Selective Targeting of Splicing Factor Mutant Hematopoietic Stem and Progenitor Cells Via STAT3 Inhibition

Myelodysplastic syndrome (MDS) is a bone marrow failure disorder driven by dysfunction of hematopoietic stem and progenitor cells (HSPCs). Patient sequencing studies over the last decade have revealed that mutations in splicing machinery predominate in MDS, thus selective targeting of these cells is therapeutically attractive. STAT3 inhibition has been explored previously as a means to eradicate HSPCs in MDS. While efficacy was demonstrated in a subset of samples, the underlying mechanism for this selectivity remains unknown. We examined RNAseq of MDS CD34+ HSPCs with splicing factor mutations versus wildtype, finding alternative splicing and differential expression of STAT3 pathway components. Functionally, we explored if STAT3 signaling represents a novel vulnerability in SF3B1 mutant HSPCs using a multi-model approach of in vivo zebrafish and mouse systems, and in vitro assays of CRISPR-engineered human leukemia K562 cells and primary MDS samples. Utilizing the small molecule STAT3 inhibitor STATTIC, we found that human cells carrying MDS-associated SF3B1 point mutations had heightened sensitivity to STAT3 inhibition compared to wildtype controls. To evaluate the activity of STAT3 inhibition in vivo, we utilized an Mx1-cre conditional knock-in mouse model of mutant SF3B1 (Sf3b1^+/K700E). We demonstrated that in vivo STATTIC treatment selectively depleted Sf3b1 mutant cells over wildtype in vivo. RNAseq of sf3b1 homozygous mutantzebrafish cells revealed conserved dysregulation of STAT3 pathway splicing and target expression. Diminishing Stat3 (via morpholino knockdown, stable mutants, or STATTIC treatment) decreased HSPCs in sf3b1 heterozygotes but not wildtype embryos, demonstrating synthetic lethality between Sf3b1 and Stat3. Our data indicate that SF3B1 heterozygosity, regardless of the type of mutation, confers a heightened sensitivity to STAT3 inhibition in zebrafish, mouse, and human HSPCs. Critically, our data indicate that SF3B1-mutant cells can be selectively killed in vivo while sparing wildtype cells. We sought to rescue HSPCs in sf3b1 homozygous mutant zebrafish, however overexpression of ligands Osm and Il6 or wildtype Stat3 was insufficient. Instead, overexpression of constitutively-active Stat3 partially restored HSPCs, indicating that functional Stat3 signaling downstream of Sf3b1 is critical for HSPC formation. To investigate the specificity of the synthetic lethality for SF3B1, we assessed the STAT3 synthetic lethal interaction with other mutated splicing factors in MDS. Similar to SF3B1, we demonstrated STAT3 synthetic lethality with U2AF1 and SRSF2 heterozygosity in zebrafish and human cells. RNA-sequencing analysis of STATTIC-treated K562 cells revealed an exacerbation of splicing alterations upon STAT3 inhibition that was more pronounced in SF3B1^+/K666N cells compared to wildtype. Even more strikingly, we demonstrated that constitutive activation of STAT3 could partially reverse defective splicing in zebrafish sf3b1 homozygous mutant cells. Mechanistically, these data strongly support coordinated splicing dysfunction as the underlying cause for STAT3-SF3B1 synthetic lethality. Together, we demonstrated a conserved and selective synthetic lethal interaction between STAT3 function and splicing factor defects that represents a novel liability for mutant HSPCs with important implications for MDS treatment.