Single-Cell Multi-Omics Reveals Anticipated Erythroid Differentiation and Transcriptional Regulation in Differentiation Trajectories of SF3B1 - JAK2/MPL Mutated Cells in MDS/MPN-RS-T

BACKGROUND. Myelodysplastic/Myeloproliferative Neoplasm with Ring Sideroblasts and Thrombocytosis (MDS/MPN-RS-T) is a rare hematologic malignancy with an uncertain pathobiology. Previous studies have identified coexisting driver mutations in proliferative (JAK2/MPL) and dysplastic (SF3B1) genes, contributing to its mixed myelodysplastic/ myeloproliferative neoplasm phenotype (PMID: 38714876). Despite these findings, the effects of these simultaneous mutations on the biology of hematopoietic stem cells (HSCs) and their differentiation trajectories remain unclear. The increased risk of progression to acute myeloid leukemia (AML) in patients with both SF3B1 and JAK2/MPL mutations compared to those with only SF3B1 mutations, combined with the lack of consolidated therapies, makes this a subject of significant interest for identifying molecular mechanisms that could be targeted. The aim of this study is to investigate the effect of co-mutations on clonal dynamics, differentiation trajectories, and transcriptional processes at the single-cell level.

METHODS: We used Target-seq (PMID: 30765193), a single-cell multiomics approach that combines genotyping and transcriptomics on bone marrow CD34+ cells from six MDS/MPN-RS-T patients, paired with 10X 3’ scRNA-seq and scATAC-seq.

RESULTS: Single-cell analysis revealed a significant increase in erythroid differentiation in cells carrying both driver mutations compared to wild-type (WT) cells and those with only the SF3B1 mutation (p<0.001). This phenotype was confirmed through the analysis of transcription factors (TFs) and the activity of their putative downstream target genes, known as regulons. At the level of hematopoietic stem and progenitor cells (HSPCs), we observed a decrease in the activity of classical stemness regulons such as JUN, FOS, and IRF1, accompanied by an increase in erythroid regulators like GATA1 and HES6 in cells with double mutations compared to WT cells (Wilcoxon test, p<0.01). Additionally, RUNX3 and ZBTB7A emerged as potential drivers of direct erythroid differentiation at the HSPC level. Furthermore, gene activity modeling from scATAC-seq confirmed an increase in promoter and enhancer accessibility of genes involved in erythroid differentiation.

Analyses of single-cell states revealed an additional differentiation trajectory branch from megakaryocyte-erythroid progenitor cells (MEPs) to megakaryocytes (MKs). In double-mutated cells, we identified a commitment of progenitor cells to megakaryocyte differentiation through the activation of SRF and TAL1, which could explain the presence of thrombocytosis in patients. Cell differentiation regulation was enhanced by the activation of proliferative pathways (Myc, E2F targets) and a reduction of apoptosis along the differentiation trajectory in cells carrying both mutations compared to WT and SF3B1 mutated cells (FDR-adjusted p<0.01).

Using a machine learning approach, we identified a gene signature capable of detecting double-mutated cells through these trajectories. This signature, when applied to matched 10X experiments, increased the number of genotyped cells and confirmed previous results, identifying several pathways involved in mitochondrial membrane rearrangement, potentially responsible for ring sideroblast development.

CONCLUSION. By employing an innovative single-cell approach that integrates genomics, transcriptomics, and epigenomics, we have unraveled the intricate transcriptional regulation of mixed phenotype neoplasms such as MDS/MPN-RS-T. Our study underscores the synergistic impact of JAK2/MPL and SF3B1 mutations in disease progression, primarily through an early commitment to dysplastic erythroid differentiation. This progression is orchestrated by the activation of specific transcriptional and proliferative pathways, in conjunction with oxidative stress and mitochondrial biology. Importantly, our research led to the identification of a distinct gene signature that can detect the presence of these mutations using single-cell transcriptomic data, bypassing the need for expensive single-cell genotyping approaches. These findings reveal potential molecular targets, creating new opportunities for innovative diagnostic and therapeutic strategies for MDS/MPN-RS-T.