Abstract
Myelodysplastic syndromes (MDS) are a group of neoplasms that are ineffective in generating multiple lineages of myeloid cells and have various risks to progress to acute myeloid leukemia. Recent genome-wide sequencing studies reveal that mutations in genes of splicing factors are commonly associated with MDS. However, the importance of these splicing factors in hematopoiesis has been unclear and the causal effect of their mutations on MDS development remains to be determined. One of these newly identified genes is SRSF2, and its mutations have been linked to poor survival among MDS patients. Interestingly, most of SRSF2 mutations occur at proline 95 and the majority of these mutations change this proline to histidine (P95H). Given that SRSF2 is a well-characterized splicing factor involved in both constitutive and regulated splicing, we hypothesize that SRSF2 plays an important role in normal hematopoiesis and the SRSF2 mutations induce specific changes in alternative splicing that favor disease progression. We first examined the role of SRSF2 in hematopoiesis by generating Srsf2 null mutation in mouse blood cells via crossing conditional Srsf2 knockout mice (Srsf2f/f) with blood cell-specific Cre transgenic mice (Vav-Cre). The mutant mice produced significantly fewer definitive blood cells (10% of wild type controls), exhibited increased apoptosis in the remaining blood cells, and died during embryonic development. Importantly, we detected no hematopoietic stem/progenitor cells (lineage-/cKit+) in E14 fetal livers of Vav-Cre/Srsf2f/f mice. These results indicate that SRSF2 is essential for hematopoiesis during embryonic development. We next examined the role of SRSF2 in adult hematopoiesis by injecting polyIC into mice that carry a polyIC inducible Cre expression unit. Unexpectedly, after multiple polyIC treatments, the Srsf2f/f mice stayed alive during several months of observation. Time course genotyping analyses of polyIC treated mice revealed an increased rate of incomplete Srsf2 deletion in peripheral blood cells. These observations suggest that Srsf2 ablation did not cause immediate cell lethality in differentiated blood cells, but the gene is indispensable for the function of blood stem/progenitor cells. Since mutations of splicing factors are generally heterozygous in MDS patients, we also examined mice with Srsf2+/- blood cells. No obvious defect of hematopoiesis was observed under normal conditions or in response to stress with 5-FU treatment and sublethal irradiation. To gain molecular insight into the splicing activity of MDS-associated mutant forms of SRSF2, we performed large-scale alternative splicing surveys by using RNA-mediated oligonucleotide annealing, selection, and ligation coupled with next-generation sequencing (RASL-seq) previously developed in our lab, which offers a robust and cost-effective platform for splicing profiling. Compared to vector transduction controls, we found that overexpression of both wild type and P95H SRSF2 induced many, but distinct changes in alternative splicing in lineage-negative bone marrow cells, and importantly, we noted several changes in genes with known roles in hematopoietic malignancies that were uniquely induced by the mutant SRSF2. To further link the mutations to altered splicing in MDS patients, we also applied RASL-seq to a large number of MDS patient samples with or without mutations in SRSF2 or other splicing regulators. The data revealed a specific set of alternative splicing events that are commonly linked to MDS with splicing factor mutations. These findings strongly suggest that many of these mutations in splicing regulators are gain-of-function mutations that are causal to MDS. In conclusion, we report that SRSF2 plays an essential role in hematopoietic stem/progenitor cells and that the MDS-associated mutations in SRSF2 have a dominant effect on RNA alternative splicing. These findings provide functional information and molecular basis of SRSF2 and its MDS-related mutations in hematopoiesis and related clinical disorders.
Disclosures:
No relevant conflicts of interest to declare.
Integrative Deep Models for Alternative Splicing