B cells develop from hematopoietic stem cells (HSC) through intermediates that retain various levels of multipotentiality. Commitment to differentiation into B cells occurs close to the pre-pro-B cell differentiation stage (Figure 1A). Igh rearrangements initiate in these cells with D_H rearrangements, followed by V_H rearrangements at the pro-B cells stage (Alt et al., 1984). IgH expression is a checkpoint during B cell development. Only IgH-expressing pro-B cells differentiate to pre-B cells where Ig light chain genes (Igk and Igl) rearrange. Expression of light chain permits immature B cells to express membrane antibody molecules of the IgM isotype.

The 3′ Igh domain, extending from the intergenic control region 1 (IGCR1) to 3′ CTCF binding elements (3′CBE), within which the first rearrangements occur is marked by several regions of high chromatin accessibility in pro-B cells (Figure 1C, left panel). At the 5′ boundary two CTCF binding sites within IGCR1 regulate Igh rearrangements. Their mutation or deletion accentuates use of the closest V_H gene segments thereby severely compromising Igh diversity, as well as permits V_H rearrangements to unrearranged D_H gene segments (Featherstone et al., 2010; Guo et al., 2011; Lin et al., 2015; Qiu et al., 2018; Giallourakis et al., 2010). The intronic enhancer, Eµ, and a promoter (PQ52) associated with the 3′-most D_H gene segment, DQ52, are marked by two closely associated ATAC-sensitive regions. In pro-B cells Eµ regulates histone modifications in the 3′ Igh domain (Chakraborty et al., 2009), induces transcription of the unrearranged (germline) locus and activates rearrangements. Deletion of Eµ reduces D_H recombination substantially (∼80%) and virtually abolishes V_H recombination (Chakraborty et al., 2009; Perlot et al., 2005; Afshar et al., 2006; Bolland et al., 2007).

Several additional chromatin accessible sites are evident moving 3′ from Eµ (Figure 1C, left panel). hRE1 is an enhancer located between Cγ1 and Cγ2b IgH isotypes (Medvedovic et al., 2013; Predeus et al., 2014). hRE1 is not required for Igh rearrangements in pro-B cells but promotes class switch recombination (CSR) to IgG3, IgG2b and IgG2a isotypes during immune responses (Amoretti-Villa et al., 2019). The 3′ regulatory region (3′RR), located 3′ of the last Cα exons, consists of a cluster of four B cell-specific transcriptional enhancers that span 28 kb (Lieberson et al., 1991; Giannini et al., 1993; Matthias and Baltimore, 1993; Madisen and Groudine, 1994; Michaelson et al., 1995). Like hRE1, the 3′RR contributes primarily to CSR, but not to control of Igh rearrangements in pro-B cells (Vincent-Fabert et al., 2010; Rouaud et al., 2012; Saintamand et al., 2015; Bruzeau et al., 2021; Oudinet et al., 2020). Finally, 3′CBE is a cluster of CTCF binding chromatin accessible regions that mark the 3′ boundary of the Igh topologically associated domain (TAD) (Garrett et al., 2005; Vian et al., 2018). Accordingly, its deletion leads to transcriptional activation of genes located further 3′ that are not normally expressed in pro-B cells (Volpi et al., 2012; Zhang et al., 2021). Recently, an additional enhancer has been identified within the V_H genes. Its deletion affects recombination of closely positioned V_H gene segments (Bhat et al., 2023). Thus, Eµ constitutes the only validated regulatory sequence associated with both Igh locus activation and rearrangements in pre-pro- and pro-B cells.

Though originally identified as a transcriptional enhancer and proposed to be important for transcription of rearranged Igh alleles, several lines of evidence suggest that Eµ is not required for Igh expression in mature B cells or during immune responses. Eckhardt and colleagues first demonstrated that VDJ ‘knock in’ Igh alleles that lacked Eµ permitted normal B cell development and function (Li and Eckhardt, 2009; Li et al., 2010). Other studies in germline Eµ-deficient mice indicate that immune responses are not affected substantially despite smaller numbers of mature B cells in these strains (Perlot et al., 2005; Marquet et al., 2014). However, Eμ is essential for B cell-specific transcription of functionally rearranged Igh transgenes in mice. Eμ has also been shown to be a diversity activating sequence (DIVAC) that promotes activation-induced deaminase dependent somatic hypermutation of Ig sequences, a process that occurs only during peripheral immune responses (Buerstedde et al., 2014). We believe that additional studies of the role of Eμ in mature B cells are warranted.

Eµ binds many transcription factors (Figure 1D) (Kumari et al., 2018; Kleiman et al., 2016; Ernst and Smale, 1995; Henderson and Calame, 1998; Lin et al., 2010). However, none of these easily explain lineage- or developmental stage-specificity of Igh activation. Most of these proteins are widely expressed in hematopoietic cells, such YY1, RUNX family members, ETS proteins, OCT proteins and bHLH-zip proteins. Others, like PU.1 and E2A have more restricted tissue distribution. PU.1 is expressed at highest levels in myeloid cells where it has been proposed to act as a pioneer factor (Heinz et al., 2010). Lower levels of PU.1 present at early hematopoietic stages (such as HSC, CMP and CLP) are maintained throughout B cell development but extinguished during T cell differentiation (Heinz et al., 2010; Iwasaki et al., 2005; Dakic et al., 2007; Pang et al., 2018). E2A is also expressed in HSC through CLP stages but further up-regulated in pre-pro-B cells and thereafter (Semerad et al., 2009; Fischer et al., 2020; Aubrey et al., 2022). Based on early transfection experiments it was proposed that Eµ function is generated by combinatorial activity of different Eµ-binding proteins, especially combinations of ETS proteins and E2A (Ernst and Smale, 1995; Nelsen and Sen, 1992; Nelsen et al., 1993). Which combinations are most important in the endogenous context have not been identified.

Eμ function is also modulated by 5′ and 3′ flanking matrix attachment regions (MARs) (Scheuermann and Garrard, 1999). These are A/T-rich sequences that bind several transcription factors, including Cux/CDP, Satb1 and Bright (Romig et al., 1992; Weitzel et al., 1997; Dickinson et al., 1992; Herrscher et al., 1995; Alvarez et al., 2000; Dobreva et al., 2003). In transgenic studies MAR sequences are essential to fully reveal Eμ activity as reflected in transcription activation, especially at a distance, and extension of chromatin accessibility (Jenuwein et al., 1997; Forrester et al., 1999; Fernández et al., 2001; Forrester et al., 1994). These effects of MARS are independent of transcription factor binding to Eμ, suggesting that factor binding is not sufficient for Eμ function (Fernández et al., 2001). Unlike the effects of deleting Eμ, however, deletion of one or both Igh MARs from the endogenous locus does not affect V(D)J recombination regulated by Eμ or B cell development (Sakai et al., 1999). The role of MARs and their relationship to enhancer function awaits further studies.

While Eµ is both necessary and sufficient for gene activation, examination of chromatin accessibility throughout hematopoiesis reveals hitherto unstudied complexities. First, accessibility at closely positioned Eµ and DQ52 promoter is evident in HSC and maintained throughout developmental stages that precede B lineage commitment (Figure 1C, left panel). Absence of the DQ52 ATAC peak in pro-B cells likely reflects loss of that region by D_H and V_H recombination. The most parsimonious explanation is that Eµ is not B lineage specific though it drives transgenic expression largely in B and T lymphocytes. Alternatively, it is possible that though Eµ is accessible at earlier stages, it does not have enhancer activity until pre-pro- or pro-B cell stages. This is reminiscent of poised enhancers that are chromatin accessible but lack H3K27ac. Consistent with the latter hypothesis, transcription (Figure 1C, right panel) and H3K27ac modifications (not shown) are higher in pro-B cells compared to HSC (Choukrallah et al., 2015). We surmise that binding of hematopoietic transcription factors to Eµ marks this site for later activation in B lineage cells. Comprehensive analyses of enhancer function throughout hematopoiesis will be required to understand the underlying mechanisms.

Lack of simple concordance between ATAC sensitivity and enhancer function raises two questions. First, what transforms a chromatin accessible, but transcriptionally silent, enhancer into a functionally active enhancer in the B lineage? Our working hypothesis is that chromatin accessibility throughout hematopoiesis marks a poised enhancer that is transcriptionally activated in pre-pro- and pro-B cells. How many transcription factor binding sites are required to poise the enhancer for B lineage-specific activation and which factors convert a pre-marked but inactive enhancer to an active one remain to be discovered. Second, why does Eµ bind so many different transcription factors? One possibility is that some of these factors may suppress enhancer activity in other hematopoietic lineages. The most closely related one is developing T cells where Tcr genes recombine. Eµ chromatin accessibility and associated transcription is considerably lower in T lineage precursors (DN1-DP stages) compared to pro-B cells (Figure 1C, left panel), though they express many of the same transcription factors. This reduced accessibility drives low levels of Igh D_H rearrangements in DP cells. We hypothesize that sub-optimal activation of Eµ in T cell precursors precludes compete V(D)J recombination and, thereby, the possibility of co-expressing functional IgH and TCRα/β chains in the same cell.

Surface Ig heavy chain expression in pro-B cells triggers a proliferative burst that culminates with production of pre-B cells (Figure 1A). The bulk of Ig light chain gene rearrangements take place in these cells. In this review we focus on the Igk locus that encodes more than 90% of light chain protein in mice. The mouse Igk locus spans 3.2 Mb, most of which encodes 96 functional Vκ gene segments (Figure 2A). Clustered at the 3′ end are 4-5 Jκ gene segments, one exon encoding the constant region of Igκ (Cκ) and several regulatory sequences (Figure 2A). One recombination event creates VκJκ junctions that encode Igκ.

Multiple enhancers control Igk recombination. The intronic enhancer, iEκ, appears most important. iEk is marked by repressive H3K27me3 modification in pro-B cells that is replaced by acetylated histone 4 in pre-B cells where Igk genes recombine (Inlay et al., 2006). In its absence Vκ recombination is reduced about 10-fold (Xu et al., 1996). Residual recombination is lost upon additionally deleting 3′Eκ (Inlay et al., 2002), though deletion of 3′Eκ by itself has little/no effect (Inlay et al., 2002). A third enhancer, Ed, located further 3′ also has little effect when deleted alone, but the double deletion of Ed with 3′Eκ abolishes Igk rearrangements (Zhou et al., 2010). Thus, iEκ is insufficient for recombination, whereas 3′Eκ plus Ed together are relatively weak activators of Igk rearrangements. 3′Eκ and Ed may function as shadow enhancers alongside an active iEκ (Hobert, 2010). Loss of 3′Ek is accompanied by substantially reduced H3 acetylation at the Jκ gene segments in pre-B cells. Ed deletion has a smaller effect, but loss of both 3′Ek and Ed abolishes H3ac at Jκs (Zhou et al., 2010). Though histone acetylation status of iEκ deleted alleles has not been characterized, these results cumulatively suggest that enhancer activation reflected in active histone modifications correlates with induction of rearrangements. To what extent such modifications contribute to V(D)J recombination per se remains to be determined. Moreover, because different measures of locus activation have been used (many of which preceded the chromatin immunoprecipitation (ChIP) era), systematic studies of WT and mutated loci are needed to understand molecular connections between enhancers, chromatin states, transcriptional activation and recombination.

In addition to the classical enhancers, two elements, the contracting element for recombination (Cer) and silencer in the intervening sequence (Sis), regulate Igk rearrangements (Xiang et al., 2011; Xiang et al., 2013). Deletion of these elements skews the rearranged repertoire of Vκ genes to those located closer to the 3′ end of the locus (proximal Vκ genes). Specifically, Vκ genes lying within 100 kb of Jκs account for 25% of the repertoire in the absence of Sis and 62% of the repertoire in the absence of Cer, compared to 10% in the wild type context (Xiang et al., 2011; Xiang et al., 2013). The effects are even more pronounced when both elements are missing (Xiang et al., 2014). Their activity is likely conferred by oriented CTCF binding sites in each element, much like IGCR1 in Igh. However, the role of CTCF has not been directly confirmed by point mutations of these sites. Like Igh, additional enhancers within the Vκ region promote recombination of nearby gene segments (Barajas-Mora et al., 2019; Barajas-Mora et al., 2023). The greater effects of enhancers located near the joining gene segments (J_H and Jκ) at both immunoglobulin loci for gene rearrangements may be via their role in establishing RAG1/2-rich recombination centers at which antigen receptor gene rearrangements are initiated.

To what extent can developmental timing of Igk rearrangements be explained by enhancer activation by transcription factors? iEκ has been best studied in this regard. Early studies showed that mutating the NF-κB binding site in iEκ had little effect on rearrangements, whereas mutating two ‘E’ elements significantly reduced iEκ function (Inlay et al., 2004). These E elements bind E2A and bHLH-zip proteins such as TFE3 in vitro (Staudt and Lenardo, 1991). However, the importance of these motifs for iEκ function does not provide a ready explanation for developmental timing of Igk activation, in part because similar motifs are also found in Eμ that is activated at an earlier developmental stage. Additionally, genes encoding E2A and bHLH-zip proteins are not selectively expressed in pre-B cells. Timing of iEκ activation is now attributed to a combination of chemokine and cytokine activity that moves pre-B cells away from an IL-7-rich milieu essential for pro-B cell differentiation and proliferation. Clark and colleagues have proposed that STAT5 activated in response to IL-7 signals binds to iEκ and suppresses its activity in pro-B cells by competing with E2A binding and/or recruitment of EZH2, a writer of repressive H3K27me3 modification (Mandal et al., 2009; Malin et al., 2010; Mandal et al., 2011). In the IL-7-poor pre-B cells niche, phospho-STAT5 binding is reduced, permitting iEκ activation (Clark et al., 2014). STAT5 is also implicated in repressing BRWD1, a transcription factor that is up-regulated in the transition to pre-B cells, which was recently shown to bind iEκ and regulate Igk locus contraction (Mandal et al., 2015; Mandal et al., 2024). Loss of STAT5 from iEκ also coincides with recruitment of Ikaros to this enhancer which may also regulate its developmental stage-specific activation (see below). Finally, iEκ is also associated with a MAR on its 5′ flank (Cockerill and Garrard, 1986). Though systematic transgenic studies have not been done to investigate the contribution of this MAR to iEκ function, deletion of the MAR from the endogenous locus does not affect Igk rearrangements or B cell development (Yi et al., 1999).

Less is known about factors that activate 3′Eκ and Ed. ChIP studies show 3′Eκ binds to PU.1, E2A, IRF and PAX5 in pro- and pre-B cells (Figure 2C) (Lin et al., 2010; Shaffer et al., 1997; Lu et al., 2003; Ochiai et al., 2012; Revilla-I-Domingo et al., 2012; Schwickert et al., 2014; Stadhouders et al., 2014). Because 3′Eκ function has only been demonstrated in combination with either iEκ or Ed, it is not clear at which developmental stage 3′Eκ is activated. Published studies show that Ed binds PU.1, E2A, and IRF4 in pro-B cells (Lin et al., 2010; Shaffer et al., 1997; Lu et al., 2003; Schwickert et al., 2014; Stadhouders et al., 2014), though it is uncertain if this binding occurs in pre-B cells as well. In contrast, Ed binds selectively to Ikaros and PAX5 in pre-B cells (Ochiai et al., 2012; Revilla-I-Domingo et al., 2012; Schwickert et al., 2014; Ferreiros-Vidal et al., 2013). How these dynamically shifting interactions contribute to Ed function is not understood. An interesting unifying feature of all three Igk-associated enhancers is that they bind Ikaros selectively in pre-B cells (Ribeiro de Almeida et al., 2015). Ikaros has been shown to regulate pre-B cell differentiation and proposed to induce an enhancer hub in the 3′ Igk locus (Hu et al., 2023). This hub promotes interactions with Vκ gene segments leading to locus contraction required for distal Vκ rearrangements. Spatial proximity of 5′ Vκs and the 3′ Igk region is reduced in Ikaros-deficient pre-B cells with concomitant reduction in Vκ gene rearrangements. These observations are consistent with Ikaros playing a crucial role in timing Igk locus activation and recombination in pre-B cells (Hu et al., 2023). The possible interplay between BRWD1 and Ikaros, both of which induce locus contraction of Igk, remains to be discovered.

A global view of the chromatin structure of the Igk locus reveals similarities and dissimilarities with Igh. Like Igh, chromatin accessibility of all three recombination-related enhancers precedes developmental stage-specific functional activation. This is reflected in strong ATAC peaks at iEκ, 3′Eκ and Ed in pro- as well as pre-B cells (Figure 2B, left panel). However, activity as reflected by germline transcription is most prominent in pre-B cells (Figure 2B, right panel). Thus, developmental timing is strictly enforced at the level of function but not at the level of chromatin accessibility. Unlike Igh, however, Igk enhancers are not pre-marked in early hematopoiesis (Figure 2B, left panel). 3′Eκ gains accessibility first at the CLP stage whereas iEκ and Ed are most prominently accessible at pro- and pre-B cell stages (Figure 2B, left panel). The close coincidence of transcriptional activation with Ikaros binding strongly suggests that transformation of accessible but functionally inactive enhancers to a transcriptionally active state is mediated by recruitment of Ikaros to pre-marked chromatin regions.

It is interesting to note that mutation of E elements in iEκ or absence of Ikaros in pre-B cells attenuate iEκ function (Inlay et al., 2004; Hu et al., 2023). E2A binds to iEκ in both pro- and pre-B cells, suggesting that pre-marking of iEκ in pro-B cells may be mediated by this protein. Furthermore, a two-fold increase of E2A transcripts was observed in pre-B cells compared to pro-B cells (ImmGen) (Heng et al., 2008), indicating a greater abundance of E2A at iEκ. The mechanism of Ikaros recruitment to iEκ could be via direct interactions with E2A or to the altered chromatin state induced by E2A. Chromatin structure analyses of E-mutated iEκ in pro-B cells and point mutation of Ikaros binding sites will help to disentangle functions of these proteins at iEκ. It is also interesting that the spectrum of transcription factors that bind to 3′Eκ is very similar to that at iEκ. Yet, 3′Eκ does not effectively compensate for genetic deletion of iEκ. The basis for this distinction is not clear but may relate to its location beyond Cκ and associated inefficiency in inducing a recombination center near the Jκ gene segments.

In summary, multiple enhancers regulate developmental stage-specific transcription and rearrangements of Igh and Igk loci. Of these, Eμ and iEκ enhancers appear to be the most important because deletion of either element alone substantially impairs activation of the associated locus. A comparison of the two enhancers highlights several features:

1) Chromatin accessibility and transcriptional activation are temporally distinct for both enhancers. Accessibility at Eμ is evident throughout early hematopoiesis, whereas iEκ gains most accessibility in B lineage committed cells. Yet, transcriptional activation occurs precisely at pro- and pre-B stages, respectively (Figures 1C, 2B).

Multipotential cells that migrate to the thymus undergo sequential rearrangements of Tcrb and Tcra genes. Developmental stages in the thymus are defined based on the expression of CD4 and CD8 cell surface proteins. The earliest stages lack both and are referred to as double negative (DN) cells (Figure 1A). DN cells are further divided into DN1, 2a/b, 3a/b and 4 (Krangel, 2009; Rodriguez-Caparros et al., 2020; Hosokawa and Rothenberg, 2021) subsets. DN1 cells are a mixed population that include multipotential cells. Commitment to T cell differentiation program is imposed in DN2 subsets by the combined action of transcription factors TCF1, BCl11b and GATA3 (Rothenberg and Taghon, 2005; Hosokawa and Rothenberg, 2021). Tcrb rearrangements initiate in DN2 cells and are completed by the DN3 stage. Only cells expressing TCRβ protein differentiate to CD4⁺CD8⁺ double positive (DP) stage via the intermediate DN4 stage. Tcra rearrangements occur in DP cells leading to generation of T cell receptor-expressing progenitor cells.

Organization of the germline Tcrb locus is shown in Figure 3A. The Tcrb locus spans approximately 0.65 Mb of mouse chromosome 6. The 5′ part contains 33 Vβ gene segments (Trbv) of which 21 are functional. Multiple trypsinogen genes are located between Vβ and Dβ genes. The 3′ end contains two Dβ-Jβ-Cβ clusters. Each cluster has one Dβ gene segment (Dβ1 or Dβ2), six Jβ gene segments and exons encoding the constant parts of TCRβ chains (Cβ1 or Cβ2). One Vβ gene segment, Vβ31, is located beyond the Dβ-Jβ-Cβ clusters. Despite being organized differently from Igh, V(D)J recombination proceeds similarly at both loci. Dβs rearranges first in DN2 cells followed by Vβ rearrangements to DJβ junctions in DN3 cells (Krangel, 2009). Three regulatory sequences control Tcrb rearrangements. Eβ, an enhancer located between Cβ2 and Vβ31, is essential for Tcrb recombination. Its deletion abrogates all Tcrb rearrangements (Bories et al., 1996; Bouvier et al., 1996). Absence of Eβ leads to loss of activating histone modifications (H3ac and H3K4me2) in a 25 kb region extending to a boundary element located 5′ of Dβ (Majumder et al., 2015; Mathieu et al., 2000; Carabana et al., 2011). Coordinately, this region gains repressive histone modifications, H3K9me2 and H3K27me3, on Eβ-deleted alleles (Majumder et al., 2015; Carabana et al., 2011). One side of Eβ contains a MAR, however, its deletion from the locus does not affect Tcrb transcription in thymocytes (Chattopadhyay et al., 1998).

PDβ1 and PDβ2 are promoters located 5′ of the respective Dβ gene segments. Deletion of PDβ1 promoter attenuates recombination of the nearest Dβ1 gene segment (Whitehurst et al., 1999; Whitehurst et al., 2000). PDβ1 lies within the domain of influence of Eβ, and its histone modification state is altered on Eβ-deleted Tcrb alleles (Spicuglia et al., 2002). The effects of mutating or deleting PDβ2 alone have not been investigated. 5′PC, located within the trypsinogen cluster, is a CTCF binding site. Deletion of a large genomic region between Dβ2-Jβ2 and Trbv5 gene segments that includes 5′PC permits rearrangement of the remaining Trbv5 to Dβ2. The authors concluded that 5′PC plays a regulatory role in ordered assembly of Tcrb genes (Senoo et al., 2003).

ATAC profile of the 3′ end of Tcrb shows that Eβ is accessible in bone marrow progenitors as noted above for Eµ (Figure 3B, left panel). Accessibility of 5′PC in HSC and all developmental stages leading to B and T lymphocytes probably reflects lineage non-specific binding of CTCF. Accessibility at a region near the Vβ31 promoter parallels the pattern of Eβ (Figure 3B, left panel). This region also contains a CTCF binding site which, unlike 5′PC, may require Eβ accessibility for CTCF binding. Coincident with T cell precursors reaching the thymus (DN1 cells), PDβ1 and PDβ2 become accessible. We surmise this represents Eβ activation and Eβ-dependent activation of Dβ-associated promoters. Accordingly, non-coding transcription of Tcrb alleles is also first evident in DN1 cells (Figure 3B, right panel). This is reminiscent of the distinction between enhancer accessibility and enhancer activation observed with Eµ. Thereafter, PDβ1 and PDβ2 remain accessible in DN2 cells as Tcrb rearrangements proceed. Loss of ATAC accessibility of these regions in DN3 cells likely represents their loss from the genome by Vβ recombination. Eβ accessibility and activity is maintained in further differentiated T cells, but not in other mature hemopoietic lineages such as B cells and myeloid cells (Figure 3B, left panel). Taken together, these observations are consistent with developmental timing of Tcrb transcription and rearrangements being determined by Eβ activation.

The most prominent transcription factor motifs identified within Eβ are two composite ETS/RUNX binding sites (Figure 3C) (Hollenhorst et al., 2009). Targeted mutation of both RUNX sites abolishes enhancer activity and blocks T cell development as seen with Eβ deletion (Majumder et al., 2015). Oltz and colleagues dissected the requirements for ETS and RUNX proteins using mutated enhancers in which RUNX binding sites were replaced with GAL4 binding sites. Recruitment of RUNX1/GAL4 fusion proteins to the mutated enhancer activated Eβ-like function with regard to PDβ1 and PDβ2 transcription even in the absence of the adjacent ETS binding sites (Zhao et al., 2017). The ETS family member proposed to work at Eβ is ETS1, a gene whose expression is increased in T cell progenitors undergoing Tcrb rearrangements (Rodriguez-Caparros et al., 2020; Rothenberg et al., 2008; Cauchy et al., 2016). ETS1 and RUNX proteins bind cooperatively to ETS/RUNX composite motifs via neutralization of autoinhibitory domains in each factor leading to the following model for Eβ activation by these factors (Wotton et al., 1994; Kim et al., 1999; Goetz et al., 2000; Gu et al., 2000). RUNX1 is expressed throughout hematopoiesis, however it does not bind and activate Eβ until ETS1 levels rise in DN2 cells close to T lineage commitment (Rothenberg et al., 2008). Cooperative RUNX1-ETS1 binding to Eβ permits recruitment of co-activators such as CBP/p300 to establish transcriptional competence (Hollenhorst et al., 2009; Yang et al., 1998).

Several questions remain to be addressed. First, it has not been established whether the two ETS/RUNX motifs are sufficient for Eβ activity. Other transcription factors, such as E2A and the related bHLH protein HEB, have been shown to bind Eβ (Spicuglia et al., 2002) (Figure 3C), however their functional significance in the context of the Tcrb locus has not been explored. Second, it is not clear what makes Eβ T lineage specific because ETS and RUNX family proteins are widely expressed in hematopoietic cells. One possibility is that negative regulatory elements within Eβ suppress its activity in the wrong lineage. However, such elements have not been identified. Third, what factors make Eβ ATAC sensitive in bone marrow progenitors? Amongst key Eβ binding proteins identified to date the obvious candidate is RUNX1 (or a related RUNX family member). However, if RUNX1 proteins can bind and increase chromatin accessibility, then what prevents them from activating transcription as shown by the GAL4 fusion recruitment studies? Perhaps RUNX proteins bind with other (non-ETS) proteins in uncommitted progenitors in a configuration that precludes transcriptional coactivator recruitment. ETS1 (or other functionally similar ETS proteins) may replace these factors in DN1/2 cells to cooperatively activate Eβ with pre-bound RUNX proteins. Alternatively, progenitor accessibility may be mediated by currently unknown Eβ binding proteins. We expect that additional functional Eβ binding proteins will be identified.

Productive rearrangement of Tcrb alleles in DN cells leads to proliferation and differentiation of TCRβ chain-expressing T cell precursors to the CD4⁺CD8⁺ (double positive, DP) stage (Figure 1A). Tcra gene rearrangements take place in DP cells, poising these cells to express the complete αβ T cell receptor. Following additional selection events in the thymus, CD4⁺ and CD8⁺ single positive cells capable of mounting immune responses emerge. Tcra genes arise from a single recombination event between one of ∼80 Vα gene segments (Trav) and one of approximately 60 Jα gene segments (Traj) located close to exons that encode the constant part (Cα, Trac) of TCRα chains (Figure 4A). A 1.7 Mb region of mouse chromosome 14 houses the Tcra locus with interspersed gene segments that will recombine to generate TCRδ chains of the γδ T cell receptor (Figure 4A). Tcrd diversity (Dδ, Trdd), joining (Jδ, Trdj) and a few variable (Vδ, Trdv) gene segments, along with constant parts of the TCRδ chains (Cδ, Trdc), are embedded between Traj and most of the Tcra variable gene segments (Trav) (Figure 4A). Despite being located within the Tcra locus, Tcrd gene rearrangements occur in DN cells guided by the Eδ enhancer located near Trdc. Loss of Eδ selectively abolishes Tcrd rearrangements without affecting Tcra rearrangements (Monroe et al., 1999). Though Eδ will not be further discussed in this review it is interesting to note that its activity, as measured by Trdc transcription, is restricted precisely to DN cells where Tcrd rearrangements occur (Figure 4B).

The Eα enhancer located 3′ of Cα is essential for Tcra rearrangements. This enhancer is marked by an ATAC peak at all hematopoietic stages, including HSC (Figure 4B, left panel). Indeed, accessibility appears stronger in B cell precursors than in T cell precursors. Yet, its activity, as inferred by Cα RNA, is exquisitely specific to the developmental stage (DP cells) at which Tcra genes rearrange. Eα and approximately 80 kb of 5′ sequence that include Cα exons and all Jα gene segments gain histone H3 acetylation during transition to DP cells (McMurry and Krangel, 2000), closely correlating with transcriptional activation of Tcra. This domain of active histone modifications is lost in DP cells from mice that lack Eα, showing that the enhancer regulates long-range chromatin state that correlates with gene recombination (McMurry and Krangel, 2000). A MAR has not been identified near Eα (Scheuermann and Garrard, 1999). The ATAC pattern is reminiscent of those at Igh and Tcrb loci and distinct from that at Igk, insofar as enhancers in Igk are not accessible in early bone marrow precursors. Thus, locus activating enhancers at three out of four major antigen receptor loci are pre-marked early in hematopoiesis but activated in the appropriate lineage at the correct developmental stage.

In addition to Eα, several CTCF-binding regulatory elements modulate Tcra rearrangements. The T early alpha (TEA) promoter guides use of Jα gene segments. Its deletion results in reduced rearrangements of several 5′ Jα segments closest to TEA, but no effect on Jα segments located further 3’ (Villey et al., 1996). Intergenic CBEs 1 and 2 (INT1 and INT2) separate the large genomic region that contains Trav and Trdv gene segments from the rest of the Tcra-d locus. Orientation of CTCF binding sites in each element define an 80 kb domain that extends till the CTCF binding site in TEA (Chen et al., 2015). This domain, containing Dδ, Jδ and Cδ under control of Eδ, is thereby effectively segregated from the rest of the locus. Double deletion of INT1 and INT2 alters the Tcra repertoire and leads to defective γδ T cell development. The two elements are partially redundant as mutation of INT2 alone has minor effects (Chen et al., 2015).

Many transcription factors bind Eα both in DN cells (where it is inactive) and in DP cells (where it is active, Figure 4C) (Hernandez-Munain, 2015). Amongst these are the first examples of T lineage-restricted factors such as LEF1/TCF1 and GATA3. These factors are expressed concomitant with T cell commitment in DN2 cells simultaneously with Tcrb activation by Eβ. However, their binding (along with other proteins) is apparently insufficient to activate Eα. By contrast, three factors that are widely associated with cell activation and pre-T cell receptor (pre-TCR) signaling, NFAT, EGR and AP1, bind selectively to Eα in late DN4 cells and DP cells (King et al., 1999; Aifantis et al., 2001; Carter et al., 2007; del Blanco et al., 2012). Hernández-Munain and colleagues proposed that constitutive but lymphoid-restricted transcription factors, such as E2A and ETS1, pre-mark Eα before its activation. The pre-marked enhancer recruits additional factors induced by pre-TCR signaling, as well as the CREB-binding protein/p300 coactivators, to fully activate Eα (del Blanco et al., 2012).

RORγt and Ikaros are two other proteins that were recently shown to bind to Eα in DP cells by ChIP experiments (Schwickert et al., 2014; Naik et al., 2024); whether they bind to Eα at earlier stages has not yet been explored. RORγt is important for lifetime of DP cells which, in turn, impacts the Vα repertoire (Sun et al., 2000; Guo et al., 2002). However, a direct role for RORγt in regulating recombination via Eα activity has not been demonstrated. It is noteworthy that Ikaros binds to both late-activated antigen receptor gene enhancers (iEκ and Eα) at the appropriate developmental stage. Because Ikaros deficiency perturbs T cell development at the earliest stages (Georgopoulos, 2002), it is not known to what extent Ikaros binding to Eα is required for enhancer function in DP cells. Point mutational analyses of Eα will be needed to sort through transcription factor dynamics and functions in developing T cells.

Unlike Igh and Igk where multiple enhancers guide developmental stage-specific locus activation, Tcrb and Tcra loci each rely on only one (known) enhancer to initiate developmentally appropriate transcription and rearrangements. A comparison of Eβ and Eα is therefore pertinent for identifying mechanisms that guide T cell stage-specific gene expression. Additionally, a comparison of BCR- and TCR-associated enhancers provides a perspective into lineage-restricted gene expression. Several interesting features can be highlighted:

1) Both Eβ and Eα are ATAC sensitive from early hematopoietic stages. Thus, chromatin accessibility can be dissociated from transcriptional activation at all antigen receptor gene enhancers. This raises interesting questions regarding the role of pioneer factors in lineage- and stage-specific activation of antigen receptor gene enhancers. It is possible that these enhancers are atypical because of their role in recombination regulation beyond classical transcription activation. It is also interesting to note that both Eβ and Eα remain accessible in B lineage precursors, whereas both Eμ and iEκ are mostly inaccessible in T cell precursors.

It is interesting to examine the mechanisms by which Eμ and Eβ are transcriptionally activated coincident with B or T lineage commitment from multipotential progenitors. Both Eμ and Eβ contain multiple binding sites for bHLH, RUNX and ETS proteins, yet each largely activates transcription and recombination in distinct lineages. One possibility is that this is due to different organization of binding sites within each enhancer. For example, only one of the four RUNX binding sites in Eμ has the configuration of ETS/RUNX composite elements that dominate Eβ. ETS binding sites in Eμ are also distinct in both family member selectivity (Eμ has two PU.1 binding sites whose sequence specificities are different from that of ETS1 and related factors) and distribution across the enhancer. Both Eμ and Eβ are also pre-marked by chromatin accessibility much earlier in hematopoiesis than their functions are manifest (Figures 1C, 3B). We hypothesize that pre-marking identifies genomic sites at which functional factors will be recruited in the appropriate lineage and at the appropriate developmental stage. It will be interesting to identify other such regulatory sites that are pre-marked early in hematopoiesis for later functional activation to understand the underlying regulatory logic. It will also be interesting to compare such regulatory sequences to those at which accessibility and functional activation coincide. In the context of pioneer factors these observations suggest that factors that drive Eμ and Eβ activity may not be capable of pioneering correctly.