Somatic Genetic and Epigenetic Architecture of Myelodysplastic Syndromes Arising from GATA2 Deficiency

The emergence of GATA2 deficiency as a germline predisposition to myeloid malignancies raises questions about the nature of acquired secondary genetic and epigenetic events facilitating leukemogenesis. Previously, mutations in ASXL1 were implicated as a possible somatic driver in single cases of GATA2-related MDS. However the landscape of secondary changes had not yet been systematically examined in larger MDS cohorts, and accounting for confounding factors. In this study, we used next-generation genomic platforms to investigate targeted mutational landscape and global epigenetic profiles in patients with GATA2 deficiency.

In a large cohort of consecutively diagnosed children with MDS we had initially established that GATA2 deficiency accounts for 7% of primary MDS cases. Exploring the known association between GATA2 mutated (GATA2mut) cases and monosomy 7 (-7), the prevalence of GATA2 deficiency was very high in patients with -7 (37%), reaching its peak in adolescence (>70%). We next tested 60 GATA2-deficient patients with MDS for the presence of secondary mutations using targeted NGS for genes involved in myeloid malignancies. Somatic status was confirmed by matched analysis of fibroblasts, hair follicles or T-cells. Single hematopoietic CFU colonies were sequenced to identify subclonal patterns. For comparison, a GATA2 wildtype (GATA2-WT) cohort of 422 children and adolescents with MDS enrolled in the studies of the European Working Group of Childhood MDS were analyzed by targeted NGS. Somatic mutations were detected in 45% (27/60) of GATA2mut as compared to 19% (82/422) GATA2-WT MDS cases (p<0.0001). Recurrently mutated genes in the GATA2mut group included SETBP1, ASXL1, STAG2, RUNX1, CBL, EZH2, NRAS/KRAS, JAK3, and PTPN11. No mutations were found in TP53, BCOR/BCORL and a number of other oncogenes. Because -7 karyotype was significantly overrepresented in GATA2mut cases with somatic mutations (78%), we next focused on this cytogenetic category. Within the -7 subgroup the rate of somatic mutations was the same in GATA2mut (56%) and GATA2-WT (58%) subgroups. However, hotspot SETBP1 mutations were overrepresented in GATA2-deficient patients with -7 (50%) vs. GATA2-WT MDS cohort (22%, p<0.05). Furthermore, STAG2 mutations were found frequently in the GATA2mut group (10%, 6/60) as opposed to only 0.2% (1/422) of the total GATA2-WT cohort (p<0.0001). Next, we aimed to define the clonal hierarchy of concurrent mutations by longitudinal NGS-analysis during disease course in selected patients. Our results indicate that somatic SETBP1 lesions precede the development of ASXL1 mutations. Remarkably, this model of clonal evolution does not depend on preexisting germline GATA2 lesion, as confirmed by sequencing of single CFU colonies cultivated from the bone marrow of 3 GATA2mut and 3 GATA2-WT MDS patients. Finally, to elucidate the epigenetic effects, we compared methylation patterns using methyl-CpG-immunoprecipitation and Illumina-NGS in 25 GATA2mut to 17 GATA2-WT patients and 10 healthy controls. Based on the degree of global methylation, there were no significant alterations allowing for the discrimination of GATA2-deficient patients from the total MDS cohort, when accounted for bias arising from cytogenetic and morphologic subgroups.

In summary, somatic SETBP1 and STAG2 mutations are associated with MDS arising from GATA2 deficiency. The remaining targeted clonal landscape is essentially determined by the presence of monosomy 7. Similarly, the global epigenetic changes correlate with morphological and cytogenetic subgroups, rather than with germline GATA2 status. The prospect of potential drug targetability of mutations frequently found in children, particularly in the SETBP1 oncogene, and in histone modifiers ASXL1 and EZH2, warrants further biological studies.