Whole Exome Sequencing Reveals Acquired SF3B1 Mutations Defining Patients with Acquired Idiopathic Sideroblastic Anaemia

Abstract 2793

Whole Exome Sequencing Reveals Acquired SF3B1 Mutations Defining Patients with Acquired Idiopathic Sideroblastic Anaemia.

Refractory anaemia with ringed sideroblasts (RARS) is an acquired form of MDS characterized by accumulation of mitochondrial iron that gives rise to ringed sideroblasts. The molecular pathogenesis of RARS and hence the cause of abnormal mitochondrial iron is unknown. To address these questions, whole exome sequencing (Illumina) was performed on CD34+ cells from 8 patients all of whom had >50% ringed sideroblasts, comprising 5 RARS, 1 RARS-T, 1 tMDS and 1 RCMD-RS patient. Following filtering against variations found in available paired constitutional samples (CD3+ T cells or skin) and published polymorphisms, candidate mutations were initially identified for commonality between patients. Aberrations in one particular gene, splicing factor 3b subunit 1 (SF3B1), a component of the major and minor spliceosomes, stood out at this stage. Mutations in SF3B1 exons 14 –16 were observed in CD34+ cells in 7/8 cases, but not in the single case of tMDS, which carried mutant TP53. Follow-up amplicon sequencing (Roche 454 and Sanger sequencing) for a total of 27 cases with >50% ringed sideroblasts, using total bone marrow tissue, identified acquired SF3B1 mutations in 93% (14/15) RARS, 67% (6/9) RCMD-RS, 100% (1/1) AML with a preceding RARS and 0% (0/2) tMDS cases. All mutations were heterozygous substitutions and found in the highly-conserved protein c-terminal HEAT motifs, associated with snRNP stabilization within the U2 snRNP complex of the major spliceosome. However, gene expression microarray analysis (Affymetrix exon arrays) showed no significant differences between cases with (n=5) and without (n=3) mutations.

All the patients with SF3B1 mutations had normal cytogenetics except one patient with trisomy 8. There was no difference in overall survival or leukaemic transformation when comparing patients with or without SF3B1 mutation. Furthermore, in a single RARS case with available sequential samples, the SF3B1 mutation burden was maintained at the same level from diagnosis, at progression to AML and at time of remission following intensive chemotherapy (FLAG). SF3B1 mutations were also detected in glycophorin A + erythroid cells. Interestingly, no ‘hotspot’ SF3B1 mutations were seen in a single case of congenital sideroblastic anaemia.

A possible mutual exclusivity between SF3B1 and TP53 mutations (p<0.03), which is often associated with disease progression, was also observed (2/2 tMDS cases with wildtype SF3B1). Conversely, mutations in TET2 (2 cases), IDH1 (1 case) and DNMT3A (2 cases), which are known to highly prevalent in haematological disease but not proven prognosticators in MDS, were observed with coinciding SF3B1 aberrations.

Exome sequencing also revealed additional novel candidate mutations, including ones in previously uncharacterized zinc finger proteins, and candidate mutations in factors which suggest defects in cell proliferation, DNA maintenance and cytoskeletal organization. In addition to known mutations mentioned above, between 30–70 well-supported candidate mutations were identified per individual exome. However, candidate mutations in the haem-synthesis pathway enzymes were rare (1/8 cases) and are therefore unlikely to contribute to the RARS disease phenotype.

We therefore show that whole exome sequencing is effective in identifying novel mutations in a well-defined patient cohort. We subsequently demonstrate that mutations in the fundamental spliceosome component SF3B1, is acquired in a majority of cases of sideroblastic anemia