Memory B cells (MBCs) have enhanced capabilities to differentiate to plasma cells and generate a rapid burst of Abs upon secondary stimulation. To determine if MBCs harbor an epigenetic landscape that contributes to increased differentiation potential, we derived the chromatin accessibility and transcriptomes of influenza-specific IgM and IgG MBCs compared with naive cells. MBCs possessed an accessible chromatin architecture surrounding plasma cell–specific genes, as well as altered expression of transcription factors and genes encoding cell cycle, chemotaxis, and signal transduction processes. Intriguingly, this MBC signature was conserved between humans and mice. MBCs of both species possessed a heightened heme signature compared with naive cells. Differentiation in the presence of hemin enhanced oxidative phosphorylation metabolism and MBC differentiation into Ab-secreting plasma cells. Thus, these data define conserved MBC transcriptional and epigenetic signatures that include a central role for heme and multiple other pathways in augmenting MBC reactivation potential.

This article is featured in Top Reads, p.1407

Long-lived humoral immunity is realized through the joint functions of memory B cells (MBCs) and plasma cells (PCs). Naive B cells (nBs) and MBCs have similar fates upon T-dependent stimulation, although MBCs respond more quickly and with greater magnitude to antigenic challenge (1, 2). For the most part, MBCs tend to be affinity matured with class-switched BCRs (35) and gain unique surface markers, such as CD73, PD-L2, and CD80 (6, 7). It is now also appreciated that IgM+ MBCs are also key players in memory responses (8, 9). For class-switched MBCs (SM), enhanced immune responses have been attributed in part to the signaling capacity of the IgG versus IgM BCR (4). However, naive cells genetically engineered to express the IgG BCR respond similarly to IgM+ nBs (10). Together, these data suggest that alternative mechanisms function to promote heightened MBC functional responses.

One possibility is that MBCs have acquired epigenetic programming that distinguishes them from their naive precursors. Evidence supporting this hypothesis includes the role of the monocytic leukemia zinc finger protein (MOZ), a histone acetyltransferase that targets H3K9 and is required for high affinity SM (11). Human MBCs harbor a unique DNA methylation profile compared with naive cells, germinal center (GC) B cells, and PCs (12, 13). In a similar manner, memory CD8 T cells maintain a primed chromatin accessibility architecture at key effector genes (14). Changes in the levels of transcription factors also correlate with MBC properties. For example, Krüppel-like factor 4 (KLF4) restricts the proliferative response and is decreased in expression in MBCs compared with naive. Similarly, downregulation of BACH2, a repressor of BLIMP-1 and the PC fate, has been proposed as one mechanism for enhanced MBC differentiation potential (10). In contrast, upregulation of ZBTB32 restrains MBC secondary responses (15). Despite this understanding, the molecular programming of MBC subsets remains incomplete.

In this study, we define the transcriptional and chromatin accessibility profiles of IgM and IgG MBC subsets compared with nBs. We found that MBCs harbor distinct epigenetic programming from nBs that is conserved between the MBCs of mice and humans. This programming includes the priming of accessible chromatin surrounding PC-specific genes, some of which are BLIMP-1 targets. Moreover, we found that heme biosynthesis was a conserved key feature of the MBC state in which augmented differentiation to PCs was observed in the presence of the iron-containing complex hemin. In summary, our findings describe unique epigenetic parameters and features of MBC programming that allow these cells to respond more quickly and vigorously to secondary immune challenges.

Sequencing data generated in this study have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) and are available under accession GSE164173. Code used to analyze the data and generate the figures is available from http://github.com/cdschar and upon request.

C57BL/6J wild-type mice were obtained from The Jackson Laboratory (000664). All experiments were performed with a mix of male and female mice that were between 8 and 10 wk of age. All animals were housed in specific pathogen-free cages by the Emory Division of Animal Resources, and all protocols were approved by the Emory Institutional Animal Care and Use Committee. Both male and female mice that were between 8 and 10 wk of age were infected with influenza A/PR8/34 (PR8) at 15,000 virus-forming units per dose. After 35 d postinfection, mice were sacrificed or challenged with a 0.1 median lethal dose (LD50) of influenza A/HK-X31 (×31). For all infections, mice were anesthetized with vaporized isoflurane, and 30 μl of PR8 or 100 μl of ×31 was administered intranasally.

Splenic cell suspensions were made by mechanically forcing spleens through a 40-μm filter and lysing RBCs with ammonium–chloride–potassium lysis buffer (0.15 M NH4Cl, 10 mM KHCO3, and 0.1 mM EDTA) for 3 min before quenching the reaction with four volumes of complete RPMI media: RPMI 1640 (50-020-PC; Corning Cellgro), 10% heat-inactivated FBS (Sigma-Aldrich), 1% MEM nonessential amino acids (ENBF3930-01; Sigma-Aldrich), 100 μM sodium pyruvate (RNBF6686; Sigma-Aldrich), 10 mM HEPES (SH30237; HyClone), 1× penicillin–streptomycin–glutamine (10378016; Life Technologies), and 0.0035% 2-ME (Sigma-Aldrich). nBs were isolated using negative selection for CD43 (130-090-862; Miltenyi Biotec) following the manufacturer’s protocol. Purity was confirmed by flow cytometry. B cell Ag tetramers were described previously (16). Influenza nucleoprotein (NP)–specific MBCs were enriched using B cell tetramer pull-downs. Cells were stained with 1:100 NP-PE for 30 min, washed, and incubated with anti-PE microbeads (130-105-639; Miltenyi Biotec) before positive selection with magnetic columns. MBCs for ex vivo cultures were enriched using the Memory B Cell Isolation Kit, Mouse (130-095-838; Miltenyi Biotec), following manufacturer’s protocol.

nBs and total enriched MBCs were isolated as described above, resuspended at 0.5 × 106 cells/ml, and stimulated with 500 ng/ml soluble CD40L (8230-CL; R&D Systems), 10 ng/ml IL-4 (404-ML; R&D Systems), and 10 ng/ml IL-5 in complete RPMI 1640. Cultures were fed on each subsequent day with 10 ng/ml IL-4 and 10 ng/ml IL-5 until harvest at the indicated time points. For hemin experiments, 60 μM hemin (Sigma-Aldrich) or vehicle (NH4OH) was added to cultures on day 1.

PBMCs were obtained from deidentified healthy donors according to approved protocols from the Emory Institutional Review Board. For each isolation, 40 ml of blood was diluted 1:2 with PBS, followed by layering over a 50% volume of Ficoll (17-1440-02) and centrifuged at 1850 rpm for 30 min with minimal acceleration and deceleration. The lymphocyte layer was removed and human MBCs enriched using the human MBC kit (130-093-546; Miltenyi Biotec) following the manufacturer’s protocol.

Enriched human MBCs were cultured at 0.2 × 106 cells/ml in complete RPMI 1640 supplemented with 1 mg/ml R848 (InvivoGen), 1 mg/ml soluble CD40L (Fitzgerald Industries), 10 ng/ml IL-21, 50 U/ml IL-2, 10 ng/ml BAFF (PeproTech), and 0.8 μl/ml human insulin for 6 d. On day 2 of the culture, vehicle or 60 μM hemin was added as indicated.

Staining panels for mouse included anti-Fc (anti-CD16/CD32) (2.4G2; Tonbo Biosciences), anti-CD11b (M1/70; Tonbo Biosciences), anti-CD11c (N418; Tonbo Biosciences), anti-F4/80 (BM8; BioLegend), and CD90.1 (30-H12; BioLegend) conjugated to APC-Cy7 each at a concentration of 0.25 μg per 1 × 106 cells to remove dendritic cells, macrophages, and T cells. The following stains and Ab–fluorophore combinations were used to assess cellular phenotype: anti-B220–PE-Cy7 or –Alexa Fluor 700 (RA3-6B2; Tonbo Biosciences) at 0.5 μl per 1 × 106 cells; anti-CD138–APC or –BV711 (281-2; BD) at 0.125 μl per 1 × 106 cells; anti-GL7–PerCp-Cy5.5, –PE, or –e450 (GL7; BioLegend) at 0.25 μl per 1 × 106 cells; anti-CD38–Pacific Blue (90; BioLegend); anti-IgD–BV711 (11-26c.2a; BD Biosciences); anti-IgM–FITC (II/41; Invitrogen); anti-IgG–PerCp-Cy5.5 (Poly4053; BioLegend); Zombie Yellow (BV570) (77168; BioLegend); and CellTrace Violet (C34557; Life Technologies). CellTrace Violet was used at 10 μM per 1 × 107 cells/ml. Human staining panels included anti-CD3–Alexa 700 (UCHT1; Invitrogen), anti-CD19–APC (SJC25C1; BioLegend), anti-CD27–BV786 (L128; BD), anti-IgD–APC-Cy7 (IA6-2; BioLegend), anti-CD38–PE-Cy7 (HIT2; BioLegend), and anti-CD138–PerCp-Cy5.5 (MI15; BD Biosciences) or anti-CD138–Alexa 700 (MI15; BioLegend), all used at 5 μl per 1 × 106 cells. Intracellular heme oxygenase-1 (HO-1) staining for human and mouse was performed using the Invitrogen FIX & PERM Kit (GAS003) using HO-1 PE at 1:100 dilution (ab83214; Abcam) and PE IgG2b isotype (ab91532; Abcam) for gating. Staining panels included fluorescence minus one controls to ensure that correct compensation and gating were applied. BD LSR II or BD LSRFortessa X-20 was used for analysis, and an FACSAria II was used for sorting (BD Biosciences). All flow cytometry data were analyzed with FlowJo v10.

nBs were isolated from the draining lymph node (dLN) of age-matched naive mice and sorted as follows: Live CD11bF4/80Thy1.2B220+CD38+GL7IgM+. MBCs were isolated from four pooled memory mice and sorted as follows: Live CD11bF4/80Thy1.2B220+CD38+GL7NP-PE+NP-APC+IgM+ or IgG+. Five hundred cells were isolated via FACS into 300 μl RLT buffer (QIAGEN) containing 1% 2-ME. At the time of RNA isolation, 3 μl of a 1:400,000 dilution of External RNA Controls Consortium synthetic mRNAs (Thermo Fisher Scientific) were added, and total RNA was isolated using a Quick-RNA MicroPrep Kit (11-328M; Zymo Research) and all subsequent RNA used for cDNA synthesis using the SMART-seq v4 kit (Takara) and 10 cycles of PCR. To generate final libraries 200 pg of cDNA was used as input for the NexteraXT DNA Library kit (Illumina) using 10 additional cycles of PCR. An Agilent Bioanalyzer was used to quality control check each library. Libraries were pooled at an equimolar ratio and sequenced on a NextSeq 500 (Illumina) using 75-bp paired-end chemistry.

Raw sequencing reads were mapped to the mm10 genome using STAR (17). Gene counts were determined with the Bioconductor package GenomicAlignments (18) using the mm10 transcriptome database. The Bioconductor package edgeR (19) was used to determine differentially expressed genes (DEGs; false discovery rate [FDR] ≤0.05, absolute log2 fold-change (log2FC) ≥1). All DEGs are listed in Supplemental Table I. Hierarchical clustering bootstrapping analysis was performed using the pvclust v2.2-0 R package with 1000 replications. Gene ontology (GO) analysis was performed with the Database for Annotation, Visualization and Integrated Discovery (20). Preranked gene set enrichment analysis (GSEA) (21) was performed on all detected transcripts ranked by multiplying the sign of fold change by the −log10 of the edgeR p value. Normalized mRNA content was calculated as previously described (22). To compare gene expression between species the ENTREZ gene identifier for both human and mouse was used as input for the National Center for Biotechnology Information HomoloGene database (23). Only unique matches were used for downstream analysis, resulting in 8639 unique genes that were detected in both species.

Naive and MBCs were sorted as defined above. At least 1000 cells for each cell type were isolated via FACS for the assay for transposase-accessible chromatin sequencing (ATAC-seq). Libraries were generated as described in detail previously (24). Briefly, cells were centrifuged at 500 × g for 10 min at 4°C. The supernatant was removed with a pipette, and cells were resuspended in 25 μl Tn5 Reaction Buffer (2.5 μl Tn5, 12.5 μl 2× TD buffer, 5 μl molecular grade H2O, 2.5 μl 1% Tween 20, and 2.5 μl 0.2% digitonin) and incubated at 37°C for 1 h. Cells were lysed by adding 2 μl 10 mg/ml Proteinase K and 23 μl lysis buffer (326 mM NaCl, 109 mM EDTA, and 0.63% SDS) and incubated for 30 min at 40°C. Large m.w. DNA was removed by addition of 0.7× volumes of AMPure XP SPRI beads (Agencourt, A63880; Invitrogen) and low m.w. DNA isolated for downstream analysis using a 1.2× volumes AMPure XP SPRI beads and eluted in 15 μl EB buffer (QIAGEN). ATAC-seq libraries were amplified using 5 μl each of the i5 and i7 Nextera Indexing primers and 25 μl of 2× HiFi HotStart ReadyMix (KK2601; Roche Diagnostics). Final libraries were purified using 1× volumes of AMPure XP SPRI beads and sequenced using 75-bp paired-end reads on an Illumina NextSeq 500.

Raw sequencing reads were mapped to the mm10 genome using Bowtie (25). Peaks were called using MACS2 (26) and annotated to the nearest transcription start site using HOMER (27). Data were normalized to reads per peak per million as previously described (28), and all samples demonstrated a fraction of reads in peaks score of 21% or greater. The coverage for all peaks was annotated using the GenomicRanges (18) package, and edgeR (19) was used to determine differential accessibility between samples (FDR ≤ 0.05 and absolute log2FC ≥ 1). All differentially accessible regions (DAR) are listed in Supplemental Table II. Motif analysis was performed with the HOMER program findMotifsGenome.pl using randomly generated genomic sequences as background. PageRank analysis (29) was performed using the default parameters and transcription factors with a PageRank score >0.001 in at least one sample group of the indicated comparison were included for all downstream analysis.

Cells were lysed in 300 μl RLT buffer (QIAGEN) containing 1% 2-ME, and total RNA was isolated using the Quick-RNA MicroPrep Kit (11-328M; Zymo Research). All subsequent RNA was used for reverse transcription using random hexamer and oligo dT primers with SuperScriptII (18064022; Thermo Fisher Scientific). cDNA was diluted 1:5 and gene expression quantitated by real-time PCR using SYBR Green incorporation on a Bio-Rad iCycler. Gene expression was quantitated as a percentage of 18S rRNA. All primer sets are listed in Supplemental Table III.

For ELISPOT assays, multiscreen HA plates (MAHAS4510; MilliporeSigma) were coated with purified NP at 1 μg/ml in PBS overnight at 4°C. Plates were washed four times with PBS and blocked with complete RPMI 1640. Single-cell suspensions were washed, diluted in complete RPMI 1640, and cultured on coated plates overnight in a 37°C incubator with 5% CO2. Cells were aspirated, and plates were washed with 0.2% Tween 20 in PBS. Bound IgG was detected using alkaline phosphatase–conjugated goat anti–mouse κ (diluted 1:1000; SouthernBiotech) in 0.5% BSA and 0.05% Tween 20 in PBS for 1 h at 37°C. Plates were washed with 0.2% Tween 20 in PBS and developed with 5-bromo-4-chloro-3-indolyl phosphate/NBT (Moss Substrates) substrate for 1 h. ImmunoSpot S6 ULTIMATE Analyzer (Cellular Technology Limited) and quantified using ImmunoSpot software v5.0.9.21 using the Basic Count feature with the following parameters: sensitivity, 165; background balance, 162; diffuse processing, largest; spot size range, 0.0201–9.626 mm2; and spot separation, 1.

A FluxPack cartridge was hydrated for 12 h prior to each assay with 200 μl dH2O at 37°C in a non-CO2 incubator. Approximately 1 h prior to each assay, dH2O was replaced with prewarmed Seahorse calibrant solution (103059-000; Agilent Technologies). Cell cultures were washed in Seahorse XF Assay media supplemented with 1 mM sodium pyruvate, 2 mM l-glutamine, and 5.5 mM glucose. Cells concentrations were determined by flow cytometry using AccuCheck counting beads (PCB100; Invitrogen). Cell-Tak (354420; Corning) was diluted in sterile 1× PBS to a final concentration of 22.4 μg/ml and 25 μl added to each well of a Seahorse XFe96 cell culture plate and incubated for 20 min at room temperature. Following the incubation, Cell-Tak was washed out with 200 μl dH2O, and 250,000 cells were plated per well. The resulting plate was incubated at 37°C in a non-CO2 incubator for 45 min. Each drug was diluted in Seahorse media and loaded onto the following ports: port A, 2 μM oligomycin (75351; Sigma-Aldrich); port B, 2.5 μM carbonyl cyanide-4-phenylhydrazone (C2920; Sigma-Aldrich); and port C, 1 μM rotenone (R8875; Sigma-Aldrich) and 1 μM Antimycin A (A8674; Sigma-Aldrich). Following the incubation, the Mitochondrial Stress Test was performed on a Seahorse Bioanalyzer XFe96 instrument.

The statistical tests and exact group sizes are indicated in each figure legend. Statistical tests were performed in Prism v8.4.1, Excel v16.35, or R v3.5.3. Analysis of DEGs or accessible regions was performed using edgeR (19) and the Bonferroni FDR correction in R. DEGs and DAR were determined based on both an FDR ≤0.05 and absolute log2FC ≥1. GSEA p values are determined by permutation testing. Significance between time course quantitative RT-PCR data were determined by two-way ANOVA with Tukey post hoc correction.

The murine model of influenza infection using heterosubtypic A/PR8/34 (PR8) and A/HK-X31 (×31) viral strains allows for the tracking of MBC responses to the shared NP segment of influenza (16, 30, 31). To characterize the differentiation capacity between MBCs and nBs, C57BL/6J mice were infected with a primary dose of PR8 and allowed to rest for 35 d to generate Ag-experienced memory mice. Memory and naive C57BL/6J mice were then challenged with ×31, and the early differentiation kinetics (days 4–7) were monitored by flow cytometry. Memory mice showed enhanced generation of GL7+FAS+ GC B cells by both frequency and cell number in the spleen and mediastinal dLN at all time points assayed (Fig. 1A, Supplemental Fig. 1). PC differentiation was low in both naive and memory mice at day 4, but significantly increased in memory mice compared with naive at days 5 and 6 after ×31 challenge. Naive mice showed similar frequencies and numbers of CD138+ PCs at later time points in the spleen and dLN.

FIGURE 1.

Memory mice show increased kinetics of differentiation and generate more Ag-specific, isotype-switched PCs. (A) Frequency of FAS+GL7+ GC B cells or CD138+ PCs from naive or memory mice challenged with 0.1 LD50 influenza ×31 at the indicated time point. Data were collected from the mediastinal dLN or spleens (SPL) and are representative of two independent experiments with four to five mice per group. (B) ELISPOT analysis of NP-specific IgM (right) or IgG (left) Ab-secreting cells at days 4 or 5 after ×31 challenge. SPL and dLN were pooled, and 500,000 total cells were plated per well. (C) Quantitation of IgM and IgG NP-specific ELISPOTS per 500,000 cells. Significance determined using two-tailed Student t test. See also Supplemental Fig. 1. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 1.

Memory mice show increased kinetics of differentiation and generate more Ag-specific, isotype-switched PCs. (A) Frequency of FAS+GL7+ GC B cells or CD138+ PCs from naive or memory mice challenged with 0.1 LD50 influenza ×31 at the indicated time point. Data were collected from the mediastinal dLN or spleens (SPL) and are representative of two independent experiments with four to five mice per group. (B) ELISPOT analysis of NP-specific IgM (right) or IgG (left) Ab-secreting cells at days 4 or 5 after ×31 challenge. SPL and dLN were pooled, and 500,000 total cells were plated per well. (C) Quantitation of IgM and IgG NP-specific ELISPOTS per 500,000 cells. Significance determined using two-tailed Student t test. See also Supplemental Fig. 1. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Close modal

To further elucidate differences in influenza-specific PC differentiation, naive and memory mice were infected with ×31, and NP-specific ELISPOTs were performed at days 4 and 5 postinfection. NP-specific IgM spots were similar between naive and memory mice at day 4 postchallenge (Fig. 1B, 1C). However, there were significantly more NP-specific IgG spots from pooled spleens and dLN isolated from PR8-primed memory mice than from naive mice. Thus, as expected from the reports using model Ags and malaria infection (6, 8, 10, 32), the system employed in this study allows the elucidation of MBCs that form following a viral infection.

To investigate the molecular properties that underlie enhanced MBC differentiation compared with nBs; at day 35 after PR8 infection, influenza-specific MBCs were isolated using NP B cell tetramers (16). MBCs were subset into those that expressed either an IgM or IgG BCR (Supplemental Fig. 2). In addition, follicular IgM+ nBs from the dLN were isolated from naive mice and the transcriptome of each population determined by RNA sequencing (RNA-seq). Principal component analysis (PCA) of all 10,324 detected genes from the RNA-seq data showed that each of the three subsets segregated into discrete clusters (Fig. 2A). Principal component 1 separated the two MBC populations from the nBs and principal component 2 distinguished IgM MBCs from the other two subsets. Hierarchical clustering followed by bootstrapping analysis of the DEGs (absolute log2FC >1 and FDR <0.05) revealed that IgM MBCs were a transcriptionally distinct MBC population with both unique and shared gene expression patterns compared with both nBs and IgG MBCs (Fig. 2B, Supplemental Table I). Analysis of the DEG overlap between MBCs and nBs showed 191 signature genes that were similarly expressed in both IgM and IgG MBCs (Fig. 2C). Synthetic spike-in mRNAs were used to normalize and quantitate the global mRNA content (22) and revealed a significant increase in IgM and IgG MBCs as compared with nBs (Fig. 2D).

FIGURE 2.

MBCs have distinct transcriptional signatures. (A) PCA of 10,324 detected genes in follicular IgM+ nB (n = 3) from naive mice and influenza NP-specific IgM (n = 3) and IgG (n = 2) MBC 35 d after PR8 infection. See also Supplemental Fig. 2. (B) Heatmap analysis of 901 DEGs. The asterisk (*) indicates 100% reproducibility based on bootstrapping analysis. (C) Venn diagram showing the overlap of 287 IgM versus nB DEG (blue circle) with 478 IgG versus nB DEG (green circle). (D) Normalized mRNA content calculated as the sum of all transcripts in a given subset. (E) Volcano plots for each comparison with relevant genes upregulated in the MBC subsets indicated. (F) GO analysis of DEGs in every comparison. Bar plots for key genes highly expressed in both MBC populations compared with nB (G), IgM or IgG MBC (H), or downregulated in both MBC populations (I). (J) GSEA of the REACTOME Cell Cycle Mitotic gene set in each comparison. The asterisk (*) in (G)–(I) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). See also Supplemental Fig. 3. NES, normalized enrichment score.

FIGURE 2.

MBCs have distinct transcriptional signatures. (A) PCA of 10,324 detected genes in follicular IgM+ nB (n = 3) from naive mice and influenza NP-specific IgM (n = 3) and IgG (n = 2) MBC 35 d after PR8 infection. See also Supplemental Fig. 2. (B) Heatmap analysis of 901 DEGs. The asterisk (*) indicates 100% reproducibility based on bootstrapping analysis. (C) Venn diagram showing the overlap of 287 IgM versus nB DEG (blue circle) with 478 IgG versus nB DEG (green circle). (D) Normalized mRNA content calculated as the sum of all transcripts in a given subset. (E) Volcano plots for each comparison with relevant genes upregulated in the MBC subsets indicated. (F) GO analysis of DEGs in every comparison. Bar plots for key genes highly expressed in both MBC populations compared with nB (G), IgM or IgG MBC (H), or downregulated in both MBC populations (I). (J) GSEA of the REACTOME Cell Cycle Mitotic gene set in each comparison. The asterisk (*) in (G)–(I) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). See also Supplemental Fig. 3. NES, normalized enrichment score.

Close modal

To further explore the specific differences between subsets, the relationship between fold change and significance (FDR) was plotted, and GO (33) analysis was performed on the DEG sets from each comparison group. Both IgM and IgG MBCs showed upregulation of GO terms describing activation, proliferation, and immune system processes, highlighted by Tbx21 and Cd80 (Fig. 2E, 2F). Tbx21, which encodes the transcription factor T-bet, promotes the expression of the chemokine receptor Cxcr3 (34), which was also upregulated in both MBC subsets, and is required for secondary PC responses to influenza (35). Additional genes that have similar expression patterns in both IgM and IgG MBCs included Traf1, Vcam1, and Il9r (Fig. 2G). Traf1 is a member of the TNFR family and heterodimerizes with TRAF2 to mediate TNF-α–induced NF-κB (36). Similar to Traf1, the adhesion molecule Vcam1 is upregulated in response to NF-κB signaling (37). IL-9R was previously shown to be critical for MBC differentiation from the GC (38). Intriguingly, Il9r expression was found to be >3-fold higher in IgM compared with IgG MBCs in our data.

DEGs uniquely enriched in MBC subsets included Tlr3 and S1pr1 in IgM and Nte5, Mt1, and Pik3r6 in IgG MBCs (Fig. 2H). TLR-3, encoded by Tlr3, senses intracellular dsRNA and is upregulated in murine B cells upon BCR stimulation (39). S1PR1 is important for egress of lymphocytes from the lymph nodes (40), as well as release of developing B cells from the bone marrow (41). Consistent with its expression on SM (32), Nt5e, which encodes CD73, was found to be more highly expressed by IgG MBCs as compared with both nBs and IgM MBCs. Mt1 plays a role in iron ion homeostasis (42), and Pik3r6 regulates PI3K signaling, which is activated downstream of BCR triggering (43). The unique expression of Pik3r6 in IgG MBCs suggests a differential role for this protein downstream of IgM and IgG BCRs (2, 3). Further functional exploration of the DEG using GSEA (44) identified that cell cycle genes were upregulated in IgG MBCs compared with both IgM MBCs and nBs, underscoring their rapid responses to stimulation (Fig. 2J). Examples include Ube2c, a ubiquitin-conjugating enzyme important for degrading target proteins during cell cycle progression (45) and cell cycle genes Cdc20 and Mcm10 (Supplemental Fig. 3A).

Genes downregulated in both MBC subsets compared with nBs functioned in protein phosphorylation and cellular response to cytokine stimulus (Fig. 2F) and included Bach2, Cxcr4, and Foxo1 (Fig. 2I). Reduced expression of Bach2 in MBCs confirms previous observations using model Ags (10). Cxcr4 is a homing receptor for migration to the bone marrow and between the dark and light zones of the GC. Foxo1 is critically important for B cell development and maintenance (46). Together, these data indicate that MBCs harbor a gene expression program that is distinct from nBs and sustained following Ag clearance.

ATAC-seq (47) was performed to interrogate the epigenetic landscape of MBCs and nBs on the same three populations of cells as above. PCA of the 42,231 accessible regions identified revealed that principal component 1 separated MBCs from nBs and principal component 2 separated IgM-expressing populations from IgG (Fig. 3A). A heatmap of all significant DAR (absolute log2FC >1 and FDR <0.05) demonstrated a large number of regions in MBC with gains in accessibility compared with nBs (Fig. 3B, Supplemental Table II). Annotation of the chromatin accessibility for 2 kb surrounding all DAR revealed that IgM and IgG MBCs both had overall increased accessibility compared with nBs, with IgG MBCs also displaying a small, but significant, increase in accessibility over IgM MBCs (Fig. 3C). Zbtb32, which has been implicated in MBC reactivation (15), showed increased accessibility and a corresponding increase in expression that peaked in IgG MBCs (Supplemental Fig. 3B). Additionally, consistent with gene expression changes, DAR were observed surrounding the genes Tbx21, Nt5e (CD73), and Cxcr3.

FIGURE 3.

MBCs are epigenetically primed to differentiate to PCs. (A) PCA of 42,231 accessible regions for nB (n = 3) and IgM (n = 3) and IgG (n = 2) MBC defined in Fig. 2, Supplemental Fig. 2. (B) Heatmap showing 6948 DAR. (C) Histogram of all DAR identified in nB, IgM, and IgG MBCs (left) and boxplot quantitating chromatin accessibility (right). (D) Genome plots of accessible chromatin (left) and gene expression (right) for select genes. DAR are indicated by a horizontal bar and box, and the upper scale limit was set to show the changes at DAR in MBCs. PC RNA-seq (22) and ATAC-seq (49) datasets were described previously. (E) GSEA plots analyzing the Shi et al. (48) PC signature gene set showing enrichment for IgM (right) or IgG (left) MBCs compared with nBs. (F) Boxplot of z-score normalized accessibility for all DAR that map to the Shi et al. PC genes from (E). (G) Genome plots for indicated gene locus showing accessibility (left) and gene expression (right). BLIMP-1 chromatin immunoprecipitation sequencing (ChIP-seq) data were described previously (50). The pound sign (#) denotes genes not detected by RNA-seq for the indicated cell type. DAR are indicated by a horizontal bar and box, and the upper scale limit was set to show the changes at DAR in MBCs. (H) Scatterplot of the PageRank fold change for the IgM versus nB compared with IgG versus nB. DEG up in nB are indicated in blue, DEG up in IgG only are indicated in purple, and DEG up in both IgM and IgG are indicated in red. (I) Histogram (left) and boxplot (right) of accessibility at all BACH2:AP1 binding motifs identified in nB, IgM, and IgG MBCs. Significance for (C), (F), and (I) determined by two-tailed Student t test. See also Supplemental Fig. 3.

FIGURE 3.

MBCs are epigenetically primed to differentiate to PCs. (A) PCA of 42,231 accessible regions for nB (n = 3) and IgM (n = 3) and IgG (n = 2) MBC defined in Fig. 2, Supplemental Fig. 2. (B) Heatmap showing 6948 DAR. (C) Histogram of all DAR identified in nB, IgM, and IgG MBCs (left) and boxplot quantitating chromatin accessibility (right). (D) Genome plots of accessible chromatin (left) and gene expression (right) for select genes. DAR are indicated by a horizontal bar and box, and the upper scale limit was set to show the changes at DAR in MBCs. PC RNA-seq (22) and ATAC-seq (49) datasets were described previously. (E) GSEA plots analyzing the Shi et al. (48) PC signature gene set showing enrichment for IgM (right) or IgG (left) MBCs compared with nBs. (F) Boxplot of z-score normalized accessibility for all DAR that map to the Shi et al. PC genes from (E). (G) Genome plots for indicated gene locus showing accessibility (left) and gene expression (right). BLIMP-1 chromatin immunoprecipitation sequencing (ChIP-seq) data were described previously (50). The pound sign (#) denotes genes not detected by RNA-seq for the indicated cell type. DAR are indicated by a horizontal bar and box, and the upper scale limit was set to show the changes at DAR in MBCs. (H) Scatterplot of the PageRank fold change for the IgM versus nB compared with IgG versus nB. DEG up in nB are indicated in blue, DEG up in IgG only are indicated in purple, and DEG up in both IgM and IgG are indicated in red. (I) Histogram (left) and boxplot (right) of accessibility at all BACH2:AP1 binding motifs identified in nB, IgM, and IgG MBCs. Significance for (C), (F), and (I) determined by two-tailed Student t test. See also Supplemental Fig. 3.

Close modal

The rapid formation of PCs in memory mice suggested the hypothesis that MBCs might harbor a distinct programming related to PC differentiation. In support of this, DAR were observed that mapped to key PC transcription factors: Xbp1, Prdm1, and Irf4 (Fig. 3D, Supplemental Fig. 3C). To examine other genes, a transcription signature derived from genes expressed in PC (48) was used in a GSEA. This analysis revealed that PC genes were significantly enriched in both IgM and IgG MBC populations compared with nB (Fig. 3E). The change in chromatin accessibility at DAR mapping to PC signature genes was compared between nBs, MBCs, and previously generated PC data (49). nBs displayed the lowest accessibility with both IgM and IgG MBC subsets exhibiting intermediate levels compared with PC (Fig. 3F), indicating that the MBCs may be primed for expression of PC genes. Examples include the PC-expressed genes Prdm1 and Xbp1, which displayed regions with higher accessibility in MBCs and little to no accessibility in nBs. Although annotated as DEG in our dataset, the overall MBC expression for these genes was low compared with PCs, suggesting chromatin accessibility changes precedes expression changes.

To further examine how the observed accessibility patterns corresponded to PC-programming events, BLIMP-1 chromatin immunoprecipitation sequencing data (50) were analyzed along with accessibility data in nBs and MBCs. Both repressed and activated BLIMP-1 target genes displayed a pattern of increased accessibility at the same genomic location that contained a BLIMP-1 binding site and where PCs also have open chromatin (Fig. 3G, Supplemental Fig. 3B). The BLIMP-1–repressed genes Gypc and Aicda displayed a binding site and open chromatin at a regulatory site proximal to the promoter. Il10 and Tigit each contained loci that were increased in accessibility in MBC that matched the PC profile. Both of these genes are induced by BLIMP-1 in PC, and despite the MBC accessibility profile, expression was not detected in either nBs or MBCs.

To assess the kinetics of gene expression during reactivation MBCs were enriched along with nB and stimulated to differentiate into PC ex vivo using CD40L, IL-4, and IL-5. At early time points poststimulation, RNA was extracted, and the expression of a set of genes displaying MBC-specific accessible chromatin was assessed by quantitative RT-PCR. Consistent with a role for epigenetic priming, Zbtb32, Aicda, Il10, and Xbp1 were significantly induced more rapidly and to higher levels in stimulated MBCs compared with nB (Supplemental Fig. 3D). Importantly, the control gene Gapdh was equally induced in both conditions. Thus, these data indicate that MBCs harbor accessibility changes that resemble the PC chromatin architecture and may prime gene expression programs for rapid induction.

To determine additional transcription factor networks that were differentially regulated in MBCs compared with nBs, the PageRank (29) algorithm, which integrates RNA-seq and ATAC-seq data to define transcription factor importance based on changes in target gene expression, was used. For each transcription factor, the fold change in PageRank score between IgG and nB was correlated with the change between IgM and nB to define common changes. Both MYC and MAFB had a high PageRank score in both MBC subsets (Fig. 3H). Interestingly, Myc was not identified as a DEG in either comparison despite its target genes and network being activated in MBCs. Consistent with their transcriptional downregulation, BACH2 and FOXO1 were less important in MBC transcriptional networks. In memory CD8 T cells, BACH2 has been shown to repress the activity of AP-1 factors through binding to a similar consensus motif (51). This suggested that lower expression of BACH2 may allow for increased AP-1 activity and therefore accessibility at these motifs. When accessibility was annotated for all instances of the BACH2:AP-1 composite motif in DAR, IgG MBCs displayed significantly increased accessibility compared with both IgM MBCs and nBs (Fig. 3I), suggesting there may be a similar mechanism at play in MBCs as in memory CD8 T cells.

Intriguingly, both GSEA and GO analysis revealed an enrichment of the gene sets related to iron and heme metabolism as being enriched in MBCs compared with nBs (Figs. 2F, 4A). Examples of iron pathway genes include Slc40a1, Trf, and Hmox1, which were all significantly upregulated and expressed in both MBC subsets with near background expression in nBs (Fig. 4B). Additionally, Hmox1 displayed promoter chromatin accessibility increases that correlated with its unique expression in MBCs (Fig. 4C). HO-1, encoded by Hmox1, is a heme-response gene that facilitates its recycling to biliverdin, ferrous iron, and CO (52) and can be used as a readout for intracellular heme levels (53). In agreement with the interpretation that the heme uptake pathway is upregulated in MBCs, paramagnetic bead–enriched, SM displayed significantly higher levels of HO-1 than nBs (Fig. 4D), indicating that MBCs retain higher levels of intracellular heme.

FIGURE 4.

Hemin increases PC formation of murine MBCs. (A) GSEA showing the enrichment of the iron ion transport gene set for IgM (blue) and IgG (green) MBCs compared with nBs. (B) Bar plot showing expression of select genes from the iron ion transport gene set in (A). The asterisk (*) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). (C) Genome plot of the Hmox1 locus showing the location of DAR compared with nB. DAR are indicated by a horizontal bar and yellow shading. The upper scale limit was set to show the changes at DAR in MBCs. (D) Flow cytometry (mean fluorescence intensity [MFI]) quantitation of HO-1 in nBs isolated from naive mice and isotype-switched MBCs that were enriched from PR8-primed mice. (E) nBs and MBCs were stimulated with CD40L, IL-4, and IL-5 for 3 d with vehicle or hemin added to the cultures at day 1 as indicated. The formation of CD138+ PC was analyzed at day 3 by flow cytometry. (F) Quantitation of CD138+ PC from (E). Data are from three independent experiments with n = 11. (G) Representative ELISPOT wells showing Ab spots per 6250 cells plated from day 3 cultures as in (E). (H) Quantification of ELISPOT data from (G). Data are from three independent experiments with three to four samples per group. Significance for (D), (F), and (H) determined by two-tailed Student t test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 4.

Hemin increases PC formation of murine MBCs. (A) GSEA showing the enrichment of the iron ion transport gene set for IgM (blue) and IgG (green) MBCs compared with nBs. (B) Bar plot showing expression of select genes from the iron ion transport gene set in (A). The asterisk (*) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). (C) Genome plot of the Hmox1 locus showing the location of DAR compared with nB. DAR are indicated by a horizontal bar and yellow shading. The upper scale limit was set to show the changes at DAR in MBCs. (D) Flow cytometry (mean fluorescence intensity [MFI]) quantitation of HO-1 in nBs isolated from naive mice and isotype-switched MBCs that were enriched from PR8-primed mice. (E) nBs and MBCs were stimulated with CD40L, IL-4, and IL-5 for 3 d with vehicle or hemin added to the cultures at day 1 as indicated. The formation of CD138+ PC was analyzed at day 3 by flow cytometry. (F) Quantitation of CD138+ PC from (E). Data are from three independent experiments with n = 11. (G) Representative ELISPOT wells showing Ab spots per 6250 cells plated from day 3 cultures as in (E). (H) Quantification of ELISPOT data from (G). Data are from three independent experiments with three to four samples per group. Significance for (D), (F), and (H) determined by two-tailed Student t test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Close modal

To determine if heme modulated PC differentiation, MBCs and nBs were isolated for ex vivo culture and stimulated using CD40L, IL-4, and IL-5 to model T-dependent differentiation conditions. Cultured MBCs differentiated into CD138+ PCs to a higher degree than nBs over 3 d (Fig. 4E, 4F). To determine the role of heme in PC differentiation, nBs and enriched MBCs were stimulated in the presence of hemin as above. A 3-fold greater induction of CD138+ PCs than matched vehicle-treated controls was observed for both nBs and MBCs (Fig. 4E, 4F). Additionally, each culture was subjected to ELISPOT analysis to ensure that hemin treatment produced functional PCs. The addition of hemin significantly enhanced PC formation in MBCs (Fig. 4G, 4H). These data indicate that hemin can augment B cell differentiation and enhance the formation of PCs ex vivo.

To determine if the epigenetic and transcriptional changes observed in mice were also observed in humans, RNA-seq and ATAC-seq data from resting nBs (rN) and SM from healthy subjects (54) were compared with the data generated in this study. Of the 10,324 genes expressed in mouse nBs and MBCs, 84% were also detected in the human SM and rN RNA-seq data (Fig. 5A, pie chart insets). Reciprocally, 76% of the genes expressed in the human cell types had direct homologs that were detected in the mouse. To determine the concordance of the transcriptional changes, the fold change of mouse IgG MBCs versus nBs was plotted against human SM versus rN B cells. Eighty-five genes were called DEG in both comparisons, with 53 having a positive fold change (Fig. 5A, red) and 32 having a negative fold change in both species (Fig. 5A, blue). Notably, HMOX1 was among the conserved upregulated genes along with ZBTB32, CXCR3, and SLC11A1. BACH2, CXCR4, and IL21R highlighted the downregulated conserved genes.

FIGURE 5.

MBC epigenetic signatures are conserved between mouse and human. (A) Scatter plot of the fold change for all detected genes between mouse IgG versus nBs and human isotype-switched memory (SM) versus resting naive (rN) reported previously (54). Pie chart indicates the percentage of genes with homologs that were also detected as expressed in the corresponding species. DEG annotations are denoted by symbols shown in the key. (B) Heatmap of corresponding chromatin accessibility in human (left) and mouse (right) cell types for peaks surrounding select gene sets. Genes are subset based on seven classifications for changes in expression as indicated. The frequency of accessible regions that are annotated as a DAR for each category is shown to the right of each heatmap. (C) PageRank analysis (29) was performed on each cell subset, and the resulting transcription factors ranked by fold change in memory versus naive. The x-axis scale indicates PageRank log2FC, with the dotted line indicating the transition from positive (+) to negative (−) fold changes in the ranked list. Select transcription factors and their location in each list are annotated. (D) Histogram (left) and boxplot (right) of accessibility at BACH2 motifs and surrounding sequences. rppm, reads per peak per million. (E) Average gene expression of log2FPKM normalized data for all BACH2 target genes (55). (F) Gene expression bar plots for examples of two BACH2 target genes in rN and SM. The asterisk (*) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). (G) Genome plots displaying DAR in human (top row) and mouse (bottom row). Highlighted by a box are homologous regions that were identified as DAR. Significance for (D) and (E) determined by two-tailed Student t test.

FIGURE 5.

MBC epigenetic signatures are conserved between mouse and human. (A) Scatter plot of the fold change for all detected genes between mouse IgG versus nBs and human isotype-switched memory (SM) versus resting naive (rN) reported previously (54). Pie chart indicates the percentage of genes with homologs that were also detected as expressed in the corresponding species. DEG annotations are denoted by symbols shown in the key. (B) Heatmap of corresponding chromatin accessibility in human (left) and mouse (right) cell types for peaks surrounding select gene sets. Genes are subset based on seven classifications for changes in expression as indicated. The frequency of accessible regions that are annotated as a DAR for each category is shown to the right of each heatmap. (C) PageRank analysis (29) was performed on each cell subset, and the resulting transcription factors ranked by fold change in memory versus naive. The x-axis scale indicates PageRank log2FC, with the dotted line indicating the transition from positive (+) to negative (−) fold changes in the ranked list. Select transcription factors and their location in each list are annotated. (D) Histogram (left) and boxplot (right) of accessibility at BACH2 motifs and surrounding sequences. rppm, reads per peak per million. (E) Average gene expression of log2FPKM normalized data for all BACH2 target genes (55). (F) Gene expression bar plots for examples of two BACH2 target genes in rN and SM. The asterisk (*) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). (G) Genome plots displaying DAR in human (top row) and mouse (bottom row). Highlighted by a box are homologous regions that were identified as DAR. Significance for (D) and (E) determined by two-tailed Student t test.

Close modal

The global chromatin accessibility profiles between the two species were analyzed by clustering the genes that were conserved between mice and humans into seven transcriptional bins based on fold change. For each bin, the chromatin accessibility fold change between memory and nBs was calculated and averaged for all peaks. A concordance between chromatin accessibility and transcription was observed for genes conserved between the MBCs of humans and mice (Fig. 5B). Additionally, up to ∼20% of accessible regions were annotated as DAR in the up and down DEG bins. Importantly, the genes with no expression changes had the lowest enrichment for accessibility changes.

The PageRank algorithm was applied to the human SM versus rN comparison and the mouse IgG and IgM MBCs versus nBs. Each transcription factor was ranked from the most important in MBCs to nBs. Many transcription factors showed a similar ranking of importance in each of the three comparisons (Fig. 5C). Among those that had high importance in MBCs included MYC, BLIMP-1, EGR1, RFX3, MAX, and EGR2. Among those that had high importance in nBs were BACH2, FOXO1, and JUND (Fig. 5C). To further explore the change of BACH2 transcriptional networks in MBCs, all peaks were annotated for the presence of a BACH2 binding site motif and the chromatin accessibility in SM and rN calculated (Fig. 5D). We observed significantly more accessibility at BACH2 binding motifs in human SM compared with rN. Additionally, the overall expression of genes that were predicted to be BACH2 targets (55) was increased in SM (Fig. 5E). For example, the BACH2 targets, IL2RB, a component of the IL-2R, and SLC7A5, a subunit of the neutral amino acid transporter CD98, were both increased in SM (Fig. 5F). Because BACH2 is described as a transcriptional repressor and its expression is decreased in MBCs, this suggests that the BACH2 binding sites, which are also shared by other members of the AP-1 family of factors, are derepressed.

Examples of regions that are directly conserved between mice and humans and have similar changes in programming include Tox2 and Tcf7 (Fig. 5G). Both of these genes display DAR that have increased accessibility in MBCs compared with naive cells at syntenic genomic regions. Tox2 drives T follicular helper cell differentiation (56) and is regulated by BACH2 in B cells (55). Tcf7 encodes the gene TCF1 that is required for CD8 memory formation and marks resource stem cells that are capable of reseeding the effector population (57, 58). Thus, human and mouse MBCs share a conserved programming with respect to chromatin accessibility and gene expression.

Consistent with mouse MBCs, the iron ion homeostasis gene set was significantly enriched in MBCs from humans (Fig. 6A). Examples of these genes include HMOX1, HFE, and SLC11A1 (Fig. 6B). HO-1 expression was used as a readout for intracellular heme levels (53), and as in the mouse, SM displayed increased expression of HO-1 compared with rN (Fig. 6C). The role of heme in MBC reactivation was evaluated by stimulating SM with a cytokine mixture that promotes PC differentiation over 6 d (59) in the presence of vehicle or hemin. The MBC cultures that were treated with hemin exhibited increased formation of CD38+CD27+ PC (Fig. 6D, 6E). Additionally, the formation of a more mature population of PC that were CD38+CD138+ (60) was increased in the presence of hemin (Fig. 6F, 6G). ELISPOT analyses of these cultures confirm increased Ab secretion in cultures treated with hemin (Fig. 6H). Thus, hemin directly influences the ex vivo generation of PCs from human MBCs.

FIGURE 6.

Hemin augments human PC differentiation and mitochondrial metabolism. (A) GSEA for the iron ion homeostasis gene set between human SM versus rN B cells. (B) Bar plots showing gene expression for the indicated genes from the iron ion homeostasis gene set in (A). The asterisk (*) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). (C) Mean fluorescence intensity (MFI) quantitation of HO-1 in rN and SM isolated from total PBMCs from 13 human healthy donors. (D) Representative flow cytometry plots of SM from a single donor stimulated with R848, CD40L, IL-21, IL-2, BAFF, and human insulin as described previously (59) and supplemented with vehicle control or hemin showing the formation of CD38+CD27+ PC. (E) Quantitation of CD38+CD27+ PC from six individual donors differentiated as in (D). (F) Representative flow cytometry of CD38+CD138+ PC from cultures differentiated as in (D). (G) Quantitation of CD38+CD138+ PC from six individual donors from (F). (H) Representative ELISPOT showing PC Ab spots from 3125 cells plated from the cultures described in (D) (left) and quantitation of spots from nine individual donors (right). (I) MBCs were differentiated as in (D), then harvested and washed in Seahorse media before being subjected to a Mitochondrial Stress Test on a Seahorse Bioanalyzer. (J) Spare respiratory capacity was quantified as the difference between first measurement following carbonyl cyanide-4-phenylhydrazone injection and the last measurement before oligomycin injection and shown as a bar plot. Data are representative of five patients from two independent experiments. Significance in (C), (E), (G), (H), and (J) determined by two-tailed Student t test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 6.

Hemin augments human PC differentiation and mitochondrial metabolism. (A) GSEA for the iron ion homeostasis gene set between human SM versus rN B cells. (B) Bar plots showing gene expression for the indicated genes from the iron ion homeostasis gene set in (A). The asterisk (*) indicates significant differential expression (FDR < 0.05 and absolute log2FC > 1). (C) Mean fluorescence intensity (MFI) quantitation of HO-1 in rN and SM isolated from total PBMCs from 13 human healthy donors. (D) Representative flow cytometry plots of SM from a single donor stimulated with R848, CD40L, IL-21, IL-2, BAFF, and human insulin as described previously (59) and supplemented with vehicle control or hemin showing the formation of CD38+CD27+ PC. (E) Quantitation of CD38+CD27+ PC from six individual donors differentiated as in (D). (F) Representative flow cytometry of CD38+CD138+ PC from cultures differentiated as in (D). (G) Quantitation of CD38+CD138+ PC from six individual donors from (F). (H) Representative ELISPOT showing PC Ab spots from 3125 cells plated from the cultures described in (D) (left) and quantitation of spots from nine individual donors (right). (I) MBCs were differentiated as in (D), then harvested and washed in Seahorse media before being subjected to a Mitochondrial Stress Test on a Seahorse Bioanalyzer. (J) Spare respiratory capacity was quantified as the difference between first measurement following carbonyl cyanide-4-phenylhydrazone injection and the last measurement before oligomycin injection and shown as a bar plot. Data are representative of five patients from two independent experiments. Significance in (C), (E), (G), (H), and (J) determined by two-tailed Student t test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Close modal

The mitochondrion is the primary place that iron resides within cells, where it is integral to the functions of the electron transport chain and the ability of cells to perform oxidative phosphorylation (52). Previous observations showing that oxidative phosphorylation promotes PC differentiation (61) suggested that the increased PC differentiation in the presence of heme may be accompanied by metabolic changes. To test this, SM were enriched from human PBMCs and stimulated to differentiate (59) in the presence or absence of hemin as above, and the resulting cultures were analyzed for their usage of oxidative phosphorylation by extracellular flux assay on a Seahorse bioanalyzer. Hemin-treated samples showed significantly higher basal and maximal respiration (Fig. 6I), which also resulted in increased spare respiratory capacity (Fig. 6J). These data indicate that hemin facilitates increased mitochondrial metabolism and oxidative phosphorylation ex vivo.

To distinguish the epigenetic and transcriptional properties contributing to B cell memory, the chromatin accessibility and transcriptome profiles of influenza-specific IgM and IgG MBCs were determined and compared with those of nBs. An overall increase in MBC accessibility was identified that correlated with PC programming and indicated that prior B cell activation imparted an epigenetic state that was distinct from Ag-inexperienced nBs. IgM and IgG MBCs were also distinct from each other, agreeing with previous observations that each MBC subset may be predisposed for different reactivation fates (62). Importantly, MBC programming was largely conserved between mice and humans, with the heme and iron homeostasis pathways significantly upregulated in both species. Increased MBC heme content was validated and ex vivo differentiation in the presence of hemin enhanced PC differentiation. Together, these data identify an epigenetic programming that is unique to MBCs and highlight mechanisms for enhanced recall responses upon secondary Ag exposure.

The chromatin accessibility data provide evidence for a primed epigenetic architecture as one mechanism that allows for more rapid responses for MBCs compared with nBs. The maintenance of accessible chromatin at specific loci in MBCs may act to reduce the number of reprogramming steps necessary for many functions, including the formation of GC and PC following reactivation. Indeed, MBC displayed higher expression of genes related to the cell cycle, B cell activation, signal transduction, and migration. Many of these genes showed a graded expression that was high in IgM and maximal in IgG MBCs compared with nBs. Of particular interest was the enhanced expression and accessibility surrounding genes unique to PC in MBCs. PageRank analysis of both mouse and human MBCs revealed BLIMP-1 as an enriched regulator of the MBC transcriptional network, indicating that aspects of the PC program is partially engaged or primed in MBCs. In support of this, the three-dimensional architecture of the Prdm1 locus, which encodes BLIMP-1, is reorganized in activated B cells prior to Prdm1 expression in PC (63). Chromatin accessibility priming was also observed in memory CD8 T cells and helped facilitate rapid induction of effector responses upon secondary stimulation (14). Similar to memory CD8 T cells, a set of primed genes demonstrated rapid induction and higher expression in MBC compared with nB. Therefore, in a similar manner, MBCs maintain an accessibility programming that serves as a molecular signature of prior activation and may improve the fidelity and speed of differentiation during secondary Ag encounters.

The presence of differentially accessible chromatin implies that transcription factor activity is different between MBCs and nBs. Consistent with previous findings (10), both human and mouse MBCs displayed lower expression of Bach2. BACH2 functions by partnering with various AP-1 factors and represses its target genes (51). Both species displayed increased accessibility at BACH2/AP-1 binding motifs. Intriguingly, such target genes such as Prdm1 (mouse) and SLC7A5 and IL2RB (human) were increased in expression, indicating BACH2 activity was reduced in MBCs and that other AP-1 family members were dictating the change in expression. This was supported by PageRank analysis, which found both MYC and a set of AP-1 factors to be more important in MBCs and might allow more rapid induction of metabolic changes, proliferation, and differentiation to PC. Altered MBC metabolic states were indicated by enhanced mRNA and heme content compared with nBs. Additionally, activation of the AKT pathway, which regulates cellular metabolism through mTOR, was indicated by reduced expression and network importance of Foxo1 (64). Furthermore, Zbtb32, which acts to restrict secondary responses (15), was upregulated in IgM and was maximally expressed in IgG MBC. Thus, altered MBC transcription factor networks provide mechanisms to rapidly modify metabolism, differentiate to PC, and ultimately regulate humoral responses.

One of the most striking conserved differences between MBCs and nBs was the increased expression of heme biosynthetic genes in MBCs. Heme enhanced the formation of PC from both human and mouse MBCs during ex vivo differentiation. Heme is synthesized and primarily found in the mitochondria, where it is an essential component of cytochrome proteins that function in the electron transport chain during oxidative phosphorylation (52). BACH2 and BACH1 both negatively regulate heme biosynthesis in part through direct binding to the Hmox1 promoter (55, 65). Heme can also directly bind BACH2, leading to its degradation and derepression of the heme biosynthesis pathway (53). Heme binding is not the only mechanism controlling BACH2 activity, as the PI3K pathway can also lead to phosphorylation of and decreased nuclear localization of BACH2 (66). Given that heme is an essential component of the electron transport chain and our previous findings linking PC differentiation to oxidative phosphorylation (61), this relationship was explored ex vivo. Differentiation in the presence of heme resulted in increased oxidative phosphorylation. In T cells, iron is important for activation-induced proliferation and mitochondrial function (67). Thus, by upregulation of the components to import and generate heme, MBCs are primed to rapidly increase their metabolism toward modes that promote PC differentiation and support the demands of Ab secretion.

The importance of heme and iron biogenesis for MBCs is also observed in many human disorders. Patients with combined immunodeficiency caused by a missense mutation in TFRC, the transferrin receptor, have reduced MBC frequencies as a proportion of total B cells (68). TFRC is required for iron uptake, and this observation suggests that iron is important in either the formation or maintenance of MBCs. In multiple myeloma, the iron exporter ferroportin-1 (SLC40A1) is decreased, resulting in increased cell-associated iron and poor prognosis (69, 70). Consistent with previous reports (53), administration of heme to mice had no measurable effect on the immune response to influenza (data not shown). However, mice deficient in HO-1, which degrades heme, have increased serum Ab titers (71), indicating heme regulates B cell function in vivo. Thus, targeted modulation of iron metabolism and biogenesis may be a mechanism to promote Ab responses, as well as limit PC differentiation and cell survival in cancer and potentially autoimmune settings. In summary, these data define a conserved epigenetic state centered around primed chromatin accessibility in MBC that provides a molecular history of prior activation and likely contributes to enhanced functions and responses upon secondary challenge.

We thank R. Butler for animal husbandry and care, R. Karaffa for cell sorting at the Emory Flow Cytometry Core, and Drs. E. Zumaquero and F. E. Lund (University of Alabama, Birmingham) for guidance with human ex vivo MBC cultures and differentiation.

The sequences presented in this article have been submitted to the National Center for Biotechnology Information Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE164173.

This work was supported by National Institute of Allergy and Infectious Diseases, National Institutes of Health Grants 1 R01 AI148471 and 5 P01 AI125180 (to J.M.B.), 1 R01 AI148471 (to C.D.S.), and 1 F31 AI138391 (to M.J.P.).

The online version of this article contains supplemental material.

Abbreviations used in this article:

ATAC-seq

assay for transposase-accessible chromatin sequencing

DAR

differentially accessible region

DEG

differentially expressed gene

dLN

draining lymph node

FDR

false discovery rate

GC

germinal center

GO

gene ontology

GSEA

gene set enrichment analysis

HO-1

heme oxygenase-1

log2FC

log2 fold-change

MBC

memory B cell

nB

naive B cell

NP

nucleoprotein

PC

plasma cell

PCA

principal component analysis

rN

resting nB

RNA-seq

RNA sequencing

SM

class-switched MBC.

1
Kurosaki
,
T.
,
K.
Kometani
,
W.
Ise
.
2015
.
Memory B cells.
Nat. Rev. Immunol.
15
:
149
159
.
2
Wienands
,
J.
,
N.
Engels
.
2016
.
Control of memory B cell responses by extrinsic and intrinsic mechanisms.
Immunol. Lett.
178
:
27
30
.
3
Moens
,
L.
,
A.
Kane
,
S. G.
Tangye
.
2016
.
Naïve and memory B cells exhibit distinct biochemical responses following BCR engagement.
Immunol. Cell Biol.
94
:
774
786
.
4
Wakabayashi
,
C.
,
T.
Adachi
,
J.
Wienands
,
T.
Tsubata
.
2002
.
A distinct signaling pathway used by the IgG-containing B cell antigen receptor.
Science
298
:
2392
2395
.
5
Wan
,
Z.
,
X.
Chen
,
H.
Chen
,
Q.
Ji
,
Y.
Chen
,
J.
Wang
,
Y.
Cao
,
F.
Wang
,
J.
Lou
,
Z.
Tang
,
W.
Liu
.
2015
.
The activation of IgM- or isotype-switched IgG- and IgE-BCR exhibits distinct mechanical force sensitivity and threshold.
Elife
4
: e06925.
6
Tomayko
,
M. M.
,
N. C.
Steinel
,
S. M.
Anderson
,
M. J.
Shlomchik
.
2010
.
Cutting edge: hierarchy of maturity of murine memory B cell subsets.
J. Immunol.
185
:
7146
7150
.
7
Zuccarino-Catania
,
G. V.
,
S.
Sadanand
,
F. J.
Weisel
,
M. M.
Tomayko
,
H.
Meng
,
S. H.
Kleinstein
,
K. L.
Good-Jacobson
,
M. J.
Shlomchik
.
2014
.
CD80 and PD-L2 define functionally distinct memory B cell subsets that are independent of antibody isotype.
Nat. Immunol.
15
:
631
637
.
8
Krishnamurty
,
A. T.
,
C. D.
Thouvenel
,
S.
Portugal
,
G. J.
Keitany
,
K. S.
Kim
,
A.
Holder
,
P. D.
Crompton
,
D. J.
Rawlings
,
M.
Pepper
.
2016
.
Somatically hypermutated plasmodium-specific IgM(+) memory B cells are rapid, plastic, early responders upon malaria rechallenge.
Immunity
45
:
402
414
.
9
Reynaud
,
C. A.
,
M.
Descatoire
,
I.
Dogan
,
F.
Huetz
,
S.
Weller
,
J. C.
Weill
.
2012
.
IgM memory B cells: a mouse/human paradox.
Cell. Mol. Life Sci.
69
:
1625
1634
.
10
Kometani
,
K.
,
R.
Nakagawa
,
R.
Shinnakasu
,
T.
Kaji
,
A.
Rybouchkin
,
S.
Moriyama
,
K.
Furukawa
,
H.
Koseki
,
T.
Takemori
,
T.
Kurosaki
.
2013
.
Repression of the transcription factor Bach2 contributes to predisposition of IgG1 memory B cells toward plasma cell differentiation.
Immunity
39
:
136
147
.
11
Good-Jacobson
,
K. L.
,
Y.
Chen
,
A. K.
Voss
,
G. K.
Smyth
,
T.
Thomas
,
D.
Tarlinton
.
2014
.
Regulation of germinal center responses and B-cell memory by the chromatin modifier MOZ.
Proc. Natl. Acad. Sci. USA
111
:
9585
9590
.
12
Lai
,
A. Y.
,
D.
Mav
,
R.
Shah
,
S. A.
Grimm
,
D.
Phadke
,
K.
Hatzi
,
A.
Melnick
,
C.
Geigerman
,
S. E.
Sobol
,
D. L.
Jaye
,
P. A.
Wade
.
2013
.
DNA methylation profiling in human B cells reveals immune regulatory elements and epigenetic plasticity at Alu elements during B-cell activation.
Genome Res.
23
:
2030
2041
.
13
Kulis
,
M.
,
A.
Merkel
,
S.
Heath
,
A. C.
Queirós
,
R. P.
Schuyler
,
G.
Castellano
,
R.
Beekman
,
E.
Raineri
,
A.
Esteve
,
G.
Clot
, et al
.
2015
.
Whole-genome fingerprint of the DNA methylome during human B cell differentiation.
Nat. Genet.
47
:
746
756
.
14
Scharer
,
C. D.
,
A. P.
Bally
,
B.
Gandham
,
J. M.
Boss
.
2017
.
Cutting edge: chromatin accessibility programs CD8 T cell memory.
J. Immunol.
198
:
2238
2243
.
15
Jash
,
A.
,
Y.
Wang
,
F. J.
Weisel
,
C. D.
Scharer
,
J. M.
Boss
,
M. J.
Shlomchik
,
D.
Bhattacharya
.
2016
.
ZBTB32 restricts the duration of memory B cell recall responses.
J. Immunol.
197
:
1159
1168
.
16
Allie
,
S. R.
,
J. E.
Bradley
,
U.
Mudunuru
,
M. D.
Schultz
,
B. A.
Graf
,
F. E.
Lund
,
T. D.
Randall
.
2019
.
The establishment of resident memory B cells in the lung requires local antigen encounter.
Nat. Immunol.
20
:
97
108
.
17
Dobin
,
A.
,
C. A.
Davis
,
F.
Schlesinger
,
J.
Drenkow
,
C.
Zaleski
,
S.
Jha
,
P.
Batut
,
M.
Chaisson
,
T. R.
Gingeras
.
2013
.
STAR: ultrafast universal RNA-seq aligner.
Bioinformatics
29
:
15
21
.
18
Lawrence
,
M.
,
W.
Huber
,
H.
Pagès
,
P.
Aboyoun
,
M.
Carlson
,
R.
Gentleman
,
M. T.
Morgan
,
V. J.
Carey
.
2013
.
Software for computing and annotating genomic ranges.
PLOS Comput. Biol.
9
: e1003118.
19
Robinson
,
M. D.
,
D. J.
McCarthy
,
G. K.
Smyth
.
2010
.
edgeR: a Bioconductor package for differential expression analysis of digital gene expression data.
Bioinformatics
26
:
139
140
.
20
Wei Huang
,
D.
,
B. T.
Sherman
,
R. A.
Lempicki
.
2009
.
Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources.
Nat. Protoc.
4
:
44
57
.
21
Mootha
,
V. K.
,
C. M.
Lindgren
,
K. F.
Eriksson
,
A.
Subramanian
,
S.
Sihag
,
J.
Lehar
,
P.
Puigserver
,
E.
Carlsson
,
M.
Ridderstråle
,
E.
Laurila
, et al
.
2003
.
PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes.
Nat. Genet.
34
:
267
273
.
22
Barwick
,
B. G.
,
C. D.
Scharer
,
A. P. R.
Bally
,
J. M.
Boss
.
2016
.
Plasma cell differentiation is coupled to division-dependent DNA hypomethylation and gene regulation.
Nat. Immunol.
17
:
1216
1225
.
23
Sayers
,
E. W.
,
R.
Agarwala
,
E. E.
Bolton
,
J. R.
Brister
,
K.
Canese
,
K.
Clark
,
R.
Connor
,
N.
Fiorini
,
K.
Funk
,
T.
Hefferon
, et al
.
2019
.
Database resources of the National Center for Biotechnology Information.
Nucleic Acids Res.
47
(
D1
):
D23
D28
.
24
Guo
,
M.
,
M. J.
Price
,
D. G.
Patterson
,
B. G.
Barwick
,
R. R.
Haines
,
A. K.
Kania
,
J. E.
Bradley
,
T. D.
Randall
,
J. M.
Boss
,
C. D.
Scharer
.
2018
.
EZH2 represses the B cell transcriptional program and regulates antibody-secreting cell metabolism and antibody production.
J. Immunol.
200
:
1039
1052
.
25
Langmead
,
B.
,
C.
Trapnell
,
M.
Pop
,
S. L.
Salzberg
.
2009
.
Ultrafast and memory-efficient alignment of short DNA sequences to the human genome.
Genome Biol.
10
:
R25
.
26
Zhang
,
Y.
,
T.
Liu
,
C. A.
Meyer
,
J.
Eeckhoute
,
D. S.
Johnson
,
B. E.
Bernstein
,
C.
Nusbaum
,
R. M.
Myers
,
M.
Brown
,
W.
Li
,
X. S.
Liu
.
2008
.
Model-based analysis of ChIP-Seq (MACS).
Genome Biol.
9
:
R137
.
27
Heinz
,
S.
,
C.
Benner
,
N.
Spann
,
E.
Bertolino
,
Y. C.
Lin
,
P.
Laslo
,
J. X.
Cheng
,
C.
Murre
,
H.
Singh
,
C. K.
Glass
.
2010
.
Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities.
Mol. Cell
38
:
576
589
.
28
Scharer
,
C. D.
,
E. L.
Blalock
,
B. G.
Barwick
,
R. R.
Haines
,
C.
Wei
,
I.
Sanz
,
J. M.
Boss
.
2016
.
ATAC-seq on biobanked specimens defines a unique chromatin accessibility structure in naïve SLE B cells.
Sci. Rep.
6
:
27030
.
29
Zhang
,
K.
,
M.
Wang
,
Y.
Zhao
,
W.
Wang
.
2019
.
Taiji: system-level identification of key transcription factors reveals transcriptional waves in mouse embryonic development.
Sci. Adv.
5
: eaav3262.
30
Kees
,
U.
,
P. H.
Krammer
.
1984
.
Most influenza A virus-specific memory cytotoxic T lymphocytes react with antigenic epitopes associated with internal virus determinants.
J. Exp. Med.
159
:
365
377
.
31
Ballesteros-Tato
,
A.
,
B.
León
,
B. O.
Lee
,
F. E.
Lund
,
T. D.
Randall
.
2014
.
Epitope-specific regulation of memory programming by differential duration of antigen presentation to influenza-specific CD8(+) T cells.
Immunity
41
:
127
140
.
32
Pape
,
K. A.
,
J. J.
Taylor
,
R. W.
Maul
,
P. J.
Gearhart
,
M. K.
Jenkins
.
2011
.
Different B cell populations mediate early and late memory during an endogenous immune response.
Science
331
:
1203
1207
.
33
Ashburner
,
M.
,
C. A.
Ball
,
J. A.
Blake
,
D.
Botstein
,
H.
Butler
,
J. M.
Cherry
,
A. P.
Davis
,
K.
Dolinski
,
S. S.
Dwight
,
J. T.
Eppig
, et al
The Gene Ontology Consortium
.
2000
.
Gene ontology: tool for the unification of biology.
Nat. Genet.
25
:
25
29
.
34
Lord
,
G. M.
,
R. M.
Rao
,
H.
Choe
,
B. M.
Sullivan
,
A. H.
Lichtman
,
F. W.
Luscinskas
,
L. H.
Glimcher
.
2005
.
T-bet is required for optimal proinflammatory CD4+ T-cell trafficking.
Blood
106
:
3432
3439
.
35
Stone
,
S. L.
,
J. N.
Peel
,
C. D.
Scharer
,
C. A.
Risley
,
D. A.
Chisolm
,
M. D.
Schultz
,
B.
Yu
,
A.
Ballesteros-Tato
,
W.
Wojciechowski
,
B.
Mousseau
, et al
.
2019
.
T-bet transcription factor promotes antibody-secreting cell differentiation by limiting the inflammatory effects of IFN-γ on B cells.
Immunity
50
:
1172
1187.e7
.
36
Carpentier
,
I.
,
R.
Beyaert
.
1999
.
TRAF1 is a TNF inducible regulator of NF-kappaB activation.
FEBS Lett.
460
:
246
250
.
37
Shu
,
H. B.
,
A. B.
Agranoff
,
E. G.
Nabel
,
K.
Leung
,
C. S.
Duckett
,
A. S.
Neish
,
T.
Collins
,
G. J.
Nabel
.
1993
.
Differential regulation of vascular cell adhesion molecule 1 gene expression by specific NF-kappa B subunits in endothelial and epithelial cells.
Mol. Cell. Biol.
13
:
6283
6289
.
38
Takatsuka
,
S.
,
H.
Yamada
,
K.
Haniuda
,
H.
Saruwatari
,
M.
Ichihashi
,
J. C.
Renauld
,
D.
Kitamura
.
2018
.
IL-9 receptor signaling in memory B cells regulates humoral recall responses.
Nat. Immunol.
19
:
1025
1034
.
39
Browne
,
E. P.
2012
.
Regulation of B-cell responses by toll-like receptors.
Immunology
136
:
370
379
.
40
Matloubian
,
M.
,
C. G.
Lo
,
G.
Cinamon
,
M. J.
Lesneski
,
Y.
Xu
,
V.
Brinkmann
,
M. L.
Allende
,
R. L.
Proia
,
J. G.
Cyster
.
2004
.
Lymphocyte egress from thymus and peripheral lymphoid organs is dependent on S1P receptor 1.
Nature
427
:
355
360
.
41
Allende
,
M. L.
,
G.
Tuymetova
,
B. G.
Lee
,
E.
Bonifacino
,
Y. P.
Wu
,
R. L.
Proia
.
2010
.
S1P1 receptor directs the release of immature B cells from bone marrow into blood.
J. Exp. Med.
207
:
1113
1124
.
42
Alam
,
J.
,
A.
Smith
.
1992
.
Heme-hemopexin-mediated induction of metallothionein gene expression.
J. Biol. Chem.
267
:
16379
16384
.
43
Wang
,
Y.
,
S. R.
Brooks
,
X.
Li
,
A. N.
Anzelon
,
R. C.
Rickert
,
R. H.
Carter
.
2002
.
The physiologic role of CD19 cytoplasmic tyrosines.
Immunity
17
:
501
514
.
44
Subramanian
,
A.
,
P.
Tamayo
,
V. K.
Mootha
,
S.
Mukherjee
,
B. L.
Ebert
,
M. A.
Gillette
,
A.
Paulovich
,
S. L.
Pomeroy
,
T. R.
Golub
,
E. S.
Lander
,
J. P.
Mesirov
.
2005
.
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles.
Proc. Natl. Acad. Sci. USA
102
:
15545
15550
.
45
Stegmeier
,
F.
,
M.
Rape
,
V. M.
Draviam
,
G.
Nalepa
,
M. E.
Sowa
,
X. L.
Ang
,
E. R.
McDonald
III
,
M. Z.
Li
,
G. J.
Hannon
,
P. K.
Sorger
, et al
.
2007
.
Anaphase initiation is regulated by antagonistic ubiquitination and deubiquitination activities.
Nature
446
:
876
881
.
46
Dengler
,
H. S.
,
G. V.
Baracho
,
S. A.
Omori
,
S.
Bruckner
,
K. C.
Arden
,
D. H.
Castrillon
,
R. A.
DePinho
,
R. C.
Rickert
.
2008
.
Distinct functions for the transcription factor Foxo1 at various stages of B cell differentiation.
Nat. Immunol.
9
:
1388
1398
.
47
Buenrostro
,
J. D.
,
P. G.
Giresi
,
L. C.
Zaba
,
H. Y.
Chang
,
W. J.
Greenleaf
.
2013
.
Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position.
Nat. Methods
10
:
1213
1218
.
48
Shi
,
W.
,
Y.
Liao
,
S. N.
Willis
,
N.
Taubenheim
,
M.
Inouye
,
D. M.
Tarlinton
,
G. K.
Smyth
,
P. D.
Hodgkin
,
S. L.
Nutt
,
L. M.
Corcoran
.
2015
.
Transcriptional profiling of mouse B cell terminal differentiation defines a signature for antibody-secreting plasma cells.
Nat. Immunol.
16
:
663
673
.
49
Scharer
,
C. D.
,
B. G.
Barwick
,
M.
Guo
,
A. P. R.
Bally
,
J. M.
Boss
.
2018
.
Plasma cell differentiation is controlled by multiple cell division-coupled epigenetic programs.
Nat. Commun.
9
:
1698
.
50
Minnich
,
M.
,
H.
Tagoh
,
P.
Bönelt
,
E.
Axelsson
,
M.
Fischer
,
B.
Cebolla
,
A.
Tarakhovsky
,
S. L.
Nutt
,
M.
Jaritz
,
M.
Busslinger
.
2016
.
Multifunctional role of the transcription factor Blimp-1 in coordinating plasma cell differentiation.
Nat. Immunol.
17
:
331
343
.
51
Roychoudhuri
,
R.
,
D.
Clever
,
P.
Li
,
Y.
Wakabayashi
,
K. M.
Quinn
,
C. A.
Klebanoff
,
Y.
Ji
,
M.
Sukumar
,
R. L.
Eil
,
Z.
Yu
, et al
.
2016
.
BACH2 regulates CD8(+) T cell differentiation by controlling access of AP-1 factors to enhancers.
Nat. Immunol.
17
:
851
860
.
52
Igarashi
,
K.
,
M.
Watanabe-Matsui
.
2014
.
Wearing red for signaling: the heme-bach axis in heme metabolism, oxidative stress response and iron immunology.
Tohoku J. Exp. Med.
232
:
229
253
.
53
Watanabe-Matsui
,
M.
,
A.
Muto
,
T.
Matsui
,
A.
Itoh-Nakadai
,
O.
Nakajima
,
K.
Murayama
,
M.
Yamamoto
,
M.
Ikeda-Saito
,
K.
Igarashi
.
2011
.
Heme regulates B-cell differentiation, antibody class switch, and heme oxygenase-1 expression in B cells as a ligand of Bach2.
Blood
117
:
5438
5448
.
54
Scharer
,
C. D.
,
E. L.
Blalock
,
T.
Mi
,
B. G.
Barwick
,
S. A.
Jenks
,
T.
Deguchi
,
K. S.
Cashman
,
B. E.
Neary
,
D. G.
Patterson
,
S. L.
Hicks
, et al
.
2019
.
Epigenetic programming underpins B cell dysfunction in human SLE.
Nat. Immunol.
20
:
1071
1082
.
55
Hipp
,
N.
,
H.
Symington
,
C.
Pastoret
,
G.
Caron
,
C.
Monvoisin
,
K.
Tarte
,
T.
Fest
,
C.
Delaloy
.
2017
.
IL-2 imprints human naive B cell fate towards plasma cell through ERK/ELK1-mediated BACH2 repression.
Nat. Commun.
8
:
1443
.
56
Xu
,
W.
,
X.
Zhao
,
X.
Wang
,
H.
Feng
,
M.
Gou
,
W.
Jin
,
X.
Wang
,
X.
Liu
,
C.
Dong
.
2019
.
The transcription factor Tox2 drives T follicular helper cell development via regulating chromatin accessibility.
Immunity
51
:
826
839.e5
.
57
Zhou
,
X.
,
S.
Yu
,
D. M.
Zhao
,
J. T.
Harty
,
V. P.
Badovinac
,
H. H.
Xue
.
2010
.
Differentiation and persistence of memory CD8(+) T cells depend on T cell factor 1.
Immunity
33
:
229
240
.
58
Im
,
S. J.
,
M.
Hashimoto
,
M. Y.
Gerner
,
J.
Lee
,
H. T.
Kissick
,
M. C.
Burger
,
Q.
Shan
,
J. S.
Hale
,
J.
Lee
,
T. H.
Nasti
, et al
.
2016
.
Defining CD8+ T cells that provide the proliferative burst after PD-1 therapy.
Nature
537
:
417
421
.
59
Zumaquero
,
E.
,
S. L.
Stone
,
C. D.
Scharer
,
S. A.
Jenks
,
A.
Nellore
,
B.
Mousseau
,
A.
Rosal-Vela
,
D.
Botta
,
J. E.
Bradley
,
W.
Wojciechowski
, et al
.
2019
.
IFNγ induces epigenetic programming of human T-bet hi B cells and promotes TLR7/8 and IL-21 induced differentiation.
Elife
8
: e41641.
60
Garimalla
,
S.
,
D. C.
Nguyen
,
J. L.
Halliley
,
C.
Tipton
,
A. F.
Rosenberg
,
C. F.
Fucile
,
C. L.
Saney
,
S.
Kyu
,
D.
Kaminski
,
Y.
Qian
, et al
.
2019
.
Differential transcriptome and development of human peripheral plasma cell subsets.
JCI Insight
4
: e126732.
61
Price
,
M. J.
,
D. G.
Patterson
,
C. D.
Scharer
,
J. M.
Boss
.
2018
.
Progressive upregulation of oxidative metabolism facilitates plasmablast differentiation to a T-independent antigen.
Cell Rep.
23
:
3152
3159
.
62
Harms Pritchard
,
G.
,
M.
Pepper
.
2018
.
Memory B cell heterogeneity: remembrance of things past.
J. Leukoc. Biol.
103
:
269
274
.
63
Johanson
,
T. M.
,
A. T. L.
Lun
,
H. D.
Coughlan
,
T.
Tan
,
G. K.
Smyth
,
S. L.
Nutt
,
R. S.
Allan
.
2018
.
Transcription-factor-mediated supervision of global genome architecture maintains B cell identity.
Nat. Immunol.
19
:
1257
1264
.
64
Boothby
,
M.
,
R. C.
Rickert
.
2017
.
Metabolic regulation of the immune humoral response.
Immunity
46
:
743
755
.
65
Sun
,
J.
,
H.
Hoshino
,
K.
Takaku
,
O.
Nakajima
,
A.
Muto
,
H.
Suzuki
,
S.
Tashiro
,
S.
Takahashi
,
S.
Shibahara
,
J.
Alam
, et al
.
2002
.
Hemoprotein Bach1 regulates enhancer availability of heme oxygenase-1 gene.
EMBO J.
21
:
5216
5224
.
66
Ando
,
R.
,
H.
Shima
,
T.
Tamahara
,
Y.
Sato
,
M.
Watanabe-Matsui
,
H.
Kato
,
N.
Sax
,
H.
Motohashi
,
K.
Taguchi
,
M.
Yamamoto
, et al
.
2016
.
The transcription factor Bach2 is phosphorylated at multiple sites in murine B cells but a single site prevents its nuclear localization.
J. Biol. Chem.
291
:
1826
1840
.
67
Yarosz
,
E. L.
,
C.
Ye
,
A.
Kumar
,
C.
Black
,
E. K.
Choi
,
Y. A.
Seo
,
C. H.
Chang
.
2020
.
Cutting edge: activation-induced iron flux controls CD4 T cell proliferation by promoting proper IL-2R signaling and mitochondrial function.
J. Immunol.
204
:
1708
1713
.
68
Jabara
,
H. H.
,
S. E.
Boyden
,
J.
Chou
,
N.
Ramesh
,
M. J.
Massaad
,
H.
Benson
,
W.
Bainter
,
D.
Fraulino
,
F.
Rahimov
,
C.
Sieff
, et al
.
2016
.
A missense mutation in TFRC, encoding transferrin receptor 1, causes combined immunodeficiency.
Nat. Genet.
48
:
74
78
.
69
Gu
,
Z.
,
H.
Wang
,
J.
Xia
,
Y.
Yang
,
Z.
Jin
,
H.
Xu
,
J.
Shi
,
I.
De Domenico
,
G.
Tricot
,
F.
Zhan
.
2015
.
Decreased ferroportin promotes myeloma cell growth and osteoclast differentiation.
Cancer Res.
75
:
2211
2221
.
70
Zhan
,
X.
,
W.
Yu
,
R.
Franqui-Machin
,
M. L.
Bates
,
K.
Nadiminti
,
H.
Cao
,
B. A.
Amendt
,
Y.
Jethava
,
I.
Frech
,
F.
Zhan
,
G.
Tricot
.
2017
.
Alteration of mitochondrial biogenesis promotes disease progression in multiple myeloma.
Oncotarget
8
:
111213
111224
.
71
Kapturczak
,
M. H.
,
C.
Wasserfall
,
T.
Brusko
,
M.
Campbell-Thompson
,
T. M.
Ellis
,
M. A.
Atkinson
,
A.
Agarwal
.
2004
.
Heme oxygenase-1 modulates early inflammatory responses – evidence from the heme oxygenase-1-deficient mouse.
Am. J. Pathol.
165
:
1045
1053
.

The authors have no financial conflicts of interest.