BCL11B, the Cerberus of human leukemia

In this issue of Blood, Di Giacomo et al provide support for BCL11B as an important oncogene for early human acute leukemia with both myeloid and lymphoid features.¹ 

In Greek mythology, Cerberus (in Greek, κερβερος; Kérberos) is the 3-headed watchdog who protects the dead from leaving the Underworld. Recent findings on lineage-promiscuous acute leukemia reveal the BCL11B locus as a multiheaded beast that can drive a variety of immature leukemias, including undifferentiated acute leukemia (AUL), minimally differentiated acute myeloid leukemia (AML M0/M1), mixed-phenotype acute leukemia with T myeloid features (T/M MPAL), early thymic progenitor (ETP) acute lymphoblastic leukemia ALL (ALL), and T-cell ALL (T-ALL) by virtue of different pathogenic mechanisms. T/M MPAL and ETP ALL are closely related, lineage-promiscuous diseases, both on the (epi)genetic and transcriptomic levels, that may originate from a common precursor cell.²  T/M MPAL and ETP ALL tumors are frequently characterized by ETV6, NUP214-ABL1, or CALM-AF10 fusion products and mutation rates of WT1, FLT3, RAS, and JAK/STAT, which are estimated to be the same, whereas NOTCH1-activating mutations and CDKN2A/B loss-of-heterozygosity aberrations are nearly absent, unlike in patients with T-ALL.^2,3 

Since its initial discovery as a translocation partner, recurrent chromosomal rearrangements between the T-cell lineage commitment factor BCL11B and the TLX3 oncogene or, less frequently, NKX2-5, NKX2-1, or PU.1 have been found in patients with T-ALL. These rearrangements highjack the BCL11B enhancer that is normally activated during early T-cell development in the thymus. This enhancer, located 1 Mb distal of the BCL11B gene, drives ectopic expression of these oncogenes as initiating events of T-ALL leukemogenesis. These rearrangements thereby inactivate 1 functional BCL11B allele. Approximately 16% of T-ALL patients accumulate inactivating mutations in BCL11B, strongly implying that BCL11B acts as an important tumor-suppressor gene during early T-cell pathogenesis.^4,5 

In recent studies, novel t(2;14)(q22;q32) translocations that affect the BCL11B locus were identified in patients with lineage-promiscuous AUL, T/M MPAL, or ETP ALL, as well as in patients with poorly differentiated AML, in whom T-cell markers may also be expressed.^2,6-9  In contrast to the BCL11B enhancer highjacking by oncogenes, as seen in patients with T-ALL, the t(2;14) translocation yields an in-frame ZEB2-BCL11B fusion product that is driven by the ZEB2 promoter.^2,6,8  It leads to the misexpression of BCL11B in early progenitor cells where the BCL11B enhancer is not normally active. Although ZEB2 activation in an early T-cell context drives T-cell pathogenesis with enhanced interleukin-7 signaling in mice,⁹ ZEB2-BCL11B fusions in lineage-ambiguous leukemia patients are chiefly found in combination with activating FLT3 mutations.^2,8 

Di Giacomo et al provide further support for an oncogenic role of BCL11B in a subgroup of patients with lineage-ambiguous leukemia. Novel recurrent rearrangements were identified in which transcriptional regulatory sequences of various loci were highjacked by the BCL11B gene itself in nearly 4% of AML M0/M1 and 3.6% of patients with T-ALL, including those with T/M MPAL, ETP ALL, and rare cases of T-ALL.¹  In addition to (2;14)(q22.3;q32), other rearrangements that were identified included t(6;14)(q25.3;q32), t(8;14)(q24.2;q32), and t(7;14)(q21.2;q32), which repositioned ARID1B, BENC-cMYC, or CDK6 regulatory sequences upstream or downstream of the BCL11B gene, respectively. All these rearrangements result in high expression of BCL11B. These patients with so-called BCL11B-activated acute leukemia share a unique expression signature, including a JAK/STAT signature, that is different from that of other patients with AML, ETP ALL, and T-ALL. Leukemia cells from BCL11B-activated patients express stem cell antigens, including HLA-DR, CD117, and/or CD34, together with T-cell and myeloid markers, such as CD2, CD7, CD13, and/or CD33. In addition to previously reported findings,^2,8  all BCL11B-activated patients carried mutations in FLT3 in the absence of NOTCH1-activating mutations. Recurrent mutations in WT1, DNMT3A, and TET2 were found in 44%, 33%, and 22% of BCL11B-activated patients, respectively.

This study integrates the data presented by Montefiori et al¹⁰  at the late-breaking abstract session during the 2020 meeting of the American Society of Hematology. Genome-wide RNA and whole-exome sequence analyses of a large pan–acute leukemia cohort comprising 2573 patients led to the identification of 60 patients who had a similar profile (typically CD7⁺, CD2⁺, CD5⁻, CD1a⁻, cCD3⁺, CD8⁻, cMPO^+/−, and myeloid/stem cell marker positive). These 60 patients included 25 T/M MPAL, 20 ETP ALL, 8 AML, and 2 UAL, 80% of whom harbored FLT3 alterations. Ninety-three percent of the patients carried rearrangements that produced ZEB2-BCL11B or RUNX1-BCL11B fusion transcripts or activated BCL11B by rearrangements to regulatory sequences of ARID1B, the BENC distal cMYC enhancer at CCDC26, CDK6, ETV6, or SATB1. Another 21% of the patients identified had multicopy tandem duplications of a 2.5-Kb region, denoted as the BCL11B enhancer tandem amplification, that is located 700 Kb distal of the BCL11B gene that activates BCL11B transcription.¹⁰ 

These combined studies show that these novel BCL11B gene aberrations can drive oncogenesis, which yields lineage-promiscuous acute leukemias with myeloid and T-cell lymphoid features, including UAL, T/M MPAL, AML, and ETP ALL in patients (denoted as BCL11B-activated patients; see figure). BCL11B therefore seems an essential oncogene for these BCL11B-activated patients that may originate from early hematopoietic progenitor cells, in contrast to T-ALL, which originates from recent thymic progenitor cells or its offspring cells in the thymus. In patients with lineage-ambiguous acute leukemia, BCL11B is activated upon its repositioning in close proximity to promoters or enhancers that are active at the precursor cell stage and that remain active during early myeloid/lymphoid stages. Although BCL11B-activated patients in general were more resistant to genotoxic agents, FLT3 tyrosine kinase inhibitors or JAK/STAT inhibitors may provide the clinical power of Hercules to slay the deranged BCL11B Cerberus locus in these lineage-promiscuous leukemias.¹