AFFILIATIONS
1 Fred Hutch Cancer Center, Seattle, Washington, United States, University of Washington, Seattle, Washington, United States
2 Fred Hutch Cancer Center, Seattle, Washington, United States
3 Fred Hutch Cancer Center, Seattle, Washington, United States
4 Fred Hutch Cancer Center, Seattle, Washington, United States
5 Fred Hutch Cancer Center, Seattle, Washington, United States
6 Fred Hutch Cancer Center, Seattle, Washington, United States
7 Fred Hutch Cancer Center, Seattle, Washington, United States
8 Fred Hutch Cancer Center, Seattle, Washington, United States
9 University of Washington, Seattle, Washington, United States
10 Fred Hutch Cancer Center, Seattle, Washington, United States, University of Washington School of Medicine, Seattle, Washington, United States
11 Fred Hutch Cancer Center, Seattle, Washington, United States
12 Fred Hutch Cancer Center, Seattle, Washington, United States, University of Washington School of Medicine, Seattle, Washington, United States
13 Van Andel Institute, Grand Rapids, Michigan, United States
14 Fred Hutch Cancer Center, Seattle, Washington, United States
15 Fred Hutch Cancer Center, Seattle, Washington, United States
16 Fred Hutch Cancer Center, Seattle, Washington, United States, University of Washington School of Medicine, Seattle, Washington, United States
17 Fred Hutch Cancer Center, Seattle, Washington, United States, University of Washington School of Medicine, Seattle, Washington, United States
18 Fred Hutch Cancer Center, Seattle, Washington, United States, University of Washington School of Medicine, Seattle, Washington, United States
ABSTRACT
Introduction: Reciprocal chromosomal translocations involving lysine methyltransferase 2A (KMT2A) are prevalent driver mutations in infant acute lymphoblastic leukemia (ALL), yet how hematopoietic stem/progenitor cells (HSPCs) harboring these fusions progress from latency to overt leukemia remains poorly understood. Prior models have largely relied on artificial overexpression systems that fail to capture the temporal dynamics and cellular heterogeneity of leukemic evolution from endogenous translocations. We leveraged an alternative CRISPR-based model to generate the endogenous KMT2A-AFF1 gene fusion in human cord blood HSPCs, enabling investigation of the cellular and molecular changes driving transformation in a system that more faithfully recapitulates the native timing, dosage, and cellular context of leukemia initiation and progression.
Methods: CRISPR-Cas9-edited cord blood cells fusing KMT2A exon 11 to AFF1 exon 4 engrafted in NSG mice, producing a bone marrow phenotype consistent with pediatric B-ALL (CD19+CD34+CD10-NG2+), confirmed by flow cytometry and Wright-Giemsa staining. Serial transplantation into secondary recipients accelerated disease kinetics, enabling us to model progressive leukemogenesis. When projected onto published single-cell datasets of patient KMT2A-AFF1 and other fusion-driven leukemias, both primary and secondary engrafted cells clustered with KMT2A-AFF1 patient cells and were classified as KMT2A-AFF1 by label transfer, supporting that our model recapitulates a clinically relevant disease state.
Results: CITE-seq (Cellular Indexing of Transcriptomes and Epitopes by Sequencing) revealed distinct population shifts between primary and secondary transplants, including loss of CD34 in secondary transplants. Projection onto the BoneMarrowMap reference of normal B-cell development (Iacobucci*, Zeng*, Gao*, Garcia-Prat*, et al., Nature Cancer 2025) confirmed that primary and secondary transplants represent distinct, developmentally divergent KMT2A-AFF1 disease states, with secondary transplants showing a more differentiated transcriptional signature. Whole-genome sequencing further revealed CDKN2A silencing in secondary transplants due to a de novo genomic deletion not present in primary transplants, mirroring a recurrent, prognostically unfavorable lesion in KMT2A-rearranged B-ALL patients. Despite this more mature transcriptional signature, secondary transplant cells retained the ability to propagate disease upon serial transplantation, indicating retained self-renewal capacity and a partial rather than complete differentiation block.
Additionally, canonical KMT2A-AFF1 targets HOXA9 and IRX1 were significantly downregulated in the secondary transplants, consistent with low-HOXA9/IRX1 leukemia subsets described in patients. Previous studies have linked low HOXA9 expression to KMT2A breakpoints within the telomeric exon 11-intron 12 region. Although our model shares this same breakpoint, it exhibits progressive, transplant generation-dependent HOXA9 silencing, indicating that the breakpoint position alone is insufficient to determine HOXA9 expression and implicating additional epigenetic regulation. Fiber-seq-based chromatin architecture profiling indicated that the HOXA9 locus was more accessible in primary transplants but gained 5mC methylation in secondary transplants, suggesting progressive epigenetic silencing as a potential mechanism underlying the observed disease state transition. Notably, targeted pull-down enrichment detected the KMT2A-AFF1 fusion transcript at low but consistent levels in both transplant groups despite declining target gene expression, raising the possibility that the fusion protein is redistributed to alternative genomic sites as disease progresses. CoCUT&Tag profiling will map KMT2A-AFF1 co-occupancy with fusion complex-specific partners across primary and secondary samples to directly test this hypothesis.
Conclusions: Altogether, our integrated approach combining an endogenous CRISPR-generated translocation, serial transplantation, and multiomic profiling reveals key genomic and epigenetic changes underlying KMT2A-AFF1 leukemic transformation and disease progression, providing a physiologically relevant system to identify molecular vulnerabilities for early intervention.
DISCLOSURES
Mark A Mendoza, MS: No relevant conflicts of interest to declare.