Background and Purpose: The molecular pathogenesis of progression of myelodysplastic syndromes (MDS) to secondary acute myeloid leukemia (sAML) remains incompletely understood. We studied genemutations involving spliceosome in the transformation of MDS to sAML by comparing matched paired MDS/sAML bone marrow samples to determine the roles of SRSF2, U2AF1 SF3B1, and ZRSR2 mutations in the evolution of MDS to sAML.

Patients and Methods: One hundred and forty-nine de novo MDS patients (2 RCUD, 5 RARS, 27 RCMD, 52 RAEB-1, and 63 RAEB-2) were examined forspliceosome mutations at initial diagnosis and 93 patients progressed to sAML with a median follow-up of 16.4 months. Seventy one of them had paired MDS/sAML bone marrow samples available for comparative analyses. Mutational analyses of spliceosome were performed by pyrosequencing with a detection sensitivity of 5% for U2AF1 (exons 2 and 6) and SRSF2 (P95 of exon 1) mutations and by direct sequencing for SF3B1 (exons13-16) and ZRSR2 (whole coding exons 1-11) mutations. Additional 29 gene mutations known to involve in myeloid neoplasms were also analyzed by direct sequencing or next-generation sequencing (NGS, Ion Torrent PGM) followed by Sanger sequencing validation; NGS was mainly used for mutation detection of epigenetic regulators and cohesin complex. The allele burden of targeted genes was determined by pyrosequencing and/or NGS.

Results: The frequencies of U2AF1, SRSF2, SF3B1, and ZRSR2 mutations in the 149 MDS patients were 14.8% (22/149), 12.8% (19/149), 12.1% (17/141), and 7.5% (10/133), respectively. Together, spliceosome mutations occurred in 51.5% of MDS patients at initial diagnosis and were mutually exclusive. All the 5 RARS patients had SF3B1 mutations. Co-existed mutations with epigenetic regulators were detected in 36 patients (52.9%), with RUNX1 in 13 (19.1%), with cohesin complex including STAG2, SMC3, RAD21,or SMC1A, in 13 (19.1%), with signaling pathways including RAS, PTPN11, JAK2, or FLT3-TKD in 4 (5.9%), BCOR in 2, and one each with CEBPa and SETBP1. There were no differences in blood counts, percentage of blasts in bone marrow and blood, WHO subtype, cytogenetic risk group, and IPSS-R between spliceosome-mutated and -unmutated patients. Of the patients carrying spliceosome mutations, only SRSF2 mutation had prognostic impact on predicting a higher risk of sAML transformation (P = 0.025) and sAML-free survival (median 10.8 months, 95% CI 5.1-16.5 months) compared to SRSF2-unmutated patients (median 17 months, 95% CI 8.5-25.5 months, P = 0.050). Of the 71 paired MDS/sAML samples, 37 (52.1%) had spliceosome mutations at diagnosis; the mutational status and patterns remained unchanged in all the 37 matched sAML samples but allele burden was apparently increased in the sAML samples with SRSF2 (P = 0.017) or SF3B1 (P = 0.015) mutations. In addition, one ZRSR2 mutant clone and two SRSF2 mutant clones evolved during sAML progression. Acquisition of other mutated genes was found in 37 spliceosome-mutated patients at sAML phases, including RUNX1 in 5, N-RAS in 5, CEBPa in 4, K-RAS in 3, FLT3-ITD in 3, and one each with ASXL1, TET2, STAG2, WT1, PTPN11, CBL, FLT3-TKD, and C-FMS. Notably, of the 3 patients acquiring spliceosome mutations, none gained other mutated genes during sAML transformation.Clonal expansion of other mutated genes were observed in 4 cases with RUNX1, in 2 with TET2, and one each with N-RAS, CEBPa, ASXL1, SMC1A, and STAG2.

Conclusions: Our results showed that spliceosome mutations occurred in more than half of de novo MDS patients at initial diagnosis. Clonal expansion, evolution, or unchanged allele burden of spliceosome mutations might occur during sAML transformation with frequent acquisition of additional mutated genes. SRSF2 mutation predicted a higher risk and more rapid sAML transformation. (Grants support: NHRI-EX103-10003NI, MOHW103-TD-B-111-09 and CMRPG3D1532)

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution