Suppression by T_FR cells leads to durable and selective inhibition of B cell effector functions

B cells suppressed by T_FR cells undergo early activation

To define mechanisms of suppression of T_FH and B cells by T_FR cells, we utilized an in vitro suppression assay in which T_FR cells are cultured with T_FH and B cells, leading to suppression of both T_FH and B cell responses^{12, 14, 16} T_FR (CD4⁺ICOS⁺CXCR5⁺CD19⁻FoxP3⁺) cells were cultured with B (CD19⁺) and T_FH. (CD4⁺ICOS⁺CXCR5⁺CD19⁻FoxP3⁻) cells from NP-OVA immunized Foxp3^GFP mice along with anti-CD3 and anti-IgM (Supplementary Fig. 1a–b). We measured robust upregulation of the GC B cell marker GL7 on B cells, CSR to IgG1, and substantial quantities of secreted IgG when B cells were cultured with T_FH cells alone (Fig. 1a). When T_FR cells were added, CSR, GL7 expression and secreted of antibody were diminished. CD4⁺CXCR5⁻ICOS⁻FoxP3⁺ T_reg cells were not able to suppress as efficiently as T_FR cells (Fig. 1b), as previously reported¹². T_FR cells also suppressed CSR of B cells in response to specific antigen (Supplementary Fig. 1c). Furthermore, T_FR cell suppression of CSR required cell contact, as supernatant from T_FR cultures could not suppress B cells (Fig. 1c). Time-lapse microscopy of cultures revealed that T_FR cells closely interacted with both T_FH and B cells (Fig. 1d and Supplementary video 1). These data suggest that T_FR cells may physically disrupt T_FH and B cell linked recognition during suppression^{2, 17}. Therefore, our culture system is a robust model for studying synchronized T_FR cell-mediated suppression of T_FH and B cells.

We first investigated whether T_FR cells prevent initial B cell activation by using cell trace violet (CTV) dye dilution of B cells to assess proliferation. Despite a decrease in the number of cell divisions, most B cells proliferated at least one cell cycle when T_FR cells were present (Fig. 1e). T_FR cells could not inhibit proliferation of B cells in co-cultures of only T_FR and B cells (without T_FH cells) and LPS + IL4 (Fig. 1f). We also analyzed CD69 expression and found that CD69 was upregulated in B cells whether or not T_FR cells were present (Fig. 1g). Studies with the active caspase reagent zVAD revealed that T_FR cells do not enhance B cell apoptosis as a means of suppression (Fig. 1h). These data demonstrate that early activation of B cells still occurs even in the presence of T_FR cells.

To determine if T_FR cells alter somatic hypermutation, we performed our in vitro suppression assay using NP-OVA. We found a low frequency of mutations in B cells when cultured with T_FH cells, and that T_FR cells reduced somatic hypermutation in B cells (Fig. 1i). T_FR cells (but not T_reg cells) also suppressed T_FH cell activation as evidenced by reduced Bcl6 and Ki67 expression (Fig. 1j). Taken together, these data demonstrate that T_FR cells allow initial B cell activation, but suppress downstream effector responses including CSR and antibody production.

T_FR cells inhibit specific genes but not effector programs

Our findings suggested that T_FR cells suppress B cells in a manner that allows B cells to receive activation signals, but not perform effector functions. To further characterize suppressed B and T_FH cells, we performed RNA-seq transcriptional analysis. We cultured B and T_FH cells (from NP-OVA immunized FoxP3^GFP mice) with or without T_FR cells (from NP-OVA immunized Actin^CFP Foxp3^GFP mice) and NP-OVA. B (CD19⁺IA⁺CD4⁻) and T_FH (CD4⁺IA⁻CD19⁻CFP⁻) cells were sorted from the ‘activated’ culture (T_FH and B cells) or from the ‘suppressed’ culture (T_FH, B and T_FR cells) and transcriptional analysis was performed (Supplementary Figure 2a–b). Principal component analysis (PCA) demonstrated modest separation of activated and suppressed B cells, but not activated or suppressed T_FH cells (Fig. 2a). We identified 1171 differentially expressed genes (FDR-adjusted P value <0.05) between activated and suppressed B cells, and 407 genes between activated and suppressed T_FH cells (Fig. 2b and Supplementary Figure 2c–e).

We next investigated whether T_FR suppression alters T_FH cell identity. We compiled a curated list of “T_FH genes” that encoded regulators of T_FH cell differentiation/function. The expression of many T_FH genes, including the essential T_FH transcription factors Bcl6 and Ascl2, were not changed during T_FR cell-mediated suppression (Fig. 2c). Bcl6 and Cxcr5 were slightly elevated in T_FH cells upon suppression. In contrast, Il4 and Il21 transcripts were markedly reduced in T_FH cells during T_FR cell suppression, consistent with previous findings¹². These data suggest that suppressed T_FH cells still retain their T_FH program, and either fail to upregulate, or actively downregulate, specific effector molecules. To confirm that suppressed T_FH cells still maintain a T_FH transcriptional program, we performed single sample gene set enrichment analysis (ssGSEA) with transcriptional signatures of T_FH cells¹⁸. Suppressed T_FH cells qualitatively retained their T_FH-like transcriptional signature (Fig. 2d). To determine if the suppressed state in T_FH cells resembles anergy, senescence or exhaustion, we compared activated and suppressed T_FH cells using GSEA, Activated T_FH cells were enriched for effector (NES=1.72, FDR<0.001) and senescence signatures (NES=1.72, FDR<0.001), but suppressed T_FH cells did not show substantial enrichment for anergy (NES=−1.45, FDR=0.110) nor exhaustion signatures (NES=1.03, FDR=0.664)(Fig. 2e). Therefore, T_FR cell suppression of T_FH cells results in a unique suppressed state that does not strongly resemble anergy or exhaustion.

Next we examined whether B cells retain their effector program during suppression by T_FR cells. We found lower expression of Ighg1, Ighg2c and Igha (Fig. 2f–g). Ighg2b transcripts were more abundant upon suppression, however IgG2b was not increased at the protein level (data not shown). Three of the most attenuated B cell transcripts in B cells suppressed by T_FR cells were Pou2af1 (encoding a transcription factor essential for GC B cell formation¹⁹), Xbp1 (encoding a protein important for secretion of antibody²⁰) and Aicda (which encodes AID, the enzyme responsible for initiating CSR^{21, 22}). Pax5, Bach2 and IRF8, genes downregulated upon differentiation to plasma cells^{23, 24}, were more highly expressed in B cells suppressed by T_FR cells, suggesting that T_FR cells may prevent differentiation of B cells to plasma cells. However, Mxd4, which is highly expressed in plasma cells, was upregulated in suppressed B cells²³. These data suggest that the suppressed B cell signature is more complex than inhibition of B cell subset differentiation.

Additional comparative analyses of activated and suppressed B cells revealed that suppressed B cells still retained their GC B cell signature (Fig. 2h). Moreover, suppressed B cells did not enrich for exhaustion (NES=−0.69, FDR=0.994), senescence (NES=1.05, FDR= 0.561), nor anergy signatures (NES=1.27, FDR=0.837) (Fig. 2i). Together, these data indicate that B cells and T_FH cells maintain their transcriptional signature when suppressed by T_FR cells, but specific effector molecules are actively downregulated.

We next used GSEA to determine if any non-effector subset related pathways were altered in B and T_FH cells during suppression by T_FR cells. Activated B cells showed strong enrichment of gene sets for Myc targets, MTORc1 signaling, oxidative phosphorylation and glycolysis compared to suppressed B cells (Table 1). Activated T_FH cells also showed enrichment of E2F targets, glycolysis and MTORc1 signaling compared to suppressed T_FH cells; however, this enrichment was not as strong as in B cells (Table 1). Thus, T_FR cell suppression leads to marked changes in pathways associated with metabolism in B and T_FH cells.

T_FR suppression alters Myc and mTOR pathways in B cells

We next tested whether Myc pathway genes were altered in B cells suppressed by T_FR cells, since this was one of the most attenuated gene sets in suppressed B cells (NES=2.77, FDR<0.0001), has roles in metabolism²⁵, and is essential for GC reactions^26,27. Most Myc target genes in B cells showed lower expression during T_FR cell suppression (Fig. 3a). To determine if suppressing Myc signals in B cells could recapitulate suppression of B cells by T_FR cells, we cultured B cells and T_FH cells with the Myc inhibitor 10058-F4 (F4)²⁸. F4 slightly reduced CSR, but robustly attenuated secreted IgG, to a similar extent as T_FR cells (Fig. 3b–c). We also tested whether Myc overexpression could result in resistance of B cells to suppression by T_FR cells. We cultured B cells from WT or Igh^Myc (Myc overexpressing²⁹) mice with T_FH and T_FR cells. Myc overexpression did not rescue CSR nor secreted antibody suppression (Fig. 3d–e). However, GL7 expression in B cells was partially rescued by Myc overexpression, suggesting that Myc overexpression may prevent some minor aspects of T_FR cell suppression of B cells.

Since we found lower expression of the mTORc1 pathway in B and T_FH cells during suppression by T_FR cells (Fig. 3f and Table 1) and the mTOR pathway promotes protein synthesis during activation and has been linked to enhancing cellular metabolism³⁰, we also investigated whether blocking the mTOR could suppress antibody production. We cultured B and T_FH cells with either the mTORc1 inhibitor rapamycin or T_FR cells. Rapamycin potently diminished CSR and antibody production similarly to T_FR cells (Fig. 3g–h). The mTORc1/mTORc2 inhibitor PP242 similarly reduced CSR and antibody production (Fig. 3i–j). Therefore, inhibiting the mTOR pathway suppresses B cell responses similarly to T_FR cells.

T_FR cells suppress metabolism in B cells

Since our GSEA analysis revealed that the expression of genes associated with glycolysis, oxidative phosphorylation, Myc and mTOR pathways was lower in B cells suppressed by T_FR cells, we further analyzed metabolic alterations. Expression of a number of metabolic pathways including serine biosynthesis, purine metabolism, one-carbon metabolism, TCA cycle, subunits of the mitochondrial electron transport chain, and glycolysis in suppressed B cells was reduced (Fig. 4a and Supplementary Figures 3–4), suggesting that central metabolic and anabolic pathways are defective in B cells suppressed by T_FR cells.

We examined the effects of T_FR cell suppression on glycolysis since this pathway is essential for antibody production³¹. We first compared expression of the glucose transporter Glut1 in B cells suppressed by T_FR cells. T_FR cells (but not T_reg cells) suppressed Glut1 expression in B cells, suggesting that T_FR cells suppress B cell glycolysis (Fig. 4b). Suppression of Glut1 expression (and CSR) in B cells by T_FR cells was not due to an increase in non-dividing cells (which are low expressers of Glut1) (Fig. 4c and Supplementary Figure 5) because comparisons of B cells that have undergone the same number of cell divisions revealed reduced Glut1 and CSR in the suppressed B cells. In addition, T_FR suppression of class switching and metabolism occurred prior to changes in B cell proliferation. When we analyzed B cells that had been added to activated or suppressed cultures and harvested 20 hours later (before the first cell division), B cells in suppressed cultures had lower Glut1 and IgG1 expression (Fig. 4d and Supplementary Figure 5). These studies decouple CSR and metabolism from proliferation, demonstrating T_FR suppression of CSR and metabolism in B cells can occur independently of changes in proliferation. T_FR cells (but not T_reg cells) also caused lower expression of Glut1 on T_FH cells (Fig. 4e).

We next analyzed glucose uptake in the culture supernatants. Glucose utilization was markedly decreased in cultures containing T_FR cells (Fig. 4f). We also measured lactate production since a large fraction of glucose-derived carbon is secreted from cells in the form of lactate. Lactate production was markedly reduced in cultures with T_FR cells (Fig. 4g). We next investigated whether inhibiting glycolysis could recapitulate T_FR cell suppression. We cultured B and T_FH cells with 2-deoxyglucose (2DG), a glucose analog that inhibits glycolysis. 2DG robustly suppressed antibody production, similar to T_FR suppression of B cells (Fig. 4h–i). In addition, glutaminolysis was lower in the presence of T_FR cells and inhibition of glutaminolysis resulted in diminished antibody production (Supplementary Figure 5).

Since one-carbon, serine and purine metabolism were attenuated during T_FR cell-mediated suppression, we assessed these pathways in more detail. In proliferating cells, folate-mediated one-carbon metabolism pathway catabolizes serine to generate one-carbon units (10-formyl THF) for de-novo purine biosynthesis. Genes encoding enzymes in one-carbon metabolism and serine biosynthetic pathway were all significantly lower in suppressed B cells (Fig. 4j and Supplementary Figure 3). Reduced expression of one-carbon metabolism genes in B cells suppressed by T_FR cells was not due to altered proliferation, because Shmt1 (P=0.0167) and Shmt2 (P=0.0104)(early upregulated cytosolic and mitochondrial enzymes in one-carbon metabolism, respectively³²) were attenuated before the first cell division (Fig. 4k). Inhibitors of purine metabolism could recapitulated T_FR cell suppression. Methotrexate (a purine synthesis inhibitor targeting Dhfr) or azathioprine (a purine analog) robustly suppressed antibody production when added to B and T_FH cell cultures (Fig. 4l–m). These data demonstrate that T_FR cells suppress multiple metabolic pathways in B cells, and inhibiting these pathways can recapitulate suppression of antibody production by T_FR cells.

T_FR suppression of B cells results in epigenetic changes

Next we examined whether the effects of T_FR cell suppression on B cells are durable and persist in the absence of T_FR cells. After 3 days of culture, we sorted B cells from activated (“act B”) or suppressed B (“supp B”) conditions, and cultured these B cells with new T_FH cells in a secondary culture (Supplementary Figure 6). We compared B cells from these secondary cultures with B cells from primary cultures (Fig. 5a). T_FH cells cultured with supp B cells in the secondary cultures co-expressed much higher levels of Ki67 and Bcl6 and were more numerous compared to T_FH cells cultured with B and T_FR cells in the primary cultures (Fig. 5a–b). Thus, B cells suppressed by T_FR cells are able to facilitate T_FH cell expansion after T_FR cells are no longer present.

We also analyzed IgG1 in B cells from these secondary cultures. Supp B cells cultured with T_FH cells were still severely defective in their ability to undergo CSR (Fig. 5c). Although most cultures contained less secreted antibody, in some cases antibody could be found in these cultures (Supplementary Figure 6), suggesting that a small population of B cells may escape suppression. Glycolysis was also defective upon restimulation of supp B cells. Glut1 expression was lower in supp B cells that were reactivated in secondary cultures with T_FH cells compared to activated B cells from primary cultures (Fig. 5d). Additionally, glucose uptake from suppressed B cell secondary cultures was much lower than activated primary cultures (Fig. 5e). Together, these data indicate that B cells suppressed by T_FR cells have defects in CSR and metabolism, which continue in the absence of continued T_FR cell contact.

To investigate whether the durable defects in B cell effector function and metabolism may be caused by epigenetic changes enforced by T_FR cells during suppression, we assessed chromatin accessibility using ATAC-sequencing (ATAC-seq). Comparisons of activated versus suppressed B cells from primary cultures revealed 2,334 genes with evidence of less accessibility in the suppressed B cell condition (Fig. 5f). Of these genes, 114 also had lower expression in RNA-seq analysis (Fig. 5g). Genes encoding B cell functional proteins and upstream regulators of metabolism, such as Aicda, Myc and Pou2af1, showed evidence of chromatin inaccessibility in the suppressed condition, suggesting epigenetic modification (Fig. 5g).

To explore how these genes may be regulated, we overlayed our ATAC-seq data with the B cell regulome annotated using ChIA-pet³³ (which identifies interactions of promoters with long range enhancers). The Aicda locus showed less accessibility in two regions, ~8kb and ~ 21kb, upstream of the Aicda transcriptional start site (TSS) (Fig. 5h). These enhancer regions are essential for AID expression^{33, 34}. We identified 7 putative enhancer regions in the Myc locus that were less accessibile in B cells suppressed by T_FR cells (Fig. 5i). We found one peak that was less accessibile in the Pou2af1 locus, which was located in an intronic region in between exons 1 and 2 (Fig. 5j). We noticed that many of the less accessible regions in suppressed B cells were not at the gene TSS, but at sites of long-range enhancers. When we quantified the distribution of all ATAC-seq regions in relation to the gene TSS, we found that less accessible regions in B cells suppressed by T_FR cells tended to be further away from the TSS compared to all regions (Fig. 5k). These data indicate that critical B cell function genes, but not the key metabolic enzymes, show evidence of epigenetic regulation during T_FR cell suppression.

IL-21 can overcome T_FR cell-mediated suppression of B cells

We tested whether IL-21 can overcome T_FR cell suppression of B cells because IL-21 is essential in the GC reaction, is suppressed by T_FR cells, can coordinate lipid and glucose metabolism in fat tissue, and can inhibit T_reg cells^{35, 36}. The addition of IL-21 to suppression cultures rescued attenuation of B cell proliferation and CSR by T_FR cells (Fig. 6a–b). IL-21 and IL-6, but not IL-4, rescued IgG secretion in cultures with T_FR cells (Fig. 6c). Rescue of CSR by IL-21 and IL-6 was not due to enhanced proliferation (Supplementary Figure 7).

We reasoned that if IL-21 could rescue antibody production, it would also rescue defects in metabolism during T_FR suppression since our studies show that metabolism and B cell function are interconnected. Glut1 expression was rescued in both B and T_FH cells when IL-21 was added to cultures containing T_FR cells (Fig. 6d and data not shown). Moreover, IL-21 rescued glucose uptake (Fig. 6e), and both IL-21 and IL-6 increased lactate production in cultures with T_FR cells (Fig. 6f). IL-21 rescue required glycolysis, as 2DG completely prevented IL-21 rescue of antibody production in suppressed cultures (Fig. 6g). Therefore, IL-21 renders B cells resistant to T_FR cell-mediated suppression of B cells at least in part by enhancing glycolysis.

To identify which B cell transcripts were rescued by IL-21, we performed RNA-seq analysis on B cells from activated (T_FH + B cells), suppressed (T_FH, B cells + T_FR cells), and IL-21 rescue (T_FH, B, T_FR cells + IL-21) cultures. Transcripts for IgG isotypes showed evidence of significant upregulation in IL-21 rescue cultures compared to suppressed cultures (Fig. 6h). Only 12 genes were differentially expressed in B cells from both suppressed versus IL-21 rescue cultures and suppressed versus activated cultures (FDR<0.05), and of these genes, only Ighg1 and Ighg2c were rescued with IL-21 (Fig. 6i and Supplementary Figure 7). Metabolic pathways such as Myc, glycolysis and oxidative phosphorylation showed evidence of some rescue by IL-21 (Fig. 6j).

We next determined whether IL-21 rescued antibody production by acting directly on B cells by performing suppression assays using B cells lacking the receptor for IL-21 (Il21r^−/−). Although baseline antibody responses were lower in cultures of Il21r^−/− B cells with T_FH cells, we found no evidence that antibody responses were rescued by the addition of IL-21 (Fig. 7a). Loss of IL-21R did not abolish increases in the number of B cells, yet prevented rescue of class switch and Glut1 expression (Supplementary Figure 7). Since GL7 upregulation is a robust indicator of B cell activation, we compared GL7 upregulation in suppressed cultures with addition of IL-21. IL-21 rescued GL7 expression when WT B cells, but not when Il21r^−/− B cells, were present (Fig. 7b–c), suggesting that IL-21 signaling into B cells is essential for the rescue of B cell activation.

To test if IL-21 also affected T_FR cells in the cultures, we compared T_FR cells from cultures. T_FR cells expressed much less Ki67 in the presence of IL-21 and there were significantly fewer T_FR cells in suppressed cultures with IL-21, suggesting IL-21 inhibits cell cycling in T_FR cells (Fig. 7d–e). Glut1 expression was higher on T_FR cells in the presence of IL-21 suggesting that IL-21 can alter T_FR cell metabolism as well as activation (Fig. 7f). These data demonstrate that IL-21 can act on both B cells and T_FR cells to overcome suppression of B cells by T_FR cells.

Mice

WT C57Bl/6J and Igh-Myc mice were purchased from Jackson Laboratories. FoxP3^IRES–GFP (FoxP3^GFP) on the C57Bl/6 background have been published previously⁴⁸. Il21r^−/− mice on the C57Bl/6 background were from the Kuchroo Lab. Actin^CFPFoxP3^GFP mice on the C57Bl/6 background have been published previously¹². All mice were between 6 and 8 weeks of age at the time of experiments and were housed in a SPF facility. Each individual experiment contained one sex of mice, but replicates were performed with males or females. All mice were used according to the Harvard Medical School Standing Committee on Animals and National Institutes of Health Guidelines.

Immunizations

Mice were immunized with 100μg NP-OVA (Biosearch Technologies) emulsified in H37RA CFA s.c. in the mouse flanks as previously described^{12, 16}. Mice were sacrificed 7 days later and inguinal lymph nodes were harvested.

Antibodies

The following antibodies were used for surface staining at 4°C: anti-CD4 (Biolegend, 1:200, RM4-5)¹², anti-ICOS (Biolegend,1:200, 15F9)¹², anti-CD19 (Biolegend,1:200, 6D5)¹², anti-CXCR5 biotin (BD Biosciences,1:100, 2G8)¹², GL7 (BD Biosciences,1:200, GL-7)¹², CD69 (Biolegend,1:200, H1.2F3)¹², and anti-IA (Biolegend,1:200, M5/114.15.2)¹². For further CXCR5 detection, streptavidin-BV421 (Biolegend, 1:400, 405225) was used at 4°C. For intracellular staining, samples were fixed with the Foxp3 Fix/Perm buffer set according to the manufacturer’s instructions (eBioscience). Samples were then intracellularly stained with anti-IgG1 (BD Biosciences, 1:200, A85-1)¹², anti-FoxP3 (eBiosciences, 1:200, FJK-16S)¹², anti-Ki67 (BD Biosciences, 1:100, B56)¹², anti-Glut1 (Abcam, 1:200, EPR3915)³¹, anti-Shmt2 (Abcam, 1:200, ab64417,)³² or anti-Shmt1 (Novus Biologicals, 1:200, NBP2-32173,)³² at 4°C. In some cases, a donkey anti-rabbit BV421 secondary was used (Biolegend, 1:400, 406410)

Sorting

Single cell suspensions were diluted in PBS supplemented with 1% FBS with 1mM EDTA. T_FH and T_FR cells were isolated by enriching with CD4⁺ cells by magnetic positive selection (Miltenyi Biotec). CD4⁺ enriched cells were then stained and sorted as follows: T_FH (CD4⁺ICOS⁺CXCR5⁺FoxP3⁻CD19⁻), T_FR (CD4⁺ICOS⁺CXCR5⁺FoxP3⁺CD19⁻). In some cases, an alternative gating strategy was used; T_FH (CD4⁺ICOS⁺CXCR5⁺GITR⁻CD19⁻), T_FR (CD4⁺ICOS⁺CXCR5⁺GITR⁺ CD19⁻) as we have previously described^{12, 13}. B cells were isolated from flow-through from CD4⁺ selection, which was then positively selected using CD19 beads (Miltenyi Biotec). For WT versus Igh-Myc B cells, B cells were enriched with magnetic selection and then sorted as CD19⁺GL7^– cells to rule out effects of spontaneous activation in Igh-Myc mice.

Suppression assays

In vitro suppression assays were performed as described previously^{12, 14, 16}. Briefly, FoxP3^GFP reporter mice were immunized with NP-OVA and 7 days later dLN were harvested and CD19⁺ B cells and CD4⁺CXCR5⁺ICOS⁺FoxP3⁻CD19⁻ T_FH cells were purified by cell sorting. B and T_FH cells were cultured alone (“Activated”) or with CD4⁺ICOS⁺CXCR5⁺CD19^-FoxP3⁺ T_FR cells (“Suppressed”) sorted from NP-OVA immunized FoxP3^GFP or FoxP3^GFP Actin^CFP mice. 5×10⁴ B cells, 3×10⁴ T_FH cells and/or 1.5×10⁴ T_FR cells were plated in 96-well plates along with 2μg/ml anti-CD3 (BioXcell) and 5μg/ml anti-IgM (Jackson Immunoresearch). For some experiments, 20μg/ml NP-OVA was added to wells instead of anti-CD3 and anti-IgM. For some studies B cells were pre-labeled with the proliferation dye cell trace violet (Thermo Scientific). Cultures were harvested after 2–6 days as described. For analysis, B cells were gated as CD19⁺IA⁺CD4⁻ cells, T_FH cells were gated as CD4+IA−CD19−FoxP3- cells, and T_FR cells were gated as CD4⁺IA⁻CD19⁻FoxP3⁺ cells. For resorting for RNA-seq analysis of NP-OVA containing cultures, B cells were gated as CD19⁺IA⁺CD4⁻CFP^– and T_FH cells were gated as CD4⁺IA⁻CD19⁻CFP⁻. For ATAC-seq analysis B cells were harvested after 4 days of culture with NP-OVA and were gated as CD19⁺IA⁺CD4⁻ cells. For Igh sequencing, B cells were harvested after 6 days of culture with NP-OVA and were gated as CD19⁺IA⁺CD4⁻ cells. For reculturing of suppressed cells, suppressed cultures (which contained anti-CD3/IgM) were harvested after 3 days and B cells were sorted as CD19⁺IA⁺CD4⁻. 2.5×10⁴ suppressed B cells (or freshly isolated B cells) were then cultured with freshly isolated 3×10⁴ T_FH cells, prepared as above, in the presence of anti-CD3/IgM. In some studies, small molecule compounds were added to cultures: Myc inhibitor 10058-F4 (100μM), mTOR inhibitors rapamycin (10nM) and PP242 (200nM), methotrexate (1μM), azathioprine (50μM) (all from Tocris Bioscience), glycolysis inhibitor 2-deoxyglucose (2DG; 500μM) and glutaminolysis inhibitor BPTES (8μM) (both from Sigma). In some experiments recombinant IL-21, IL-4 or IL-6 (Peprotech) were added to wells (at 60–90ng/ml).

RNA-seq

Samples were sorted as described above and each replicate indicates a biological replicate that was prepared using different sets of mice on different experimental days. RNA-seq library preparations were performed (Chevrier, N. manuscript in preparation). Briefly, RNA was isolated using MyOne Silane Dynabeads (Thermo Fisher Scientific). RNA was fragmented, and barcoded using 8bp barcodes in conjunction with standard Illumina adaptors. Primers were removed using Agencourt AMPure XP bead cleanup (Beckman Coulter/Agencourt) and samples were amplified with 14 PCR cycles. Libraries were gel purified and quantified using a Qubit high sensitivity DNA kit (Invitrogen) and library quality was confirmed using Tapestation high sensitivity DNA tapes (Agilent Technologies). RNA Sequencing reactions were sequenced on an Illumina HiSeq 2000 or Illumina NextSeq sequencer (Illumina) according to manufacturer’s instructions, sequencing 50bp reads. Analysis was performed using the CLC Genomics Workbench version 8.0.1 RNA-seq analysis software package (Qiagen). Briefly, reads were aligned (mismatch cost=2, insertion cost=3, deletion cost=3, length fraction=0.8, similarity fraction=0.8) to the mouse genome and differential expression analysis was performed (total count filter cutoff=5.0). Results were normalized to reads per million. Gene-e (Broad Institute) was used to generate heatmaps.

ATAC-seq

50,000 sorted cells per biological replicate were washed once in cold PBS and lysed in 50μL cold lysis buffer (10 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl2, 0.1% IGEPAL CA-630). Lysed nuclei were incubated in Tn5 transposition reaction mix as described⁴⁹ and purified using MinElute Reaction Cleanup kit (Qiagen). ATAC-seq fragments were size-selected for fragments between 115 and 600 bp using Pippin Prep 2% Agarose Gel Cassettes and the Pippin Prep DNA Size Selection System (Sage Science). Post size-selection, ATAC libraries were amplified and Nextera sequencing primers ligated using Polymerase Chain Reaction (PCR). Finally, PCR primers were removed using Agencourt AMPure XP bead cleanup (Beckman Coulter/Agencourt) and library quality was verified using a Tapestation machine. Samples were sequenced on an Illumina NextSeq sequencer. Peak calling and analysis was performed using CLC Genomics Workbench 8.0.1 ChipSeq analysis software (Qiagen) by combining all samples (maximum P value for peak calling=0.05). Gene-peak associations were determined using the GREAT software package utilizing the basal extension method⁵⁰. Visualization was performed using Integrated Genomics Viewer (Broad Institute) using the mm10 mouse genome build. Gene tracks were annotated with previously published ChIA-pet gene-enhancer associations in B cells³³.

Microscopy

In vitro suppression assays were performed as above except B cells were pre-labeled with eFluor670 proliferation dye (eBioscience) and cells were cultured in cytek imaging chambers coated with celltak (VWR). After 4 days of culture, cell mask orange (Thermo Fisher Scientific) was added to the imaging chambers to detect membrane of B, T_FH and T_FR cells. Samples were imaged on a Nikon spinning disk confocal microscope using a 43x objective and standard lasers/filters. Z-stacks were converted to projections using ImageJ software. B cells were identified as blue (eFluor670) and red (cell mask orange), T_FR cells were identified as green (FoxP3^IRES-GFP) and red (cell mask orange), and Tfh cells were identified as red (cell mask orange). Movie shown was representative of imaging studies that were repeated in three different biological replicates.

IgH Sequencing

B cells were sorted from suppression assays as above, and IgH sequencing was performed using the Adaptive Biotechnologies ImmunoSeq mouse IgH sequencing platform (Adaptive Biotechnologies).

Gene Set Enrichment Analysis (GSEA)

For GSEA analysis, RNA-seq data were converted to human nomenclature and compared to GSEA mSigDatabases including Hallmarks, C3 and C5 collections using standard settings (Broad Institute). For enrichment plots of specific gene sets, pathways were analyzed along with 20 randomized gene sets to ensure specificity. Single sample GSEA was performed in GenePattern using indicated gene sets (Broad Institute).

ELISA

ELISA measurements of total IgG from culture supernatants were performed as described previously^{12, 16}.

Statistics

Most statistical tests were performed using Prism 6.0 (GraphPad) utilizing Student’s two-tailed unpaired t test or 1-way ANOVA with Tukey’s correction as specified assuming Gaussian distribution. Statistics for RNA-seq and ATAC-seq were performed using CLC Genomics Workbench (Qiagen). Statistics for gene set enrichment was performed in GSEA (Broad Institute). Statistics for Volcano plots were performed in Microsoft excel using a Chi² test. Samples were not randomized or blinded, and there was no exclusion of data. Sample sizes were chosen as the minimum number of mice (typically 20) to sort enough cells to perform 3–4 technical replicate stimulation/suppression assays.