Abstract
Abstract 273
Myelodysplastic syndromes (MDS) are a heterogeneous group of myeloid neoplasms showing a frequent transition to acute myeloid leukemia. Although they are discriminated from de novo AML by the presence of a preleukemic period and dysplastic cell morphology, the difference in molecular genetics between both neoplasms has not been fully elucidated because of the similar spectrum of gene mutations. In this regards, the recent discovery of frequent pathway mutations (45∼90%) involving the RNA splicing machinery in MDS and related myeloid neoplasm with their rare mutation rate in de novo AML provided a novel insight into the distinct molecular pathogenesis of both neoplasms. Thus far, eight components of the RNA splicing machinery have been identified as the targets of gene mutations, among which U2AF35, SF3B1, SRSF2 and ZRSR2 show the highest mutation rates in MDS and CMML. Meanwhile, the frequency of mutations shows a substantial variation among disease subtypes, although the genetic/biological basis for these differences has not been clarified; SF3B1 mutations explain >90% of the spliceosome gene mutations in RARS and RCMD-RS, while mutations of U2AF35 and ZRSR2 are rare in these categories (< 5%) but common in CMML (16%) and MDS without increased ring sideroblasts (20%). On the other hand, SRSF2 mutations are most frequent in CMML (30%), compared with other subtypes (<10 %) (p<0.001) (Yoshida K, et al, unpublished data).
So to obtain an insight into the genetic basis for these difference, we extensively explored spectrums of gene mutations in a set of 161 samples with MDS and related myeloid neoplasms, in which mutations of 10 genes thus far identified as major targets in MDS were examined and their frequencies were compared with regard to the species of mutated components of the splicing machinery. The mutation status of the 161 specimens was determined using the target exon enrichment followed by massively parallel sequencing.
In total, 86 mutations were identified in 81(50%) in the 8 components of the splicing machinery. The mutations among 4 genes, U2AF35 (N = 20), SRSF2 (N = 31), SF3B1 (N = 15) and ZRSR2 (N = 10), explained most of the mutations with a much lower mutational rate for SF3A1 (N = 3), PRPF40B (N = 3), U2AF65 (N = 3) and SF1 (N = 1). Conspicuously, higher frequency 4 components of the splicing machinery were mutated in 76 out of the 161 cases (47.2%) in a mutually exclusive manner. On the other hand, 172 mutations of the 10 common targets were identified among 117, including 41 TET2 (25%), 32 RUNX1 (20%), 26 ASXL1 (16%), 24 RAS (NRAS/KRAS) (15%), 22 TP53 (14%), 17 IDH1/2 (10%), 10 CBL (6%) and 10 EZH2 (6%) mutations. We examined the difference between the major spliceosome mutations in terms of the number of the accompanying mutations in the 10 common gene targets. The possible bias from the difference in disease subtypes was compensated by multiple regressions. The SRSF2 mutations are more frequently associated with accompanying gene mutations with a significantly higher number of those mutations (N=29; OR 6.2; 95%CI 1.1–35) compared with that of the U2AF35 mutations (N=14) (p=0.038).
Commonly involving the E/A splicing complexes, these splicing pathway mutations lead to compromised 3' splice site recognition. However, individual mutations may still have different impacts on cell functions, which could contribute to the determination of discrete disease phenotypes. It was demonstrated that SRSF2 was involved in the regulation of DNA stability and that depletion of SRSF2 can lead to DNA hypermutability, which may explain the higher number of accompanying gene mutation in SRSF2-mutated cases than cases with other spliceosome gene mutations.
In conclusion, it may help to disclosing the genetic basis of MDS and related myeloid neoplasms that highly paralleled resequencing was confirmed SRSF2 mutated case significantly overlapped common mutations.
Disclosures:
No relevant conflicts of interest to declare.
Efficient genotype compression and analysis of large genetic variation datasets