SAMD9 and SAMD9L Germline Disorders in Patients Enrolled in Studies of the European Working Group of MDS in Childhood (EWOG-MDS): Prevalence, Outcome, Phenotype and Functional Characterisation

Hereditary predisposition has been ever since implicated in the etiology of childhood myelodysplastic syndromes (MDS). Until recently, GATA2 deficiency prevailed as a major germline cause in pediatric primary MDS. In the past 2 years, we and others identified germline mutations in paralogue genes SAMD9 and SAMD9L residing on chromosome 7q21.2 as new systemic diseases with high propensity for MDS with monosomy 7. Although initially, mutations in SAMD9 and SAMD9L genes were associated with MIRAGE and Ataxia-Pancytopenia syndromes, respectively, with recent reports the phenotypes are becoming more intertwined. Nevertheless, the predisposition to MDS with monosomy 7 (-7) remains a common clinical denominator. Both genes are categorized as negative regulators of cellular proliferation and mutations were shown to be activating. Because of their high evolutionary divergence, classical in silico prediction is erratic, thereby establishing in vitro testing as the current gold standard for pathogenicity evaluation. The objectives of this study were to define the prevalence of SAMD9/9L germline mutations in primary pediatric MDS, and to describe the clinical phenotype and outcome. In addition, we aimed to characterize the somatic mutational architecture and develop a functional scoring system.

Within the cohort of 548 children and adolescents with primary MDS diagnosed between 1998 and 2016 in Germany, 43 patients (8%) carried SAMD9/9L mutations that were mutually exclusive with GATA2 deficiency and known constitutional bone marrow (BM) failure. MDS type refractory cytopenia of childhood was diagnosed in 91% (39/43), and MDS with excess blasts in 9% (4/43) of mutated cases. Karyotype at diagnosis was normal in 58%, and -7 was detected in 37% of SAMD9/9L cohort. Within MDS subgroup with -7 (n=74), SAMD9/9L mutations accounted for 22% of patients. Notably, the demographics, familial disease, diagnostic blood and BM findings, overall survival (OS) and the outcome after HSCT were not influenced by mutational status in our study cohort (n=548). At the last follow up, 88% (38/43) of SAMD9/9L MDS patients were alive; 35/43 had been transplanted with a 5-year-OS of 85%.

Next, we added 26 additional cases with SAMD9/9L mutations diagnosed in Europe within EWOG-MDS studies. In the total cohort of 69 germline mutated patients we found a total of 75 SAMD9/9L mutations, of which 67 were novel. Of those we tested 47 using a HEK293 cell in vitro system and 45/47 mutants inhibited proliferation. While 53/69 patients carried only single germline mutations (missense in 50/53 and truncating in 3/53), in the remaining 16 patients, 11 additional truncating and 7 missense mutations were found. We did not observe an association between germline mutation and phenotype. Immunological issues (e.g. recurring infections, low Ig) were described in 32%/50% of SAMD9/9L-mutated patients, while physical anomalies were very heterogeneous and reported in ~50% of patients in both mutational groups. Intriguingly, genital phenotypes occurred in 40% of SAMD9L, while neurological problems were present in 30% of SAMD9 - mutational subgroups.

To elucidate the somatic mutational landscape, we performed whole exome and deep sequencing of 58 SAMD9/9L patients and identified recurrent somatic mutations in known oncogenes that were earlier associated with pediatric MDS: SETBP1 (10%), RUNX1 (7%), ASXL1 (5%), EZH2 (5%), CBL (3%). The identified somatic mutations occurred in association with monosomy 7 background (18/20). Finally, we utilized the results from functional testing of the 47 SAMD9/9L variants as our test cohort to develop combinatorial in silico scoring. The rationale was to decrease the dependency on functional validation. Based on the results of 20 in silico tools we could concatenate a matrix of 5 algorithms to resolve the pathogenicity of >80% of variants. Using this model, all variants predicted as pathogenic showed also growth-restrictive effect in vitro. In summary, pathogenic SAMD9/9L germline mutations account for 8% of primary pediatric MDS and 22% of MDS/-7. The mutations identified are heterogeneous and their effect can be predicted using a combinatorial in silico - in vitro approach. Finally, the clinical outcome and somatic mutational landscape are not influenced by the mutational status.