Association of SF3B1 with Ring Sideroblasts in patients, In Vivo, and In Vitro models of Spliceosomal Dysfunction

Abstract 457

Ring sideroblasts (RS) are abnormal nucleated erythroblasts characterized by iron granules in mitochondrial cristae. RS are seen in acquired and congenital sideroblastic anemia. Dysfunction in mitochondrial metabolism has been implicated in the pathogenesis of RS. However, the true mechanism leading to RS formation remains elusive. Clonal sideroblastic anemias are usually acquired in the context of MDS: 15% or more RS in the bone marrow (BM) with appropriate morphologic and cytogenetic criteria in MDS is best classified as RARS, but varying quantities of RS (<15%) can also occur in other myeloid malignancies. RARS presenting with thrombocytosis (RARS-T) is a form of MDS/MPN often associated with JAK2, TET2, and MPL mutations. To date, no recurrent mutations or chromosomal defects have been found. We hypothesized that they likely exist in RARS and therefore systematic unbiased application of whole exome next generation sequencing may lead to identification of cryptic clonal mutations of pathogenetic significance. Initially, we applied whole exome sequencing to 15 patients with MDS and found a new somatic mutation in SF3B1 (splicing factor 3b, subunit 1) on chromosome 2q in a case of RARS with high platelets. SF3B1 is a component of the U2-small nuclear ribonucleoprotein complex (U2 snRNP), a part of the U2-dependent spliceosome in eukaryotes. The index mutation resulted in a lysine to glutamic acid substitution (K700E) and its somatic nature was confirmed in non-clonal CD3 cells. Further screening of patients with a similar phenotype (N=81) revealed somatic missense mutations in 9/14 (64%) and 13/18 (72%) patients with RARS and RARS-T, respectively. No mutations were detected in 49 MDS and MDS/MPN patients with <15% RS. All SF3B1 mutations were heterozygous affecting mostly exons 15 [15%] and 14 [7%]; no corresponding hemizygous deletions were found (N=430). Similarly, no mutations were detected in congenital sideroblastic anemias. SF3B1 mutations were associated with JAK2 V617F (4/13; 30%), MPL (3/13; 23%), DNMT3A (1/18; 5%), TET2 (1/18; 5%), ASXL1 (1/18; 5%), and LNK (1/18; 5%) in RARS-T and TET2 (1/14; 7%) and DNMT3A (4/14; 29%) in RARS. There was no difference, adjusted for IPSS in terms of survival and AML progression between wild type (WT) and carriers of SF3B1 but the latter showed a higher frequency of thrombotic events in (p=.04). To confirm the pathogenetic role of SF3B1 mutations and their impact on phenotype, we investigated an engineered SF3B1^+/− B6 knockout mouse model. While no overt anemia and only mild leukopenia were observed, BM from SF3B1^+/− mice showed numerous RS compared to WT B6 controls, further confirming that SF3B1 alterations may lead to RS. Since around 95% of multiexonic genes are differentially spliced, errors in splicing mechanism are important in maintaining genomic diversity and may lead to cancer; in fact SF3B1 mutations have been previously described in ovarian cancer and melanoma. To assess if other members of the spliceosome machinery are also affected, SF314 and SF3B4 were tested, but to date no mutations were found. Since DYRKA1 encodes a protein that phosphorylates SF3B1 and its mutation could also affect SF3B1 function, we performed Sanger sequencing and found no mutations. Altogether these observations led us to speculate that disruption or mutations in SF3B1 may have a key role in the manifestation of RS phenotype. To examine the functional consequences of defects in SF3B1, we utilized a spliceosome inhibitor that targets SF3B1, Meayamycin (MM). Healthy donors' BM cells were cultured with EPO and MM for 7 days. Vehicle treated cells served as controls. Erythroblast cultures treated with MM showed numerous RS, absent in controls. Furthermore, cultures with MM displayed marked dyserythropoiesis. We concluded that MM induces RS by blocking SF3B1 and produces a phenotype similar to patients with SF3B1 mutations. Quantification of the levels of unspliced and spliced U2-dependent introns using real-time PCR-based spliceosome assays are ongoing but preliminary results suggest differences in the rate of splicing between mutant and WT patients. In conclusion, our findings suggest that defects in the spliceosome machinery contribute to pathogenesis of MDS. In particular, SF3B1, a member of the spliceosome complex, is a novel tumor suppressor gene frequently mutated in patients with RS.