Effect of germline ATM mutations on clonal hematopoiesis.

Abstract

1509 Background: Clonal hematopoiesis (CH) in myeloid related-genes is associated with development of primary and secondary leukemia and atherosclerotic disease, as well as, decreased overall survival. Identification of factors beyond age and cytotoxic exposures that predispose to CH may be useful to both recognize individuals at increased risk for CH and to better understand how CH develops. We have previously shown that germline mutations in the DNA repair gene ATM may predispose to CH. We hypothesized here that heterozygous ATM germline mutation carriers would have higher rates of CH in myeloid genes compared to controls. Methods: Germline DNA samples from 34 heterozygous ATM germline mutation carriers (cases) and 22 controls without ATM germline mutations were sequenced on an Illumina 2500 using a custom 79-gene-myeloid-CH-coding-exon-amplicon-based Qiaseq panel. Read depth averaged 130x. Pathogenic and likely pathogenic CH variants (PV) above an allele fraction of 2% were used for analyses. Cases and controls were compared using a rank-sum test. Results: Cases had a higher median age (56 years, range 30-82) than controls (48 years, range 5-72). Cases and controls were similar in solid tumor cancer history and known exposure to cancer cytotoxic therapy; 73.5% vs 86.4%, and 18.1 vs 20.6%, respectively. The number of CH PV was similarly associated with age in both cases and controls (cor = 0.31, p = 0.01). Cases displayed more CH PVs than controls (total 62 vs 3 PVs, median 2 PVs vs 0, p = 10-6). Of note, cases frequently had a concomitant second (n = 10; 29% of cases) or third (n = 4; 11.8% of cases) unique ATM CH PV, whereas no ATM CH PVs were seen in controls. Even after excluding ATM CH PVs, CH PVs were more frequent in cases (p = 0.00003). After ATM CH PVs, the most frequent CH PVs in cases were in NF1 (5 PVs), BCORL1 (4 PVs), and DMNT3A (4 PVs). Conclusions: Our study supports ATM as a strong predisposition locus for myeloid gene CH. CH in ATM germline mutation carriers frequently involved unique low allele fraction PVs in ATM, suggesting ATM germline PVs are driving production of likely bi-allelic ATM inactivation in white blood cells, or complete ATM loss. Complete ATM loss may be a nidus particularly for lymphocytic leukemia, as bi-allelic ATM inactivation is a frequent somatic finding.