Monoallelic Germline Pathogenic Variants in DNA Damage Repair Genes and Their Impact on Post-Hematopoietic Cell Transplantation Outcomes in Severe Aplastic Anemia

Background Recipients of HCT are under significant stressors that may predispose them to DNA damage and unfavorable outcomes. Patients with inherited defects in DNA damage response or repair (DDRR), such as Fanconi anemia, need special conditioning regimens to minimize toxicity. HCT outcomes in patients carrying pathogenic or likely pathogenic (P/LP) variants in DDRR genes is unknown. Methods We used exome sequencing data and clinical information from the NCI-CIBMTR collaborative Transplant Outcomes in Aplastic Anemia (TOAA) project. This analysis includes 683 patients with acquired severe aplastic anemia (SAA) who received HCT between 1989-2015. We curated germline variants in 127 DDRR genes according to the ACMG/AMP criteria. We used Kaplan-Meier estimator and Cox proportional hazard models. DDRR genes were evaluated for presence or absence, by pathway (base excision repair (BER), nucleotide excision repair, homologous recombination, Fanconi anemia, mismatch repair, or nonhomologous end joining, and by genes (focusing on those affecting ³ 5 patients). Results We identified 106 pathogenic or likely pathogenic (P/LP) variants, all monoallelic, in 54 DDRR genes, present in 128 patients (18.9%). The majority of those identified (83%) were loss of function variants. The most frequent P/LP variants in the SAA patients were in MUTYH (N=14, 2%), POLG (N=9, 1.3%), FANCM (N=8, 1.2%), and CHEK2 (N=6, 0.9%). No difference in patient demographics or transplant-related factors were noted between patients with and without P/LP variants in DDRR genes. The median age was 22.4 and 19.9 years in the two groups, respectively. The majority of the patients received reduced intensity regimens (60.9% and 62.3%, in those with and without DDRR, respectively). Multivariable models showed no statistically significant difference in patient overall survival (OS), by the presence or absence of DDRR variants (HR=1.20, 95% CI=0.87-1.67, p=0.26). But, in analysis focusing on DDRR pathways, patients with P/LP variants in the BER genes (N=43; 6.3%) had statistically significant inferior OS compared with those without BER variants (1-year OS=63% vs. 74%, in those with and without BER variants, respectively; adjusted HR=2.03, 95% CI=1.30-3.15, p=0.002) (Figure 1). Gene-based analysis identified ERCC3, FANCD2, or MUTYH (affecting 3.4% of the patients) as high-risk genes; 1-year OS=48% in patients with P/LP in high-risk genes vs. 74% in patients with P/LP in other DDRR genes or no DDRR variants, p log-rank=0.0004) (Figure 2). Compared with patients without P/LP variants in high-risk genes, the adjusted HR for all-cause mortality in patients carrying P/LP variants in any of the risk genes was 2.74 (95% CI=1.61-4.67, p=0.0002). Conclusion These results highlight the importance of germline testing before HCT. Patients with P/LP variants in high-risk DDRR genes may need tailored regimens.