Whole Transcriptome Analysis Identifies Distinct Gene Expression Profiles between SF3B1^mut and SF3B1 ^wt Myelodysplastic Syndrome with Ring Sideroblasts

Background and aims - The 2016 revised WHO classification incorporated somatic mutation in SF3B1 spliceosome gene within the diagnostic criteria of myelodysplastic syndrome (MDS) with ring sideroblasts (RS). However, SF3B1^wt MDS-RS display significantly different clinical features and outcome from those of SF3B1^mut MDS-RS. Recently, the recognition of SF3B1-mutant MDS as a distinct nosologic entity has been proposed to overcome this limitation.

Methods - To evaluate the biological relevance of this proposal, we studied a consecutive cohort of 132 MDS with RS >5% using a pangenomic approach (targeted-DNA sequencing, genome-wide copy number variation analysis and bulk RNA-sequencing of CD34+ bone marrow mononuclear cells). 16 age-matched healthy individuals and 43 MDS-SLD/MLD negative for both splicing mutation and RS were included in this study as controls.

Results - Unsupervised clustering analysis based on mutation profiles identified two major clusters predicted by SF3B1 mutation (87 MDS-RS-SF3B1^mut and 45 MDS-RS-SF3B1^wt). The most recurrently mutated genes in MDS-RS-SF3B1^wt were TP53(40%), SRSF2(38%), TET2 (33%), ASXL1 (21%) and DNMT3A (12%). SRSF2 and TP53 mutations were found to be mutually exclusive with SF3B1 (p-value <0.05), whereas no difference was found in TET2, DNMT3A and ASXL1 frequencies between MDS-RS-SF3B1 ^mutand MDS-RS-SF3B1^wt subgroups. TP53-mutated MDS-RS exhibited shorter overall survival (median 1.3 years, log-rank p-value<0.0001) compared to SF3B1^mut (median 7.6 years), SRSF2^mut (median 3.4 years) and MDS-RS without the abovementioned aberrations (median 4.4 years). Notably, allelic imbalances analysis of oncogenic variants identified a significant enrichment of TP53 biallelic inactivation in MDS-RS-SF3B1^wt (64% vs 0, p<0.01).

Differential gene expression analysis results were incorporated into a specific expression signature highly predictive of MDS-RS-SF3B1 ^mutand MDS-RS-SF3B1^wt subgroups (Figure 1). The resulting gene clusters were classified in RS-specific genes (cluster 1 and 2) and SF3B1-specific genes (cluster 3 and 4). RS-specific genes comprising heme and hypoxia genes were enriched (Figure 2AB) and correlated with RS percentage (p<0.01). Pathway analysis revealed a specific downregulation of adhesion molecules and an upregulation of G-protein coupled receptor signaling molecules in MDS-RS-SF3B1^mut. Among SF3B1-specific genes (Figure 2CD), we confirmed ABCB7 downregulation and identified new molecular targets that may concur to the pathophysiology of SF3B1-driven myeloid neoplasms.

Conclusions - This study contributes to unveil molecular features of SF3B1-mutant MDS and provides further evidence to support recognition of somatic SF3B1 mutation as a disease-defining genetic lesion.