EBF1 Is Essential for B-Lineage Priming and Establishment of a Transcription Factor Network in Common Lymphoid Progenitors¹

Production and development of B lineage cells is a precisely controlled process originating from rare hematopoietic stem cells. Lineage commitment of stem cells to the B lineage involves a stepwise process progressing through a series of intermediate progenitor stages. Studies of knockout (KO)⁴ mice models and gain of function experiments has revealed indispensable functions for transcription factors such as Pax5, E2A, PU.1, and Ebf1 in this process, but the exact developmental stage as well as the genetic networks surrounding these factors is still under investigation. The Helix loop helix protein EBF1 (1), within the hematopoietic system expressed in lymphoid progenitors and B-lineage cells (2, 3), has been suggested to be one of the key regulators of early B cell development (4). Disruption of EBF1 function in mice leads to complete block in the formation of peripheral B-lymphoid cells as well as dramatic disturbances of several immature B-lineage progenitor populations in the bone marrow (4). The observed block in differentiation appears to be earlier than what is found in PAX5-deficient mice (5) because, in contrast to the EBF1 mutants (4), expression of multiple B-lineage-restricted genes can be found in the bone marrow of PAX5 KO mice (6). EBF1 deficiency rather resembles that observed in mice deficient of the E2A-encoded E47 and E12 proteins (7, 8). It has been shown that expression of EBF1 and E2A is crucial for initiation of B cell development and these factors have been shown to directly collaborate with in the regulation of target genes (9, 10, 11, 12). Furthermore, mice transheterozygous for mutations in the Ebf1 and E2A genes display an enhanced block of B-lineage development as compared with the single heterozygote mice (13). Several lines of investigations have revealed genetic targets directly regulated by EBF1 (12, 13, 14, 15), leading to the idea that EBF1 acts as a pleiotropic activator of B-lineage genes. Additionally, studies on the CD79a promoter suggest that EBF1 acts as a pioneer transcription factor with crucial roles in the establishment of the epigenetic program in the developing B lymphocyte (16). EBF1 has also been shown able to rescue B cell development in the absence of E2A (17), Pu.1 (18), or Il-7Ra (19, 20) and been suggested to restrict the development potential of hematopoietic progenitors (21). However, even though EBF1 has an undisputable role in B cell development, the exact function of this transcription factor remains unclear. The presence of candidate pre-pro-B cells in an EBF1-deficient mouse bone marrow suggested that the initial phases of B lineage commitment is not contingent upon EBF1 activity (4). However, the lack of expression of B cell-restricted genes and IgH recombination (4) argued against the formation of the pre-pro-B cell compartment in the absence of EBF1. Furthermore, even though a number of target genes have been identified, a complete explanation of the phenotype found in the EBF1-deficient mice is still lacking. In vitro differentiation experiments suggested that the developmental block in EBF1-deficient pre-pro-B cells is cell autonomous (18), however the identification of EBF1 proteins in other bone marrow populations (22, 23) also opens for alternative and more or less complex explanations of the phenotype observed in the absence of EBF1.

We report in this study that EBF1-deficient hematopoietic cells display a cell autonomous block of B cell priming already at the common lymphoid progenitors (CLP) stage. Reduced expression of genes associated with B-lineage development was observed in EBF1-deficient CLPs both at the population and single cell level. Global gene expression analysis also revealed a reduced expression of the transcription factors Pou2af1 (OcaB) and Foxo1 in EBF1-deficient CLPs and analysis of the promoter elements flanking these genes support the idea of them being direct targets for EBF1. Thus, we conclude that EBF1 deficiency results in a prominent block of B cell development already at the CLP stage and that EBF is directly involved in the establishment of a B-lineage-restricted transcription factor network.

Animal procedures were performed with consent from the local ethics committee at Lund University (Lund, Sweden). To obtain E14.5 fetal livers (FLs), heterozygous EBF-KO mice (on C57BL/6 background, CD45.2) were mated early evening and females checked the following morning for the presence of a vaginal plug designated as embryonic day 0.5 (E0.5). Embryos were collected from pregnant mice at E14.5 and feta liver dissected. Recipient mice (C57SJL, CD45.1) were lethally irradiated 2 h before transplantation of 2.5 × 10⁶ unfractionated FL cells and 0.25 × 10⁶ unfractionated bone marrow (BM) support cells (C57SJL, CD45.1). Transplanted mice were analyzed at least 16 wk after transplantation. BM B-cells, BM CLPs, and thymocytes from respective tissues were stained as follows: B cell staining: purified Ter119 (Ter119), Gr1 (RB6-8C5) and CD11b (M1/70) (visualized with GAR-QD605), subsequently followed by Fc-Block (CD16/CD32, 2.4G2) and CD45.1 (A20) FITC, CD43 (S7) PE, CD3 (145-2c11) PE-Cy5, CD45.2 (104) PE-Cy5.5, B220 (RA3-6B2) PE-Cy7, CD19 (1D3) allophycocyanin, CD4 (RM4-5) allophycocyanin-Alexa750, Nk1.1 (PK136) Pacific Blue, AA4.1 (AA4.1) biotin (visualized with Streptavidin QuantumDot (QD) 655), and propidium iodine (PI); CLP staining: Fc-Block (CD16/32, 2.4G2) followed by CD45.1 (A20) FITC, FLT3 (A2F10) PE, TER119 (Ter119) PECy5, CD3 (145-2c11) PECy5, Gr1 (RB6-8C5) PECy5, CD11b (M1/70) PE-Cy5, CD45.2 (104) PE-Cy5.5, B220 (RA3-6B2) PE-Cy7, CD19 (1D3) allophycocyanin, KIT (2B8) allophycocyanin-Cy7, SCA1 (D7) Pacific Blue, IL7r (A7R34) biotin (visualized with Streptavidin QD655), and PI; and Thymocyte staining: CD45.2 (104) Alexa488, CD45.1 (A20) PE, CD8a (53-6.7) PE-Cy5.5, CD4 (RM4-5) allophycocyanin-Alexa750, and PI.

Analysis and cell sorting was done on a FACSAria (BD Biosciences). For isolation of B cell progenitors and CLPs, BM cells were subjected to lineage depletion and enrichment of Kit⁺ cells respectively as previously described (53).

Cells were sorted directly into buffer-RLT and frozen at −80°C. RNA extraction and DNase treatment was performed with the RNeasy Micro Kit (Qiagen) according to the manufacturer’s instructions for samples containing ≤10⁵ cells. Eluted RNA samples were reverse transcribed using SuperScript III and random hexamers (Invitrogen) according to protocol supplied by the manufacturer. Real-time quantitative PCR (Q-PCR) reactions were performed by mixing 2 × TaqMan universal PCR master mix, 20 × TaqMan primer/probe mix, RNase-free H₂0 and 5 μl of cDNA for a final reaction volume of 20 μl. Assays-on-Demand probes used were: Pax5; Mm00435501_m1, Hprt; Mm00446968_m1, Ebf11; Mm00432948_m1, Cd79a; Mm00432423_m1, Cd79b; Mm00434143_m1, Cd19; Mm00515420_m1, Foxo1; Mm00490672_m1, Pou2af1; Mm00448326_m1, Rag1; Mm01270936_m1, Rag2; Mm00501300_m1, Sequences for the Assay-by-Design probe used for λ5 (Igll1) detection were: forward 5′GGAACAACAGGCCTAGCTATGG, reverse 5′CTCCCCGTGGGATGATCTG, probe 5′CCGGCAGCTCCTGTTC and sterile IgH: forward 5′GGACTTTGGGATGGGTTTGGTT, reverse 5′CCCTGGTCCTAGACATCAGAGTAAT, prob 5′CCCAGATGAAGGGCTAC.

All experiments were performed in triplicates and differences in cDNA input were compensated by normalizing against Hprt expression levels. Values shown in figures are arbitrary numbers calculated as ΔCt, which is basically 2^{(Ct GOI-Ct Hprt)}, where Ct GOI is the mean of Cts for gene of interest. To validate the data with regard to significant differences, values are subjected to Student’s t test.

For evaluating B cell and T cell potential, cells were respectively seeded (using a FACSAria) onto preplated OP9 (supplemented with 10 ng/ml FLT3L Amgen and 10 ng/ml IL7 PeproTech) and OP9Δ1 (supplemented with 10 ng/ml FLT3L) stroma layers for 14 days. Cultures were substituted with new cytokines every 7 days.

OP9 cocultures were evaluated by FACS staining with CD19 (1D3) PE, Gr1 (RB6-8C5) PE-Cy5, CD11b (M1/70) PE-Cy5, B220 (RA3-6B2) allophycocyanin and PI. OP9DL1 cocultures were evaluated by FACS staining with CD25 (7D4) FITC, CD19 (1D3) PE, CD90.2/Thy1.2 (53-2.1) allophycocyanin and PI.

Samples were analyzed on a BD FACSCalibur (BD Biosciences).

RNA was extracted from purified adult BM subsets as described for Q-RT-PCR. RNA was labeled and amplified according to the Affymetrix GeneChip Expression Analysis Technical Manual. Chips were scanned using a GeneChip Scanner 3000 and scaled to a median intensity of 100. RNA from cells sorted on separate occasions was separately hybridized for each investigated population. Probe level expression values were calculated using RMA and further analysis was done using dChip (http://biosun1.harvard.edu/complab/dchip/). Array data will be deposited in Geo upon acceptance of the report for publication.

Principal component analyses was performed using the online NIA Microarray analysis tools provided by national institute for aging (http://lgsun.grc.nia.nih.gov/ANOVA) (24).

Genomic DNA was prepared using TRIzol (Life Technologies) according to the manufacturer’s instructions. IgH D-J rearrangements were amplified by 38 cycles (94°C, 30 s; 60°C, 45 s; and 72°C, 1 min) using the DH and J3 primer at 1 μM. The PCR products were blotted onto Hybond N⁺ nylon membranes (Amersham Biosciences) using capillary blotting. Membranes were prehybridized in 5 × Denhardts (Ficoll 10g/L, polyvinylpyrridol 10g/L, BSA 10g/L), 6 × saline sodium phosphate EDTA, 0.1% SDS, and 100 μg/ml salmon sperm DNA, at 45°C for 60 min and hybridized with a specific γ³²[P]-labeled oligonucleotide for 12 h at 45°C in the same solution. Membranes were washed at room temperature in 2 × SSC (0.015M Na citrate (pH 7.0) containing 0.15 M NaCl) supplemented with 0.1% SDS for 15 min and 0.1 × SSC with 0.1% SDS for 5–10 min. The hybridized membrane was subjected to autoradiography.

The following oligonucleotides were used for PCR: DH: 5′-GGAATTCG(A/C)TTTTTGT(C/G)AAGGGATCTACTACTGTG-3′, J3: 5′-GTCTAGATTCTCACAAGAGTCCGATAGACCCTGG-3′. The following oligonucleotide was used for hybridization: JH3: 5′-AGACAGTGAC CAGAGTCCCTTGG-3′.

The OPN promoter or a part of the HPRT gene was amplified as control for that the samples contained amplifiable material.

Multiplex single-cell RT-PCR analysis was performed as previously described (25, 26). In brief, single cells were deposited into 96-well PCR plates containing 4 μl lysis buffer and frozen at −80°C. Cell lysates were reverse-transcribed using gene-specific reverse primers and 50 U MMLV-RT (Invitrogen) in a final volume of 10 μl. First-round PCR was performed by the addition of 40 μl PCR mix containing 1.25 U Taq polymerase (TaKaRa Bio) and gene-specific forward primers to the RT reaction and the mix subsequently subjected to 35 cycles of PCR. One-microliter aliquots of first-round PCRs were further amplified for 35 cycles using nested gene-specific primers. Second-round PCR products were subjected to gel electrophoresis and visualized by ethidium bromide staining. Primers sequences as follows: Hprt; 1) 5′GGGGGCTATAAGTTCTTTGC; 2) 5′GTTCTTTGCTGACCTGCTGG; 3) 5′TGGGGCTGTACTGCTTAACC; 4) 5′TCCAACACTTCGAGAGGTCC. Rag1; 1) 5′CCAAGCTGCAGACATTCTAGCACTC; 2) 5′CAGACATTCTAGCACTCTGG; 3) 5′GCTTGACTTCCCATCAGCATGGA; 4) 5′CAACATCTGCCTTCACGTCGATCC. Pax5; 1) 5′CTACAGGCTCCGTGACGCAG; 2) 5′ATGGCCACTCACTTCCGGGC, 3) 5′GTCATCCAGGCCTCCAGCCA; 4) 5′TCTCGGCCTGTGACAATAGG. IgII1 (λ5); 1) 5′AGTTCTCCTCCTGCTGCTGC; 2) 5′GGGTCTAGTGGATGGTGTCC; 3) 5′CAAAACTGGGGCTTAGATGG; 4) 5′CCCACCACCAAAGACATACC. Ebf1; 1) 5′CCCTCTTATCTGGAACATGC; 2) 5′CTACTCCCTGTATCAAAGCC; 3) 5′TGTACGACAGTGTGACTTCC; 4) 5′TAAGGATCACTTCCTTTGGC. Cd79a (Mb-1); 1) 5′CCTCCTCTTCTTGTCATACG; 2) 5′AAACAATGGCAGGAACCC; 3) 5′TGATGATGCGGTTCTTGG; 4) 5′GAACAGTCATCAAGGTTCAGG. Foxo1: 1) 5′TCATCACCAAGGCCATCGA; 2) 5′TGTCGCAGATCTACGAGTGGAT; 3) 5′GTTCCTTCATTCTGCACTCGAA; 4) 5′GTGATTTTCCGCTCTTGCCTC; Pou2af1: 1) 5′ACGCCCAGTCACATTAAAGAAG 2) 5′AGCCAGTGAAGGAGCTACTGAG 3) 5′ATGTTCCCTCCTCTGTCACTGT 4) 5′TTAGGAAGAAAGGAAGCCTGTG; Sterile IgH: 1) 5′GCTAGCCTGAAAGATTACC; 2) 5′TGAGTTTCTGAGGCTTGG 3) 5′AGTAGCACAGTCTCTGTTCTGC; 4) 5′AAGGACTGACTCTCTGAGG Primers 1 plus 4 and 2 plus 3 are outer and inner primer pairs respectively.

BaF/3 cells and transduced Baf/3 lines (12) as well as HeLa cells were grown at 37°C and 5% CO₂ in RPMI 1640 supplemented with 10% FCS, 10 mM HEPES, 2 mM pyruvate, 50 μM 2-ME, and 50 μg/ml gentamicin (all purchased from Life Technologies). The medium for the BaF/3 cells was supplemented with 10% of conditioned medium from confluent WEHI3 cells as a source of IL-3. ΟΠ9/OP9Δ1 cells were grown in OPTIMEM supplemented with 10% (v/v) FCS, 50 μM 2-ME, and 50 μg/ml gentamicin (Invitrogen) at 37°C and 5% CO₂.

Nuclear extracts were prepared according to Schreiber et al. (27) DNA probes were labeled with γ [³²P] ATP by incubation with T4 polynucleotide kinase (Roche), annealed with the complementary strand and purified on a microspin column (Roche). Five to ten micrograms of nuclear extract was incubated with the labeled probe (20,000 cpm, 3 fmol) for 30 min at room temperature in binding buffer (10 mM HEPES (pH 7.9); 70 mM KCl; 1 mM DTT; 1 mM EDTA; 2,5 mM MgCl₂; 5% glycerol, 1 mM ZnCl₂) with 0.75 μg Poly dI/dC (Amersham Biosciences). DNA competitors were added 10 min before addition of the labeled probe. The samples were separated on a 6% acrylamide TBE gel, which was dried and subjected to autoradiography. Competitors based on synthetic oligo-nucleotides were added at molar excesses indicated in the respective figure legends. Super shifts were performed under the same conditions but with additional presence of 2 μl polyclonal anti-EBF (SC-15333) or anti-actin (SC-1616) (Santa Cruz Biotechnology) as a negative control.

Oligo-nucleotides used for EMSAs were: Cd79a sense: AGCCACCTCTCAGGGGAATTGTGG; Cd79a antisense: CCACAATTCCCCTGAGAGGTGGCT; mutated Cd79a sense: AGCCACCTCTCAGCCGTTTTGTGG; mutated Cd79a antisense: CCACAAAACGGCTGAGAGGTGGCT; Pou2af1 EBF sense: ATAGTGACCCCTGGGAATCGCAT; Pou2af1 antisense: ATGCGATTCCCAGGGGTCACTAT; Foxo1 EBF1 sense: TGCCGCCTCCCGAGGGAAGGCAGC; Foxo1 EBF1 antisense: GCTGCCTTCCCTCGGGAGGCGGCA; Foxo1 EBF2 sense: CACCGACTTCCCGAGAGCTGCGCGA; Foxo1 EBF1 antisense: TCGCGCAGCTCTCGGGAAGTCGGTG; Rag-E sense: ACAATGCTAAGCCCTCGGGTTCT GTCTAA; Rag-E antisense: TTAGACAGAACCCGAGGGCTTAGC ATTGA.

The EBF-1 expression plasmid has been described before (11). The reporter plasmids were based on the luciferase reporter plasmid pGl-3 (Promega). The FoxO1 and Pou2af1 promoter reporter plasmids were generated by blunt end cloning of PCR amplified promoter elements into the SmaI site of pGl-3. All constructs were verified by sequencing. The PCR products were generated using the following primers: Pou2af1 −560sense: TCAGACGCGTGGAGCCGTTACATTTTAAAGTGG; Pou2af1 antisense: TCGACTCGAGGAACCACCGTTTACTCTGAGG; Foxo1 −499 sense: TCGAACGCGTCCCACCGACTTCCCGAGAGCTGC; Foxo1 antisense: AGCTCTCGAGGACTGAAGCGACTCCGCAGACTCC.

HeLa cells were seeded into a 24-well tissue culture plate so that they reached 60–80% confluence for transfection. A total of 400 ng of DNA was used to transfect the cells in each well. In brief, 50 ng of each reporter construct was cotransfected with 600 ng expression vector and all transfections included 50 ng pRL-0 Renilla luciferase, received from Dr. Björn Olde (Lund University, Sweden), as internal control, and used for normalization of the luciferase activities. The DNA was first added to 50 μl of serum free medium OPTIMEM (Invitrogen), mixed with 7 μl of PLUS Reagent (Invitrogen), and incubated for 15 min at room temperature. In brief, 0.75 μl of Lipofectamin reagent (Invitrogen) was diluted in 25 μl of serum free medium, incubated for 30 min in room temperature, mixed with the diluted DNA, and then incubated for another 15 min at room temperature. Meanwhile, the complete RPMI medium was replaced with 300 μl of serum-free medium, and the DNA plus lipofectamin mixture was added to the cells that then were incubated at 37°C in 5% CO₂ for 3 h, after which each well was supplemented with 1 ml complete RPMI. Cells were harvested 48 h after transfection and protein extracts were prepared directly in the 24-well plates by adding 100 μl cell lysis buffer.

All statistical analyses have been done either in Microsoft Excel and GraphPad Prism or appropriate software for microarray analysis (dChip and NIA microarray analyses).

Expression of EBF1 is within the hematopoietic system closely related to B lineage development (3). However, when comparing the changes in global gene expression patterns between multipotent lympho myeloid primed progenitors (LMPPs) (28) and the more developmentally restricted CLPs (29) by microarray analysis, we noted a striking difference in the expression levels of Ebf1 (53). These experiments suggested that the multipotent CLP expressed ∼20% of the mRNA level observed in Cd19⁺B220⁺CD43⁺ pro-B cells, opening the possibility that EBF1 may be involved in the development of CLPs. Due to a high early lethality of EBF1 deficient mice, and due to the reported function and expression of EBF1 proteins in bone marrow stroma cells (22), we investigated the ability of EBF1-deficient day 14.5 FL hematopoietic progenitor cells to generate hematopoietic cells after transplantation (Fig. 1, A–D). In mice analyzed 16 wk or more posttransplantation, the overall reconstitution of CD45.2⁺ cells out of the total CD45⁺ BM cells in mice receiving wild-type (WT) FL cells was >95% while in those transplanted with the EBF-deficient FL cells, it was between 75 and 95% (Fig. 1,A). Among the CD45.2⁺ cells, the fraction of early progenitors including the Lin⁻ScaI+Kit+ (LSK) and the CLPs (Fig. 1,B) was comparable as was the distribution of the major T cell subpopulations in the thymus (Fig. 1,C), suggesting that EBF is not crucial for early lymphoid development or thymopoiesis. Investigations of early bone marrow B cell compartments, however, revealed a complete lack of pre- or pro-B cells derived from EBF1 deficient cells. In contrast, we detected slightly increased numbers of B220⁺CD43⁺AA4.1⁺ cells, reported to represent the pre-pro-B cell compartment (Fig. 1,D), (30, 31) supporting the previously raised idea that deficiency in EBF1 results in a lineage restricted early block in B cell differentiation. To verify the apparently normal ability of EBF1-deficient fetal liver cells to generate CLPs with lymphoid potential, we cultured CLPs (lineage negative, Il7R⁺FLT3⁺SCA1^lowKIT^low, sorted from the FL transplanted mice) on either OP9 or OP9Δ1 stroma cells (see Materials and Methods). This revealed that T cell development of EBF1-deficient CLPs progressed normally in vitro while development of CD19⁺ cells was completely abrogated (Fig. 1 E). These data suggest that EBF1-deficient hematopoietic progenitor cells have a B-lineage-specific cell autonomous defect and suggest that a CLP compartment forms independently of EBF1.

The apparent block of B cell development in fraction A, according to the Hardy nomenclature (30, 31) would suggest that EBF1 is involved in regulating critical genes in the pre-pro-B cell compartment defined by expression of B220, CD43, and AA4.1 on the cell surface. To identify potential target genes of EBF1, we performed microarray gene expression analyses on sorted CD45.2⁺ CLPs and B220⁺CD43⁺AA4.1⁺ cells from mice reconstituted with either WT or EBF1-KO FL cells (Fig. 2,A). To place the gene expression patterns observed into a context, expression data from LMPPs as well as CD19⁺CD43^lowAA4.1⁺ pro B cells, B220⁺CD43⁻CD19⁺IgM⁻ pre-B cells and IgM⁺B220⁺ spleen B cells sorted from WT mice were included in the analysis. These experiments suggested that WT and EBF1 KO B220⁺CD43⁺AA4.1⁺ cells, generated in vivo in mice reconstituted with FL cells, expressed very low levels of B-lineage-associated transcripts, indicating that in this setting and from the gene expression point of view, this population of cells cannot be linked to progressive B lineage development. This was also supported by a global principal component analysis because CLPs displayed a higher overall similarity to pro-B cells than the B220⁺CD43⁺AA4.1⁺ population (data not shown). The principal component analysis also supported the idea that the EBF1-deficient cells displaying a CLP surface phenotype indeed represent a CLP population (data not shown). The microarray data suggested that there were no significant alterations in the expression of TCF3, Id, or Ikaros proteins in WT or EBF-deficient CLPs (Data not shown), however, we could detect reductions of gene expression of B-lineage-associated genes including Pou2af (OcaB), Pax5, Vpreb1, and Blnk, in the cells generated from EBF1-deficient fetal livers. To verify a reduced expression of B-lineage-associated genes in the EBF1-deficient CLPs from transplanted mice, we performed Q-PCR analysis that supported reductions in Cd79a, Pou2af, Pax5, and Igll1 (λ5) expression (Fig. 2,B). We could also detect reductions in the levels of Cd79b and Foxo1 message, suggesting that normal expression of these genes were dependent of functional EBF1. To investigate these alterations in gene expression patterns at a single cell level, WT and EBF1-KO CLPs (sorted from FL reconstituted mice), were analyzed for gene expression patterns with multiplex single cell PCR (Fig. 2 C). This revealed that while a small subpopulation of the WT CLPs coordinately expressed Mb-1, Pax-5, and Pou2af (p < 0.03), these cells are absent in the CLPs from EBF1-deficient mice (Fishers exact test, p < 0.03). We were also able to detect transcripts from the WT as well as the mutated Ebf1 gene in a large number of the cells. These findings support the idea that in the absence of EBF1, B cell development is blocked already at the level of the CLP.

It has been reported that a fraction of the CLPs in the mouse BM has undertaken the first steps of the IgH recombination process by joining D segments to J segments (31). To investigate whether EBF1 has a direct role in recombination of the IgH genes, we extracted genomic DNA from WT as well as EBF1-deficient CLPs obtained by transplantation as above, and investigated the presence of IgH D-J recombination in these cells (Fig. 3,A). Recombination events were easily detected in the WT CLPs,while they appeared to be dramatically reduced in CLPs generated from EBF1 deficient cells. To approximate the differences in recombination efficiency, we performed a semiquantitative PCR (Fig. 3,A). This suggested that the difference was in the order of 2–4 cycles, or 12.5–25% of WT values. To further investigate whether D-J recombination occurs in the absence of EBF1, we sorted CD3⁺ cells generated by transplantation of WT or EBF1 KO fetal liver cells as above and investigated IgH recombination in this compartment (Fig. 3,B). This indicated that IgH D-J recombination could be detected supporting the idea that IgH D-J recombination occur, at least in the developing T cell compartment, in the absence of EBF1. Thus, even though DJ recombination may occur in the absence of EBF1, this event is largely reduced even at the level of the CLP. The global gene expression analysis suggested that EBF1-deficient CLPs expressed a lower level of Rag1 essential for IgH recombination. To verify this finding and expand the analysis, we investigated the expression levels of Rag1, Rag2, and sterile IgH transcripts in EBF1 deficient CLPs (Fig. 3,C). We found that the Rag1 expression was only ∼25% of that in the WT cells while the expression of Rag2 and sterile IgH transcripts were only marginally reduced. Because the single cell PCR analysis suggested that a comparable number of CLPs express Rag1 either in the absence or presence of EBF1 (Fig. 2,C), this differential level is likely to be reflected in a reduction of transcription rather than a loss of a given population. This could indicate that Rag1 would be a direct target for EBF1, but we were unable to identify any EBF binding sites in the published Rag1 promoter. To investigate whether an alternative promoter was used in the CLP, we performed 5′ RACE analysis using CLP mRNA. This did, however, result in identification of the identical promoter region arguing against this explanation. Another possibility would be a direct involvement of EBF in the regulation of the E-Rag enhancer (32). Targeted deletion of this element has a rather dramatic effect on the transcription of Rag1 while the effect on Rag2 transcription is comparably small (32). Investigation of the E-Rag enhancer sequence revealed a potential EBF binding site, and using EMSA we were able to show that this site had the ability to compete for binding of an CD79a promoter EBF site to EBF1 in nuclear extracts from a mouse pre-B cell line (Fig. 3,D). Furthermore, upon incubation with nuclear extracts, this site was able to form a complex that could be supershifted with anti-EBF Ab (Fig. 3 D). Therefore, we conclude that EBF1 is of large importance for IgH D-J recombination in the CLP and that lack of EBF1 is associated with a reduced expression of Rag1, possibly as a direct result of EBF activity at the E-rag enhancer or via reduced levels of Fox proteins acting at the same element (see below).

Knowing that the expression of Pou2af and Foxo1 is reduced in EBF-deficient CLPs, we wanted to examine a potential direct role of EBF1 in the regulation of these genes. Because expression of the EBF gene is not perfectly correlated to the expression of its target genes (Fig. 2,C, 53), we sorted CLPs from mice carrying an Igll1 promoter-controlled human CD25 gene (33). The Igll1 promoter is a direct target for EBF and these mice can therefore be used to identify early cells that have initiated the B lineage program (53). Q-PCR analysis suggested that the increased expression of Igll1 was correlated to that of Pax5, Pou2af1, Foxo1, and Ebf1supporting the idea that these genes are coordinately regulated in early B cell development (Fig. 4,A). In addition, we were able to detect low but significant amounts of both Pou2af1 and FoxO1 transcripts in Ba/F3 progenitor cells after stable ectopic expression of the combination of EBF and E2A (12) (Fig. 4,B). These data suggest that EBF1 might be directly involved in the regulation of both the Pou2af1 and Foxo1 genes. RACE analysis supported the idea that the published promoter sequences (Ref. 34, 35 and NCBI-Gene) were valid also in immature cells (data not shown) and inspection of the 5′ flanking regions revealed two potential EBF1 sites in the Foxo1 and one potential site in the Pou2af1 promoters (Fig. 4,C). To investigate whether these sites were able to interact with EBF1, we performed EMSA experiments using nuclear extracts from a mouse pre-B cell line and competed for complex formation between EBF1 and the high affinity binding site from the Cd79a promoter (Fig. 4,C). This revealed that both the Foxo1 and Pou2af1 promoters contained high affinity binding sites able to compete for EBF1 binding. To verify the ability of these binding sites to interact with EBF1 in a nuclear extract from a pre-B cell line, we performed a second round of EMSAs using labeled Foxo1 or Pou2af1 EBF binding sites. Both gave rise to one major complex that was supershifted by the addition of anti-EBF Ab (Fig. 4 C). To investigate the ability of EBF1 to directly activate the Foxo1 and Pou2af1 promoters, we cloned these regulatory elements in front of a luciferase reporter gene and transfected them either with empty or EBF1 encoding expression vector into epithelial HeLa cells lacking endogenous expression of EBF1. This revealed that both promoter elements responded to the ectopic expression of EBF1, supporting the idea that the Foxo1 and Pou2af1 genes are direct targets for EBF1.

In this article, we present data suggesting that even though EBF1 expression is dramatically increased in the transition from the multipotent LMPP stage to the lymphoid-restricted CLP stage, the transcription factor is not critical for the development of cells with a general CLP phenotype. The presence of surface proteins and general gene expression patterns as well as preserved potential for development into T-lineage cells in vitro all support this conclusion. This appears to be in contrast to what has been reported for TCF3 (TCFe2a) because mice lacking this factor display a rather dramatic reduction in the number of CLPs (36). This would indicate that even though EBF1 and TCF3 act in synergy in the earliest stages of B cell development (9, 13), TCF3 might have an EBF1-independent function in the generation of CLPs. This would be in line with the finding that while TCF3 target genes can be detected in LMPPs, these cells lack transcription of EBF target genes (28). However, even though CLP like cells were generated from EBF1-deficient fetal liver cells, they lacked the transcription of B-lineage associated genes normally found in the CLP compartment. Even though we were able to detect cells with surface expression of AA4.1, B220, and CD43, indicative of pre-pro-B cells, we were unable to obtain evidence that these cells would represent a homogenous population of B-lineage cells and we conclude that EBF1-deficient progenitors are severely compromised in their ability to take on B lineage fate. A role for EBF1 in lineage commitment is also supported by the finding that the factor restricts the lineage choices of PAX5 deficient progenitor cells (37). The finding that EBF1-deficient CLPs lack expression of B-lineage genes found in normal CLPs supports the notion that EBF1 is a key coordinator of the transcriptional program that drives B-lymphocyte development already in the CLP. This activity could be explained by the apparent ability of EBF1 to pioneer transcription by mediation of epigenetic changes (18) but it could also be contributed to EBF1s ability to directly activate a set of B-lineage associated transcription factors (Fig. 5). EBF1 has previously been reported to be involved in the activation of the Pax5 gene (13) coding for a key transcription regulator in early B cell development. PAX5 has in turn been suggested to regulate a large set of genes including the expression of the transcription factor Lef1 (38) and Ebf1 itself (3). PAX5 has also been suggested to positively regulate expression of Cd19, Cd79a, and Rag2 (38, 39, 40) and to repress expression of Notch1 (41). In addition, we report in this study that EBF1 is directly involved in the regulation of Pou2af1 transcription. This protein was originally identified as a regulator of OCT protein activity in B-lymphoid cells and even though the effect of targeted mutation of the Pou2af1 gene is most pronounced in late stages of B cell differentiation (42, 43, 44), reduced levels of POU2AF1 have been reported to result disturbances also in early B cell development (45). In addition, POU2AF1 has been reported to regulate the expression of the Ets transcription factor SpiB (46) as well as of other genes encoding proteins involved in signaling and cell cycle regulation (47). EBF1 also appear to be directly involved in the regulation of the Foxo1 gene that might act in a large network of redundant FOXO family transcription factors (48). Among other FOX proteins, Foxp1 has been shown to be important for early B cell development (49), to some degree possibly via its ability to modulate Rag1 transcription though a direct interaction with the E-Rag enhancer. Several of the FOX family members also share binding sites, and it is possible that there exist a large redundancy among these proteins similar to what has been reported for basic-helix-loop-helix transcription factors (50, 51). Thus, it is possible that the link between EBF1 and the FOX-factor network is mediated via FOXO1. Using available information about target genes for this network of transcription factors allows for the creation of a scheme where several genes known to define the B lineage can be incorporated (Fig. 5).

It is also notable that with the possible exception of Ig H chain DJ recombination, the activation of EBF1 target genes is not directly linked to the transcription of the Ebf1 gene itself. A majority of the CLPs express Ebf1 message, while only a small percentage of the cells express known target genes. We have investigated this in more detail using mice transgenic for a Igll1 promoter-controlled reporter gene, and it appears as if the activation of EBF1 target genes is linked to a 4-fold increase in Ebf1 message in a B cell committed subpopulation of the CLPs (53). We believe these data indicate that transcription of the Ebf1 gene is a precommitment process, crucial for but nor directly coupled to activation of the B-lineage program. This could be a result of a direct dose-dependent function, translational inhibition by inhibitory RNA or posttranslational modifications, for instance due to active Notch signaling (52). Regardless, because we find expression of Ebf1 even in the absence of the expression of other B-lineage genes, this provides strong support to the idea of a defined precommitment stage of B cell development represented by the expression of Ebf1. Because Ebf1 expression is low in LMPPs (53), a cell type that contains Tdt, sterile Ig, and Rag1 transcripts (28), this could indicate that the pathway to B cell commitment is achieved at least at three different levels. In the LMPP, lymphoid priming is established, represented by expression of genes common for B and T lymphoid cells. Upon transition into the CLP compartment, Ebf1 is up-regulated, making the cell competent to develop into B-lineage, but as the full function of EBF1 would be regulated at a posttranscriptional level, cells still remain uncommitted even though they undergo Ig DJ recombination, facilitated by a low EBF1 activity incapable of activating B cell specific target genes. The final step is then associated with the functional activation of EBF1 resulting in the transcription of a set of B lineage specific genes, including Pax5, Foxo1, and Pou2af1.

Thus, it may be so that there is no single crucial target gene for EBF1 but rather that EBF1 is the key regulator of the B-lineage transcriptional program and commitment through its ability to activate B-lineage genes as well as a set of transcription factors capable to strengthen the B cell identity, possibly already at the level of CLP.

We thank R. Grosschedl for the gift of EBF deficient mice, R. A. Cumano and J. C. Zuniga-Pflücker for providing the OP9 and OP9Δ1 stromal cell lines, and expert advice regarding the usage of these, Ragnar Mattson for help with investigations of EBF1 deficient embryos. The expert advice and assistance of Liselotte Lenner and Gerd Sten, and the SweGene affymetrix facility in assistance with the micro arrays is highly appreciated.

The authors have no financial conflict of interest.