Recurrent Mutations of Multiple Components of Cohesin Complex in Myeloid Neoplasms

Abstract 782

Recent genetic studies have revealed a number of novel gene mutations in myeloid malignancies, unmasking an unexpected role of deregulated histone modification and DNA methylation in both acute and chronic myeloid neoplasms. However, our knowledge about the spectrum of gene mutations in myeloid neoplasms is still incomplete. In the previous study, we analyzed 29 paired tumor-normal samples with chronic myeloid neoplasms with myelodysplastic features using whole exome sequencing (Yoshida et al., Nature 2011). Although the major discovery was frequent spliceosome mutations tightly associated with myelodysplasia phenotypes, hundreds of unreported gene mutations were also identified, among which we identified recurrent mutations involving STAG2, a core cohesin component, and also two other cohesin components, including STAG1 and PDS5B.

Cohesin is a multimeric protein complex conserved across species and is composed of four core subunits, i.e., SMC1, SMC3, RAD21 and STAG proteins, together with several regulatory proteins. Forming a ring-like structure, cohesin is engaged in cohesion of sister chromatids in mitosis, post-replicative DNA repair and regulation of gene expression.

To investigate a possible role of cohesin mutations in myeloid leukemogenesis, an additional 534 primary specimens of various myeloid neoplasms was examined for mutations in a total of 9 components of the cohesin and related complexes, using high-throughput sequencing. Copy number alterations in cohesin loci were also interrogated by SNP arrays. In total, 58 mutations and 19 deletions were confirmed by Sanger sequencing in 73 out of 563 primary myeloid neoplasms (13%).

Mutations/deletions were found in a variety of myeloid neoplasms, including AML (22/131), CMML (15/86), MDS (26/205) and CML (8/65), with much lower mutation frequencies in MPN (2/76), largely in a mutually exclusive manner. In MDS, mutations were more frequent in RCMD and RAEB (19.5%) but rare in RA, RARS, RCMD-RS and 5q- syndrome (3.4%). Cohesin mutations were significantly associated with poor prognosis in CMML, but not in MDS cases.

Cohesin mutations frequently coexisted with other common mutations in myeloid neoplasms, significantly associated with spliceosome mutations. Deep sequencing of these mutant alleles was performed in 19 cases with cohesin mutations. Majority of the cohesin mutations (16/19) existed in the major tumor populations, indicating their early origin during leukemogenesis.

Next, we investigated a possible impact of mutations on cohesin functions, where 17 myeloid leukemia cell lines with or without cohesin mutations were examined for expression of each cohesin component and their chromatin-bound fractions. Interestingly, the chromatin-bound fraction of one or more components of cohesin was substantially reduced in cell lines having mutated or defective cohesin components, suggesting substantial loss of cohesin-bound sites on chromatin. Finally, we examined the effect of forced expression of wild-type cohesin on cell proliferation of cohesin-defective cells. Introduction of the wild-type RAD21 and STAG2 suppressed the cell growth of RAD21- (Kasumi-1 and MOLM13) and STAG2-defective (MOLM13) cell lines, respectively, supporting a leukemogenic role of compromised cohesin functions.

Less frequent mutations of cohesin components have been described in other cancers, where impaired cohesion and consequent aneuploidy were implicated in oncogenic action. However, 23 cohesin-mutated cases of our cohort had completely normal karyotypes, suggesting that cohesin-mutated cells were not clonally selected because of aneuploidy. Alternatively, a growing body of evidence suggests that cohesin regulate gene expression, arguing for the possibility that cohesin mutations might participate in leukemogenesis through deregulated gene expression. Of additional note, the number of non-silent mutations determined by our whole exome analysis was significantly higher in 6 cohesin-mutated cases compared to non-mutated cases. Since cohesin also participates in post-replicative DNA repair, this may suggest that compromised cohesin function could induce DNA hypermutability and contribute to leukemogenesis.

In conclusion, we report a new class of common genetic targets in myeloid malignancies, the cohesin complex. Our findings highlight a possible role of compromised cohesin functions in myeloid leukemogenesis.