The case for targeting enhancer regions in high-risk TCF3::HLF-positive B-cell acute lymphoblastic leukemia

Abstract

Overall survival for children diagnosed with acute lymphoblastic leukemia (ALL) is now above 90%. However, this statistic can be misleading because not all genetic subtypes of ALL patients share in this success. Amongst the worst are children diagnosed with TCF3::HLF-positive B-cell ALL (B-ALL), arising from the translocation t(17;19)(q22;p13). Although rare at less than 1% of cases, these patients have a 5-year event-free survival of approximately 25%, respond poorly to induction chemotherapy and essentially fatal at relapse.^1-3 This TCF3::HLF fusion has been shown to rewire the transcriptional program in committed lymphoid progenitors towards a stem-like, myeloid-biased state, overwhelmingly by binding to enhancer elements in intronic and intergenic regions rather than at promoters.^{2, 4} However, direct links between TCF3::HLF-bound enhancers and target genes have remained unclear. Furthermore, progress on understanding the early steps of TCF3::HLF-driven leukemogenesis has in part been due to the inability to generate a genetically engineered mouse model.⁵ Now two new studies^{6, 7} provide new mechanistic insight on how the TCF3::HLF fusion drives leukemogenesis and supports cell proliferation.

In the first study by Priebe et al.,⁶ the authors provide a comprehensive three-dimensional map of the TCF3::HLF regulome. Using TCF3::HLF-HiChIP in HAL-01 cells, they demonstrate that 94% of bound sites were within intronic or intergenic regions enriched at chromatin with H3K27ac and H3K4me1 that demarcate enhancers. The authors then ranked the interactions by cumulative contact frequency and found MEF2C (myocyte enhancer factor 2C) at the very top of this list. The functional importance of this interaction and MEF2C was validated using CRISPR-mediated disruption of the HLF binding motif within the MEF2C intragenic enhancer. This selectively decreased the proliferation of HAL-01 cells but not in 697 cells harboring the related TCF3::PBX1 fusion. Direct MEF2C knockout reproduced the same block in HAL-01 and a second TCF3::HLF-positive cell line YCUB-2 and was equivalent to loss of TCF3::HLF itself. This new data suggests that a single enhancer–promoter interaction in the MEF2C locus sits at the apex of the TCF3::HLF oncogenic hierarchy, sustaining a self-renewal program as essential to leukemia maintenance as the fusion protein itself (Figure 1A).

In the second study by Suzuki et al.,⁷ the authors independently address two other critical knowledge gaps: the lack of a faithful murine model and the unexplained clinical association of TCF3::HLF B-ALL with hypercalcaemia and osteolytic bone disease. Using induced leukocyte stem cells transduced with TCF3::HLF, the authors generated a mouse model that recapitulates aggressive pro-B-like leukemia and also an osteolytic bone phenotype.

Transcriptomic profiling identified interleukin-1β (IL-1β) as the most specifically upregulated cytokine in TCF3::HLF versus TCF3::PBX1 B-ALL. Consistent with the work by Priebe et al., this expression is also driven by an enhancer bound by TCF3::HLF, this time within the intronic region of the IL1B gene locus. This increased IL-1β expression is implicated in not only increasing cell proliferation but also upregulating receptor activator of nuclear factor κB (NF-κB) ligand (RANKL) that is responsible for osteoclast activation and the observed bone destruction (Figure 1B). Whilst this second study does not show direct enhancer–promoter looping, indirect evidence comes from CRISPR deletion of this enhancer region that abolished IL-1β expression and reduced leukemic cell proliferation. Furthermore, IL1B knockout in leukemic blasts transplanted into mice showed prolonged survival, prevented bone destruction, and hematopoietic niche disruption. Taken together, these studies reveal that TCF3::HLF creates at least two distinct, leukemia-specific regulatory hubs. One centered on MEF2C and one on IL-1β that are both driven by intragenic enhancer elements uniquely accessible in TCF3::HLF B-ALL.

How does this new knowledge provide an opportunity to improve the survival of patients diagnosed with TCF3:HLF-positive B-ALL? One testable strategy would be to use IL-1β blocking antibodies that would have the added benefit of addressing the bone disease that now has a mechanistic explanation. Unfortunately, directly targeting MEF2C is more complicated. It is an essential transcription factor for cardiac morphogenesis and cortical neuronal development^8-10 and therefore global MEF2C inhibition would almost certainly be prohibitively toxic.

One exciting development is the potential to target the short non-coding enhancer RNAs (eRNAs) that are transcribed by active enhancers.^{11, 12} These eRNAs would be unique and therefore minimize on-target off-tumor toxicity related to direct targeting of MEF2C. Proof-of-concept for this strategy comes from an unrelated study also published this year on sickle cell disease. Here, Wang et al.¹³ show that the erythroid-specific +58 intronic enhancer of BCL11A, the same locus targeted by the first approved CRISPR gene therapy for sickle cell disease, expresses unique eRNA transcripts that maintain a higher order chromatin rosette protecting BCL11A from Polycomb-mediated silencing. Depletion of this eRNA using modified antisense oligonucleotides (ASOs) triggers H3K27me3 spreading and silencing of BCL11A expression without genome editing, achieving meaningful fetal hemoglobin reactivation in primary erythroblasts from sickle cell patients.

Translating this to the TCF3::HLF context would require the MEF2C and IL1B enhancer eRNAs to be precisely characterized using techniques such as PRO-cap and PRO-seq thereby enabling specific ASOs to be designed (Figure 1). The critical experiment would then be to determine whether these ASO targeting the unique eRNAs phenocopy the anti-proliferative effect of CRISPR-mediated enhancer disruption by decreasing MEF2C or IL1B gene expression. If so, this could then help inform novel combinations that combine eRNA-directed silencing with small molecule inhibitors such as venetoclax that blocks BCL2 and itself also a direct TCF3::HLF transcriptional target. Perhaps then, this high-risk subtype might also be associated with overall survival above 90% seen more broadly across B-ALL.

Zahra Sheybani: Conceptualization; writing original draft; writing review editing. Charles E. de Bock: Conceptualization; writing review editing; writing original draft; supervision.

The authors declare no conflicts of interest.

No funding was received for this publication.

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.