IRF8 regulates B-cell lineage specification, commitment, and differentiation

Abstract

PU.1, IKAROS, E2A, EBF, and PAX5 comprise a transcriptional network that orchestrates B-cell lineage specification, commitment, and differentiation. Here we identify interferon regulatory factor 8 (IRF8) as another component of this complex, and show that it also modulates lineage choice by hematopoietic stem cells (HSCs). IRF8 binds directly to an IRF8/Ets consensus sequence located in promoter regions of Sfpi1 and Ebf1, which encode PU.1 and EBF, respectively, and is associated with transcriptional repression of Sfpi1 and transcriptional activation of Ebf1. Bone marrows of IRF8 knockout mice (IRF8(-/-)) had significantly reduced numbers of pre-pro-B cells and increased numbers of myeloid cells. Although HSCs of IRF8(-/-) mice failed to differentiate to B220(+) B-lineage cells in vitro, the defect could be rescued by transfecting HSCs with wild-type but not with a signaling-deficient IRF8 mutant. In contrast, overexpression of IRF8 in HSC-differentiated progenitor cells resulted in growth inhibition and apoptosis. We also found that IRF8 was expressed at higher levels in pre-pro-B cells than more mature B cells in wild-type mice. Together, these results indicate that IRF8 modulates lineage choice by HSCs and is part of the transcriptional network governing B-cell lineage specification, commitment, and differentiation.

Figure 1

IRF8 suppresses PU.1 expression. (A) IRF8 is present at the promoter region of PU.1 in vivo. ChIP analyses were performed with NFS-202 cells or purified spleen B cells. Protein-DNA complexes were immunoprecipitated by addition of antibody to IRF8 and analyzed by PCR for the presence of PU.1 promoter sequences. (B) qPCR analysis of IRF8 and PU.1 expression in cells expressing a repressive IRF8 siRNA. IRF8 Nos. 2 and 5 are 2 clones from IRF8 siRNA–treated cells. (C) Western blotting analysis of IRF8 and PU.1 expression in IRF8 siRNA–expressing cells. (D) Western blotting analysis of IRF8 and PU.1 expression in NFS-203 cells expressing pCDNA-Irf8 or an empty pCDNA vector as indicated. The numbers are arbitrary units of protein intensities normalized by β-actin. All data are representative of 2 to 4 independent experiments.

Figure 2

IRF8 regulates EBF expression. (A) Schematic arrangement of IRF8-binding sites (IECS, EICE, and ISRE) in the promoter of the Ebf1 gene. +1 corresponds to the first nucleotide of exon 1 of the gene. The nucleotides that were mutated at each IRF8-binding site are shown on the top. WT indicates wild type. (B) ChIP analysis of IRF8 binding with Ebf1. In vivo cross-linked IRF8-chromatin complexes from NFS 202 cells were analyzed using PCR primers that span the IRF8 sites in the Ebf1 gene. (C) IRF8 and PU.1 regulate EBF expression in a luciferase reporter assay. HeLa cells were cotransfected with a promoter reporter pGL4-Ebf1-WT and vectors expressing IRF8, PU.1, or both. An empty vector was used as a control. (D) Mutation of IRF8-binding sites in the Ebf1 gene impaired the expression of the Ebf1 reporter. The Ebf1 promoter reporter constructs containing mutated IRF8-binding sites were generated as illustrated in panel A and were cotransfected with plasmids expressing IRF8 and PU.1. Luciferase activities were measured after 22 hours. All data represent 3 to 4 independent experiments.

Figure 3

Expression of Irf8, Pu.1, Ebf, Pax-5, and E2a in HSCs, CLPs, MPs, and various B-cell subsets. Each cell population was sort-purified from a pool of 5 to 6 mice. RNA was extracted and reverse transcribed. qPCR was used to quantitate Irf8 (A), Pu.1, Ebf, E2a, and Pax-5 (B) expression levels. Data are mean plus or minus SEM of 3 independent experiments. UD indicates undetectable.

Figure 4

Bias of MP development in IRF8^−/− mice. (A) BM cells from IRF8^−/− and IRF8^+/+ littermate mice (n = 8) were stained with antibodies against Lineage panel (Lin), IL-7Rα, c-Kit, and Sca-1. The numbers represent percentage of total events. (B) The frequency (left panel) and absolute cell numbers (right panel) of HSCs, CLPs, and MPs in IRF8^+/+ and IRF8^−/− mice. Data are mean plus or minus SEM of 5 mice per group. *P < .05; **P < .001 compared with controls. (C) Sorted HSCs from IRF8^+/+ and IRF8^−/− mice were injected into lethally irradiated CD45.2⁻ B6 mice. Two months later, BM cells were analyzed by FACS and the donor cells (CD45.2⁺) were identified by an anti-CD45.2 antibody. The frequency (left panel) and absolute cell numbers (right panel) of donor HSCs, CLPs, and MPs were shown as mean plus or minus SEM of 6 mice per group. *P < .05 compared with controls.

Figure 5

Impaired B-cell development in the BM of IRF8^−/− mice. (A) BM cells from the indicated mice (n = 8) were stained for B220, IgM, CD43, BP-1, and HSA. The percentage of B cells falling within each gate is given. Hardy Frs A to F are indicated. L indicates large; S, small. (B) The frequency and absolute numbers of each B-cell subset. Data are mean plus or minus SEM of 6 mice. * indicates P value less than .001 compared with control groups; **, P value less than .05 compared with Fr A cells. (C) IL-7 cultured BM cells were stained with antibodies against B220, IgM, and pre-BCR. The anti–pre-BCR antibody (clone SL156) specifically recognizes an epitope composed of the μH chain and the surrogate L chain. The percentage of B cells falling within the gate is given. Data represent 1 of 3 mice with similar results. (D) Sorted HSCs from IRF8^+/+ and IRF8^−/− mice were injected into lethally irradiated CD45.2⁻ B6 mice. Two months later, BM cells were analyzed by FACS and the donor cells (CD45.2⁺) were identified by an anti-CD45.2 antibody. The frequency (left panel) and absolute cell numbers (right panel) of donor B cells were shown as mean plus or minus SEM of 6 mice per group. * P < .05; **P < .001 compared with the control group.

Figure 6

The failure of IRF8-deficient HSCs to differentiate into B cells is cell intrinsic. (A) Sorted HSCs from IRF8^+/+ and IRF8^−/− mice were induced to differentiate in OP9 cocultures in the presence of SCF, IL-7, and Fl3tL for 7 days. The cells were stained with antibodies recognizing B220, CD19, and CD11b and analyzed by flow cytometry. Cells were gated on 7AAD⁻ living cells. The numbers are percentages of cells falling in each gate. Data represent 1 of 3 independent experiments. (B) Sorted HSCs from each group of mice were cultured in the presence of SCF, Flt3L, IL-3, and IL-6 for 18 hours before they were infected with retroviral vectors encoding an IRF8-GFP fusion protein, an IRF8 mutant K79E-GFP fusion protein, or GFP only for 24 hours in the presence of SCF, Flt3L, IL-3, IL-6, IL-7, GM-CSF, and 4 μg/mL polybrene. GFP⁺ cells were resorted and plated onto OP9 cell layer in the presence of SCF, IL-7, and Flt3L for 7 days. The cells were analyzed by flow cytometry. The numbers are percentages of cells falling in each gate. Data represent 1 of 4 independent experiments. (C) The sorted GFP⁺ cells were cultured with OP9 cells at the same conditions as in panel B and were pulsed with BrdU for 40 minutes at day 4 and analyzed by FACS. The numbers are percentages of cells falling in each gate. Data represent 1 of 2 independent experiments.