Integrative scATAC-seq and mtDNA mutation analysis reveals disease-driven regulatory aberrations in AML

Abstract

Disturbances in transcriptional regulation during hematopoiesis can lead to aberrant hematopoietic differentiation and potential leukemogenesis. Given the high relapse risk of acute myeloid leukemia (AML), understanding its progression mechanisms and identifying prognostic biomarkers are crucial for improving treatment outcomes. Here, we applied a mitochondrial single-cell assay for transposase-accessible chromatin with sequencing mitochondria single-cell ATAC-seq (mtscATAC-seq), supplemented with single-cell RNA-seq, to comprehensively characterize AML cells. By constructing a single-cell hematopoiesis reference and mapping tumor cells to it, we identified significant alterations in chromatin accessibility of cis-regulatory elements (CREs) associated with AML differentiation. Using an in-house developed algorithm, cisGRN, we analyzed CRE dynamics and discovered that mutations in the zinc finger domain of WT1 are associated with decreased chromatin accessibility and hypermethylation at the regulatory regions of their target genes, leading to gene downregulation. Functional validation confirmed the pathogenic effects of five WT1 zinc finger mutations: R467Q, R467L, R467W, H470Y, and H470R. Additionally, we identified a CRE mutation that emerges early in hematopoiesis, introducing a new binding motif for CCAAT enhancer binding protein beta (CEBPB), which facilitates its binding to the enhancer, activating enhancer activity and GATA binding protein 4 (GATA4) expression, thereby promoting AML proliferation. These effects were validated through luciferase reporter assays and prime editing experiments. Using mitochondria DNA tracing and machine learning modeling, we identified relapse-associated clones resembling leukemia stem cells (LSCs) and uncovered marker genes in the clones that more accurately predict drug resistance and relapse risk. Collectively, our findings underscore the utility of combining mtscATAC-seq, lineage tracing, and gene regulatory network (GRN) analysis to uncover aberrant regulatory changes and track relapse-prone AML clones.