Genomic Instability in Early-Onset Myelodysplastic Syndrome Patients Harboring Rare Germline Variants in NHEJ DNA Repair Genes

Introduction

Myelodysplastic syndrome (MDS) in young adults (<60 yo) accounts for approximately 10% of all MDS patients. Comprehensive sequencing studies have recognized an increasing number of cases attributed to germline predisposition and have expanded the spectrum of predisposition genes. Recently, it has been demonstrated that the DNA repair gene ERCC6L2, involved in the non-homologous end joining (NHEJ) pathway, may carry deleterious germline variants in this subset of patients. The Homologous Recombination Deficiency score (HRD score) is used to determine the double strand DNA repair pathways state. A 2020 study showed that ERCC6L2 harmful variants resulted in an aberrant HRD score (Liu X, et al. Cell Res. 2020). We aim to expand the list of potentially pathogenic variants in NHEJ pathway genes in MDS cases and develop tools to help categorize their pathogenicity.

Methods

A prospective study was conducted with a cohort of 235 patients diagnosed with MDS between the ages of 16 and 60, recruited from 32 centers from the Spanish Group of MDS (GESMD). Whole exome sequencing (WES) was performed on paired tumor-germline samples using the NovaSeq6000-Illumina platform, with variant categorization according to the ACMG criteria. Genomic instability analysis was performed using WES data through the following tools: FastQ file quality control (Fastp), alignment (bwa-mem2), integrity assessment (GATK and samtools), genomic segmentation and ploidy estimation (Sequenza), component determination, and genomic instability score summation (scarHRD). The three components considered to obtain the HRD score were: loss of heterozygosity (LOH), large-scale transitions (LST), and sub-telomeric regions with allelic imbalance (TAI).

Results

In the total cohort of 235 cases of adults diagnosed with MDS at a young age (median 49 years; p25-p75, 38-55), we found eight cases with deleterious germline variants in NHEJ repair pathway genes: two pathogenic variants (LIG4, MRE11), three likely pathogenic variants (2 ERCC6L2, 1 MRI1), and three variants of undetermined significance (VUS) prioritized due to a REVEL score >0.8 (1 BAZ1A, 2 ERCC6L2). Additionally, twelve patients had other VUS. These 22 patients significantly differed from the rest of the cases by a higher co-occurrence of acquired TP53 mutations (p=0.03), and a shorter time to progression to acute myeloid leukemia (p=0.009).

Regarding genomic instability status, the HRD score was calculated from WES data in 116 cases: 14 patients with variants in NHEJ genes were compared with a cohort of 101 MDS patients without germline variants in these genes. The sum of the three scores, the HRDsum, was significantly higher in the group with NHEJ variants (median 58.8 vs 39.6, p=0.041). When deconstructing the score, the subgroup with NHEJ variants had significantly higher TAI (median 28.9 vs 19.4, p=0.007). However, neither the LOH component nor the LST one showed significant differences between the two groups (median 6.2 vs. 4.8, p=0.592; median 23.6 vs 16.9, p=0.140, respectively). Regarding the ERCCL6 subset (n=9) they also showed a higher TAI (median 28.6 vs. 19.9, p=0.039), with no significant differences regarding the rest of the parameters.

Conclusions

In this study, we show how adult patients with MDS diagnosed at unusual young age present an enrichment in germline variants in genes involved in the alternative DNA double-strand break repair pathway, NHEJ. This subset of patients is depicted by a particular clinical-molecular profile, with a higher frequency of acquired mutations in TP53 and a worse prognosis. Furthermore, we describe a specific genomic instability “scar” of this subgroup, with a predominance of sub-telomeric regions with allelic imbalance defect, which emerges as a potential marker of pathogenicity.