Properties of Regulatory B Cells Regulating B Cell Targets

Abstract

Regulatory B cells (Bregs) have shown promise as anti-rejection therapy applied to organ transplantation. However, less is known about their effect on other B cell populations that are involved in chronic graft rejection. We recently uncovered that naïve B cells, stimulated by TLR ligand agonists, converted into B cells with regulatory properties (Bregs-TLR) that prevented allograft rejection. Here we examine the granular phenotype and regulatory properties of Bregs-TLR cells suppressing B cells. Co-cultures of Bregs-TLR with LPS-activated B cells showed a dose-dependent suppression of targeted B cell proliferation. Adoptive transfers of Bregs-TLR induced a decline in antibody responses to antigenically disparate skin grafts. The role of Breg BCR specificity in regulation was assessed using B cell deficient mice replenished with transgenic BCR (OB1) and TCR (OTII) lymphocytes of matching antigenic specificity. Results indicated that proliferation of OB1 B cells, mediated through help from CD4⁺ OTII cells, was suppressed by OB1 Bregs of similar specificity. Transcriptomic analyses indicated that Bregs-TLR suppression is associated with a block in targeted B cell differentiation controlled by PRDM1 (Blimp1). This work uncovered the regulatory properties of a new brand of Breg cells and provided mechanistic insights on potential applications of Breg therapy in transplantation.

1. Introduction

Numerous studies have focused on the function of B cells as antibody-secreting cells (ASCs), antigen-presenting cells (APCs), and mediators of antibody-mediated rejection (ABMR).^1–3 B cells can be activated and differentiated into ASCs after stimulation by T cell-dependent (TD) or T-independent (TI) antigens. PRDM1 (PR/SET Domain 1, also known as Blimp1), which drives the maturation of B cells into ASCs,⁴ has also been shown to be critical for the commitment of B cells to a plasma cell fate.⁵ Our previous published data suggest that blocking TGF-β accelerates the rejection of Ovalbumin-expressing (Ova) skin grafts in mice expressing Ovalbumin-specific B cell receptor (OB1 mice).⁶ TGF-β also plays a crucial role in inducing anergy in CD8⁺ T cells.⁷ And yet, the specific mechanisms underlying B cell regulation have remained elusive.

A particular subset of B cells, regulatory B cells (Bregs) have recently shown promise as a cell therapy by inhibiting the proliferation of cytotoxic and helper T cells while promoting the induction of regulatory T cells in murine models of autoimmune diseases and transplantation.^8–10 Our recently published work showed that several B-cell-related gene sets predicts allograft tolerance after kidney transplantation with a better performance.¹¹ Thus far, multiple subsets of Bregs, such as CD5⁺CD1d^hi B (B10) cells,⁸ Tim-1⁺ B cells,¹² and CD138⁺ plasma cells,¹³ have been reported in mice. Our group has recently characterized Bregs-TLR, resulting from the application of a TLR-agonist in vitro to naïve B cells. Bregs possess the capacity to interact with other lymphocytes, and similarly, Bregs-TLR, pre-stimulated via TLR receptors, acquired potent suppressive function toward naïve T cell proliferation. We also showed that B cells from mice tolerant of allogeneic islets, via a combined anti-CD45RB and anti-Tim-1 treatment, could transfer tolerance to naïve recipients sufficient in Nk1.1⁺ cells.¹⁴ Additionally, Bregs can inhibit the interactions of T cells with dendritic cells (DCs), thereby preventing the secretion of proinflammatory cytokines.¹⁵ Indeed, human Bregs can inhibit CD4⁺ T cell differentiation into follicular helper T (Tfh) cells in vitro.¹⁶ Decreased production of antibodies mediated by Tfh cells suggests that Bregs may also possess the ability to regulate B cell activity indirectly. While substantial research has been conducted on the modulatory effects of Bregs on other T cells, much less is known regarding Bregs’ ability to directly suppress and influence the activity of other B cells. For instance, tolerance in heart transplantation has been shown to result in tolerant donor-specific B cells that can suppress de novo alloreactive B cells.¹⁷ Nonetheless, the suppressive potentiality of Bregs on other B cell subtypes has not been characterized.

Herein, we investigated the in vitro and in vivo potential of Bregs-TLR to modulate naïve B cell activation and proliferation. We find that B6-derived Bregs-TLR suppress syngeneic B cells activated by bacterial lipopolysaccharide (LPS) and regulate in vitro the differentiation of suppressed B cells in a dose-dependent manner. Interestingly, Bregs-TLR suppressed LPS-B cell activation not through programmed cell death, but via inactivation of gene pathways involved in cellular proliferation. We also demonstrated that Bregs-TLR cells inhibit B cell proliferation in vivo and substantially decreased the production of IgG antibodies following transplantation of antigenically disparate skin grafts. Results are also provided in support of the role of BCR specificity in Breg-mediated suppression. Transcriptomic analyses of RNAseq and scRNAseq data revealed important biomarkers of Breg functions and underlined the potential role of PRDM1, a key controller of B cell differentiation. This study reports a previously unrecognized regulatory capacity of Bregs-TLR to modulate the function of other B cells.

2. Materials and Methods

2.1. Mice

Wild-type C57BL/6 (WT B6), B6μMT^−/− (B cell-deficient), Ova-transgenic B6 (Ova⁺), OTII (Ova-specific CD4⁺ T cells), and homozygote mb1-Cre (mb1) mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). OB1 (Ova-specific B cells, B6 background) mice were generously provided by Dr. Hidde Ploegh from the Whitehead Institute for Biomedical Research.¹⁸ All mice were housed under specific pathogen-free barrier conditions in the animal facility of Massachusetts General Hospital (MGH). All procedures were performed following the principles of laboratory animal care and approved by the Institutional Committee for Research Animal Care.

2.2. Cell transfer and suppression model development

Splenic B6/OB1 B cells and CD4⁺ T cells were selected using EasySep™ Mouse Naïve CD19⁺ B cell or CD4⁺ T Cell Isolation Kit (Stem cell technologies, Cambridge, MA). The purity of the isolated CD19⁺ B cell population was ~ 99%. Isolated B6 B cells were stimulated with CpG ODN 1668 (10ug/ml, InvivoGen, San Diego, CA, USA) for 3 days to induce Bregs-CpG, and further LPS (10ug/ml), PMA (50ng/ml), and ionomycin (1ug/ml) stimulation for the last 5 hours to generate Bregs-TLR. Complete RPMI 1640 cell culture medium was prepared using 10% fetal bovine serum (HyClone FetalClone III, Thermo Scientific), 50 μM 2-mercaptoethanol (ACROS Organics), 1 mM sodium pyruvate, 1x MEM Eagle’s non-essential amino acids, 2 mM L-glutamine, 100 IU/mL Penicillin, and 100 μg/mL Streptomycin, all from MP Biomedicals).¹⁰ Bregs-CpG/Bregs-TLR and naïve B6 B cells were then labelled with Cell Trace Violet (CTV, CellTrace™ Violet Cell Proliferation Kit, Invitrogen) and CFSE (CellTrace™ CFSE Cell Proliferation Kit, Invitrogen) respectively. For in vitro studies, after 2–4 days of co-culture with and without LPS, proliferation and activation were measured by flow cytometry. For Transwell experiments, 96-well plates (0.4 μm pore, Corning® HTS, Sigma-Aldrich) were used. Bottom compartment contained CFSE-labelled B cells, pulsed with LPS, and incubated for 72h with or without Bregs-TLR or Bregs-CpG control (top compartment). Antibodies were detected in sera from WT B6 recipients, transplanted with Ova-skin grafts using methods previously reported by our group.⁶ Briefly, WT B6 recipients were adoptively transferred with 5M Bregs-TLR on D-1, transplanted with Ova-skin on D0. Sera were collected on D14, D21, and D28 for analyses of immunoglobulin production. For donor-specific experiments, irradiated Ova splenocytes (irra-OvaSPC, 25M) were suspended in PBS and adoptively transferred by intraperitoneal (IP) injection to μMT, mb1, or B6 recipients. Recipients were then transplanted with Ova-skin grafts.⁶ Lastly, OB1 B cells (5M), OTII CD4⁺ T cells (1M), and/or OB1 Bregs-TLR (5M) were transferred through IV injection.

2.3. Flow cytometry

Splenocytes (SPC) were obtained from B6, Mb1 or μMT mouse recipients that were treated by the adoptive transfer of irra-OvaSPC. Cells were also collected from the draining lymph node (DLN) and non-draining lymph node (Non-DLN) and stained with the following antibodies: CD19-PB (6D5, Biolegend), IgM-PE (R6–60.2, BD Biosciences), IgD-Percp/Cy5.5 (11–26c.2a, Biolegend), CD21-APC (eBio4E3, eBioscience), CD23-PE/Cy7 (B3B4, Biolegend), CD80-PE (16–10A1, Biolegend), CD86-APC/Cy7 (GL-1, Biolegend), CD19-APC (6D5, Biolegend), Blimp1-PE (5E7, Biolegend), LAP-PE (TGF-β, TW7–16B4, Biolegend), CD365/Tim-1 PE (RMT1–4, Biolegend). Detection of antibodies was done in sera from B6 mice that were incubated with Ova⁺ SPC, and stained with CD19-PB, IgM-PE, and IgG-FITC (Poly4060, Biolegend). All results were analyzed using the FlowJo software (Treestar, Ashland, OR, USA).

2.4. RNA-sequencing and data analysis

RNA extraction from naïve B cells (10M), Bregs-CpG (10M), and Bregs-TLR (10M) was performed respectively, per manual manufacturer’s instructions (Qiagen RNeasy Mini kit, Qiagen, Germantown, MD, USA). RNA concentration and integrity were then measured using a Nanodrop ND-1000 (v3.8.1, Thermo Scientific, Waltham, MA, USA). Samples were sequenced at MGH nextGen Sequencing Core. Briefly, sequencing was performed on an Illumina HiSeq 2500 instrument, resulting in approximately 30 M of 50 bp reads per sample. Sequencing reads were then splice-aware mapped to the mouse reference transcriptome (mm9 assembly) using STA.¹⁹ Read counts over transcripts were calculated using HTSeq ²⁰ based on the Ensembl gene annotation system for GRCm37/mm9 assembly. The transcripts were then analyzed with EdgeR in R programming.²¹ genes with the cutoffs of 2-fold change in expression value and a false discovery rate (FDR) P-value < 0.05 were identified as differentially expressed genes (DEGs). All data are available in the Gene Expression Omnibus (GEO number: GSE178713). To investigate changes within B cells undergoing the process of LPS stimulation, original data of naïve B6 B cells, LPS-B for 24h and LPS-B for 72h (accession numbers: GSE35998), and scRNA-seq data of LPS-B (accession numbers: GSE136376) were downloaded from GEO database ^13,22 and analyzed using edgeR, clusterProfiler, Seurat, ComplexHeatmap packages in R programming.^21,23–25 Gene Set Enrichment Analysis (GSEA) was then performed on all DEGs using GSEA software (v4.0.3, BROAD, MIT, MA, USA); the enriched pathways with |normalized enrichment score| >1, and FDR-adjusted Q-value < 0.25 were considered significant.²⁶

2.5. Statistical analysis

Data were analyzed by Prism 6 (Graph Pad, San Diego, CA, USA) and R language (R version 3.6.3). Shapiro-Wilk normality test or Kolmogoov-Smirnov test was used for test if the values came from a Gaussian distribution. The Wilcoxon Rank sum test or student’s t-test was used to measure quantitative differences between each group. Differences among three or more groups were analyzed using the one-way ANOVA analysis or Kruskal-Wallis test. FDR-adjusted P values of < 0.05 were considered significant.

3. Results

3.1. B cell proliferation can be downmodulated by Bregs in vitro

Activation and proliferation of enriched naïve B cells were induced in vitro by the addition of the T-cell independent mitogen LPS. Co-cultures of LPS-activated, CFSE-labelled B cells with unlabeled syngeneic Bregs-TLR were then analyzed for B cell proliferation after 2–4 days. As seen in Figure 1A–C, B cell proliferation was gradually inhibited by increasing doses of Bregs-TLR. Optimal suppression (70–75%) was achieved by 48 hrs of culture at a 1:2 B cell/Breg ratio. To exclude the possibility of Bregs-TLR competing with normal B cells for culture medium nutrients, we added a control condition in which Bregs-TLR were substituted with CpG-stimulated B cells, an intermediate cell subset occurring during the development of Bregs-TLR cells. As seen in Figure 1D, Bregs-TLR suppressed the proliferation of B cells, whereas Bregs-CpG had no significant effect even after 96 hours of cell culture (FDR-adjusted P < 0.001).

Figure 1. In vitro impacts of Bregs-TLR suppression on proliferation and activation of B cell subsets.

(A) Representative flow cytometry (FACS) histograms of B cells suppressed by Bregs in 72hr cultures of CFSE-labelled naïve or LPS-activated B cells. Co-culture conditions of CD19⁺ naïve B cells, B cells + LPS and Breg cells, are depicted in panel B. Suppression of B cell proliferation by Bregs-TLR (% of proliferative B cells = number of dividing cells / total CD19⁺ CFSE⁺ cells), relative to culture times (C) and Bregs-CpG controls (72 hr cocultures (D)). Representative FACS panels and cumulative data of phenotypic changes observed among naive / transitional (E), follicular (FO) / marginal zone (MZ) (F); and activated B cell subsets (G). Each dot represents single sample from 3–5 independent experiments, and data are presented in box plots, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; NS: not significant.

3.2. Effect of Bregs-TLR suppression on the phenotype of suppressed B cells.

Marginal-zone (MZ) B cells and follicular (FO) B cells are the two main cell subsets generated during B cell development. Our investigations into the MZ and FO B cell percentages in Bregs-TLR suppression cocultures, showed increase in the proportions of both the CD19⁺IgM^loIgD^hi naive and CD19⁺CD21⁺CD23^hi follicular phenotypes. This feature was associated with a concomitant decrease in B cells with the CD19⁺IgM^hiIgD^lo transitional and CD19⁺CD21^hiCD23^neg MZ B cell phenotypes (FDR-adjusted P < 0.05, Figure 1E and F). Interestingly, when the dose of Bregs-TLR was increased, CD19⁺CD21^hiCD23^neg MZ B cells increased by about 30% after 48 hours of stimulation and about 10% after 72h. Such increases in MZ B cell proportions inversely correlate with decreases of CD19⁺CD21⁺CD23^hi FO B cells (30% after LPS stimulation for 48h and about 10% after 72h). To study whether Bregs-TLR could control B cell activation, expression of two key markers, CD80 and CD86, was evaluated in 24 to 96 hours cocultures. In a dose-independent manner, as seen in Figure 1G, the presence of Bregs-TLR coincided with a decrease in the percentage of CD19⁺CD86⁺CD80⁺ B cells while also increasing the percentage of CD19⁺CD86⁺CD80^neg B cells (Figure 1G), suggesting that Bregs-TLR can inhibit B cell activation as seen by differential expression of only CD80.

3.3. Cellular mechanisms of Breg-mediated suppression of B cell targets.

To formally assess whether Bregs-TLR suppress LPS-B cell activation via Breg/B cell contacts, we conducted transwell experiments that physically separated the suppressor from the target cells in culture medium. Results presented in Figure 2A clearly indicated that Bregs-TLR were still suppressive of LPS-B cells and that suppression is mediated by soluble factors. Quantitative assessment of the numbers of B cells in successive mitotic step indicated that Bregs-TLR kept most of suppressed B cells within the first two division stages (FDR adjusted P < 0.05). There was no significant difference by increasing the doses of Bregs-TLR (1:1 vs 1:2). In contrast, Bregs-CpG had, again, no significant suppressive effect on B cell proliferation or on the number of cells engaged in the first or second division post CFSE staining. We also investigated whether Bregs-TLR suppression resulted in apoptosis of LPS-B cells in cultures (Figure 2B). Surprisingly, neither the early nor the late apoptotic cell population was significantly affected by Bregs-TLR activity as their numbers were unchanged between the two Breg doses tested.

Figure 2. Analysis of Breg-TLR mechanism of suppression in vitro.

(A) Transwell experiment: bottom compartment contains CFSE-labelled B cells, pulsed with LPS, and incubated for 72h with or without TLR or CpG control Bregs (top compartment). Suppression tests are reported in CFSE dilution histograms (left) and in dot plots of percentages of cells that have accomplished 1 or 2 divisions (div. right). (B) CFSE-labelled B cells, stimulated with LPS, incubated for 72h with or without Bregs-TLR. CFSE-labelled B cells were then stained with 7AAD and Annexin V to detect apoptotic cells. The percentage of early apoptotic cells (Annexin V⁺ 7AAD^neg B cells) and live cells (Annexin V^neg, 7AAD^neg B cells) are reported in contour plots. Each dot represents duplicates; values are means with *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; NS: not significant.

The in vivo relevance of these initial observations was then evaluated in a mouse model of skin transplantation. First, we developed an original model of antigen-specific B cell proliferation in which the effects of adoptively transferred Bregs could be tested. Several B cell deficient mouse strains, including μMT and mb1, were tried as potential B cell recipients, according to the protocol depicted in Supplementary Figure 1. In these settings, BCR transgenic B cells (OB1) of uniform specificity for an ovalbumin (Ova epitope, were injected together with antigen presenting cells presenting Ova (splenocytes from the act-Ova mouse) and help from cognate Ova-specific OTII CD4⁺ T cells. Results from Supplementary Figure S1B–C showed that OB1 cells survived better in μMT recipients than they do in B6 control mice. The presence of Ova-specific Th cells also allowed for OB1 cell proliferation (panel C). Interestingly, no OB1 cells were detected in the B-less, homozygote mb1-Cre mouse strain, even in the presence of antigen-cognate helper T cells, suggesting that early steps of B cell development that involve the mb1 gene, are implicated in survival of mature B cells.

The suppressive functions of OB1 derived Breg-TLR cells was assessed after boosting OB1 B cell activation with an Ova-skin graft (Figure 3A). By day 9 after cell transfers, we observed that OTII CD4 Th cells divided mostly in draining lymph nodes (Figure 3B, top panels), a proliferation which was significantly inhibited by antigen cognate OB1-derived Bregs (P < 0.05; Figure 3B, bottom panels. Detailed stats not shown). On the other hand, OB1 Bregs-TLR were able to inhibit Ova driven proliferation of OB1 B cells residing mainly in the spleen and, to a lesser extent, in draining lymph nodes (Figure 3C). Similar results were obtained using immunodeficient SCID mice as recipients of Breg cells (results not shown).

Figure 3. Analysis of Breg-TLR mechanism of suppression in vivo.

(A) Outline of experimental protocol. Day −1: transplantation of OVA-skin and injection of 25 million (M) irradiated (irra) OVA splenocytes. Day 0: co-injection of 5M OB1 cells + 1 M OTII CD4 T cells (both labelled with CTV) and 5M OB1 Bregs-TLR. Day 9: analysis by flow cytometry of adoptively transferred OTII CD4⁺ T cells (gated on dividing cells, B), and CD19⁺ OB1 B cells (C) Cells numbers are reported in dot plots with each dot representing duplicates from 4 independent experiments. *P < 0.05, NS: not significant. SPC: splenocytes; DLN: draining lymph nodes; Non-DLN: Non-draining lymph nodes. CTV: Cell Trace Viole. (D) Experimental outline for assessing IgG production in sera from mice recipient of OVA skin grafts and injected with Breg-TLR. Serum samples were analyzed for IgM and IgG contents by flow cytometry using isotype-specific antibodies (MFI: mean fluorescence intensity of bound antibodies).

Given that graft-specific antibodies remain the main cause of chronic allograft rejection and that Bregs-TLR can inhibit B cell activation and proliferation in vivo, we tested whether suppressive Bregs would inhibit the production of graft-specific antibodies. Sera from B6 mice treated with B6 Bregs-TLR cells and grafted with Ova-skin grafts, were examined for the presence of Ova-specific IgM and IgG antibodies. No IgM antibodies were detected in this model at any time points starting 14 days after Breg injection (Figure 3D). In contrast, production of anti-Ova IgG was seen on day 14, 21 and 28. Adoptively transferred B6 Bregs-TLR, but not control B6 Bregs-CpG, significantly inhibited the production of anti-Ova IgG by day 21 (FDR-adjusted P < 0.05).

3.4. Transcriptome sequence analyses unveiled Bregs-TLR related genetic pathways.

Further information on the molecular mechanisms involved in Bregs-TLR suppression, was acquired by broad comparison of transcriptomes for differential gene expression in various B cell subsets. As the development of Bregs-TLR cells included an initial LPS activation step (See Material and Methods), we first compare the transcriptomes of naïve and LPS-activated B cells to pinpoint genes related to LPS activation. Presented in Figures 4A and S2A are results from computational analysis of differentially expressed genes (DEG) which revealed 29 genes differentially expressed between LPS-activated and non-activated B cells. When compared to DEG patterns from Bregs-TLR versus control Bregs-CpG, this initial gene activation signature dropped to 12 genes associated with the Breg-TLR phenotype (Figures 4A, and S2B; FDR-adjusted Q-value < 0.25). Computational interpretation of these RNAseq data by gene ontology (GO) term enrichment approach suggested that Bregs-TLR signatures included 3 immune relevant pathways potentially involved with negative regulation of leukocyte proliferation: IL-10; IL-2 RA (CD25) and PRDM1 (encoding Blimp1 protein, an important regulator of plasma cell differentiation; Figures 4B–E). The implication of the upregulation of IL-10 and IL-2 receptor gene expression in Bregs-TLR suppression was excluded in view of similar upregulation seen in both LPS-B cells and Bregs-TLR (Figure 4B). More investigations on gene pathway enrichment by gene set enrichment analysis (GSEA) broadened Bregs-TLR signature with the addition of the TGF-β and TNFα pathways (Figure 5A).

Figure 4. Bregs-TLR suppressing activated B cells: Both B cell subsets show a block in B cell differentiation controlled by PRDM1.

(A) Venn Diagram showing the distribution of sets of differentially expressed genes (DEGs) between naïve B cells and LPS-B cells stimulated for 24h or 72 h (top part). Among the 23 + 6 genes up- or down-regulated in the first screening, 12 were also differentially expressed in Bregs-TLR comparatively to control Bregs-CpG (bottom part). (B) Expression profiles (Box plots) of 3 immune-related genes (IL-10, IL2RA and PRDM1) from the set of 12 genes defined in A. Expression levels (CPM) are shown for naïve B, LPS-B cells (24h and 72h stimulation), Bregs-CpG and Bregs-TLR. (C) scRNA-sequencing analysis was performed on LPS-B and naïve B cells. Data analysis by cell type annotation defined cell clusters according to differential expression of transcripts for CD37, CCR6 (CD196), MKI67 (Ki67), PRDM1 (Blimp1), AHNAK (Neuroblast differentiation-associated protein), Fos (c-Fos), and ISG15 (Interferon-Induced 15 KD protein). See Supplementary Figure S3 for details. (D) Patterns of PRDM1 gene expression in cell clusters defined in C (Feature and violin plots). (E) Breg suppression assay (72 h cultures) of CFSE-labelled B cells, stimulated with LPS co-cultured with Bregs-TLR/Bregs-CpG. Suppression levels are reported as percentages of Blimp1⁺ suppressed B cells. Each dot represents duplicates. ****P < 0.0001; NS: not significant.

Figure 5. Transcriptomic key pathway analysis of Bregs-TLR functional gene enrichments.

(A). Gene Set Enrichment Analysis (GSEA) of compiled RNAseq data from Naïve B, LPS-B, Bregs-CpG and Bregs-TLR cells. The enriched pathways are shown in the heatmap with positive (red) and negative (blue) gene enrichments. Red arrowheads indicated potential pathways relevant to Bregs-TLR functions. Transcriptional status of the TGF-β gene in naïve B, LPS-B, and Bregs-TLR cells as revealed by RNAseq (B) and scRNAseq analysis ((C), Feature and violin plots). (D). Assessment of TGF-β active form (viewed by release of the Latency-Associated Peptide: LAP) in CFSE-labelled B cells, stimulated with LPS and incubated for 48h with CTV-labelled Bregs-TLR (E). Overall LAP expressions are reported in dot plots; Each dot represent duplicates. **P < 0.01, ****P < 0.0001; NS: not significant.

Given that TGF-β gene upregulation is also involved in other immune suppressive processes targeting lymphocytes,²⁷ we conducted an in-depth evaluation of its putative involvement in Breg suppression. We first defined the various B cell clusters emerging from LPS stimulation that could be affected by Bregs-TLR suppression. Unsupervised clustering of LPS stimulated B cells by single-cell transcriptome analysis (scRNAseq) defined eight clusters (Figure 4C), using the annotation genes shown in Supplementary Figure S3. Visible in Figure 5A, TGF-β signaling was positively enriched only in Bregs-TLR, not in LPS-B cells or Bregs-CpG, suggesting that TGF-β could be specific of Bregs-TLR suppressive function. Consistent with this, RNA-seq and scRNA-seq data showed a decreased TGF-β gene expression after LPS stimulation, while a higher TGF-β expression was noted in TLR-Bregs in comparison with Bregs-CpG (Figure 5B and C). Tracking TGF-β expression, via the expression of LAP (TGF-β latency-associated peptide), showed that Bregs-TLR induced a decrease in LAP expression in LPS-B cells (Figure 5D and E). However, LAP expression among Bregs-TLR was similar whether facing or not B cell targets (Figure 5E). Thus, it is unlikely that TBFβ is solely responsible for the suppressive activity of Bregs-TLR cells. Suppression may require additional signals beyond the activation of TGF-β. The role of TNFα in Bregs-TLR functions is presently under investigation.

3.5. Relevance of PRDM1 gene expression in Bregs-TLR function.

The study of the distribution of groups of differentially expressed genes (DEGs) between LPS activated B cell targets and Bregs-TLR has indicated earlier that the expression of PRDM1 was upregulated in activated LPS-B cells while compromised in Bregs-TLR in comparison to their naïve B and Bregs-CpG respective controls (Figure 4B, right panels). This suggested that the silencing of PRDM1 was associated with Bregs-TLR suppression. As this gene is a key regulator of plasmocyte differentiation, we wonder whether its expression could be likewise impaired in suppressed B cells. We first studied by cell cluster distribution and gene expression profiling that, as expected, PRDM1 is only expressed in plasmocytes developing from LPS activated B cells (Figure 4C and D). We then set up Bregs-TLR suppression cocultures with LPS-B cells and examined the levels of Blimp1 protein in suppressed cells by flow cytometry. As seen in Figure 4E, expression of Blimp1 in LPS-B cells was completely inhibited by Bregs-TLR cells, whereas the same amount of control Bregs-CpG cells had not effect on Blimp1. These observations that Bregs-TLR can suppress proliferation and differentiation of LPS-B cells is consistent with prior work showing a role for PRDM1 in controlling plasma cell function through the regulation of immunoglobulin secretion.¹

4. Discussion

It has been demonstrated previously that Bregs have the capacity to suppress dendritic cells, Th17 cells, Th1 cells, and cytotoxic CD8⁺ T cells, along with preventing the secretion of pro-inflammatory cytokines and inducing Foxp3⁺ T cells through IL-10, TGF-β, and IL-35.^{8,10,13,28,29} Whether Bregs can regulate B cell activation and differentiation has yet to be elucidated. Evidence is provided in this study in support of an explicit role of Bregs-TLR in the control of B cell fate.

An interesting pathway uncovered by our investigations relates to TGF-β signaling. This gene pathway was enriched in Bregs-TLR, but not in LPS-B cells or Tregs-CpG (Figure 5A). In agreement with previous reports, it is plausible that Bregs-TLR cells exert their suppressive functions, at least in part, through TGF-β signaling.^6,29 Although Bregs-TLR highly express TGF-β as well as LAP, the proxy for TGF-β, our data reveal that LPS stimulated B cells also express LAP, relegating this cytokine as a B cell activation marker rather than an element primarily involved in Breg mediated suppression (Figure 5B–D).

Antigen-specific models of B cell activation in vivo are required to evaluate whether Breg BCR antigenic specificity impacts Breg suppression of B cells of cognate BCR specificity. However, the development of such models is notoriously challenging. Attempts by our group made use of the transgenic OB1 model in which mature B cells display a unique BCR directed at an Ova epitope.⁶ Unfortunately, adoptively transferred OB1 B cells into B cell deficient recipient (μMT, Mb1) or WT B6 mice were not activated by either splenocytes or skin graft from Ova expressing mice.⁶ The present study improved the experimental conditions by adding transgenic OTII CD4⁺ T helper cells with a TCR against an Ova T epitope. As described in Supplementary Figure S1A–B, survival/proliferation of OB1 cells was observed in μMT and, to a lesser extent, in B6 mice. This latter finding being likely due to dilution of tracked OB1 and OTII cells by overwhelming numbers of B and T cells from WT B6 origin. Surprisingly, OB1 B cells did not survive in another B-less mouse model: the Mb1^cre/cre, which has an irreversible block in B cell development at early stages and consequently has no B cells.³⁰ In contrast, the μMT mouse, which has a disruption of the first transmembrane exon of the μ heavy chain, does not express the membrane form of IgM and lack mature B cells. Comparison of OB1 B cell fates, under the same experimental conditions, in these two B cell deficient models, strongly suggest that B cell homeostasis is highly dependent on the presence of B cell progenitors and/or of immunoglobulins made by other cells than mature B cells as described.³¹ Lastly, the migration of OB1, OTII CD4⁺ T, and OB1 Bregs-TLR cells to the draining lymph nodes together with the suppression of effector OB1 proliferation at location (Figure 3) validate this model in future exploration of the role of Bregs in graft-specific prolongation of skin transplants.

Results presented in Figure 4 clearly indicated that the Bregs-TLR phenotype is characterized by a significant deficiency in PRDM1 gene expression when compared to that of control Bregs-CpG. As anticipated, B cells activated by LPS do upregulate PRDM1 expression (Figure 4B and E). However, when put in culture with Bregs-TLR cells, in conditions which suppress B cell proliferation (Figure 1), LPS-B cells also extinguish the production of Blimp1, the protein encoded by PRDM1. Thus, both Bregs-TLR cells and the cells that they suppress, exhibit a Blimp1^neg/low phenotype. The PRDM1 gene has numerous key functions associated with functional B cells, including the control of plasma cell differentiation, homeostasis, and antibodies secretion.^1,32 Contrary to our results, Blimp-1 was found to be enriched in IL-10 producing Breg cells (B10), the main subset of Bregs.³³ In this report, B10 cells were generated using protocols (LPS and anti-CD40) different from ours (LPS, CpG, PMA/ionomycin). This would suggest that Bregs-TLR constitute a Breg subset distinct from B10 cells. In addition, the low expression of Blimp1 in Bregs-TLR cells suggest that these cells present a blockage in B cell differentiation similar to that seen in PRDM1 deficient mice.³³ Therefore, the experimental conditions described in this report contribute to the attractive prospective of generating, on demand, stable Bregs-TLR for clinical applications in transplantation.

Abbreviations:

ABMR

antibody-mediated rejection

APCs

antigen-presenting cells

ASCs

antibody-secreting cells

B6

C57BL/6

BCR

B cell receptor

Bregs-TLR

TLR ligand-stimulated regulatory B cells

CTV

Cell Trace Violet

Bregs-CpG

CpG-stimulated B cell

DCs

Dendritic cells

DEGs

differentially expressed genes

DLN

draining lymph node

DSA

donor-specific antigen: DSA

FDR

False Discovery Rate

FMO

Fluorescence-minus-one

FOB

follicular B cells

GEO

Gene Expression Omnibus

IP

intraperitoneal

irra-OvaSPC

irradiated Ova splenocytes

IV

intravenous

LPS

lipopolysaccharide

LPS-B cells

LPS-stimulated B cells

MGH

Massachusetts General Hospital

MZB

marginal zone B cells

Ova

Ovalbumin

SPC

splenocytes

TI

T cell-independent

TD

T cell-dependent

Tregs

regulatory T cells