Molecular Landscape of Multiple STAG2 Mutant Myeloid Malignancies

Cohesin complex mutations are observed in myeloid neoplasia (MN) with the X linked STAG2 mutations among the most frequent, accounting for 6-12% of MDS and AML. Germline STAG2 cohesinopathy syndrome leads to human growth and intellective deficits, yet no known cancer predisposition has been described. STAG2-cohesin enables enhancer and promoter interaction in cis to regulate the levels of transcription but the precise mechanism behind the preleukemic consequences of mutations is not fully resolved. In our large compendium of genomic data, we noted patients harboring multiple STAG2 mutations (consistent with subclonal/clonal mosaicism). This observation is highly suggestive of i) putative underlying germline (GL) or somatic (SM) alteration or ii) a specific milieu that provides a “fertile soil” for the emergence of multiple STAG2 mutant clones or subclones.

Through the study of GL and SM mutations in STAG2 mutant cases, we aimed to identify the factors and conditions under which multiple STAG2 mutant clones evolve and the clinical context they create and/or originate from. We analyzed molecular and clinical features of 1,243 MN patients and identified GL and SM variants according to a calling algorithm involving multiple filters and literature cross-reference.

We identified 179 patients with single and 24 with multiple STAG2 variants. In total, there were 126 STAG2 variants (48 frameshifts, 43 stop codons, 20 missense, 15 splice site): 38 confirmed somatic, 74 likely somatic, and 14 confirmed or likely germline (4/10). While GL variants were scattered along the gene, SM variants clustered in the STAG2 (n=6), stromalin conservative (n=5) and glutamine-rich (n=6) domains. Out of 4 GL variants, 3 pathogenic were stop codons and one benign variant was missense. The most frequent GL STAG2 defects were E1260D (6 patients with single hit) and R1033* (2 with single and 3 with multiple hits).

Concerning the 24 patients with multiple STAG2 variants (24/203; 12%), M/F ratio was 3:1 and median age was 71 (range 41-89, yrs). In 15 patients (M/F, 10/5) SM variants were due to clonal mosaicism. Median VAF of these multiple hits, including GL variants, was lower than the one of single GL variants (34 vs 58%, P = .03). In 9 patients (M/F, 8/1) SM plus GL STAG2 mutations were found. Trisomy 8 was present in 7/24 (29%) of multiple STAG2 variants patients and in 11/179 (6%) with single hit. No structural alteration of the X chromosome in heterozygous females was detected. The diagnosis of cases with multiple STAG2 variants included MDS and sAML (4%, both), AML NOS (2%) and CCUS (1%) while the diagnosis of cases with single hit was MDS (20%; high/low risk, 56/44%) and sAML (5%).

Two females with single STAG2 hit had sAML and none in the cohort with multiple STAG2 variants. Although our cohort is small, we noted that time to sAML progression was faster in males with multiple STAG2 variants than in males with single STAG2 hit (5.9 vs 9.8 months); suggesting that in males, double hits might be more deleterious.

The mutational spectrum of patients with multiple STAG2 variants was like the one of patients with single hit with ASXL1 (58 vs 55%), SRSF2 (54 vs 50%), RUNX1 (37 vs 31%), TET2 (33 vs 30%) and IDH2 (29 vs 25%) mutations among the top gene mutations.

Among mutations in other cohesins (SMC1A, SMC3, RAD21), one male patient showed compound cohesion mutations with an additional SMC1A (R96H) mutation, though these may represent distinct clones. None of the patients with this genetic background, in our cohort, showed any symptomatology associated with GL STAG2/cohesin hits. 5 patients (M/F, 4/1) showed SM mutations in other genes including PHF6 (1 patient), BCOR (3 patients), also associated with X linked disorders, and ZRSR2 (1 patient) and no features associated with such disorders were observed as well.

In summary, this is the first report of multiple STAG2 variants in STAG2 mutant MN. Confirmed GL multiple STAG2 variants did not cause any syndromic features or symptomatology associated with GL cohesins' disorders. Multiple STAG2 variants were observed in this cohort only in male patients with sAML, possibly suggesting that, in this context, double hits might increase clonal outgrowth. Additional retrospective and functional analysis of our large cohorts of MN patients is warranted to confirm these findings.

Viny:Arima Genomics: Consultancy, Current holder of stock options in a privately-held company, Membership on an entity's Board of Directors or advisory committees. Maciejewski:Alexion: Consultancy, Honoraria; Novartis: Consultancy, Honoraria, Speakers Bureau.