The maintenance of B cell identity requires active transcriptional control that enforces a B cell–specific program and suppresses alternative lineage genes. Accordingly, disrupting the B cell identity regulatory network compromises B cell function and induces cell fate plasticity by allowing derepression of alternative lineage-specific transcriptional programs. Although the B lineage is incredibly resistant to most differentiating factors, loss of just a single B lineage–specific transcription factor or the forced expression of individual non–B cell lineage transcription factors can radically disrupt B cell maintenance and allow dedifferentiation or transdifferentiation into entirely distinct lineages. B lymphocytes thereby offer an insightful and useful case study of how a specific cell lineage can maintain a stable identity throughout life and how perturbations of a single master regulator can induce cellular plasticity. In this article, we review the regulatory mechanisms that safeguard B cell identity, and we discuss how dysregulation of the B cell maintenance program can drive malignant transformation and enable therapeutic resistance.
Cell differentiation is the process in which a less specialized cell goes through development and maturation to acquire a specialized function. Cellular developmental trajectories are usually depicted as one-way thoroughfares with cells progressively adopting increasingly limited self-renewal and differentiation capacities, and ultimately undergoing irreversible lineage commitment. However, these models oversimplify the nature of cell fate determination, incorrectly implying that a cell, once committed, progresses along one of a few lineage-restricted unidirectional pathways. Classical reprogramming experiments using somatic cell transfer, cell fusion, or overexpression of one or more transcription factors (TFs) have demonstrated that somatic cells can alter and also reverse their lineage identity (1–4). The committed state is thus not irreversibly fixed; instead, it requires active and continuous maintenance (5).
The mammalian immune system acts as an experimentally amenable model to study the basic principles of how cell fate is established and maintained (6–12). The development and functional integrity of the immune system depend on at least four core principles: (1) segregation, where early hematopoietic progenitors differentiate into nonoverlapping immune lineages; (2) specialization, where each lineage acquires a set of unique physical, molecular, and functional features; (3) stability, where immune cell identity is stable and committed cells do not interconvert between lineages; and (4) robustness, where cells maintain a stable identity in the face of immune challenges. Together, these principles ensure that cells across all hematopoietic lineages can act in a coordinated, highly specialized, and consistent manner to execute an effective immune response.
B lymphocytes comprise one of the main immune lineages and, as described later, offer an excellent experimental system for elucidating the molecular mechanisms that functionally differentiated lineages use to establish and maintain cell fate. The factors that drive B lineage commitment in early hematopoietic progenitors have been reviewed extensively (13, 14). In this brief review, we focus on the mechanisms by which B cells safeguard their unilineage commitment. We describe the transcriptional regulation of B cell fate maintenance and discuss how B lymphocytes can renounce their lineage identity either as part of physiological progression or pathological reprogramming.
Background: B cell development and maturation
Throughout adult life, hematopoietic stem cells (HSCs) in the bone marrow (BM) generate and continuously replenish all major immune cell types. Individual HSCs can give rise to the B lineage through a series of cell fate restriction steps. First, HSCs differentiate into B cell–primed multipotent hematopoietic progenitors that, in turn, lose access to alternative immune cell fates, commit to the B lineage, and acquire B cell–specific features (13) (Fig. 1A). The earliest stages of B cell development consist of five chronological fractions known as Hardy fractions A–E (15) (Fig. 1A). Fraction A enriches for cells that have initiated a B lineage specification program, and fraction B contains cells that have stably committed to the B lineage (16). This can be revealed if the different B cell fractions are isolated from the BM and tested for their capacity to generate alternate hematopoietic cell types either in vivo via cell transfer into an irradiated mouse host or ex vivo in a culture system that supports multilineage differentiation (16). Multipotent hematopoietic progenitors possess the greatest lineage plasticity and can reconstitute all major immune lineages in vivo (16). Fraction A cells generate predominantly B cells while retaining T and NK cell differentiation capacity in vivo, as well as myeloid differentiation capacity ex vivo (16). Fraction B cells are the first population to produce a B lineage–restricted population both in vivo and ex vivo (16). Stable commitment to the B lineage thus occurs at the transition from the fraction A to the fraction B stage.
Once hematopoietic progenitors commit to the B lineage, they adopt a stable B cell identity and do not differentiate into other cell types under steady-state conditions (17–19). A stable B cell regulatory network is essential during both early development and the later stages of B cell maturation. During early lymphopoiesis, the B cell–specific regulatory program directs V(D)J gene rearrangement at the Ig loci that generates a clonally restricted BCR. B cells expressing a functional BCR exit the BM and migrate toward secondary lymphoid organs, where they complete their maturation and undergo a sequence of cell fate decisions. An example of a well-understood B cell fate choice in the periphery is the decision of splenic B cells to develop into either marginal zone or follicular B cells. This choice is established by a Notch2-dependent transcriptional shift that substantially alters the phenotype and function of cells while still maintaining a stable B lineage identity (20). On encountering their cognate Ag, follicular B cells can become activated and give rise to germinal center (GC) B cells that can ultimately differentiate into plasma cells or memory B cells (Fig. 1A). Plasma cells are terminally differentiated B lineage cells that confer immediate and long-term humoral immunity by synthesizing copious amounts of Ab. Memory B cells are responsible for long-lived humoral immunity and undergo enhanced differentiation into plasma cells on secondary immune challenge (Fig. 1A). Notably, the terminal differentiation of mature and memory B cells into plasma cells involves a striking number of transcriptional and epigenetic changes and essentially resembles a lineage switch (21). B lymphocytes thus offer an intriguing case study of a lineage where cell identity is stably maintained throughout development and late-stage maturation but is renounced during terminal differentiation.
B cell lymphopoiesis as a model system for studying lineage maintenance
B cell lymphopoiesis provides an experimentally tractable model system for elucidating the basis of lineage maintenance. One way to assess the stability of B lineage commitment is by exposing B cells to extracellular cues that promote differentiation to non–B cell immune lineages (22). A stable cell fate is defined by its resistance to alternative lineage-inducing cues; thus, stably committed cells should not be affected by the presence of alternative lineage signals. Contrastingly, cells with a compromised maintenance program might be able to respond to alternative lineage signals and, in turn, adopt a new cell fate.
B cell fate stability can be assessed ex vivo by culturing committed developing BM B cells (i.e., Hardy fractions B and C) in a mixture of B cell–specific and alternative lineage-inducing cues (Fig. 1B). A combination of the B cell–specific cytokines, IL-7, Flt3 ligand, and stem cell factor, supports the rapid expansion of early B cell progenitors (23). Adding a non–B cell lineage–inducing cue can divert unstably committed B cells to a different immune lineage. Such cues include the Notch ligand, Delta1, for T lineage differentiation; GM-CSF, M-CSF, or G-CSF for myeloid lineage differentiation; IL-2 for NK cell differentiation; and osteoprotegerin ligand for osteoclast lineage differentiation (24). B cell stability can also be assessed in vivo by transferring committed developing BM B cells or mature splenic B cells into an immunodeficient mouse. The in vivo cytokine milieu of the host can induce unstably committed B cells to adopt a new lineage identity (24) (Fig. 1B).
Loss of B cell lineage identity can be defined in several ways. These include the ability of B cells to (1) dedifferentiate or transdifferentiate into another immune lineage, (2) coexpress a hybrid combination of B and non–B lineage markers, or (3) lose expression of B lineage–defining markers (Fig. 1C). To confirm that the reprogrammed cells come from a previously committed B cell, one can perform V(D)J sequencing or Cd19-Cre–based fate mapping. Because B cells are the only known lineage to completely recombine their IgH and IgL chain loci, examining the V(D)J recombination status of the reprogrammed cells can reveal a previous B cell identity (25). V(D)J rearrangement is a highly ordered process occurring at discrete stages of B cell development (Fig. 1A). Fraction A enriches for cells undergoing DH-to-JH recombination at the IgH locus, and fractions B and C for cells that have a fully recombined VH-DH-JH IgH. Fraction D contains cells that have completed VL-to-JL joining at the IgL chain. Because VH-DH-JH IgH and VL-JL IgL chain assemblies occur exclusively in the committed B cell fractions, the two recombination events can serve as genetic “fingerprints” of committed B cells. An alternative strategy to determine whether the reprogrammed population originates from a committed B cell clone involves Cd19-Cre–based fate mapping. CD19 is a B cell–specific marker, expressed only once cells transition from uncommitted B cell progenitors (fraction A) to committed B cells (fraction B), and therefore only committed B cells and their progeny will be fate mapped by Cd19-Cre expression (12, 17, 19, 26).
Transcriptional regulation of B lineage maintenance
The Greek philosopher Aristotle posited the existence of immutable essences, which, by being distinct from continually changing matter, enforce stable identity. The TF Pax5 is essential for maintaining B lineage identity and is perhaps the best characterized “immutable essence” of any immune lineage. Within the hematopoietic system, Pax5 is exclusively expressed within B cells, beginning in early progenitors and extending to mature B cells (27). Pax5 is necessary for B lineage commitment, because Pax5-deficient B cells arrest at an early progenitor stage that can adopt alternative lineage fates ex vivo and in vivo (22, 28–30), and restoring Pax5 expression suppresses the multilineage potential of Pax5-deficient progenitors and drives B lineage commitment (22). In addition to initiating B lineage commitment, Pax5 is an essential component of the B lineage maintenance program (6, 31).
Functional evidence that Pax5 is required for B cell lineage maintenance
B lineage commitment is not irreversible; rather, it requires sustained regulation by Pax5 (6). In support of this, conditional inactivation of Pax5 in committed developing and mature B cells extinguishes the B cell gene expression program and impairs B cell function (6, 31, 32). Committed B cells that have lost Pax5 can dedifferentiate into uncommitted progenitors that can develop into functional macrophages ex vivo and functional T cells in vivo (6, 31).
Pax5 plays a dual role as a transcriptional activator and repressor (Fig. 2A). On one hand, it promotes a B cell–specific program that ensures the survival and function of B cells. Continuous Pax5 activity is required for the expression of components important for pre-BCR and BCR signaling, MHC class II presentation, cell adhesion migration, and survival (26). Pax5 activates target genes by recruiting the chromatin-remodeling complex BAF, the histone acetyltransferase CBP, and the histone methylation regulator PTIP (33). Together, these complexes induce both the gain of active and the loss of repressive histone marks at enhancers and promoters of Pax5-activated genes (Fig. 2A) (33). On the other hand, Pax5 suppresses genes that maintain an undifferentiated progenitor state or initiate commitment to other lineages. These include Flt3 and Sca1 (multipotent progenitors), Notch1 (T cells), Csf1r (macrophages), and Prdm1 (plasma cells) (31, 34). Pax5 drives gene repression by recruiting the transcriptional corepressors NCoR1 and Groucho (33, 35), and its binding at enhancers and promoters of Pax5-activated genes correlates with removal of active histone marks (Fig. 2A) (33). The mechanisms that determine whether Pax5 acts as an activator or a repressor at target genes is unclear but could involve the relative abundance of coactivators and corepressors at defined developmental stages. In this context, Pax5 controls the identity and function of B cells by regulating relatively different sets of genes in developing and mature B cells. For example, Pax5 activates the pre-BCR–specific gene, Vpreb1, in B cell progenitors, but not in mature B cells, whereas the BCR signaling components, Cd19 and Cd79a, are Pax5 target genes in both early and mature B cells (36). This is unsurprising given that the pre-BCR is required only for early B cell development, whereas CD19 and CD79a are essential for BCR-mediated signaling in both developing and mature B cells.
Apart from its role as a regulator of the B cell–specific gene expression program, Pax5 also controls the three-dimensional chromatin landscape of developing B cells. Pax5 appears to modify the genome architecture in a direct manner that is independent of ongoing transcription (37). The cohesion-release factor WAPL has been shown to be required for cohesin-mediated loop extrusion at the IgH locus (38, 39). A recent study demonstrated that Pax5 can directly regulate WAPL, and that Pax5-driven repression of Wapl is essential for efficient long-range V(D)J recombination and maintenance of the overall chromatin structure of developing B cells. Accordingly, genetic deletion of a single Pax5 binding site within the Wapl promoter led to V(D)J recombination defects and massive alterations of the chromosomal architecture of progenitor B cells (39). It was found that loss of the Pax5-mediated repression of Wapl had minimal effects on gene expression, suggesting that Pax5 controls genome architecture independently of its role as a transcriptional regulator. Given that the proper organization of higher-order chromatin is integral to cell identity, it is possible that the Pax5-enforced three-dimensional genome structure contributes not only to successful execution of the B cell developmental program but also to the overall maintenance of a stable B cell fate (40, 41).
The TF Ebf1 shares several important characteristics with Pax5: (1) it is exclusively expressed within the B lineage from early progenitors and to mature B cells (42), (2) Ebf1 deficiency arrests B cells at the uncommitted fraction A stage (43), and (3) Ebf1−/− progenitor B cells retain a multilineage developmental potential (44). Ebf1 and Pax5 play nonredundant roles during B cell commitment because neither forced overexpression of Ebf1 in Pax5−/− B cells nor that of Pax5 in Ebf1−/− B cells can restore normal B cell development (42, 44). Importantly, Ebf1 directly inhibits non–B cell fates during B cell specification and commitment, and this occurs independently of Pax5 (44). As discussed later, Ebf1 and Pax5 appear to also play nonredundant roles during B lineage maintenance.
Functional evidence that Ebf1 is required for B cell lineage maintenance
Inactivation of Ebf1 in committed B cell progenitors induces dedifferentiation to a multipotent cell state, indicating that, like Pax5, Ebf1 is required for B cell fate maintenance (45). The Ebf1-deficient multipotent cells are highly plastic and can generate functional T cells and innate lymphocytes after adoptive transfer into an alymphoid host (45). Conditional deletion of Ebf1 at later stages of B lymphopoiesis results in the loss of splenic marginal zone B cells and impairment of BCR-dependent survival and proliferation of splenic follicular B cells (42, 46). Ebf1 also plays an essential role in maintaining GC B cells (42). This suggests that Ebf1 is required both for stable B cell commitment and for the functional integrity of committed B cells. Lastly, stable B cell lineage commitment requires a critical dose of both Ebf1 and Pax5. B cell progenitors with combined heterozygous loss of Pax5 and Ebf1 (Pax5+/−Ebf1+/−) can commit to the B cell lineage, yet the committed population retains access to other lineages and can differentiate into T cells ex vivo and in vivo (47).
Similarly to Pax5, Ebf1 functions as a dual transcriptional activator and repressor (Fig. 2B). Ebf1 maintains a B cell–specific program in developing and mature B cells by inducing the expression of B cell–specific adhesion receptors, signal transducers, and TFs (42, 45, 48). The regulatory elements of Ebf1-activated genes exhibit strong correlation between Ebf1 binding and the enrichment of activating histone modifications, as well as the loss of repressive histone modifications (Fig. 2B) (48). In its role as a transcriptional repressor, Ebf1 suppresses multiple B lineage–inappropriate genes. Conditional deletion of Ebf1 in committed developing B cells causes the induction of genes associated with alternative cell states and fates, including multipotent progenitors (Flt3 and Sca1), T cells (Tcf7, Gata3, and Notch1), and innate lymphocytes (Id2) (42, 45). Ebf1 binding at enhancer and promoters of Ebf1-repressed genes correlates with both the loss of activating and the accumulation of repressive histone marks (Fig. 2B) (48). Interestingly, conditional loss of Ebf1 in mature B cells has minimal impact on the overall gene expression program (42). Thus, genes that are under active transcriptional regulation by Ebf1 in B cell progenitors may become epigenetically fixed or may be under the control of other TFs at later stages of B lymphopoiesis (42). In addition, Ebf1 may poise genes for expression at later stages (48). Several Ebf1-occupied genes exhibit an active chromatin state in early progenitors but do not become transcriptionally active until the mature B cell stage (48). Once poised, these genes may no longer require active regulation by Ebf1, which could explain why conditional loss of Ebf1 in mature B cells has minimal effect on transcription (42). The protein partners that enable Ebf1 to play a dual role as a transcriptional activator and a repressor are yet to be identified.
Loss of B cell identity during terminal differentiation into plasma cells
The principal raison d’être of B cells is to generate terminally differentiated plasma cells that secrete Abs and provide humoral immunity against invading pathogens. The transition from a mature B cell to a plasma cell is coupled to extensive transcriptional and epigenetic rewiring (49–51). Some of the changes involve suppression of genes specific to the B cell program (e.g., Pax5, Ebf1, Irf8, and Bach2) and upregulation of genes specific to the plasma cell program (e.g., Prdm1, Xbp1 Irf4, and Igj). The two programs appear to be mutually exclusive as exemplified by the cross-antagonistic relationship of Pax5 and Prdm1 (34, 52). Reactivation of at least some of the Pax5-repressed genes, such as Cd28 and Ccr2, is important for proper plasma cell function (34). Several lines of evidence suggest that plasma cell differentiation is initiated and sustained by Pax5 inhibition: (1) enforced expression of Pax5 decreases the levels of LPS-induced IgM secretion by splenic B cells ex vivo; (2) Pax5 downregulation and the derepression of several Pax5 target genes precedes induction of the plasma cell master TF, Blimp-1; and (3) loss of Pax5 in the chicken DT40 B cell line promotes plasma cell differentiation (52–54).
A more recent study challenges the idea that Pax5 repression is necessary for plasma cell development. The Busslinger laboratory examined plasma cell development in a mouse model in which Pax5 is inserted into the endogenous IgH locus (IgHPax5/+) (55). These mice continue to express high levels of Pax5 even after plasma cell differentiation because the IgH locus remains transcriptionally active in plasma cells. Interestingly, the ectopically expressed Pax5 did not interfere with the robust development of Ab-secreting plasma cells. It did, however, lead to reduced accumulation of long-term plasma cells in the BM and to a modest decrease in IgG secretion. The ability of B cells to establish a plasma cell program in the presence of sustained Pax5 expression suggests that the regulators of the plasma cell fate are dominant over Pax5. Indeed, the Pax5-expressing plasma cells showed normal levels of Prdm1, and Pax5 does not regulate Prdm1 during plasma cell differentiation (52, 55). Overall, the repression of Pax5 is dispensable for plasma cell development but is required for optimal and long-term humoral protection. It remains unclear whether Pax5 repression is sufficient to initiate plasma cell differentiation in vivo. Aicda-Cre–mediated loss of Pax5 in activated B cells decreases the frequency of GC B cells and plasma cells (32). Aicda-Cre, however, is activated early during B cell activation, often before GC formation, leading to a temporal mismatch between the loss of Pax5 and the initiation of plasma cell differentiation (56). It will be interesting to test whether conditional loss of Pax5 using an inducible GC-specific Cre during an immune response is sufficient to drive plasma cell development.
Reprogramming and transdifferentiation of B cells
Ectopic expression of selected TFs can override B cell identity and induce conversion into alternative lineages via transdifferentiation or reprogramming (Fig. 3). Transdifferentiation is the conversion of one specialized cell type into another without an intermediate stem cell–like stage. Reprogramming is the process of reverting differentiated cells into induced pluripotent stem cells (iPSCs). Later, we review several examples of TF-driven B cell transdifferentiation (driven by Hoxb5 or C/EBPα overexpression) and reprogramming (driven by overexpression of the Yamanaka TFs), and highlight the general lessons about cell fate maintenance that can be derived from such studies.
Overexpression of the HSC-specific TF, HOXB5, can convert committed developing B cells into functional T cells in vivo (57). Unlike the Pax5 and Ebf1 loss-of-function effects, this lineage switch does not involve dedifferentiation into a common T-B uncommitted progenitor. Rather, the HOXB5-overexpressing B cells convert directly into early thymic T cell progenitors (Fig. 3). HOXB5 exerts its molecular function by targeting the B cell identity regulators, Pax5 and Ebf1, as well as T cell TFs, such as Lmo2, Nfatc1, Tcf12, Prdm1, and Runx2, and chromatin modifiers, such as Kmt2a, Hdac9, Ldb1, and Smarca5. Although there is no direct evidence that HOXB5-overexpressing B cells can generate other immune cell types, it is tempting to speculate that Hoxb5 can drive conversion beyond the T cell lineage. Given that Hoxb5 is specific to HSCs and is not expressed in any mature immune lineages, it likely exerts a general inhibitory effect on B cell maintenance.
Enforced expression of the myeloid TF, C/EBPα, converts committed B cells into functional macrophages ex vivo and in vivo (12). This occurs via direct transdifferentiation into an intermediate hybrid cell population that coexpresses macrophage and B cell–specific markers (Fig. 3) (58). The C/EBPα-driven B cell-to-macrophage conversion requires the presence of the myeloid-promoting cytokines IL-3 and M-CSF and does not occur under exclusively lymphoid conditions (12, 59). This suggests that cell fate is regulated by an interplay of transcriptional regulation and extracellular signals, and that cell plasticity might be uncovered only if cells are removed from their physiological environment.
Transient expression of C/EBPα in committed B cell progenitors is sufficient for inducing a stable phenotypic conversion into functional macrophages (60). This indicates that C/EBPα reprogramming can completely extinguish and replace the B cell maintenance program with a self-sustained macrophage-specific regulatory network. This process is thought to occur in two steps: (1) C/EBPα antagonizes Pax5 expression to increase B cell plasticity; and (2) C/EBPα cooperates with PU.1, a TF common to both B cells and macrophages, to establish a macrophage expression program (12, 60). Importantly, forced expression of Pax5 decreased the efficiency of the C/EBPα-driven transdifferentiation, suggesting that Pax5 and C/EBPα can antagonize each other (60). Such cross-antagonism between lineage-specific TFs is known to constitute an important molecular mechanism during cellular commitment and to provide stability to transcriptional programs of binary cell fate choices (61).
One of the most unequivocal demonstrations that the stability of mature cells is not irreversibly fixed comes from the seminal discovery that mature fibroblasts can be reprogrammed to a pluripotent state by the expression of the TFs OCT3/4, SOX2, C-MYC, and KLF4 (OSMK; Yamanaka factors) (4). Subsequently, other somatic cells, including B cells, have been reprogrammed into iPSCs (Fig. 3) (61).
Several interesting features characterize the reprogramming of B cells into iPSCs: (1) early- and late-stage committed B cell progenitors can be efficiently reprogrammed to iPSCs by overexpression of OSMK; (2) OSMK-driven reprogramming of mature B cells requires inhibition of the B cell expression program via either C/EBPα overexpression or Pax5 knockdown; and (3) activated mature B cells are resistant to reprogramming by OSMK unless treated with a methyltransferase inhibitor (25, 62). These findings suggest that B cell development progressively stabilizes the transcriptional and epigenetic B cell program, which, in turn, imposes a greater barrier to reprogramming. This is in line with the ability of C/EBPα overexpression to induce more robust transdifferentiation in progenitor versus mature B cells and the observation that myeloid or lymphoid progenitors have a higher reprogramming efficiency than their respective mature counterparts (63). Two plausible interpretations may explain why knockdown of Pax5 or overexpression of C/EBPα is necessary for OSMK-induced reprogramming of mature B cells: these modifications either induce dedifferentiation to an earlier B cell stage that is more amenable to reprogramming or activate a previously suppressed myeloid expression program that facilitates the efficient induction of a pluripotent state. There is experimental evidence for the latter because C/EBPα has been shown to increase chromatin accessibility and mediate Tet2-dependent demethylation at loci that are important for pluripotency (64). This might also explain why myeloid progenitors are more sensitized to iPSC reprogramming than progenitor B cells (63).
B cell identity as a barrier to oncogenic reprogramming
Supervision of lineage identity serves as a crucial tumorigenic barrier, and loss of stable commitment can have two broad oncogenic consequences: (1) it can initiate dedifferentiation into a more proliferative stem-like state, and (2) it can confer an abnormal cellular plasticity that allows the survival and proliferation of cancer cells. Later, we discuss several examples of how dysregulation of the B cell maintenance program can result in the transformation and development of B cell leukemia or lymphoma.
Conditional deletion of Pax5 in mature B cells induces dedifferentiation into uncommitted hematopoietic precursors that give rise to lymphoma progenitors (6). The dedifferentiated cells owe their lymphogenic potential likely in part to their extensive self-renewal capacity, because Pax5−/− progenitor B cells can undergo >150 cell divisions (65). This suggests that Pax5 is necessary to maintain a nonproliferative state in already committed B cells. Furthermore, mutations in Ebf1 and Pax5 are commonly associated with B progenitor acute lymphoblastic leukemia (B-ALL) in humans. Hypomorphic or loss-of-function Pax5 deletions occur in ∼30% of B-ALLs (66). Somatically acquired mutations in Ebf1 are relatively rare in newly diagnosed B-ALL (4%) but occur at a higher rate (25%) in relapsed B-ALL (67). It has been suggested that Pax5 and Ebf1 exert tumor-suppressor functions by acting as metabolic gatekeepers. Pax5 and Ebf1 regulate genes involved in glucose uptake and metabolism, and Pax5 represses glucose and energy metabolism to levels that are incompatible with the high-energy requirements of leukemic cells (68).
More recently, there has been a growing appreciation that B cell neoplasms are highly plastic and can evade anticancer therapeutics via lineage infidelity. Clinical studies from B-ALL patients have shown that B-ALL cells can acquire a myeloid phenotype after immunotherapy. Treatments with anti-CD19 chimeric Ag receptor T cells or anti-CD3/CD19 bispecific T cell Abs have proven to be highly effective for refractory B-ALL, yet a small subset of patients relapse via a CD19-negative myeloid switch. In some cases, the relapsed myeloid cells carry the BCR IgH-VDJ rearrangement of the original lymphoblastic B cell, suggesting that they arose from a previously committed B cell (69–71). The majority of cases of lineage-switched leukemias harbor chromosomal rearrangements of the mixed-lineage leukemia (MLL) gene, which result in the fusion of MLL with 1 of >60 partner genes (69–72). One of the fusion oncogenes, MLL1-AF9, can reprogram committed CD19+ B-ALL clones into myeloid cells, providing a possible mechanism of how oncogenic mutations can override the B lineage–committed state and allow therapy-driven selection of transdifferentiated myeloid clones (73). Lineage plasticity might not be limited to MLL-rearranged B-ALL, and there are reports of B cell transdifferentiation occurring in other B-ALL subtypes. CD19+BCR-ABL1+ B-ALL clones have been shown to give rise to macrophage-like cells in the presence of myeloid differentiation-promoting cytokines (74). The resulting macrophages were no longer leukemogenic, suggesting that the forced conversion of BCR-ABL1+ B-ALL into terminally differentiated macrophages might have therapeutic applications (74). It would be of interest to elucidate the precise mechanisms by which the MLL and BCR-ABL1 mutations confer plasticity to the B-ALL clones, as well as to discover novel B-ALL mutations that can dysregulate the B lineage maintenance program.
Another clinical example of B cell plasticity influencing oncogenesis comes from non-Hodgkin’s B cell lymphomas that transdifferentiate into clonally related histiocytic or dendritic cell sarcomas carrying the Ig rearrangement of the original tumor (75–77). The secondary tumors appear to have completely suppressed their original B cell identity because they are uniformly negative for Pax5 expression (75). Instead, they have upregulated expression of C/EBPβ, which is essential for myeloid differentiation. Interestingly, the histiocytic and dendritic cell sarcomas can arise before any therapeutic intervention. This suggests that lineage switching of neoplastic B cells might be more common than expected and may not occur solely as an immune escape mechanism.
Finally, loss of B cell identity might influence the genesis of Hodgkin’s lymphoma (HL). Hodgkin and Reed/Sternberg cells, the hallmark cells of HL, are derived from committed mature B cells that carry fully rearranged IgH and IgL chains. However, they lack B lineage–associated markers, including the BCR, CD20, and CD79a, and a subset of them express lineage-inappropriate markers of T cells, dendritic cells, monocytes, or plasma cells (78, 79). Hodgkin and Reed/Sternberg cells nearly always express Pax5; yet, other important B cell TFs such as PU.1 are consistently repressed (80–82). It will be interesting to identify the mechanisms responsible for altering the fate of B cells during the pathogenesis of HL and to determine whether the loss of B cell identity is an essential feature of oncogenic transformation.
Throughout their life, B lymphocytes progress through various developmental stages and adopt distinct cell fates and functional specializations while maintaining a stable B lineage identity. The stability of the differentiated B cell state is tightly controlled by “active” and “passive” regulatory mechanisms that suppress alternative lineage choices and enforce a B cell–specific transcriptional program. The “active” transcriptional regulation by the TFs Pax5 and Ebf1 safeguards B cell identity in both developing and mature B cells, whereas the “passive” epigenetic-based mechanisms become more prevalent during the later stages of B cell maturation. Notably, B cell commitment is not irreversible, and B cells can renounce their stable identity either as part of physiological transition into terminally differentiated plasma cells or as a result of experimental overexpression of non–B lineage TFs, such as HOXB5, C/EBPα, or the Yamanaka TFs. The phenotypic plasticity of B cells is highly relevant to cancer biology, and loss of the B cell program can drive cancer progression and therapy evasion. Elucidating the molecular mechanisms of B cell plasticity will improve our understanding of how B lineage loss and lineage switching occur in the context of associated leukemias and lymphomas.
Although the TFs Pax5 and Ebf1 have been studied extensively, the complete array of regulatory networks that drive B cell commitment and safeguard the stability of the B lineage is yet to be elucidated. The TF Bach2 promotes B cell differentiation by suppressing the myeloid gene set program, yet its role in actively maintaining B cell identity and inhibiting reactivation of a latent myeloid program in committed B cells remains unknown (83). In addition, the role of epigenetic modifications and their cooperative interactions with the B cell–specific transcriptional program remain poorly understood. Likewise, the B cell identity maintenance roles of microRNAs, known regulators of lineage stability in other cell types (84), require further investigation. Finally, little is known about the full extent of B cell plasticity under normal physiological conditions and during immune challenges. In theory, increased B cell plasticity might enable broader phenotypic changes that contribute to the robust adaptations of activated B cells during an immune response.
We are grateful to Dr. Meinrad Busslinger for critically reading the manuscript. We would also like to thank Ryan Smolkin, all members of the Chaudhuri laboratory, and Shireen Saleque for stimulating discussions during the preparation of this review. The figures were generated using a program from Biorender.com.
This work was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health Grants RO1AI072194 and RO1AI124186.
The authors have no financial conflicts of interest.