Monoallelic NTHL1 Loss-of-Function Variants and Risk of Polyposis and Colorectal Cancer

1Department of Pathology, Leiden University Medical Center, Leiden, the Netherlands 2Department of Human Genetics, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Nijmegen, the Netherlands 3Colorectal Oncogenomics Group, Department of Clinical Pathology, Melbourne Medical School, The University of Melbourne, Parkville, Victoria, Australia 4University of Melbourne Centre for Cancer Research, Victorian Comprehensive Cancer Centre, Parkville, Victoria, Australia 5Genomic Medicine and Family Cancer Clinic, Royal Melbourne Hospital, Parkville, Victoria, Australia

approximately 0.28% of the general population. 6 Identification of monoallelic NTHL1 LoF variants currently presents a clinical conundrum regarding how best to counsel carriers with respect to their cancer risk because of the lack of published evidence. Here, we show that monoallelic LoF variants in NTHL1 are not enriched in individuals with polyposis and/or CRC compared to the general population. Furthermore, 13 colorectal tumors from NTHL1 LoF carriers did not show a somatic second hit, and we did not find evidence of a main contribution of mutational signature SBS30, the signature associated with NTHL1 deficiency, suggesting that monoallelic loss of NTHL1 does not substantially contribute to colorectal tumor development.

Methods
A total of 5,942 individuals with unexplained polyposis, familial CRC, or sporadic CRC at young age or suspected of having Lynch syndrome with CRC or multiple adenomas were included in this study and defined as case patients (individual studies and their ascertainment are described in Supplementary Methods and Supplementary Table 1). Three independent data sets were used as controls, including (1) the non-Finnish European subpopulation of the genome aggregation database (gnomAD: n = 64,328), 6 (2) a Dutch cohort of individuals without a suspicion of hereditary cancer who underwent whole-exome sequencing (WES) (Dutch WES; n = 2,329), 7 and (3) Table 1). For control individuals, all pathogenic LoF variants were retrieved from gnomAD and the Dutch WES-cohort, 6,7 and for the CCFRC control individuals, the exonic regions of NTHL1 were sequenced (Supplementary Table 1). Odds ratios between case patients and control groups were calculated and a Fisher exact test was performed to assess the significance of difference in carrier rates. Cosegregation analysis was performed by using Sanger sequencing. Two adenomas and 11 primary CRCs from NTHL1 LoF variant carriers were subjected to WES, and subsequently, mutational signature analysis was performed (Supplementary Methods  and Supplementary Table 2). For signature analysis comparison, we included 3 CRCs from individuals with a biallelic NTHL1 LoF variant.

Results
Monoallelic NTHL1 LoF variants were identified in 11 of 3,439 case patients (0.32%) and   Table 2). The NTHL1 wild-type allele was unaffected by somatic mutations or loss of heterozygosity in all tumors tested. In contrast to NTHL1-deficient tumors, in none of the tumors of the carriers was mutational signature SBS30 the main signature, because it was only present in 1 tumor, where it had a minor contribution ( Figure 1B and Supplementary Table 2). 4 These observations indicate that biallelic inactivation of NTHL1 through a somatic second hit was not evident and that monoallelic inactivation of NTHL1 was insufficient to result in the accumulation of somatic mutations that are characteristic of an NTHL1-deficiency phenotype.

Discussion
In this study, the largest investigating monoallelic LoF variants in NTHL1 to date to our knowledge, we observed no evidence of an association between carriers and the risk of polyposis and/or CRC. In our case patients, the prevalence of pathogenic NTHL1 LoF variant alleles is comparable to that of the general population. However, we cannot rule out that a small risk for CRC, similar to what is observed for MUTYH carriers, still exists.
Colorectal tumors from monoallelic NTHL1 LoF variant carriers did not show evidence of a somatic second hit in NTHL1 nor of defective base-excision repair, which is typically associated with biallelic NTHL1 inactivation. Only 1 tumor showed a minor SBS30 contribution to the mutation profile, but this contribution was far less significant compared to NTHL1-deficient CRC and is likely the result of multiple testing correction. Our data suggest that inactivation of the NTHL1 wild-type allele is a rare event in colorectal tumors, which is in agreement with the observation that loss of heterozygosity of chromosome arm 16p is not frequently observed in CRC. 8 We were unable to discriminate between individuals with polyposis or CRC due to the historical nature of the case collections. Therefore, differences in the frequencies of monoallelic NTHL1 LoF variants between control individuals and these 2 phenotypes were not made separately. However, because we identified NTHL1 LoF variants in individuals with polyposis or CRC, we do not consider a major difference between these 2 phenotypes. Because NTHL1 deficiency may also predispose to extracolonic tumors, the risk for these tumor types in monoallelic NTHL1 carriers still needs further assessment.
In conclusion, the evidence to date does not support an increased risk of polyposis and/or CRC for carriers of monoallelic NTHL1 LoF variants, and consequently, no additional surveillance is currently warranted beyond population screening for CRC, unless family history characteristics point to a reason for colonoscopy.