Early B cell Factor: regulator of B lineage specification and commitment

Abstract

B lymphocytes are generated from hematopoietic stem cells in a series of steps controlled by transcription factors. One of the most important regulators of this process is Early B cell Factor (EBF). Multiple lines of evidence indicate that expression of EBF is a principle determinant of the B cell fate. In the absence of EBF, progenitor cells fail to express classical markers of B cells, including immunoglobulins. EBF drives B cell differentiation by activating the Pax5 gene and other genes required for the pre-B and B cell receptors. New evidence suggests that expression of EBF in common lymphoid progenitors directs B cell fate decisions. Specification and commitment of cells to the B cell lineage are further established by Pax5, which increases expression of EBF. Recently, it was demonstrated that both EBF and Pax5 contribute to the commitment of cells to the B lineage. Together, these studies confirm that EBF is a keystone in a regulatory network that coordinates B cell lineage specification and commitment.

1. Introduction

The development of leukocytes from hematopoietic stem cells (HSCs) is characterized by the expression of distinct sets of genes at discrete stages of differentiation (Fig. 1). To generate early lymphoid progenitors (ELPs), HSCs with long- or short-term repopulating activities (LTRC and STRC) differentiate through intermediate stages that possess progressively restricted developmental potential. ELPs seed the thymus and may generate early T cell progenitors (ETPs; reviewed in [1]). ELPs are also the precursors of common lymphoid progenitors (CLPs), which express interleukin-7 receptors (IL-7R) and engender B and T lymphocytes, natural killer (NK) cells and dendritic cells, but lack the ability to produce other hematopoietic lineages (i.e. myeloid cells). Although this process is incompletely understood, it is proposed that the maturation of progenitor cells is a tightly controlled process governed by a select set of transcriptional regulators. These regulators activate successive developmental programs while progressively limiting potential cell fates.

Figure 1.

Schematic of B cell development. Designations of the stages of B lymphopoiesis are indicated above each cell [3, 80]. Stages at which various V(D)J rearrangements of Ig heavy (H) or light (L) chain genes occur are indicated within the cells. Select cell surface markers used to discriminate various stages of development are shown below. B lineage cells are B220⁺ at all stages of development (fr.A through plasma cells). CD117 is also known as c-kit. CD43 is expressed on progenitors (HSCs to CLPs) prior to B cell differentiation. Stage specific expression of pre-BCR, B cell receptor (IgM or IgM/IgD) and secreted immunoglobulins are indicated. Stages at which developmental arrest occurs due to loss of transcription factors are indicated.

Differentiation of progenitors to B cells is heralded by the expression of cell surface markers including B220, CD43 and IL-7Rα (encoded by the Il7ra gene). Expression of these genes precedes the commencement of V(D)J recombination which results in immunoglobulin (Ig) gene rearrangements. In addition to Ig genes, the expression of accessory proteins is required for display of the pre-B and B cell receptors (pre-BCR and BCR) on the plasma membranes of pre- or immature B cells, respectively. Competent BCR complexes mediate the selection of functional B cells. These cells migrate from the bone marrow to peripheral lymphoid organs including the spleen, lymph nodes and gut-associated lymphoid tissues. At these secondary sites, stimulation of B cells by antigens results in antibody production by plasma cells. Activated B cells also generate memory cells that facilitate rapid immune responses to repeated challenges by the same antigens.

The earliest definable stages of B cell development are characterized by expression of transcriptional regulatory proteins, which initiate the B cell-specific program, or ‘transcriptome,’ via targeted gene activation and repression [2, 3]. Recent research has provided new insights as to how these proteins (including PU.1, Ikaros, EBF, E2A, and Pax5) function within an interactive network of regulators. Here, we focus on the roles of EBF, which has recently gained additional significance as a driver of both B lineage determination and commitment.

2. EBF structure, DNA binding and functions

Given its proposed role as a key determinant of the B cell fate, the biochemistry of EBF is of considerable interest. EBF and closely related proteins (EBF2, EBF3, EBF4, Collier/Knot and Unc-3) constitute a novel transcription factor family (here, termed the EBF family; also referred to as the O/E or COE family). All members of the EBF family possess a highly conserved DNA-binding domain (DBD) that is distinct from that of other known DNA-binding proteins (Fig. 2). The DBD of murine EBF comprises residues 35–251 [4, 5]. EBF binds to sequences of promoters that loosely fit the consensus 5′-CCCNNGGG-3′; however, the DBD can recognize 14 basepairs centered over this sequence [6]. Although the three dimensional structure of the EBF DBD has not been determined, a notable feature of this domain is an atypical zinc-binding motif: HEIMCSRCCDKKSC (bold residues coordinate zinc; Fig. 2B) [5]. Recent studies demonstrated the importance of this motif, termed the ‘zinc knuckle’, for binding to divergent, target-specific nucleotide sequences [7]. These studies also revealed the complexity of amino acid requirements for DNA recognition by EBF. Enforced expression of EBF (together with Pax5) in terminally differentiated 558LμM plasmacytoma cells restored expression of endogenous Cd79a (mb-1) and Vpreb1 genes. Cd79a (mb-1/Igα) and Vpreb1 encode components of the pre-BCR. Igα is also a component of the BCR. Activation of each of these genes required residues involved in zinc coordination; however, other residues within and flanking the zinc knuckle were differentially required for activating mb-1 vs. Vpreb1 transcription. These results suggest that, similar to the flexible recognition of DNA by Pax5 [8], EBF utilizes different sets of residues to bind a degenerate set of DNA sequences.

Figure 2.

The structure of EBF. A. Schematic of murine EBF. Functional domains and their positions are indicated by shading and by numbering of boundary residues. The relative position of the zinc-knuckle (Zn), which is detailed in B, is shown above the DBD. The HLH domain contains three putative α-helices indicated by boxes with graded shading. The C-terminal activation domain is not required for activation of all EBF target genes. B. The zinc-knuckle of the DBD includes four residues (bold) that coordinate zinc ions.

EBF persists as homodimers in solution and when bound to DNA. Homodimerization is a function of the helix-loop-helix (HLH) domains of EBF similar to those of the basic-HLH proteins c-Myc and MyoD1 [4, 9]. Dimerization is also facilitated by sequences within the DBD that are as yet unidentified. A dimerization function was also proposed for the ‘transcription factor-Ig-like’ (TIG) or immunoglobulin-plexin transcription factor (IPT) domain. The TIG/IPT domain links the DBD and HLH domains of EBF [10]; however, functions of this domain have not been determined. EBF also includes a C-terminal activation domain [5]. This domain is not required for transcription from the Cd79a or Vpreb1 promoters [7].

Recent studies determined that EBF functions as an epigenetic pioneer factor, which is defined as a DNA-binding protein that can direct modifications of chromatin that are permissive for transcriptional activation [11, 12]. Hypermethylation of Cd79a promoters was detected in precursor cells ex vivo, but these promoters were completely unmethylated in committed B cells [13]. Interestingly, promoters in B cell progenitors lacking EBF failed to become unmethylated. The ability of EBF to initiate DNA demethylation and enhance chromatin accessibility of Cd79a promoters was demonstrated in 558LμM plasmacytoma cells. In this context, the opening of chromatin by EBF was necessary for its cooperation with Pax5. In turn, Pax5 recruited Ets family transcription factors necessary for efficient Cd79a transcription. Thus, EBF is necessary for both the expression and function of Pax5. In more recent experiments, we confirmed that Cd79a promoter chromatin is progressively ‘opened’ in response to EBF (and Pax5) in vivo (Hua Gao, J.H., unpublished data). The pioneer effects of EBF are likely to be important for activating many genes of the B cell-specific program including the rearrangement and expression of Ig genes.

3. Temporal regulation and control of EBF expression

Because EBF is a critical determinant of B cell fate decisions, it is important to determine when EBF is first expressed during B lymphopoiesis. Using RAG1/GFP ‘knock-in’ mice, Igarashi and colleagues [14] detected Ebf1 transcripts at high levels in Lin⁻CD27⁺c-kit^LoSca-1^LoGFP⁺ bone marrow cells representing CLPs and subsequent B cell progenitors. These cells exhibited B lymphoid and NK cell potential, but had greatly reduced myeloid potential. Detection of Ebf1 transcripts was coincident with IL-7ra transcripts. Earlier stage Lin⁻CD27⁺c-kit^HiSca-1^Hi GFP⁺ cells expressed barely detectable levels of Ebf1 transcripts and gave rise to T lineage cells in addition to other cell types. These populations suggest the presence of different (or divergent) pathways to B (and NK) cells vs. T cells, with high levels of EBF promoting the B cell pathway.

Other studies have demonstrated the linkage between expression of EBF in CLPs and IL-7 signaling. CLPs from IL-7-deficient mice expressed lower amounts of Ebf1 transcripts together with greatly reduced B cell potential [15]. Restoration of EBF expression in these mice restored B cell potential in colony forming assays. Similar results were obtained using IL-7- or IL-7Rα-deficient mice in which Ebf1 transcript levels were similarly decreased in the absence of IL-7 signaling [16]. Pre-pro-B cells of IL-7-deficient mice expressed only low levels of Ebf1 transcripts. These transcripts were upregulated following the addition of IL-7 in vitro. Moreover, effects of IL-7 signaling could be mimicked by enforced expression of active STAT5 (the nuclear mediator of IL-7 signaling). Enforced expression of EBF in pre-pro-B cells of IL-7Rα-deficient mice partially rescued B cell development in the absence of IL-7 signaling. Together, these results suggest that EBF is first expressed at significant levels in early progenitors (CLPs) in response to IL-7 signaling. This hypothesis was further supported by microarray and real time PCR studies that detected Ebf1 transcripts at low levels in CLPs and at increasingly higher levels in progressively later stages of B cell development. The induction of Ebf1 transcripts was coincident with the loss of other developmental potentials [17]. Microarray studies of purified human bone marrow cell populations suggested similar patterns of EBF expression [18]. The expression of EBF in CLPs suggests that it begins the process of B lineage specification in these cells. This model is supported by recent observations suggesting that EBF drives ‘lineage priming’ in CLPs, which results in expression of B cell markers (e.g. CD79b) [17, 19, 20]. Interestingly, EBF-deficient mice (first described in [21]) retained the ability to generate CLPs and cells with pre-pro-B cell characteristics (B220⁺CD43⁺AA4.1⁺), but lacked expression of B lineage markers [19]. Thus, although EBF is detected in normal CLPs, it is not required for their development.

While the Ebf1 gene is transcribed in multiple tissues including progenitor and mature B cells, adipocytes, forebrain neurons, plasmacytoid dendritic cells, bone marrow stromal cells and developing limb buds, the pattern of expression of Ebf1 is tightly regulated [22-27]. Roessler and colleagues [28] identified two functional promoters within the Ebf1 gene, termed Ebf1α (−5113 to −4027) and Ebf1β (−1671 to +1), that function in B cells. These promoters produce distinct transcripts encoding two isoforms of EBF, EBF1α and EBF1β, which differ in their first 14 amino acids. Transactivation assays suggested that the two EBF isoforms have similar abilities to activate transcription; however, the two Ebf1 promoters are differentially regulated. Transcription from the Ebf1β promoter is dominant at each stage of normal B cell development. Experiments suggest that the Ebf1 gene’s responsiveness to cytokines is a property of the Ebf1α promoter, but not of the Ebf1β promoter. In Ba/F3 cells, enforced expression of constitutively active STAT5 or stimulation with IL-3 (which also functions via STAT5) increased expression of the endogenous Ebf1 gene by 9-fold and 66-fold, respectively. However, evidence was not obtained for direct binding of STAT5 to either Ebf1 promoter. This suggests that STAT5 may regulate Ebf1 gene transcription indirectly by enhancing the expression of another activator. The Ebf1α promoter was also activated by E47. E47 synergistically activated this promoter together with EBF (EBF1β) as part of a potential autoregulatory loop. These data are supported by previous studies linking E2A (E47) and EBF with the control of Ebf1 transcription [29]. More recently, Kwon and colleagues demonstrated that Ebf1 transcription requires E2A for both its initiation and maintenance [30].

Other data suggest that Ebf1β promoter transcription is regulated by a positive feedback loop involving PU.1, Pax5 and Ets proteins [28, 31]. Medina and colleagues demonstrated that the absence of PU.1 results in greatly reduced numbers of Flk2⁺/Flt3⁺IL-7R⁺ lymphoid progenitors and impaired expression of EBF [31]. Restoration of PU.1 expression induced expression of EBF. Furthermore, a functional PU.1 binding site was localized in the Ebf1 gene downstream of the Ebf1β promoter. Interestingly, the Ebf1β, but not the Ebf1α promoter is activated by PU.1 [28]. The Ebf1β promoter was also synergistically activated by co-expression of Pax5 and the winged helix protein Ets-1. Unlike the cooperative interactions observed between Pax5 and Ets-1 on the mb-1 promoter [32], Ets-1 bound the Ebf1β promoter independently of Pax5 in vitro. In further support of Pax5’s role in the regulation of Ebf1 transcription, Ebf1 transcripts are expressed at peak levels in pro-B and pre-B cells coincidently with the onset of Pax5 expression [17, 28]. Because EBF directly regulates Pax5 transcription, these data suggest a role for Pax5 in a positive feedback mechanism. In mice, levels of Ebf1 transcripts are reduced in Pax5-deficient proB cells [28, 33]. Pax5-dependent activation of the Ebf1 gene may be due in part to an epigenetic mechanism because expression of Pax5 directed the early replication of the Ebf1 locus in S phase [28]. The early replication of genes reflects the presence of epigenetic modifications indicative of a ‘transcription permissible’ chromatin structure.

The cascade of positive regulators of Ebf1 transcription would suggest that once EBF is made in CLPs, the cells proceed unabatedly towards the B cell fate. However, pathways that act to silence EBF expression would consign cells to other lineage fates (e.g. NK or T lymphoid fates). In this regard, the Ebf1 gene has been identified as a target of negative regulation by Notch1 signaling. The ability of Notch1 signaling to drive T cell development at the expense of B cell development is well documented (reviewed in [34]). As a possible mechanism to account for its inhibition of B cell development, expression of intracellular Notch1 in a pre-B cell line reduced both DNA binding by EBF and promoter activity of EBF-target genes [35]. EBF counters these effects through its activation of Pax5 expression, which in turn represses transcription of Notch1 [36]. The importance of this regulatory circuit is made clear by studies in Pax5-deficient mice. In the absence of continuous Pax5 expression, pro-B cells down-regulated B cell-specific genes in response to Notch signaling [37]. Furthermore, transient expression of Notch1 increased the promiscuous differentiation of Pax5-deficient pro-B cells to NK cells [38, 39]. Thus, B lineage specification is dependent on the expression of EBF, which is supported by IL-7 signaling, E2A and Pax5 in early B cell progenitors.

4. EBF drives the B cell-specific transcriptional program

The ability of EBF to initiate activation of B cell-specific genes has been explored in vivo and in vitro using EBF-deficient progenitor cells. As described above, the lack of EBF results in the early arrest of B cell development [21]. Enforced expression of EBF in HSCs prior to their transfer into irradiated mice resulted in biased developmental outcomes, as evidenced by overproduction of B cells (some myeloid cells were generated as well) [40]. In other experiments, enforced expression of EBF in PU.1-deficient multipotent progenitors activated the expression of transcripts indicative of B lineage commitment (Pax5 and Cd19)[31]. In addition, retroviral expression of EBF in B220⁺IL-7R⁺ progenitors of EBF-deficient mice similarly induced markers of active B lymphopoiesis and commitment. In contrast, enforced expression of Pax5 in such systems did not restore B lymphopoiesis. Although EBF functions cooperatively with E2A, B lymphopoiesis was largely restored following enforced expression of EBF in E2A-deficient progenitor cells [41]. Thus, EBF is critical for the generation of B cell precursors. Interestingly, a recent study demonstrated EBF’s ability to restore B cell development in LSK cells lacking Ikaros (Ikzf1^−/−) [42]. However, the resulting cells were not committed, suggesting that EBF (and Pax5, which was co-expressed) could not fully activate the entire B cell program in this context.

EBF’s ability to restore the B cell program is due partially to the fact that EBF target genes encode proteins required for survival signals that mediate developmental checkpoints (i.e. components of the pre-BCR and BCR). Biochemical studies identified functionally important EBF binding sites in promoters of the Cd79a (mb-1;Igα), Cd79b (B29; Igβ), B lymphoid kinase (Blk), Vpreb1, Igll1 (γ5) and Cd19 genes [43–48]. Many of these genes were not expressed in Ebf1 knockout mice [21]. Other identified targets of EBF included Pax5 [49]. More recently, microarray and biochemical analysis of Ba/F3 cells transduced to express EBF identified the Cd53 and carcinoembryonic antigen-related cell adhesion molecule-1 genes as targets of EBF in B cells [50]. Microarray analysis of EBF-deficient CLPs before and after reconstitution to express EBF revealed additional EBF targets including Pou2af1 (encoding the OcaB transactivator) and FoxO1 [19], a forkhead family factor involved in the control of B cell signaling and proliferation.

One of the hallmarks of EBF function is its cooperation with other regulators of B cell development. In particular, functional interactions between EBF and E2A (E47) constitute a well-documented regulatory pathway. The two proteins bind together on promoters of the Cd79a, Vpreb1 and Igll1 genes [47, 51–53]. EBF and E47 also function together in the activation of V(D)J recombination. Here, they increase the local accessibility of Ig gene segments [54, 55]. In other partnerships, EBF activates transcription of the mb-1 gene synergistically with E2A, Runx1 and Pax5 [13, 51]. A similar constellation of factors activates the Cd19 promoter together with EBF [56]. A recent report suggested that EBF regulates transcription of the Igll1 promoter by binding alternately with Ikaros family zinc finger proteins [57]. In this model, EBF appears to compete with Ikaros (Ikzf1) for Igll1 promoter binding in early pre-B cells. Silencing of the promoter is thought to occur in late pre-B cells following the up-regulation of a second Ikaros family member, Aiolos (Izkf3), in response to pre-BCR signaling. This model suggests an interesting new paradigm in which transcriptional activation or repression results from competition between multiple transcription factors for binding sites during B cell differentiation.

5. Control of B lineage commitment by EBF and Pax5

A substantial body of literature defines Pax5 as the B cell lineage commitment factor. This is largely due to the highly significant studies of the phenotype of Pax5-deficient mice. In these mice, B cell development is arrested at a pro-B cell-like stage. Interestingly, Pax5-deficient pro-B cells exhibit promiscuous gene expression and a lack of lineage commitment [58, 59]. In these studies, Pax5-deficient pro-B cells were shown to have the capacity to differentiate into a variety of non-B cell lineages, including NK cells, macrophage/myeloid cells, granulocytes, osteoclasts and T cells. Further studies revealed that continuous Pax5 expression is required to maintain B cell identity [60–62].

Previously, it was believed that the role of EBF in B lineage commitment was limited to its activation of Pax5 transcription. However, a recent study [33] revealed a new complexity of cell fate decisions during B lymphopoiesis by suggesting that EBF and Pax5 function together to mediate B cell lineage commitment (Fig. 3). In appropriate culture conditions, Ebf1-deficient lymphoid progenitor cells can be induced to differentiate into granulocytes and macrophages in vitro. Reconstitution of lethally irradiated mice with Ebf1-deficient lymphoid progenitor cells resulted in production of myeloid cells, NK cells, dendritic cells and T lineage cells in vivo. Expression of EBF blocked promiscuous differentiation in a dosage-dependent manner. This is similar to the responses of multipotent progenitors to graded levels of PU.1 [63]. Limiting dilution analysis of progenitors transduced to express EBF or Pax5 demonstrated a relatively greater ability of EBF to drive the B cell fate at the expense of the myeloid cell fate. Importantly, EBF possesses the ability to reprogram myeloid progenitors. This activity may be due to the attenuation of C/EBPα (Cebpa) and PU.1 (Sfpi1) [33], which synergistically activate myeloid differentiation [64–66]. High levels of EBF resulted in the expression of B cell-specific genes, including Igll1, Vpreb1, Cd79a and Cd79b [33]. Enforced expression of E47 or Pax5 did not restore B cell-specific gene expression efficiently in Ebf1-deficient lymphoid progenitors. Furthermore, in pre-pro-B cells of IL-7^−/− mice, the B cell program was established in a dose-dependent fashion by EBF [67]. In this study it was concluded that IL-7 is responsible for maintaining sufficient levels of EBF for B cell lineage specification.

Figure 3.

EBF activates the B cell program and commitment to the B cell lineage. EBF activates a host of genes which encode proteins 1) required for V(D)J rearrangement, 2) components of the pre- and mature BCR, and 3) transducers of BCR signals. After induction of its expression by EBF, Pax5 activates 170 lineage-appropriate genes and represses 110 lineage-inappropriate genes [68, 69], resulting in B lineage commitment. EBF also contributes to commitment (see text). EBF synergistically activates transcription of genes with various factors (indicated by *), including Pax5 (e.g. Cd19 and Cd79a), E2A (e.g. Igll1, Vpreb1 and Cd79a), Runx1 (Cd79a and Cd19) and STAT5 (Pax5)(see text for references). Similar to Pax5, EBF is hypothesized to silence transcription of lineage-inappropriate genes (e.g. Id2). Together, EBF and Pax5 function in an interdependent regulatory loop (reviewed in [30].

EBF can drive B cell differentiation, but does it actively mediate B cell lineage commitment? Recent experiments demonstrated that EBF promotes B cell fates in the absence of Pax5 [33]. Enforced expression of EBF in Pax5-deficient pro-B cells blocked promiscuous differentiation under conditions that favor myeloid cell development in vitro. Thus, EBF enforced B cell specification and blocked alternative differentiation pathways. This is likely due to EBF’s abilities to 1) activate expression of critical genes in the B cell-specific lineage program and 2) repress other lineage programs. Concerning the latter function, it was proposed that EBF directly represses transcription of genes, including Id2 and Cebpa, that promote alternative fates in Pax5-deficient pro-B cells. DNA binding of the Id2 promoter by EBF, which may result in its silencing, was detected in a B cell line. However, details of gene silencing in response to EBF are currently unknown. It will be important to determine whether EBF, like Pax5 [68, 69], has the ability to activate or repress transcription in different contexts. Together, these data demonstrate the unique functions of EBF in B lineage specification and commitment; however, other data suggest that EBF (and Pax5) may be insufficient for commitment in the absence of other factors (e.g. Ikaros; [42].

6. EBF and origins of disease

The reduction of EBF expression may result in the development of disease in B lineage cells. For example, Hodgkin lymphoma Reed-Sternberg cells have a B cell origin [70]. These tumor cells have silenced the expression of many markers of B cells, including CD79a, CD79b, Blk, CD53 and Pou2af1 [71]. Recent data suggest that the phenotype of Hodgkin lymphomas is the result of aberrant Notch signaling [72]. Consequences of the over-expression of Notch1 may include antagonism of EBF (and E2A), which in turn interferes with normal B cell identity.

Another potential link between EBF and disease was indicated by studies of the core-binding factor fusion protein Cbfβ-SMMHC, which results from a chromosomal inversion (creating a Cbfb-MYH11 fusion gene) associated with acute myeloid leukemia (AML). A conditional knock-in allele in mice that resulted in expression of Cbfβ-SMMHC in myeloid cells created an abnormal myeloid progenitor population [73]; however, Cbfβ-SMMHC also significantly reduced the production of B cells [20]. Expression of EBF target genes including Cd79a, Igll1, Vpreb1 and Blk were reduced in CLPs from these mice. One explanation for these effects was the reduction of Ebf1 transcripts by 60–80% in the mutant CLPs. Importantly, restoration of EBF expression, but not that of other factors in CLPs of Cbfβ-SMMHC-expressing mice restored B cell development. These data suggest that Cbfβ-SMMHC represses expression of EBF. The mechanism of this repression is unknown, but it may involve regulation of the Ebf1 gene by Runx family factors (Cbfβ is a co-factor of Runx proteins).

The development of leukemia may be a consequence of the loss of EBF expression. In a landmark study, eight cases of B cell progenitor acute lymphoblastic leukemia (ALL) demonstrated monoallelic deletions of EBF1 genes in tumor cells [74]. The loss of EBF1 genes was acquired somatically. In addition, the complete loss of EBF1 genes was observed in some leukemic cells. These data suggest that EBF may function as a tumor suppressor. Other recent observations identified the Ebf1 gene as a target of the activation-induced deaminase (AID) [75], which generates DNA breaks associated with translocations and lymphomagenesis [76]. These data suggest the importance of maintaining EBF expression in B cells and the potential for tumorigenesis when expression of EBF is impaired or lost.

7. Conclusions

Together, the studies described above suggest a model of B lineage specification that exploits a dynamic equilibrium between EBF and other factors in CLPs and early B lymphoid progenitors. The process begins in CLPs with the expression of low levels of EBF in response to PU.1 and IL-7 signaling. At these early stages of differentiation Ebf1 may be repressed by Notch1 signaling and/or decreased IL-7 signaling. This would turn the cell towards other, non-B cell fates. Alternatively, in a permissive environment (in the absence of Notch signaling), EBF may amplify its own expression, which is enhanced further by E47 and by Pax5 following its activation in pro-B cells. In turn, expression of Pax5 negatively regulates Notch signaling by repressing expression of the Notch1 gene itself.

Many questions concerning EBF and its functions remain to be answered. What are the details of B cell lineage specification? Which genes are activated or repressed in response to EBF? Is EBF a direct repressor of transcription, or does it activate other transcriptional repressors? What are the details of epigenetic mechanisms that are modulated by EBF? Concerning the basis of lineage decisions, what is the role of the local microenvironment within the bone marrow? If a non-B cell fate is adopted by a CLP, how is the expression of EBF extinguished? What are the roles of other transcription factors (e.g. Bcl11a/Evi9, NF-κB2/p100 and Foxp1), which are required for normal Ebf1 expression at early stages of development [77–79]? Have we only scratched the surface of relationships between EBF and human diseases? These and other questions will be addressed in the near future using a combination of biochemical and molecular methods, together with new mouse models that will enable the manipulation of EBF expression during early B cell development.