Mis-splicing of Mdm4 activates TRP53 and impairs HSPC growth in a mouse model of myelodysplastic syndromes Free

Mutations in splicing factor genes (e.g. human U2AF1, SRSF2, SF3B1 and mouse U2af1, Srsf2, Sf3b1) occur in up to 50% of patients with myelodysplastic syndrome (MDS) and 20% of patients with acute myeloid leukemia (AML) and generally confer a poor prognosis. However, in mouse and human models of MDS, we and others have shown that hematopoietic stem and progenitor cells (HSPCs) with splicing factor gene mutations have impaired growth compared to controls. This paradoxical finding suggests that for splicing factor mutant myeloid neoplasms to emerge, mutant HSPCs must first overcome growth-suppressing signals. Understanding how splicing factor gene mutations initially impair the growth of HSPCs is the first step in understanding how mutant cells later adapt, expand, and cause disease. Herein, we show that the impaired growth of primary mouse U2af1^S34F HSPCs is partially caused by hyperactive TRP53 signaling, which may be due to mis-splicing of Mdm4—a key negative regulator of TRP53.

RNAseq analysis of U2AF1^S34F bulk HSPCs showed upregulation of TP53 target genes (mouse gene name Trp53) by GSEA (HALLMARKS) in mice (2 models, 5 independent experiments; median NES=1.95 [range 1.50-2.08], FDR≤0.03) and human CD34+ progenitors (2 models; NES 1.58, 1.99; FDR<=0.01) compared to controls. Single cell RNAseq confirmed that a TRP53 target gene signature was upregulated in diverse HSPC subsets in Mx1-Cre+ U2af1^S34Fconditional knock-in mice, including neutrophil progenitors (p=3.9x10^-33), erythroid progenitors (p=3.7x10^-89), and HSCs (p=2.2x10^-48; N=2 mice/group). To determine if TRP53 activation causes the impaired growth of U2af1^S34F HSPCs, we performed competitive bone marrow transplants with U2af1^S34F donors with either active or inactive TRP53. In 4 independent experiments (mixing mutant and WT congenic cells at a 1:1 ratio), we found that deletion of Trp53 partially rescues the growth of U2af1^S34F HSPCs in vivo and increases persistence of Lin-Sca1+Kit+ progenitor chimerism at 16 weeks from 0.4% (+/-0.4%) to 21.9% (+/-4.9%) (p<0.001, N=10 mice/group). The high basal TRP53 activity in U2af1^S34F HSPCs also suggested they may be preferentially sensitive to further pharmacologic TRP53 activation. Indeed, in two independent competitive repopulation experiments, the clinical grade TRP53 activator sulanemadlin (a stapled peptide inhibitor of MDM2/MDM4) preferentially killed mouse U2af1^S34F HSPCs in vivo, compared to WT congenic controls (p<0.001, N=6-7 mice/group). Given the central role of U2AF1 in the spliceosome, we hypothesized that alternative splicing may influence TRP53 activation in U2af1^S34F HSPCs. We discovered that primary mouse U2af1^S34F HSPCs preferentially express a dysfunctional short splice isoform of Mdm4 that is unable to negatively regulate TRP53: 43% (+/-6.8%) of total Mdm4 transcripts in U2af1^S34F HSPCs vs. 14.8% (+/-1.9%) in WT controls, p<0.001, N=4 mice/group. We therefore hypothesized that the growth of U2af1^S34F HSPCs can be rescued by reducing TRP53 activity via expression of functional full-length MDM4. Indeed, overexpression of MDM4 via the Mdm4^Tg15 transgenic mouse partially rescued the growth of U2af1^S34F HSPCs in an ex vivo competitive co-culture assay: 6 days after mixing HSPCs 1:1 with CD45.1 WT competitors, the persistence of CD45.2 HSPCs increased from 16.3% for U2af1^S34F alone to 36.8% (+/- 3.3%) for U2af1^S34FMdm4^Tg15 vs. 45.1% for WT HSPCs, p<0.001, N=1-2 mice/group; representative of 2 independent experiments.

Our study reveals that splicing factor gene mutations cause impaired cell growth—in part—due to activation of TP53. We show that in mouse U2af1^S34F HSPCs, preferential expression of a dysfunctional Mdm4 splice isoform may drive TRP53 activation, whereas overexpression of full-length MDM4 partially rescues HSPC growth. While co-mutation of TP53 typically occurs in <5% of splicing factor mutant MDS patients, the observation that these patients disproportionately have duplications of chromosome 1q (which includes the MDM4 locus), suggests overexpression of MDM4 may be a mechanism for splicing factor mutant HSPCs to expand. Our ongoing studies are testing whether MDM4 overexpression rescues the growth of U2af1^S34F HSPCs in vivo, and whether splicing factor mutant MDS cells from patients have abnormal splicing of MDM4 or other TP53 regulators. By understanding what drives TP53 activation in splicing factor mutant cells, we aim to identify a vulnerability that can be exploited to kill them.