Background:U2AF1 mutations are very common in MDS, indicating that aberrant splicing of pre-mRNA driven by mutant U2AF1 plays a critical role in MDS pathogenesis. Previous studies demonstrated that U2AF1S34F impaired differentiation of erythrocyte and granulocytes, but the impact on megakaryocytes (MKs) remains unclear. Several studies reported that dysplastic megakaryocytes, especially micro-megakaryocytes, was a negative factor for outcome in MDS. However, the mechanisms underlying megakaryocytic dysplasia remain poorly understood.

Methods: We analyzed the alternative splicing events based on the bone marrow mononuclear cells (BMMNC) RNA-Seq file from healthy control (n=5), MDS without any splicing mutations (MDSWT, n=16) and MDS with U2AF1S34 mutations (MDSS34, n=15). Human BM MKs and cell lines with overexpressed U2AF1S34F were used to explore the mechanism. CHK1 inhibition was achieved by shRNA targeting on CHEK1 and a clinical CHK1 inhibitor.

Results: In our dataset of 1208 MDS patients, U2AF1 mutations were identified in 273 cases (23%), including 226 cases (83%) with the S34F/Y missense mutations. Relative to MDSWT, pathological observation showed that a significant decrease in MK diameter in MDSS34 (P=0.018). Morphological classification of MKs in BM smears also showed a significant increase in the proportion of abnormal MKs (P=0.032). Of note, the proportion of micro-megakaryocytes (diameter < 40μm,micro-MKs) significantly increased(P=0.009). Moreover, we also found that in CD41+CD42+ MKs, MDSS34 have more low-ploidy MKs (2N and 4N micro-MKs) compared to MDSWT, accompanied by less high-ploidy MKs (8N, 16N, 32N and higher-ploidy MKs) by flowcytometry, indicating a closed association between U2AF1S34 and defect in MK polyploidization. U2AF1S34F overexpressed K562 and HEL cell lines was shown to impair the phorbol myristate acetate (PMA) induced polyploidization process, which was used extensively to model megakaryocytic differentiation. The proportion of 8N cells in HEL decreased from 18.5% to 5.5% (P<0.001), whereas in K562 cells, it decreased from 6.5% to 1.9% (P<0.001). Compared to sorted CD41+U2AF1WT cells, the average expression levels of CD41 and vWF in CD41+U2AF1S34F cells were 0.52 (P=0.029) and 0.24 (P=0.008), respectively,which suggesting mutated cells were relatively immature. These results indicate that U2AF1S34 mutation impaired the MK polyploidization and block the maturation.

Transcriptome data showed that the aberrant splicing of genes related to the DNA biosynthetic process pathway were predominantly enriched in U2AF1S34 mutant cases. Splicing of CHEK1 was markedly altered, with the significant decrease of classical isoform (CHEK1-L) and increase of nonclassical isoform of CHEK1 (CHEK1-S) in MDSS34. In single CD41+ MKs from MDSS34 (n=82), there were significant CHEK1 isoform switch relative to patients without U2AF1S34 (n=50) (the mean ratio of CHEK1-S to CHEK1-L: 0.76 vs. 1.33, P=0.002), which was confirmed by RT-PCR. Immunoblot of U2AF1S34F overexpressed HEL and K562 cells showed decreased expression of CHK1 protein (0.49 and 0.53-fold to the wildtype cells, respectively). Furtherly, knockdown of CHEK1 by shRNA recapitulated the phenotype caused by U2AF1S34F. Interestingly, although the decrease of CHK1, higher level of phosphorylated CHK1 were found in K562, HEL and BMMNC from MDSS34. CHK1 inhibitor Prexasertib treatment resulted in a significant increase in the ploidy level of HEL and K562 cells with overexpressed U2AF1S34F. The percent of 8N cells in HEL increased from 12.5% to 19.15% (P=0.014), whereas in K562 cells, it increased from 5.3% to 6.9% (P=0.002).

Conclusion: As we know, we firstly reported the aberrant splicing of cell cycle related gene CHEK1 driven by U2AF1 mutation. In U2AF1S34 mutant MDS, CHEK1 isoform switching resulted aberrant phosphorylated CHK1 may arrest the cell cycle, with impairing the MK polyploidization and blocking the maturation. Therefore, CHK1 inhibitor can promote U2AF1S34F mutant megakaryocyte maturation and may serve as a novel therapy in MDS.

Disclosures

No relevant conflicts of interest to declare.

This content is only available as a PDF.
Sign in via your Institution