B cell signals underlie the generation of ABCs in WAS chimeras as were identified in prior viral and autoimmune models, suggesting that this strategy can appropriately be used to test the requirement for T-bet in ABC formation.

Figure 3.

Formation of ABCs in vivo in the absence of B cell-intrinsic T-bet expression. The generation of splenic CD11b⁺CD11c⁺ ABCs was quantified in the WAS chimera model, and in chimeras in which B cells lacked expression of MHC-II, IFN-γR, or T-bet. Bone marrow chimeras were generated as described in Methods, and sacrificed at 24 weeks post transplantation. (A) Percentage (left) and total number (right) of splenic CD11b⁺CD11c⁺ ABCs accumulating in WT, WAS, B cell-intrinsic MhcII^−/− and B cell-intrinsic Ifngr^−/− WAS chimeras. Each dot indicates an individual animal. (B) Representative histograms showing intra-nuclear T-bet expression in splenic CD21⁻CD23⁻ (dashed line) vs. CD11b⁺CD11c⁺ (solid line) ABCs derived from representative WAS chimera at 24 weeks post-transplant. Gray histogram indicates follicular mature (FM) B cells. Data shown are from one experiment, representative of three independent experiments. (C) Representative FACS plots (gated on CD19⁺ B cells) showing splenic CD11b⁺CD11c⁺ ABCs from WT, WAS and B cell-intrinsic Tbx21^−/− WAS chimeras. Number equals percentage within gate. Data shown are representative of four independent WT, WAS, and B cell Tbx21^−/− WAS chimeras. (D) Percentage (left) and total number (right) of splenic CD11b⁺CD11c⁺ ABCs in WT (white), WAS (black), B cell-intrinsic Tbx21^−/− WAS chimeras (gray). (E) Representative histograms showing intranuclear T-bet expression in Was^−/− CD11b⁺CD11c⁺ ABCs (solid line), Was^−/−.Tbx21^−/− ABCs (dashed line) and Tbx21^−/− Bulk B cells (gray shaded histogram). Data shown are representative of four independent Tbx21^−/− experiments. (A, D) Data showing percentage and total number of splenic CD11b⁺CD11c⁺ ABCs are pooled from thirteen independent experiments with WT (n = 30 mice), thirteen Was^−/− (n = 60 mice), two B cell-intrinsic MhcII^−/−.Was^−/− (n = 8 mice), two B cell-intrinsic Ifngr^−/−.Was^−/− (n = 10 mice), and four B cell-intrinsic Tbx21^−/−Was^−/− independent experiments (n = 16 mice). **p<0.01; ***p<0.001; ****p<0.0001; by one-way ANOVA, followed by Tukey’s multiple comparison test. Error bars indicate mean plus standard error of the mean (S.E.M.).

To directly address this question, we established WAS chimeras in which only B cells lacked T-bet by adoptively transferring WT, Was^−/− or Was^−/−.Tbx21^−/− bone marrow (BM) together with μMT BM (20:80 ratio) into lethally-irradiated μMT recipients. Surprisingly, ABC generation was unaffected by B cell-intrinsic T-bet deficiency, with both WAS chimeras and B cell Tbx21^−/− WAS chimeras developing an equivalent expansion of splenic CD11b⁺CD11c⁺ ABCs (Fig. 3C and D). Importantly, intranuclear staining confirmed lack of T-bet protein expression in Tbx21^−/− ABCs (Fig. 3E). Thus, although T-bet expression is a uniform feature of ABCs generated in vitro and in vivo [3, 8, 10, 11], B cell-intrinsic expression of this transcription factor is not essential for ABC formation.

Tbx21^−/− CD11b⁺CD11c⁺ B cells express ABC surface markers and produce class-switched autoantibodies

Since T-bet expression has been described as a defining characteristic of ABCs, we next examined whether Tbx21^−/− CD11b⁺CD11c⁺ B cells exhibit an ABC surface phenotype and are functional during autoimmunity. Notably, CD11b⁺CD11c⁺ B cells from T-bet-sufficient and -deficient WAS chimeras were phenotypically identical, including: increased cell size by forward scatter compared with naïve B cells; equivalent surface IgM and IgD expression; downregulation of CD21 and CD23; and CD138 staining (Fig. 4A; Supporting Information Fig. 3); a surface profile mirroring the published ABC phenotype [2, 3]. Functionally, specific roles for ABCs in the pathogenesis of SLE and other autoimmune diseases have not been completely defined. ABCs are predominantly comprised of unswitched IgM⁺ B cells that that can generate class-switched autoantibodies upon ex vivo TLR stimulation [3, 20]. In addition, ABCs can function as antigen presenting cells (APCs), a role likely facilitated by prominent expression of MHC Class II and co-stimulatory molecules CD80 and CD86 [16]. Based on these data, ABCs may function as memory B cells that can both promote T cell activation upon antigen rechallenge, or rapidly differentiate into antibody-producing effector B cells.

Figure 4.

T-bet-deficient CD11b⁺CD11c⁺ B cells exhibit an ABC surface phenotype and produce antibodies ex vivo. (A, B) Histograms showing cell surface marker expression by CD11b⁺CD11c⁺ ABCs derived from representative WAS (n = 4 mice, solid line) and B cell-intrinsic Tbx21^−/− WAS (n = 5 mice, dotted line) chimeras. Gray histograms indicate CD11b⁻CD11c⁻ B cells from WAS chimera. Data shown are from one experiment. (C, D) Total immunoglobulin (C), and anti-dsDNA IgG (D), in culture supernatants from ex vivo stimulated ABCs, FM and MZ B cells from one WAS (black; n = 4 mice) and one B cell-intrinsic Tbx21^−/− WAS (white; n = 5 mice) chimera. B cell subpopulations from individual mice were cultured in two replicate wells, with each data point representing the supernatant immunoglobulin concentration/ELISA O.D. for one well. **p<0.01; NS, not significant; by two-tailed Student’s t-test.

Since the development of SLE is a stochastic process characterized by the ongoing activation of a broad range of autoreactive B and T cell clones, defining the specific contribution of individual ABCs (in the presence or absence of T-bet) in murine lupus is technically challenging. However, as a proxy for APC functions during autoimmunity, we confirmed prominent upregulation of the co-stimulatory molecule CD86, as well as increased expression of MHC Class II by both WT and Tbx21^−/− ABCs. In addition, both T-bet-sufficient and -deficient ABCs exhibited increased expression of CD95 (FAS), a surface marker of memory B cells (Fig. 4B; Supporting Information Fig. 3) [21].

Finally, we tested whether Tbx21^−/− ABCs derived from WAS chimeras might contribute to class-switched autoantibody titers in diseased mice. To do this, we cultured sorted CD11b⁺CD11c⁺ ABCs, as well as follicular mature (FM) and marginal zone (MZ) controls, from WAS and B cell-intrinsic Tbx21^−/− WAS chimeras in media alone or with the TLR7 agonist R848. Strikingly, both T-bet sufficient and -deficient ABCs produced IgM, IgG and IgG2b subclass antibodies following R848 stimulation, with no significant differences in immunoglobulin titers noted between genotypes (Fig. 4C; Supporting Information Fig. 4). Notably, only IgG2c production was impacted by T-bet deletion, in keeping with established role for T-bet in facilitating IgG2c class-switch recombination [22, 23]. Finally, secreted IgG derived from both Tbx21^+/+ and Tbx21^−/− ABCs bound the self-ligand dsDNA (Fig. 4D; Supporting Information Fig. 4), indicating that T-bet deletion does not impact autoreactivity within the ABC compartment. Although these data do not preclude additional roles for T-bet in ABC biology during autoimmunity, our combined findings demonstrate that Tbx21^−/− ABCs exhibit an identical surface phenotype to WT controls, and likely contribute to class-switched autoantibodies in murine SLE.

Discussion

Over the past decade, independent groups have documented the expansion of phenotypically-related B cell populations in setting of chronic immune activation. In addition to “age-associated B cells (ABCs)” defined by the Marrack and Cancro laboratories, CD21^lo B cells are expanded in peripheral blood of human subjects with chronic infections, including HIV, tuberculosis and malaria infection [24–27]. In addition, patients with diverse autoimmune conditions, including systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), Sjögren’s syndrome, hepatitis C virus-associated cryoglobulinemia and common variable immunodeficiency (CVID) exhibit an expansion of B cells phenotypically similar to ABCs [28–33]. Based on functional and phenotypic characteristics, these B cells have been variably termed atypical memory B cells, tissue-like memory B cells or CD11c⁺ B cells. Although significant heterogeneity exists in the surface markers used to define these populations, a near uniform feature of CD21^lo and/or CD11c⁺ B cells is expression of the T_H1-lineage transcription factor, T-bet. Moreover, since retroviral Tbx21 overexpression promoted B cell CD11c expression, and B cell-intrinsic T-bet deletion reduced ABC numbers in murine lupus [11, 12], T-bet has been proposed to be both necessary and sufficient for ABC development [13–15]. Despite these data, we report the surprising observation that functional ABCs can be generated in the absence of B cell-intrinsic T-bet expression.

Importantly, our findings do not exclude the potential for additional B cell-intrinsic T-bet functions in ABC biology and autoimmune pathogenesis. For example, T-bet promotes IgG2a/c class-switch recombination [22, 23], and is required for the maintenance of IgG2a/c⁺ memory B cells [34]. Consistent with these data, we noted a specific defect in IgG2c production by ex vivo stimulated Tbx21^−/− ABCs. Since IgG2c is known to promote autoimmune pathogenesis by fixing complement and binding activating Fc receptors [35], B cell-specific T-bet deletion may ameliorate disease via the specific reduction in this pro-inflammatory immunoglobulin subclass. In addition, B cell T-bet expression likely exerts additional antibody-independent functions in SLE. Specifically, T-bet promotes the surface expression of CXCR3, a receptor for IFN-γ-inducible chemokines CXCL9, CXCL10 and CXCL11 [36, 37]. In this manner, T-bet signals likely drive ABC migration to inflammatory sites, and may thereby promote autoimmunity via local antigen presentation to autoreactive CD4⁺ T cells in diseased tissues. Consistent with antibody-independent roles for T-bet, B cell-intrinsic Tbx21^−/− deletion prevented viral containment in a chronic lymphocytic choriomeningitis virus (LCMV) infection model despite adoptive transfer of LCMV-specific IgG2a antibody [38]. Thus, T-bet expression likely induces a broad B cell transcriptional program resulting in both anti-viral activity and pathogenic autoimmune inflammation.

Despite these caveats, our data clearly demonstrate that functional ABCs can be formed without T-bet, raising questions as to how to reconcile our findings with previous studies. Notably, while T-bet is not absolutely required for CD11b⁺CD11c⁺ B cell formation, we show that T-bet modestly enhanced the proportion of B cells exhibiting an ABC surface phenotype in vitro. Importantly, Was^−/− B cells are hyper-responsive to both BCR and TLR stimulation [17, 39]. Thus, we hypothesize that in the WAS chimera model a threshold of B cell TLR, cytokine and co-stimulatory signals is reached to allow Tbx21-independent ABC generation. In keeping with this model, B cell-intrinsic T-bet deletion in the B6.Sle1,2,3 and Mer^−/− lupus models resulted in decreased, but not absent, ABCs, with a proportionally greater reduction in germinal center B cells relative to CD11c⁺ B cells [12].

In summary, we show that functional ABCs can be generated in the absence of B cell T-bet expression, in contrast to the prevailing model. Since ABCs have been linked with the pathogenesis of multiple human autoimmune diseases, these data emphasize that further studies are required to uncover the key transcriptional events underlying the formation of this important B cell subset.

Materials and methods

Mice

C57BL/6, μMT [40], Was [41], MhcII^−/− [42], Ifngr^−/− [43], and Tbx21^−/− [44] mice and the relevant murine crosses were bred and maintained in the specific pathogen-free (SPF) animal facility of Seattle Children’s Research Institute (Seattle, WA). All animal studies were conducted in accordance with Seattle Children’s Research Institute IACUC approved protocols.

Bone marrow transplantation

BM was harvested from C57BL/6 (“WT”), Was^−/−, Was^−/−.MhcII^−/−, Was^−/−.Ifngr^−/−, or Was^−/−.Tbx21^−/− and depleted of CD138⁺ plasma cells (Miltenyi Biotec, 130–098-257). Donor BM was mixed with μMT BM (20:80 ratio, 6 × 10⁶ total cells) and injected retro-orbitally into lethally irradiated (450cGy x 2 doses) μMT recipients. Data are representative of at least two independent experimental cohorts per genotype, sacrificed at 24 weeks post-transplant.

Flow-cytometry

Flow-cytometry was performed as described [7, 18], using the following anti-murine antibodies: B220 (RA3–6B2), CD80 (16–10A1), CD43 (S7), CD86 (GL1), CD138 (281–2), CD11b (M1/70) from BD Biosciences; CD11c (N418), CD11b (M1/70), GL7 (GL-7), T-bet (4B10), MHCII (M5/114.15.2), CD93 (AA4.1) from eBioscience; CD19 (ID3), CD21/CD35 (7E9), CD23 (B3B4), IgM (RMM-1), IgD (11–26c.2a) from BioLegend; PNA (Fl-1071) from Vector Labs; IgM (II/41), IgD (11–26c (11–26)) from Life Technologies; Fas (Jo2) from BD Pharmingen; and Alexa Fluor™ 350 NHS Ester Viability dye (Catalog number A10168 ThermoFisher Scientific).

In vitro stimulations

Murine splenic B cells were purified by CD43-microbead depletion (Miltenyi Biotec, Inc.) and cultured in RPMI at 37°C for 48 h at 1 × 10⁶ cells/well in a 96-well plate with or without: R848 (5 ng/mL); anti-mouse IgM F(ab’)₂ fragment (1 μg/mL, Jackson Immunoresearch); recombinant mouse IFN-γ (200 U/mL, Biolegend); IL-21 (50 ng/mL, PeproTech); and, anti-mouse CD40 (1 μg/mL, Southern Biotech). B cell surface markers and transcription factor expression were evaluated by flow cytometry.

Ex vivo B cell culture

Splenocytes were sorted using a FACSCalibur (BD) cell sorter based on the following cell surface markers: CD19⁺B220⁺CD11b⁺ CD11c⁺ (ABC); CD19⁺B220⁺CD21^midCD24^mid (FM); and CD19⁺ B220⁺CD21^hiCD24^hiCD23^lo (MZ). Sorted cells from individual animals were cultured in two replicate wells at 250 000 cells/mL in 96-well plates for 72 h at 37°C in RPMI with or without R848 (1 μg/mL). Antibodies in culture supernatants were determined by ELISA. For total immunoglobulin quantification, 96 well Nunc-Immuno MaxiSorp plates (Thermo Fisher) were pre-coated overnight at 4°C with goat anti-mouse IgM, IgG, IgG2b, IgG2c antibodies (1:500 dilution, SouthernBiotech) for 24 h. The antibody ELISAs were designed to measure sample concentrations in the nanograms per milliliter range, corresponding to an 11 step standard curve ranging from 400 to 0.3906 ng/mL. Supernatants for total immunoglobulin ELISAs were run as a three-step dilution series from 1:8 to 1:32, with final concentration reported as the average of each dilution. For specific autoantibodies, plates were pre-coated with calf thymus dsDNA (100 ug/mL; Sigma-Aldrich D3664–5 × 2MG) or Sm/RNP (5 ug/mL; Arotec Diagnostic Limited ATR01–10). Supernatants for autoantigen-specific ELISAs were diluted 1:4 and run in duplicate. Plates were blocked for 1 h with 1% BSA in PBS prior to addition of diluted supernatant for 2 h. Specific antibodies were detected using goat anti-mouse IgM-, IgG-, IgG2b-, IgG2c-HRP (1:2000 dilution; SouthernBiotech) and peroxidase reactions were developed using OptEIA TMB substrate (BD Biosciences) and stopped with sulfuric acid. Absorbance at 450 nm was read using a SpectraMax 190 microplate reader (Molecular Devices) and data analyzed using GraphPad Prism (GraphPad Software, Inc.).

Statistical evaluation

p-values were calculated using Prism 6 (GraphPad Software, Inc.).

Abbreviations

ABC

age-associated B cell

ANA

anti-nuclear antibody

APC

antigen-presenting cell

FM

follicular mature

MZ

marginal zone

SPF

specific pathogen-free

WAS

Wiskott–Aldrich syndrome