B Cell-Intrinsic Role for IRF5 in TLR9/BCR-Induced Human B Cell Activation, Proliferation, and Plasmablast Differentiation

Abstract

Upon recognition of antigen, B cells undergo rapid proliferation followed by differentiation to specialized antibody secreting cells (ASCs). During this transition, B cells are reliant upon a multilayer transcription factor network to achieve a dramatic remodeling of the B cell transcriptional landscape. Increased levels of ASCs are often seen in autoimmune diseases and it is believed that altered expression of regulatory transcription factors play a role in this imbalance. The transcription factor interferon regulatory factor 5 (IRF5) is one such candidate as polymorphisms in IRF5 associate with risk of numerous autoimmune diseases and correlate with elevated IRF5 expression. IRF5 genetic risk has been widely replicated in systemic lupus erythematosus (SLE), and loss of Irf5 ameliorates disease in murine lupus models, in part, through the lack of pathogenic autoantibody secretion. It remains unclear, however, whether IRF5 is contributing to autoantibody production through a B cell-intrinsic function. To date, IRF5 function in healthy human B cells has not been characterized. Using human primary naive B cells, we define a critical intrinsic role for IRF5 in B cell activation, proliferation, and plasmablast differentiation. Targeted IRF5 knockdown resulted in significant immunoglobulin (Ig) D retention, reduced proliferation, plasmablast differentiation, and IgG secretion. The observed decreases were due to impaired B cell activation and clonal expansion. Distinct from murine studies, we identify and confirm new IRF5 target genes, IRF4, ERK1, and MYC, and pathways that mediate IRF5 B cell-intrinsic function. Together, these results identify IRF5 as an early regulator of human B cell activation and provide the first dataset in human primary B cells to map IRF5 dysfunction in SLE.

Keywords: autoantibodies; differentiation; human primary B cells; immunoglobulin G; interferon regulatory factor 5; plasmablasts; toll-like receptor.

Figure 1

Toll-like receptor 9/B cell receptor stimulation induces Interferon regulatory factor 5 (IRF5) nuclear translocation. (A) Representative images of IRF5 cellular localization in human CD19⁺ B cells from a single healthy donor that were stimulated with mock, anti-immunoglobulin (Ig) M antibody, CD40 ligand (CD40L), CpG-B, or CD40L and IL21, and the combination of anti-IgM antibody, CD40L, IL21, and CpG-B for 2 h. PBMC were surface-stained with anti-CD19 antibodies, fixed and permeabilized, then stained for intracellular IRF5 and nuclear DRAQ5. Samples were then subjected to imaging flow cytometry followed by analysis in the IDEAS software suite. (B) Frequency of cells in (A) with IRF5 nuclear translocation, which was defined by an IRF5 and DRAQ5 similarity score ≥2 (one-way ANOVA with Tukey’s post hoc test; n = 4 independent donors). (C) IRF5 nuclear translocation was quantified over 12 h in isolated B cells following stimulation with anti-IgM⁺ CpG-B. Data was normalized to mock to minimize donor variability (one-way ANOVA with Tukey’s post hoc test; n = 4 independent donors). (D) Representative gating strategy for human B cells subsets in peripheral blood of healthy donors who received the influenza vaccine 7 days prior to phlebotomy. B cell populations were defined as CD19⁺CD20⁺ naive (IgD⁺CD38⁻CD27⁻CD24⁻), transitional (IgD⁺CD27⁻CD38⁺CD24⁺), non-switched memory (NSM; IgD⁺CD27⁺CD38⁻CD24⁻), switched memory (SM; IgD⁻CD27⁺CD24⁻), plasma blasts (PB; IgD⁻CD38^hiCD27⁺CD24⁻CD138⁻), and plasma cells (PC; IgD⁻CD38^hiCD27⁺CD24⁻CD138⁺). (E) Representative histograms of IRF5 protein expression in B cell subsets gated in (D). (F) Average MFI of IRF5 expression in gated B cell subsets (two-way ANOVA with Tukey’s multiple comparison post hoc test; n = 5 independent donors). Error bars represent SD. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

Figure 2

Interferon regulatory factor 5 (IRF5) is required for toll-like receptor 9/B cell receptor -induced antibody secreting cell differentiation. Isolated human naive B cells were nucleofected with 500 nM of mock, scrambled or IRF5 siRNA and mock-stimulated or stimulated with the indicated cocktails. (A) IRF5 transcript expression was quantified through qPCR on RNA isolated 48 h postnucleofection and 12 h poststimulation with anti-IgM⁺ CpG-B (two-way ANOVA with Tukey’s post hoc test; n = 4 independent donors). (B) Protein lysates were prepared from nucleofected B cells 72 h postnucleofection and 24 h poststimulation with anti-IgM⁺ CpG-B. Western blot is one representative experiment out of three performed on n = 3 independent donors. (C) Representative histograms of IRF5 expression 48 h postnucleofection. Viable cells were analyzed through live/dead staining discrimination. (D) Plotted MFI of IRF5 from (C) (one-way ANOVA with Tukey’s post hoc test; n = 4 independent donors). (E) Plotted percentage of B cell viability assessed 72 h post-nucleofection, as determined through trypan blue exclusion. Data are from n = 3 independent donors. (F) Representative dot plots from B cell apoptosis quantified following staining with Annexin V and 7 amino-actinomycin D (7-AAD). Early apoptotic events are characterized as Annexin V⁺ 7AAD⁻, whereas late apoptotic events are Annexin V⁺ 7AAD⁺. Quantitation is shown in (G). (G) Average apoptotic B cells from (F) 96 h post-nucleofection and 48 h post-stimulation (Two-way ANOVA with Tukey’s post hoc test; n = 3 independent donors). (H) Isolated naive B cells were nucleofected with 500 nM of mock, scrambled or IRF5 siRNA and stimulated with either CD40L or the combination of CD40L, IL21, anti-IgM, and CpG-B for 7 days. Plasmablast differentiation was quantified through gating of CD19⁺CD20⁺IgD⁻CD27⁺CD38⁺ B cells; final CD27⁺CD38⁺ gating is shown. Flow cytometry contour plots are representative of one experiment from a single donor. (I) Average number of plasmablasts from (H) following culture for 7 days (two-way ANOVA with Tukey’s post hoc test; n = 9 independent donors). (J) Average number of IgD⁺CD38^lo B cells from (H) following stimulation (two-way ANOVA with Tukey’s post hoc test; n = 9 independent donors). Error bars represent SD. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

Figure 3

Interferon regulatory factor 5 (IRF5) knockdown impairs B cell proliferation, activation, and immunoglobulin (Ig) G isotype secretion. (A) Representative flow cytometry histograms from cell proliferation assay as measured through dilution of the proliferation dye CFSE. Isolated primary naive B cells from a single donor were nucleofected with 500 nM of mock, scrambled or IRF5 siRNA and stimulated with either CD40 ligand (CD40L) or the combination of CD40L, IL21, anti-IgM, and CpG-B for 5 days. (B) Average percentage of proliferating B cells from (A) is shown (two-way ANOVA with Tukey’s post hoc test; n = 5 independent donors). (C) Average percentage of B cells expressing CD86. Similar to (A) except isolated naive B cells were nucleofected and then stimulated with CD40L, IL21, anti-IgM, and CpG-B for 24 h, then stained for surface CD86 (two-way ANOVA with Tukey’s post hoc test; n = 5 independent donors). (D) Representative contour plots from IgD⁻IgG⁺ B cells after gating on total CD19⁺CD20⁺ B cells. Isolated primary naive B cells were nucleofected with 500 nM of mock, scrambled or IRF5 siRNA and stimulated with either CD40L or the combination of CD40L, IL21, anti-IgM, and CpG-B for 7 days. (E) Quantification from (D) of Ig antibody class expression was determined by intracellular flow cytometery (one-way ANOVA with Tukey’s post hoc test; n = 5 independent donors). (F) Cell culture supernatants from (D) were used for ELISA to determine IgG isotype secretion. Average concentration of IgG isotype is shown (two-way ANOVA with Tukey’s post hoc test; n = 4 independent donors). (G) Frequency of activation-induced deaminase (AID) expression in B cells following nucleofection with 500 nM of mock, scrambled or IRF5 siRNA and stimulated with CD40L, IL21, anti-IgM, and CpG-B for 3 days. AID expression was determined by intracellular flow cytometry (one-way ANOVA with Tukey’s post hoc test; n = 5 independent donors). Error bars represent SD. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001.

Figure 4

Interferon regulatory factor 5 (IRF5) binds promoter regions of genes associated with antibody secreting cell (ASC) differentiation. IRF5 ChIP-Seq was performed on isolated primary naive B cells from n = 2 independent donors. B cells were either mock or anti-IgM⁺ CpG-B stimulated for 4 h. Reads were mapped through BWA and peaks called through MACs. (A) A pie chart showing representative IRF5 binding elements throughout the human primary B cell genome. (B) Common IRF5 binding motifs identified from ChIP-Seq. Motif sequences are shown in order of enrichment with associated transcription factor motif. (C) Representative peak distributions are shown for IRF4, ERK1, CASP5, and MYC. Peaks are boxed in red and were determined at a significance of p ≤ 0.0001. (D) Same as (C) except IRF5 ChIP-Seq peaks are from the Ramos B cell line. Peaks were determined at p ≤ 10⁻¹⁰. (E) Independent confirmation of IRF5 binding to IRF4, MYC, and ERK1 target sites through ChIP-qPCR in primary naive B cells mock stimulated or stimulated with anti-IgM⁺ CpG-B for 4 h (two-way ANOVA with Tukey’s post hoc test; n = 4 independent donors). Error bars represent SD. *p ≤ 0.05; ***p ≤ 0.001.

Figure 5

Identification of an interferon regulatory factor 5 (IRF5)-dependent B cell transcriptome. Isolated naive B cells were nucleofected with 500 nM of mock, scrambled or IRF5 siRNA and subsequently stimulated with anti-IgM and CpG for 6 h. RNA-seq was performed on B cells from n = 2 independent donors. (A) Significant donor correlation associated with IRF5 knockdown and stimulation. Heat map illustrates correlation coefficients for gene expression values between conditions (nucleofection and stimulation) and individual healthy donors. A strong correlation between samples is indicated by the red and pink boxes. Color legend is shown to the right indicating correlation coefficients. (B) Identification of genes with differential expression following IRF5 knockdown in mock-treated samples. Red circles indicate downregulated genes; green circles indicate upregulated genes. Differential gene expression was determined in comparison to scrambled control. (C) Same as (B) except genes were identified following IRF5 knockdown and stimulation (CpG-B⁺ anti-IgM). Differential expression was determined by comparison to stimulated scrambled control. (D) Diagram of differential expression between nucleofection and various stimulation conditions. Overlapping regions represent shared expression; single regions represent uniquely expressed genes. (E) Heat map of gene expression between the two donors based on cellular function associated with B cell activation, proliferation, and antibody secreting cell differentiation. Data are expressed as log₂RPKM. (F) Raw RPKM values of particular genes relevant to IRF5 function. p values and false discovery rate (FDR) scores were obtained by multiple comparisons testing of samples indicated by line. Individual p values and FDR scores are included in each graph. (G) Independent confirmation of differential gene expression through qPCR (two-way ANOVA with Tukey’s post hoc test; n = 3 independent donors). Error bars represent SD. **p ≤ 0.01.

Figure 6

Interferon regulatory factor 5 (IRF5) regulates numerous biologic signaling pathways in human primary B cells. (A) Data from ChIP-Seq and RNA-Seq were used to identify the top 20 upregulated pathways in which there were more upregulated genes than downregulated genes as compared to Mock untreated (NT). Heat map is shown from HOMER analysis with a cutoff of log₂FC ≥ 1 representing genes with at least a twofold change in expression and an false discovery rate (FDR) ≤ 0.001. Pathways are ordered by negative log10 p-values of enrichment. (B) Same as (A) except top 20 downregulated pathways are shown. (C) Representative scheme showing enrichment of the upregulated NF-ĸB survival signaling pathway after IRF5 knockdown. Boxes representing genes are color-coded based on the log2 fold-change values. Upregulated (red) or downregulated (blue) genes are shown as compared to Mock NT. Expression is shown by color code in four sections of each gene box. The gene box is divided into sections with one to one mapping comparisons of IRF5KD_NT, IRF5KD_anti-IgM⁺ CpG, Scr_anti-IgM⁺ CpG, and Mock_anti-IgM⁺ CpG vs. reference Mock_NT, as shown in the Legend.

Figure 6

Interferon regulatory factor 5 (IRF5) regulates numerous biologic signaling pathways in human primary B cells. (A) Data from ChIP-Seq and RNA-Seq were used to identify the top 20 upregulated pathways in which there were more upregulated genes than downregulated genes as compared to Mock untreated (NT). Heat map is shown from HOMER analysis with a cutoff of log₂FC ≥ 1 representing genes with at least a twofold change in expression and an false discovery rate (FDR) ≤ 0.001. Pathways are ordered by negative log10 p-values of enrichment. (B) Same as (A) except top 20 downregulated pathways are shown. (C) Representative scheme showing enrichment of the upregulated NF-ĸB survival signaling pathway after IRF5 knockdown. Boxes representing genes are color-coded based on the log2 fold-change values. Upregulated (red) or downregulated (blue) genes are shown as compared to Mock NT. Expression is shown by color code in four sections of each gene box. The gene box is divided into sections with one to one mapping comparisons of IRF5KD_NT, IRF5KD_anti-IgM⁺ CpG, Scr_anti-IgM⁺ CpG, and Mock_anti-IgM⁺ CpG vs. reference Mock_NT, as shown in the Legend.

Figure 7

Interferon regulatory factor 5 regulates B cell proliferation and cell cycle. Similar to Figure 6C showing the overall downregulation of the cell cycle and DNA damage signaling pathway from ChIP-Seq and RNA-Seq datasets. Genes are shown as being either upregulated (red) or downregulated (blue) when compared to Mock untreated (NT). Representative expression is shown in four sections of each gene box as depicted in the Legend.