Sudemycin Selectively Inhibits Growth of Primary Murine Hematopoietic Cells Expressing Mutant U2AF1

Abstract 554

Core components of the pre-mRNA splicing complex (the “spliceosome”) are targets of recurrent somatic mutation in patients with myelodysplastic syndrome (MDS). Using whole genome and targeted resequencing, we previously identified recurrent missense mutations in U2AF1 (encoding U2AF35, the 35 kDa U2 accessory factor) at serine 34 (S34F or S34Y) or glutamine 157 (Q157R) in 8.7% of de novo MDS patients. The mechanism by which mutations in U2AF1 and other recurrently mutated spliceosome components contribute to MDS pathogenesis is unknown, but recurrent mutations in this pathway suggest that it may provide a novel therapeutic target. Pladienolide B and FR901464 are two bacterial products that modulate activity of the spliceosome. The derivatives, E7107 and spliceostatin A, and synthetic analogs (sudemycins), bind SF3B, a complex that includes SF3B1 and six other subunits of the spliceosome. Treatment of cell lines with these agents causes accumulation of aberrantly spliced gene products. Sudemycins have favorable pharmacokinetic properties (improved solubility, chemical stability, and plasma half-life) and induce cell cycle arrest and apoptosis in low nanomolar concentrations in cell lines in vitro. They also inhibit growth in xenograft models in vivo, with no evidence of toxicity in mice treated with pharmacologically active doses. We hypothesize that splicing gene mutations cause alterations in splice isoforms that provide a selective advantage to tumor cells, and that perturbation of this environment with small molecule spliceosome modulators may cause a selective disadvantage for cells with these mutations. To test this hypothesis, we generated MSCV-based retroviruses encoding the wildtype (MSCV-WT-ires-GFP) or S34F mutant (MSCV-MT- ires-GFP) U2AF1 alleles. We infected c-kit+ primary C57BL/6J murine bone marrow cells with these recombinant retroviruses (or empty vector) and cultured them in vitro with cytokine-supplemented media. Consistent with previous reports, vehicle-treated cells (0.1% DMSO) expressing the mutant allele had reduced growth over 5 days, compared to cells expressing the wildtype allele or empty vector (P<0.0001, n=5 technical replicates, two experiments). At doses between 10–150 nM, sudemycin (but not an inactive sudemycin analog) induced a reproducible dose-dependent decrease in cell growth, most significant at 75 nM (p<0.0001, n=5 technical replicates, two experiments), and an increase in apoptosis (p=<0.05, n=3 technical replicates), as measured by annexin V positivity, in cells expressing the mutant, compared to the wildtype allele or empty vector. To study the effect of the U2AF1 mutation and sudemycin exposure on splicing, we analyzed RNA isoforms of Mdm2 and Fmr1, two genes known to undergo alternative splicing. After retroviral transduction, GFP+ cells were sorted and RNA was isolated for RT-PCR analysis at baseline or 2h–6h after exposure to sudemycin (75 nM) or vehicle. The U2AF1 S34F mutation increased the abundance of Fmr1 transcripts utilizing a cryptic splice acceptor, compared to cells transduced with empty vector or wildtype U2AF1 (P<0.05, n=3 technical replicates). Sudemycin treatment decreased Fmr1 cryptic splice site utilization (p<0.05, n=3 technical replicates) and increased Mdm2 exon skipping (p<0.05, n=3 technical replicates) at 2h and 6hr, compared to vehicle-treated cells, with the most significant differences in cells expressing mutant U2AF1. These experiments provide proof-of-concept that cells expressing mutant U2AF1 have increased sensitivity to spliceosome modulators and warrant further investigation as a possible treatment for patients with these mutations. Ongoing experiments include in vivo experiments in mice transplanted with transduced cells, and in vitro treatment of primary clinical samples. Small molecule modulators of the spliceosome may provide a novel therapeutic strategy for MDS/AML patients with spliceosome mutations.