A Highly Selective SF3B1-Targeted Splicing Inhibitor Reduces Human CD34+ Cell Survival and Self-Renewal In Acute Myeloid Leukemia

Myelodysplastic syndromes (MDS) result from ineffective hematopoietic stem cell (HSC) maintenance in an aged bone marrow microenvironment and have a proclivity for evolution to acute myeloid leukemia (AML). Progression to therapy resistant AML is driven by leukemia stem cells (LSC) harboring enhanced survival, dormancy and self-renewal capacity in supportive niches. Seminal next-generation DNA sequencing Results suggest that MDS evolution is controlled by mutations in splicing related genes and epigenetic modifiers of gene expression. However little is known about the cell type and context specific functional effects of these mutations on LSC transcriptional alterations that have been shown to promote MDS/AML progression and resistance to therapies such as 5-azacytidine (Vidaza). Therefore, we investigated the effect of splicing inhibitors on LSC survival and self-renewal 1) during progression of MSD to AML and 2) before and after clinical Vidaza treatment in a bone-marrow stromal co-cultures that recapitulates key aspects of the human LSC niche.

Mouse bone marrow cell lines, transfected to producehuman SCF,IL3 and G-CSF, were used as a stromal monolayers. Then human CD34⁺ cellswere selected from MDS (n=1) and AML primary samples (n=6). As normal controls, CD34⁺cells from cord blood (CB, n=3) or aged bone marrow (n=3) were utilized for the co-culture experiments. Survival and self-renewal of the CD34+ cells were investigated by colony forming and replating assays. Two SF3B1-targeted splicing inhibitors: FD 895 and a FD-analog were added at the initiation of co-culture at concentrations ranging from 0.1 to10 uM.

After 2 weeksof stromal co-culture, none of the compounds demonstrated inhibition of the cell viability. Meanwhile, the splicing inhibitors demonstratedno reduction in survival in cord blood, and minor cytotoxicity toward aged bone marrow, MDS and AML samples showed a dose- and time-dependent significant (up to 80%) inhibition of colony formation. To analyzethe effect of splicing inhibitors on LSC self-renewal, replatingassayswere performed. While compounds at high doses mediated only a slight decrease in colony formation in normal CB and a-BM samples, MDS and AML samples exhibited a dose dependent inhibition of 38.2+/-8.1% of LSC survival (p<0.001) for FD895 and considerably lower 13.8+/-3.6 % of LSC survival for FD analog (p<0.001). Analysis of pre- and post progression samples from the same patient revealed the capacity of splicing inhibitors to diminish LSC survival. In CMML, FD895 induced significantly less cytotoxicity (35% compared with 75%) after progression to AML. Notably, in aged-BM both compounds reduced only CFU-GM survival, but not HSC self-renewal. In sequential primary samples from AML patients collected before and after clinical treatment with Vidaza, naïve samples exhibited similar sensitivity to FD-895 treatment in stromal co-culture models, and LSC survival and self-renewal capacity was reduced following incubation with FD-895. In contrast, following clinical treatment in patients that were responsive to Vidaza, these samples acquired resistance to splicing inhibition. However, patients that were non-responders to Vidaza treatment retained sensitivity to FD-895 treatment.

These data indicate that RNA editing and splicing activities represent novel regulators of LSC self-renewal and survival in LSC supportive niches. These properties can be inhibited using novel splicing inhibitors with minimal toxicity toward normal progenitors.

No relevant conflicts of interest to declare.