Patient-specific enhancers in genetically uniform leukemias Free

In this issue of Blood, Smith et al¹ tackle the question of how cancers with low mutation burden can exhibit profoundly diverse transcriptional signatures and clinical phenotypes. The investigators discover that tumor cells from patients with acute lymphoblastic leukemia (ALL) harboring the KMT2A::AFF1 rearrangement exhibit a highly heterogeneous transcriptional landscape, which is dynamically shaped by the presence of unique enhancers that are patient-specific (see figure).

The KMT2A::AFF1 gene fusion is a hallmark genetic alteration and major driver of high-risk infant and pediatric ALL.² This fusion arises from the chromosomal translocation t(4;11)(q21;q23), which joins KMT2A (formerly MLL) with AFF1 (also known as AF4). Notably, cooperating mutations are uncommon in KMT2A::AFF1 ALL, and despite a low overall mutational burden, these leukemias display marked transcriptional and phenotypic heterogeneity among patients.³ Emerging evidence has highlighted the resulting chimeric protein as a central driver of leukemogenesis, in part by reprogramming enhancer landscapes to disrupt normal hematopoietic development.⁴ 

Enhancers are cis-regulatory DNA elements that augment gene expression by acting as binding platforms for multiple transcription factors and coactivators, which help recruit and assemble the transcriptional machinery at gene promoters. Active enhancers are marked by distinct chromatin features, including histone modifications such as histone H3 lysine 4 mono-methyl (H3K4me1) and H3 lysine 27 acetylation (H3K27ac). Alterations in nonprotein–coding genomic DNA sequences, which through mutation, disruption, or abnormal activation can create or disrupt enhancers and lead to dysregulated gene expression, are increasingly recognized as key contributors to cancer.^5,6 Consequently, enhancer activity and specificity play a critical role in shaping the complexity and diversity of gene regulation.

The authors initially observed marked heterogeneity in chromatin accessibility through the analysis of publicly available RNA sequencing (RNA-seq) and ATAC-sequencing (ATAC-seq) data sets from 24 samples of patients with primary ALL. Enhancer regions exhibited significantly greater variability than promoter regions across all leukemia subtypes, and this variability not only distinguished between different ALL subtypes but also revealed substantial interpatient differences. Next, the authors functionally assessed cancer-specific enhancers and found distinct usage between 2 patient-derived KMT2A::AFF1 ALL cell lines (SEM and RS4;11). In the SEM cells, an intragenic enhancer associated with GNAQ was active and essential for gene expression, but it was absent in RS4;11. Conversely, ARID1B enhancers were present in the RS4;11 cells but not in SEM. Functional CRISPR-Cas9–mediated deletion of these enhancers selectively reduced target gene expression in the respective cell lines. These results were further supported by the research group’s previously established CRISPR-based modeling in human fetal hematopoietic stem and progenitor cells, where they demonstrated that 2 KMT2A::AFF1–driven ALL models created from the same donor showed divergent H3K27ac–marked enhancer landscapes in CD19⁺ leukemic blasts. These findings provide initial evidence for the cell- and donor-specific activity of enhancers in regulating gene expression.

The advent of single-cell profiling has greatly enhanced the resolution of cancer studies, enabling deeper insights into disease-initiating mechanisms. Leveraging this approach, the authors investigated CD19⁺ blasts from 4 samples of patients with KMT2A::AFF1 ALL using single-cell ATAC-seq and RNA-seq and complemented these techniques with high-throughput bulk chromatin immunoprecipitation via the “TOPmentation” assay in 9 additional patient samples to comprehensively profile the epigenetic landscape. They found that the majority of the differentially accessible regions corresponded to putative enhancers. Integration with additional transcriptomic data sets revealed over 3000 differentially and heterogeneously expressed genes across samples. Notably, blasts from 1 patient showed elevated expression of key KMT2A::AFF1 target genes MEIS1 and RUNX2, accompanied by nearby unique enhancers enriched for KMT2A binding. Further validation using high-resolution Micro-Capture-C, a highly specific, targeted chromatin interaction assay, revealed direct physical interactions between these enhancers and the promoters of MEIS1 and RUNX2 in primary leukemia samples and recapitulated the crucial role of enhancers in mediating transcriptional diversity.

A central question emerging from this study surrounds the causes of differential enhancer activity observed across patients. The authors present evidence for the presence of patient-specific single nucleotide variants within enhancer regions; however, these variants alone do not fully account for the widespread epigenetic differences. Additional data suggest that KMT2A, either independently or in complex with cofactors such as AFF1 or PAF1, may contribute to differential enhancer utilization. These findings point to a model in which the presence and/or abundance of alternative or cooperating epigenetic and transcriptomic regulators influence the specificity and affinity of transcriptional machinery at distinct loci to modulate gene expression.⁷ 

This study also highlights the potential prognostic significance of transcriptional and epigenomic heterogeneity in KMT2A::AFF1–rearranged leukemia. The authors show that the elevated expression of key KMT2A::AFF1 target genes, such as MEIS1 and RUNX2, correlates with poorer overall survival, suggesting that enhancer activity can influence disease trajectory. Current risk stratification and treatment guidelines encompass curated genomic panels that focus on recurrent mutations, copy number alterations, and translocations as clinical biomarkers. However, the finding that identical KMT2A::AFF1 translocations can give rise to markedly diverse transcriptional and epigenomic landscapes argues for the integration of additional omics data (eg, transcriptomic and epigenomic profiling) into clinical decision-making frameworks. Furthermore, therapies that target the epigenome, such as hypomethylating agents (eg, azacitidine and decitabine) used in myelodysplastic syndromes and myeloid leukemias, and histone deacetylase inhibitors (eg, vorinostat) for the treatment of lymphomas may yield highly variable responses depending on the epigenetic context of individual patients.^8-10 Therefore, a single genetic hallmark, while informative, may not fully capture the biological complexity of leukemia.

Conflict-of-interest disclosure: O.A.-W. is a founder and scientific adviser of Codify Therapeutics; holds equity in and receives research funding from Codify Therapeutics; has served as a consultant for Amphista Therapeutics, Magnet Biomedicine, and AstraZeneca; is on the scientific advisory board of Envisagenics Inc, Harmonic Discovery Inc, and Pfizer Boulder; and has received prior research funding from Nurix Therapeutics and Minovia Therapeutics, unrelated to the current article. T.K. declares no competing financial interests.