MDS-Associated SF3B1 Mutations Promote Aberrant Hematopoietic Cell Fate Choice By Disrupting Mediator Kinase Module Component CDK8

Somatic SF3B1 mutations are believed to arise early in the pathogenesis of clonal myeloid disorders such as myelodysplastic syndromes (MDS). Studies show that SF3B1 mutations bias hematopoietic stem and progenitor cells (HSPCs) toward the myeloid lineages and impair terminal erythroid maturation. However, a direct mechanistic link between mutant SF3B1 and cell fate choice in HSPCs has yet to be established. A recent study suggested that mutant SF3B1 alters transcription via impaired transcriptional elongation and RNA polymerase II processivity. Given the central role of transcriptional regulators in modulating lineage-specific gene expression programs and cell fate determination in HSPCs, we hypothesize that mutant SF3B1 promotes aberrant splicing of core transcriptional regulators, which, in turn, disrupt gene regulatory networks and drive biased hematopoietic differentiation.

To test our hypothesis, we first performed a meta-analysis of publicly available and unpublished RNA-seq data from SF3B1-mutant MDS patient samples and cell lines, focusing on alternative splicing analysis of genes with known roles in transcriptional regulation. We found that SF3B1 mutations are associated with repression of CDK8 – a core subunit of the mediator kinase module. Mutant SF3B1 promotes the usage of a cryptic 3’ splice site located 14-basepairs upstream of the canonical splice site of exon 8 of CDK8, leading to its degradation via nonsense-mediated decay. We next evaluated the biological consequences of CDK8 loss by expressing shRNAs targeting CDK8 (CDK8 KD) or a non-targeting control (NTC) into human CD34+ HSPCs followed by functional assays. Using ex vivo colony-forming unit (CFU) and differentiation assays, we found that CDK8 KD HSPCs generated increased colony numbers, myeloid-derived colonies, and increased percentages of CD13⁺CD14⁺ myeloid cells relative to NTC HSPCs. To confirm this phenotype in vivo, we xenotransplanted CDK8 KD or NTC HSPCs into immunodeficient mice. CDK8 KD HSPC recipient mice had 14-fold reduction in human cell chimerism and increased percentages of mature myeloid cells in peripheral blood compared to controls. Additionally, these mice had reduced bone marrow cellularity, higher percentages of common myeloid and granulocyte-monocyte progenitors, and 4-fold increase in CD117 intensity relative to control HSPCs, suggesting CDK8 KD strongly skews HSPCs towards the myeloid lineage at the expense of long-term stem and engraftment potential. To define how CDK8 regulates HSPC transcriptional identity, we performed transcriptomic analysis in CD34⁺ HSPCs treated with CDK8 or NTC shRNAs. CDK8 KD HSPCs had decreased expression of HSPC homeostatic genes (e.g. ID1, ID3, JUN, FOS) and increased expression of genes associated with myeloid differentiation (e.g. CD14, CECAM6, S100A8) and oncogenesis (e.g. CD24). Importantly, our gene sets were found to be deregulated in SF3B1-mutant MDS patients, suggesting the effects of SF3B1 mutations are partly mediated through CDK8 loss. Taken together, our results suggest that CDK8 loss promotes aberrant hematopoietic cell fate choice by deregulating transcriptional networks governing lineage commitment.

Our study identifies CDK8 as an important regulator of HSPC homeostasis and cell fate determination in SF3B1-mutant MDS. CDK8 depletion not only reduces fitness and biases HSPCs towards myelomonocytic lineages but also mirrors phenotypes observed in SF3B1-mutant MDS. This finding directly connects an SF3B1-mutant splicing event to skewed hematopoietic differentiation, shedding new light on MDS etiology.