Hepatocyte Nuclear Factor 1A Is a Cell-Intrinsic Transcription Factor Required for B Cell Differentiation and Development in Mice Free

Hepatocyte NFs (HNF) are a family of heterogeneous transcription factors involved in the regulation of a wide variety of metabolic pathways and organ development (1). To date, hundreds of target genes have been identified to be regulated by HNFs—either by single family members or several of them working together in complex regulatory networks (1). The prototypical family member we have focused on, HNF1A, regulates the transcription of genes involved in gluconeogenesis, lipid, and carbohydrate metabolism in hepatocytes (1) as well as that of genes related to glucose, amino acid, and tricarboxylic acid metabolism and necessary for mitochondrial function in pancreatic island cells, respectively (2). Furthermore, HNF1A also targets genes that regulate kidney function (3). Mutations of HNF1A in humans become clinically manifest by early-in-age manifestation of diabetes mellitus (maturity onset of diabetes in the young [MODY] type 3) (4, 5). HNF1A deletion in mice results in a phenotype characterized by progressive hepatic failure, phenylketonuria, renal Fanconi syndrome (3), diabetes mellitus, and growth hormone insensitivity (6). Other members of the HNF family (HNF1B, HNF3 [renamed Foxa family of transcription factors], HNF4, HNF6) regulate a plethora of genes involved in a variety of biochemical metabolic pathways (1, 7), but also in organogenesis, cell differentiation, and tumor growth (7–12). Interestingly, two HNFs have been linked to hematopoietic functions: HNF4 is involved in the regulation of erythropoietin production (13) and hepatic parenchymal HNF6 regulates—as a non-cell–intrinsic transcription factor—B cell differentiation in the fetal liver (14). The restriction to fetal lymphopoiesis becomes evident from the disappearance of the B cell defect of newborn HNF6-deficient mice shortly after birth when hematopoiesis is taken over by the bone marrow (14).

Adult B cell lymphopoiesis begins with the differentiation of lymphoid-primed multipotential progenitor cells into common lymphoid progenitor (CLP) cells (15). By successive signaling of transcription factors and cytokines, these cells become restricted to the B lymphocyte lineage. At least the three transcription factors—E2A, early B cell factor (EBF)1, and paired box protein 5 (PAX5)—are crucially involved in the promotion of B cell lymphopoiesis (16). IL-7 and its receptor (IL-7R) play important roles in the B cell lineage commitment of CLPs (17, 18). In addition, IL-7 regulates EBF1 expression levels that are responsible for the transition of pre-pro–B cells to the pro–B cell stage (19). STAT5 phosphorylated by IL-7 can directly upregulate Pax5 transcription via EBF in early B cell progenitors (20). From the very early B cell stage on, B lineage–committed cells mature by Ig rearrangement and surface expression of specific Ags to become mature recirculating B cells (21, 22). Impaired B cell development due to a primary immunodeficiency is a rare condition in humans (5:1,000,000), which in ∼90% of the cases is caused by the X-chromosome–linked gene defect of Bruton’s tyrosine kinase and in ∼10% by autosomal recessive mutations of other genes, respectively (23). In mice, failure of B cell development is the consequence of genetic deletion of several genes (23), but none of them has yet been linked to any HNF1A functions.

In the current study, we have identified a novel cell-intrinsic role for HNF1A in adult B cell lymphopoiesis. Its deletion resulted in a decrease of B cells in peripheral blood and spleen that was caused by the dramatic reduction of B cell progenitor numbers in the bone marrow beginning from the pre-pro–/early pro–B cell stage to the immature B cell stage. We have identified the underlying defect to be a marked and transplantable irresponsiveness of developing B cell progenitors to IL-7 and have characterized the involved downstream events.

Mice heterogenous for hepatocyte NF1A (HNF1A; Institute Pasteur, Paris, France) were crossed to obtain HNF1A^+/+ and HNF1A^−/− animals. Genotypes were determined by PCR with the following primer: forward, 5′-TACCTGATGGTTGGAGAGGGTC-3′; reverse, 5′-TGAAGACCACATCTCCTAAGG-3′; control forward, 5′-TCAATCCGCCGTTTGTTCCC-3′; control reverse, 5′-GCATAACCACCACGCTCATC-3′, showing amplification products at 300 bp for HNF1A^+/+ and at 550 bp for HNF1A^−/− mice. Knockout animals showed the phenotype of failure to thrive, wasting, and progressive renal and hepatic dysfunction, as already described (3). The study had been reviewed and approved by the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen.

Mice were euthanized by inhalation of isoflurane. Blood was drawn by cardiac puncture and anticoagulated with EDTA. WBC counts were determined in whole blood with a Scil Vet ABC hematology analyzer. WBCs were isolated from whole blood by lysing the RBCs with lysis buffer (Pharmlyse; BD Biosciences) and subsequent washing of the remaining cells with stain/wash buffer (1% BSA, 0.1% sodium azide in PBS). The organs were removed from the bodies immediately and kept in R10 medium (RPMI 1640 with 10% FCS, 1% l-glutamine [200 mM, 100×] and 1% penicillin-streptomycin-amphotericin B [10,000 IE penicillin, 10 mg/ml streptomycin, 25 μg/ml amphotericin B in 0.85% NaCl, 100×], all from Life Technologies/Invitrogen) on ice. Cells were isolated from spleens and lymph nodes by disrupting the tissues manually with scalpels, and from bone marrow by flushing the bones with R10. Isolated tissue fragments were passed through cell strainers. RBCs contained in the suspension of bone marrow or splenic cells were lysed with ACK (ammonium-chloride-potassium) buffer (0.15 M NH₄Cl, 1 mM KHCO₃, 0.1 mM Na₂EDTA, pH 7.2–7.4). Cells were washed once with R10 and then resuspended in R10 and kept on ice. Numbers of cells isolated from spleens and lymph nodes were determined with a Beckman Coulter Z2 Particle Count and Size Analyzer. Numbers of bone marrow cells were counted in a Neubauer chamber. Prior to the addition of Abs, cells were washed once with stain/wash buffer and preincubated with Fc receptor block CD16/CD32 for 10 min at 4°C. After that, cells were incubated with the Abs for 30 min at 4°C in the dark, then washed once with stain/wash buffer, and analyzed for surface marker expression with a Beckman Coulter Gallios flow cytometer. Abs used were against murine CD11b, CD11c, CD49b, CD3ε, Ter119, CD19 (all from Miltenyi Biotec), B220, CD23, CD8a, CD115, IgM, IgD, CD93, CD21/CD35 (all from eBioscience), CD4 (Acris), CD45, Gr-1, CD16/CD32 (Fc receptor block), CD45.2, Sca-1, c-Kit (CD117), CD138 (all from BD Biosciences), F4/80, and CD127 (IL-7R, Ab recognizing the α-chain) (both from AbD Serotec).

Pre-enrichment of B cell progenitors from the bone marrow was performed by magnetic depletion of myeloid, dendritic, and immature/mature B cells using mouse/human CD11b microbeads, mouse CD11c microbeads, and anti-mouse IgM microbeads (all from Miltenyi Biotec), according to the manufacturer’s instructions. For subsequent B cell progenitor isolation, the cells were stained with PE-conjugated anti-B220 Ab (Miltenyi Biotec), followed by incubation with anti-PE microbeads UltraPure (Miltenyi Biotec), and magnetic cell separation was performed, according to the manufacturer’s recommendations (Miltenyi Biotec). All microbead and Ab incubations were performed on ice for 30 min.

Single-cell suspensions of bone marrow cells were prepared and counted, as previously described (24). Total bone marrow cells were depleted of mature hematopoietic cells using a mouse lineage cell depletion kit (Miltenyi Biotec), according to the manufacturer’s instructions. The remaining cells were first stained with Fc-blocking Ab (BD Biosciences) in 2% FCS/PBS staining buffer for 20 min on ice, washed once in staining buffer, followed by incubation with Abs against CD127 (IL-7R; BioLegend), CD93 (AA4.1), CD135 (Flk2), Gr-1, CD11b, CD19, CD4, and CD8 (all from BD Biosciences), Sca-1, CD117 (c-Kit), CD3ε, Ter119, Nk1.1, B220, and TCR-β (all from eBioscience) for 20 min on ice, respectively. In addition, small amounts of remaining undepleted biotinylated beads were bound to streptavidin-PerCP-Cy5.5 (eBioscience) to avoid unspecific staining. Dead cells were excluded from flow cytometric sorting by DAPI gating and cell doublets by forward light scatter width against forward light scatter area gating. CLPs were transferred directly into OP9 coculture media by a FACS Aria III (BD Biosciences).

For BrdU labeling, mice were injected i.p. once with a 10 mg/ml BrdU solution (in 0.9% NaCl, 90 mg/kg body weight; Sigma-Aldrich) and analyzed 24 h later. B cell proliferation was tested with the FITC BrdU Flow Kit from BD Biosciences (51-2354AK), according to the manufacturer’s instruction.

For phosphorylated STAT5 detection in B cell progenitors, bone marrow cells were first depleted of myeloid and dendritic cells. B cells were then magnetically isolated by anti-PE UltraPure microbeads after incubation with anti-mouse PE-CD19, according to manufacturer’s instructions (Miltenyi Biotec). Cells were then allowed to recover in RPMI 1640 containing 5% FCS and antibiotics at 37°C for 2 h, followed by stimulation with murine 10 ng/ml IL-7 (PeproTech) for 15 min. B cells were subsequently resuspended in Lyse/Fix buffer (BD Biosciences) and fixed for 12 min. After washing with 3% FCS/PBS, cells were permeabilized with Perm Buffer III (BD Biosciences), followed by washing twice with 3% FCS/PBS, according to manufacturer’s recommendations. After incubation with a Fc-blocking Ab (eBioscience) for 15 min, samples were stained for phosphorylated STAT5 (pY694; BD Biosciences) and B220 (eBioscience) for 60 min at room temperature prior to FACS analysis.

Murine IL-7 (PeproTech) was injected i.p. at 0.5 μg in 100 μl 0.15% BSA/PBS once daily for 7 d, and mice were studied 24 h later.

OP9 monolayers were cultivated in MEMα with GlutaMAX (Life Technologies/Invitrogen) containing 20% FCS (Life Technologies/Invitrogen), 1% penicillin-streptomycin-amphotericin B (Life Technologies/Invitrogen). To study B cell development, sorted CLP populations were plated onto 80–90% confluent OP9 cells and maintained in IMDM (BioWhittaker/Lonza) containing 5% FCS, 1% penicillin-streptomycin-amphotericin B (Life Technologies/Invitrogen), 50 μg/ml gentamicin (Life Technologies/Invitrogen), 2 mM glutamine (Sigma-Aldrich), 55 μM 2-ME (Life Technologies/Invitrogen), 10 ng/ml mouse stem cell factor (R&D Systems), 10 ng/ml human FMS-like tyrosine kinase 3 ligand (Flt3l); PeproTech), and murine 10 ng/ml IL-7 (PeproTech) for 6 d, respectively. B cell development was assessed by flow cytometric analyses of B220 and CD19 expression.

Bone marrow cells were isolated from donor mice after euthanasia with isoflurane. Bones (femurs and tibiae) were removed from the body, cleaned from tissue, and washed in 70% ethanol for 30 s. After rinsing with PBS, bone marrow was flushed out with R10 medium. Cells were passed through a 70 μm cell strainer and pelleted by centrifugation. RBCs were lysed with ACK buffer. After the subsequent washing with R10, cell counts were determined and the adequate cell number was resuspended in 200 μl sterile PBS. Bone marrow recipients were lethally irradiated by fractionated irradiation with 6 Gray, followed by an irradiation with 5 Gray 4 h later. On the following day, 4.5 × 10⁶ bone marrow cells in 200 μl PBS were injected i.v. into the retroorbital plexus.

RNA of bone marrow–derived B cells, thymus, and liver was isolated with the InviTrap Spin Universal RNA Mini Kit from Invitek. Copy DNA was generated from equal amounts of RNA with the RevertAid First Strand cDNA Synthesis Kit from Thermo Scientific. Real-time PCR (RT-PCR) was performed in a BioRad CFX cycler using SYBRGreen as fluorophore and a standard RT-PCR protocol (3 min at 95°C, followed by 40 cycles of 20 s at 95°C, 20 s at 55°C, and 20 s at 72°C, followed by 10 s at 95°C and the melting curve of incremental increases of 0.5°C every 5 s from 65°C to 95°C). The following primers were used: HNF1A (forward, 5′-GACCTGACCGAGTTGCCTAAT-3′; reverse, 5′-CCGGCTCTTTCAGAATGGGT-3′; Biolegio), EBF1 (forward, 5′-GCATCCAACGGAGTGGAAG-3′; reverse, 5′-GATTTCCGCAGGTTAGAAGGC-3′; BioTez), E2A (Tcfe2a; QT00098196; Qiagen), Pax5 (QT00174398; Qiagen), Ikaros (QT01550703; Qiagen), Bach2 (forward, 5′-TCAATGACCAACGGAAGAAGG-3′; reverse, 5′-GTGCTTGCCAGAAGTATTCACT-3′; BioTez), PU.1 (QT00098077; Qiagen), and GAPDH (forward, 5′-TCAAGAAGGTGGTGAAGCAG-3′ and reverse, 5′-TCGCTGTTGAAGTCAGAGGA-3′; BioTez). Relative quantification of gene expression was calculated by the 2^−ΔΔCT method using GAPDH as an endogenous reference for normalization.

Data are presented as mean ± SEM. Groups were compared by unpaired t test or if performed in sets by paired t test. The p values are understood to be strictly descriptive. Statistical significance was assumed for p < 0.05.

Analysis of peripheral blood from HNF1A^−/− mice showed a clear reduction in total B cell numbers compared with controls (Fig. 1A). In contrast, CD4⁺ and CD8⁺ T cells, monocytes, and neutrophils were unaffected (Fig. 1A and data not shown). The lower B cell number in blood was due mainly to a reduction of transitional B cells, whereas mature B cells were almost unaffected (Fig. 1B, 1C). As transitional B cells in blood are constituted of immature B cells exiting the bone marrow and bound to the spleen for further development and differentiation (25), spleens of HNF1A^−/− and control mice and their B cell populations were analyzed. The mean HNF1A^−/− spleen weight was reduced (Fig. 1D, middle), and normalization to the reduced body weight of HNF1A^−/− mice (Fig. 1D, left) still resulted in a clear decrease of relative spleen weight (Fig. 1D, right). Both absolute and normalized splenic Lin⁻B220⁺ cell numbers were reduced in HNF1A^−/− mice (74 and 46% less than controls, respectively; Fig. 1E).

Splenic B cell populations were defined by the expression and intensity of different surface markers, as described (26, 27), and distinguished in early (transitional T1 and T2) and late (follicular type I and II B cells, marginal zone precursors [MZPs], and marginal zone) B cells (Fig. 1F, 1G). Particularly, the early transitional T1 population was dramatically reduced (17-fold), followed by the developing T2 population (6-fold reduction) (Fig. 1G, left). The mature B cell populations—follicular type and marginal zone B cells—were also significantly reduced in HNF1A^−/− mice, albeit to a lesser degree (2-fold) (Fig. 1G, middle and right). This observation along with the presence of normal numbers of marginal zone precursors (MZP) and normal lymph node cell numbers (Supplemental Fig. 2A–C) argued for the presence of a compensatory mechanism accelerating B cell development starting at T2 (Fig. 1G, left). There was no difference in plasma cells (Supplemental Fig. 1) and no sign of extramedullary hematopoiesis in the spleen, respectively (data not shown).

To follow up on the reduced transitional B cell populations in spleen and blood, bone marrow from control and HNF1A^−/− mice was analyzed for developing B cell subpopulations, as described (28–31). HNF1A^−/− bone marrow showed major reductions both in immature IgD⁻ or IgD^low B cells both in percentage (Fig. 2A, 2B) and total cell numbers (Fig. 2C), offering an explanation for the reduced numbers of transitional B cell populations in spleen and blood. In addition, early B cell progenitors (Lin⁻B220⁺IgM⁻IgD⁻) defined as the sum of all pre-pro–/pro–/pre–B cell populations were also noticeably reduced in HNF1A^−/− mice (Fig. 2A–C). These findings implied the existence of a severe defect in B cell lymphopoiesis caused by HNF1A deficiency.

To pursue this further, early B cell populations were analyzed dependent on the presence or absence of c-Kit (CD117) and IL-7Rα, thereby distinguishing three consecutive groups of developing B cells, as follows: 1) pre-pro–/early pro–B cells (lin⁻B220⁺IL-7R⁺c-Kit⁺), 2) late pro–/early pre–B cells (lin⁻B220⁺IL-7R⁺c-Kit⁻), and 3) late pre–/immature B cells (lin⁻B220⁺IL-7R⁻) (Fig. 2D–F). All three groups of B cell progenitors were dramatically reduced in HNF1A^−/− mice compared with controls both in percentage and total cell counts (Fig. 2E, 2F). Accordingly, the total population of all developing B cells (the sum of pre-pro–/pro–/pre– and immature B) was also substantially lower in HNF1A^−/− mice (Fig. 2E, 2F). Interestingly, the decrease in developing B cells was not equally prominent among the different groups: B cells at the earliest stages of B cell development (pre-pro–B/early pro–B) were reduced 2-fold, whereas later stages of B cell development (late pro–/early pre–B and late pre–/immature B) were reduced much more prominently (up to 6-fold) (Fig. 2E, 2F). This suggested that the developmental B cell defect in HNF1A^−/− mice begins at a very early B cell developmental stage and worsens further during development.

To assess earlier stages of hematopoiesis, Lineage^−/lowIL-7R⁻Sca-1⁺c-Kit⁺ (LSK) cells and CLPs (Lineage⁻IL-7R⁺Sca-1^lowc-Kit^low) were determined, as previously described (32). Both showed an increase in HNF1A^−/− bone marrow in percentage (Fig. 2G, 2H) as well as in total counts (Fig. 2I) compared with controls. This excluded a defect upstream of the B cell commitment stage and suggested either a block leading to the accumulation of LSKs and CLPs or an attempt to compensate for defects occurring during later B cell development.

To test whether the much more prominent decline in late pro–/early pre–B and late pre–/immature B cells (6-fold) as compared with that in pre-pro–/early pro–B cells (2-fold) was due to a defect in B cell progenitor proliferation, BrdU labeling was employed in vivo. Bone marrow BrdU⁺ B cell subpopulations were discriminated according to their IL-7R expression: 1) IL-7R⁺pre-pro–/pro–/early pre–B cells and 2) IL-7R⁻ late pre–/immature B cells. Both populations exhibited a marked reduction in the percentage of BrdU⁺ in HNF1A^−/− bone marrow compared with controls (Fig. 3A, 3B).

IL-7 and its receptor (IL-7R) play important roles in early murine B cell development (21, 33). Proliferation of committed B cell progenitors is dependent on IL-7R that is first expressed in pre-pro–B cells and stimulates proliferation of both pro–B and early pre–B (also referred to as large pre–B) cells (31). Flow cytometry of cell surface IL-7Rα revealed similar mean fluorescence intensities in IL-7R⁺ B cell progenitors of HNF1A^−/− mice and controls (Fig. 3C). IL-7R signaling drives B cell progenitor proliferation through STAT5 (31). To examine whether this pathway was affected by the absence of HNF1A, intracellular STAT5 phosphorylation in response to IL-7 stimulation was examined by flow cytometry in freshly isolated bone marrow CD19⁺B220⁺ (pro–/pre–/immature B) B cell progenitors. Interestingly, HNF1A^−/− B cell progenitors exhibited a marked reduction in phospho-STAT5–positive cells compared with controls (Fig. 3D, 3E).

We then examined whether the defect in HNF1A^−/− B cell lymphopoiesis in vivo may be due to a reduced response to IL-7. HNF1A^−/− mice and controls were treated with rIL-7 for 7 d, and B cell lymphopoiesis was assessed by flow cytometry. In control mice, IL-7 treatment led to a prominent increase in all developing B cells: 5-fold in pre-pro–/early pro–B cells, 6-fold in late pro–/early pre–B cells, and 6-fold in late pre–/immature B cells, respectively (Fig. 4A). In clear contrast, IL-7 treatment of HNF1A^−/− mice resulted in a severely compromised B progenitor response compared with controls, as evidenced by the only 2-fold or lesser increase in different stages of developing B cells (Fig. 4A). The response of HNF1A^−/− mice to IL-7 was so compromised that it did not even reach untreated wild-type control values (Fig. 4A). Interestingly, the B cell population that did not respond to IL-7 at all in HNF1A^−/− mice was that of the earliest pre-pro–/early pro–B cells (Fig. 4A). In agreement with the differences in the bone marrow response to IL-7, transitional B cells in peripheral blood were increased 4-fold by IL-7 in controls, but only 2-fold in HNF1A^−/− mice (Fig. 4B).

As IL-7R–dependent proliferation takes place not only in B but also during T cell development (33), we next tested whether the relative resistance of B cell progenitors to IL-7 extended to T cells in vivo. However, IL-7 increased T cell numbers both in spleen and lymph nodes to the same extent in control and HNF1A^−/− mice (Fig. 4C, middle and right) in agreement with a normal response to IL-7 (34–36). Peripheral blood T cells were even increased in HNF1A^−/− mice after IL-7 (Fig. 4C, left). This suggests that the resistance to IL-7 signaling was restricted to B cell progenitors in HNF1A^−/− mice.

As IL-7 administration alone did not rescue B cell lymphopoiesis in HNF1A^−/− mice in vivo, we compared the potential of CLPs from HNF1A^−/− and control mice to differentiate into the B cell lineage in vitro. The same number of flow cytometrically isolated CLPs (Lineage⁻IL-7R⁺CD135⁺CD93⁺Sca-1^lowc-Kit^low; Supplemental Fig. 3) from HNF1A^−/− and control mice was plated onto OP9 stromal cell monolayers, which—in the presence of IL-7, Flt3l, and stem cell factor—has been shown to result in the development of B cell progenitors with stage transition from B220⁺CD19⁻ pre-pro–B cells to B220⁺CD19⁺ pro–B cells (19). After 6-d coculture, CLPs from wild-type mice had developed into B220⁺ B cell progenitors with 50–60% of differentiated B220⁺CD19⁺ pro–B cells (Fig. 5A, 5B). In clear contrast, HNF1A^−/− CLPs were much slower to differentiate into developing B cells, as only 20% were B220⁺CD19⁺ pro–B cells and 80% remained in the pre-pro–B cell differentiation stage (Fig. 5A, 5B). Total numbers confirmed the defect as HNF1A^−/− B220⁺CD19⁺ pro–B cells were 6-fold decreased compared with wild-type controls (Fig. 5C). The defects were also visible by phase contrast microscopy, in which HNF1A^−/− B cell progenitors showed a clear impairment in growth (Fig. 5D).

To look for differences in transcription factor expression, RT-PCR was performed on isolated B cell progenitors from HNF1A^−/− and wild-type mice. E2A (3-fold), EBF1 (early B cell factor 1, 5-fold), Pax5 (4-fold), and Bach2 (5-fold) were expressed lower in B cell progenitors of HNF1A^−/− compared with controls, whereas Ikaros and PU.1 remained unchanged (Fig. 5E). These results suggest that decreased expression of several core transcription factors may be involved in impaired B cell lymphopoiesis in the absence of HNF1A.

To distinguish whether the impaired B cell lymphopoiesis of HNF1A^−/− mice was due to cell-intrinsic defects of the developing B cells or to defects of the bone marrow stroma supporting B cell differentiation (21, 22), bone marrow transplantation studies were performed to construct hematopoietic chimera. After sublethal irradiation, bone marrow from HNF1A^−/− mice was transplanted into HNF1A^+/+ mice and vice versa. A group of HNF1A^+/+ mice received HNF1A^+/+ bone marrow to serve as a control. The later featured completely reconstituted B cell lymphopoiesis after 6 wk with early and mature B cell numbers indistinguishable from untransplanted HNF1A^+/+ control mice (Fig. 6A, 6B). In contrast, none of the HNF1A^−/− mice receiving HNF1A^+/+ bone marrow survived the first few weeks after transplantation, a fact we have attributed to an aggravation of metabolic dysfunction inherent to HNF1A^−/− mice (3, 6) that must have been caused by the adverse effects of irradiation. Quite to the opposite, all HNF1A^+/+ mice that received HNF1A^−/− bone marrow survived the 6 wk until analysis. The data these chimera provided were unambiguous: compared with HNF1A^+/+→HNF1A^+/+ bone marrow chimera, HNF1A^−/−→HNF1A^+/+ bone marrow chimera had dramatically decreased numbers of pre-pro–/early pro–B cells (2-fold), late pro–/early pre–B cells, late pre–/immature B cells and all developing B cells (4-fold), and a 3-fold reduction of mature recirculating B cells (Fig. 6A). These findings were completely confirmed by analyzing B cell subpopulations as classified according to their IgM/IgD cell surface expression (Fig. 6B). Peripheral blood B cells were slightly but insignificantly reduced and CD4⁺ and CD8⁺ T cell numbers similar to those in HNF1A^+/+ bone marrow chimera, respectively (Fig. 6C). These findings were completely in line with the phenotype of regular, untransplanted HNF1A^−/− mice.

In this study, to our knowledge, we present for the first time evidence that the prototypical transcription factor HNF1A is required for normal adult B cell development. To date, HNF1A has been known for its regulation of genes involved in glucose, lipid, and amino acid metabolism in the liver, pancreas, intestine, and kidney, and is most abundantly expressed there (1, 2, 37). Nevertheless, HNF1A is also expressed in thymus and spleen (37), and we, too, have found abundant HNF1A expression in the thymus and to a lower, but well-detectable degree in isolated B cells (Supplemental Fig. 4C). Complete HNF1A deficiency leads to different metabolic disorders in mice, and heterozygous human HNF1A mutations present with early-onset diabetes mellitus type MODY3 (4, 5). In addition, human mutations have been associated with alterations in C-reactive protein (38) and the plasma glycan profile (39), increased production of bile acids (40), and higher plasma high-density lipoproteins (HDL) cholesterol levels (41), respectively. Loss-of-function HNF1A mutations have been linked to hepatic tumor development (42–44), and recent genome-wide association studies have implied HNF1A in the development of several diseases such as pancreatic cancer (45), coronary artery disease (46), and low-density lipoprotein hypercholesterolemia (47). To date, however, no B cell differentiation defects have been associated with any human or murine HNF1A^−/− mutations.

Adult B cell development takes place in bone marrow, where B cell committed progenitors arise from long-term reconstituting multipotent hematopoietic stem cells, turn into myeloid/lymphoid progenitors, and become CLP cells that subsequently differentiate into the B and T cell lineages (48, 49). We were able to trace the B cell defect caused by HNF1A deficiency to the very early B cell developmental stage beginning with the pre-pro–/early pro–B cell and following into the immature B cell stage. This was not due to increased progression through development and hence augmented departure from the bone marrow prior to the IgM-positive immature B cell stage as: 1) CD45⁺B220⁺IgM⁻IgD⁻ cells were not increased in HNF1A^−/− blood, and 2) CD45⁺B220^lowIgM⁻IgD⁻ cells in blood (the presence of which may suggest early release of immature B cell progenitors) were clearly reduced (Supplemental Fig. 4A, 4B). In addition, if progression were accelerated with fewer pro–B and pre–B cells left to respond to IL-7, then splenic T1 cells as those to first develop from any prematurely released immature bone marrow B cell progenitors would have been increased. However, precisely the splenic T1 cell population was dramatically reduced in HNF1A^−/− mice (Fig. 1G, left). Of note, the only modest decrease in mature B cells in the spleen, the normal numbers of MZP, and the unchanged mature B cells in blood and lymph nodes argue for a compensatory mechanism of splenic B cell generation taking place in HNF1A^−/− mice. In support, we have found a higher rate of BrdU incorporation in the splenic T1/T2 cell population (Supplemental Fig. 2D) with no sign of extramedullary hematopoiesis.

The bone marrow niche provides a distinct microenviroment for hematopoiesis and B cell lymphopoiesis, with stromal cells and cytokines such as IL-7 actively promoting B cell precursor development (50). Both IL-7 and its receptor IL-7Rα are indispensible for B cell development (51). Mice deficient for IL-7 or IL-7R exhibit an arrest in bone marrow B cell development at the early stage of pre-pro–B cells (IL-7R^−/−) and in the transition of pro–B to pre–B cells (IL-7^−/−) (33, 52, 53), resembling the defect we observed in HNF1A deficiency. The reduced proliferation of early B cell progenitors in HNF1A^−/− mice suggests along with their relative irresponsiveness to IL-7 a defect in IL-7 signaling. However, the cause is not lack of IL-7 as stimulation with rIL-7 did not rescue the phenotype and bone marrow chimera retained it, respectively. A further argument is the impaired ability of HNF1A^−/− CLPs to differentiate into B cell progenitors in the presence of IL-7 in vitro. The α-chain of the IL-7R mediates the specific binding of IL-7 and the cell responses (54). We found it to be present at normal levels on the surface of HNF1A^−/− early B cells. This does not exclude functional defects, for example, in dimerization with the common γ-chain, phosphorylation, recruitment of cofactors, and initiation of subsequent signaling pathways (51). Indeed, STAT5 phosphorylation after IL-7 treatment was impaired in HNF1A^−/− B cell progenitors, revealing a defect in IL-7R signaling. The defect was restricted to the B cell lineage as HNF1A^−/− T lymphocytes responded normally to IL-7 and developed regularly in HNF1A^−/− bone marrow chimera, thus excluding a general IL-7R signaling defect that would have also affected T lymphopoiesis (52, 53). STAT5 is known to induce the expression of many factors participating in B lymphopoiesis, specifically those involved in cell proliferation and survival of pro–B cells (55). Gene expression analysis from the bone marrow showed no statistically significant difference in the expression of STAT5, JAK1, or JAK3 in isolated B220⁺IgM⁻ B cell progenitors of control mice compared with HNF1A^−/− mice (data not shown). However, the essential transcription factors E2A, EBF1, and Pax5 were decreased in HNF1A^−/− B cell progenitors. B cell lymphopoiesis depends on the sequential activity of these factors during the development of CLPs to pro–B cells (16). Especially EBF1 seems to be a central factor, as its levels are considerably augmented at later stages of development through IL-7Rα– and STAT5-mediated signals (56). In addition, EBF1 and Pax5 are decreased in IL-7R^−/− pre-pro–B cells, but can be restored by the presence of constitutive active STAT5 (20, 56). Finally, STAT5a/b double knockouts showed drastically reduced numbers of B cell progenitors and mature B cells, similar to that observed in IL-7R–deficient mice (52, 57, 58). EBF1 and E2A are also interconnected as E2A initiates EBF1 expression (59). E2A also acts in concert with FoxO1 and IL-7R to establish expression of EBF1 in B cell–biased lymphoid progenitors (60). Both E2A and PU.1 upregulate the promoter activity of the EBF gene (61, 62). Pax5 is also connected to EBF1 in a positive feedback loop (63, 64), but has been better known as a regulator of B cell commitment as Pax5-deficient pro–B cells or mature B cells acquire alternative lineage potential (65, 66). Interestingly, Bach2 may also have a role in early B cell development, because Bach2 expression is activated by Pax5 in pro–B cells (67). In addition, Bach2 is a factor responsible for repression of myeloid genes in CLPs (68). The expression of PU.1 in HNF1A^−/− B progenitors was comparable to controls in agreement with the notion that the early program of B cell–specific gene expression is largely independent of PU.1 (61). PU.1 also regulates the expression of the IL-7R in lymphoid progenitors (69), providing a reason for the unaltered IL-7R expression levels in HNF1A^−/− B cell progenitors.

Recently, mice lacking apolipoprotein M (apoM)—the expression of which is completely dependent on HNF1A (37) and serves as a binding partner of sphingosine-1-phosphate (S1P) in HDL (70)—were shown to have increased LSKs and CLPs and enhanced T and B lymphopoiesis (the latter particularly evident from the increased pre-, immature, and mature B cells in their bone marrow), and this was attributed to a lack of inhibitory effect of apoM-bound S1P on lymphopoiesis (71). Although HNF1A^−/− mice also have higher LSK and CLP numbers (this study), no apoM (37), half as much S1P in plasma, and none in HDL (72)—findings all identical to those in apoM^−/− mice—their T lymphopoiesis was unaffected and B lymphopoiesis severely suppressed. Thus, any lymphopoiesis-promoting effects caused by the lack of apoM-bound S1P in HNF1A^−/− mice are invisible on the background of the potent lymphopoiesis-blocking defect of hematopoietic HNF1A deficiency.

There is precedence for another member of the HNF family, HNF6, to play a role in B cell lymphopoiesis (14), but only in fetal B cell differentiation, and its role is confined to the liver parenchymal cells providing the niche for fetal B lymphopoiesis. In contrast, we have identified HNF1A to be a cell-intrinsic factor important for adult B lymphopoiesis. Despite the pronounced B lymphopoiesis defects in HNF1A^−/− mice, there are no reports of similar findings in humans suffering from diabetes mellitus type MODY3 due to HNF1A mutations. Possible reasons may be the incomplete loss of HNF1A function and the considerably lesser dependency of human B lymphopoiesis on IL-7 in contrast to mice (21). However, B cells have not been specifically addressed in MODY3 patients and may be worth investigating. Of note, a recent study has reported a substantial upregulation of the gene network around HNF1A in patients with B cell chronic lymphocytic leukemia (73). To address the role of HNF1A in B lymphopoiesis in humans, studies on B cell populations in patients with known HNF1A mutations should be designed, and, vice versa, HNF1A mutations in patients with B cell disorders examined.

In this study, to our knowledge, we describe for the first time that the transcription factor HNF1A is required for normal adult B cell differentiation and development in mice. HNF1A^−/− mice had dramatically reduced B cell progenitors in bone marrow and less transitional B cells in blood and spleen. This was due to a functional resistance of early B cell progenitors to IL-7 as evidenced by their irresponsiveness to IL-7 stimulation, their impaired STAT5 phosphorylation, and the compromised B cell progenitor differentiation capacity of CLPs. The transplantability of the defect into HNF1A^−/− hematopoietic chimera argues strongly for a cell-intrinsic effect. The decreased expression of the transcription factors EBF1, E2A, Pax5, and Bach2 in B cell progenitors may account for the defect in HNF1A^−/− B cell development. To our knowledge, this is the first evidence of HNF1A serving as a cell-intrinsic transcription factor necessary for normal adult B lymphopoiesis.

OP9 stromal cells were a gift of J. C. Zúñiga-Pflücker to one of the co-authors (Hannes Klump). We thank K. Abou Hamed for excellent technical help and S. Weber for most helpful expertise in cell sorting.

The authors have no financial conflicts of interest.