Mono- and biallelic germline variants of DNA glycosylase genes in colon adenomatous polyposis families from two continents

Abstract

Recently, biallelic germline variants of the DNA glycosylase genes MUTYH and NTHL1 were linked to polyposis susceptibility. Significant fractions remain without a molecular explanation, warranting searches for underlying causes. We used exome sequencing to investigate clinically well-defined adenomatous polyposis cases and families from Finland (N=34), Chile (N=21), and Argentina (N=12), all with known susceptibility genes excluded. Nine index cases (13%) revealed germline variants with proven or possible pathogenicity in the DNA glycosylase genes, involving NEIL1 (mono- or biallelic) in 3 cases, MUTYH (monoallelic) in 3 cases, NTHL1 (biallelic) in 1 case, and OGG1 (monoallelic) in 2 cases. NTHL1 was affected with the well-established, pathogenic c.268C>T, p.(Gln90Ter) variant. A recurrent heterozygous NEIL1 c.506G>A, p.(Gly169Asp) variant was observed in two families. In a Finnish family, the variant occurred in trans with a truncating NEIL1 variant (c.821delT). In an Argentine family, the variant co-occurred with a genomic deletion of exons 2 - 11 of PMS2. Mutational signatures in tumor tissues complied with biological functions reported for NEIL1. Our results suggest that germline variants in DNA glycosylase genes may occur in a non-negligible proportion of unexplained colon polyposis cases and may predispose to tumor development.

Keywords: DNA glycosylase; MUTYH; NEIL1; NTHL1; OGG1; exome sequencing; germline variant; polyposis.

Figure 1

(A) Flowchart of the investigation including the Finnish and South American cohorts. Het denotes heterozygosity and hom homozygosity for the variants. (B) Detailed information of germline variants discovered in DNA glycosylase genes.

Figure 2

(A) Pedigrees of the polyposis families with NEIL1 variants. Numbers below the symbols are patient identifiers. Arrow denotes the index person. Zygosity of NEIL1 variants is shown (+/− heterozygous). Tumor manifestations and age at diagnosis (years) are given below the patient symbol. Mets refers to metastasis. Nonessential pedigree features were removed or modified to protect confidentiality. (B) Locations of the variants relative to the main functional domains of the DNA glycosylase genes. Zn denotes the metal binding sites in MUTYH as listed in the Uniprot database (www.uniprot.org; Q9UIF7). The pedigrees were generated with Pedigree Chart Designer and the lollipop diagrams with MutationMapper.

Figure 3

Expressional consequences of the NEIL1 c.506G>A, p.(Gly169Asp) variant. (A) ASE analysis based on the NEIL1 c.506G>A variant. Longer peak (G) represents the wild type sequence. The control individual is homozygous for the wild-type allele. The index individual of FAP104 is heterozygous: Allele A corresponds to the NEIL1 c.506A missense variant, whereas the G allele is known to have a frameshift variant (c.821delT) in a downstream position. This individual displays ASE with the peak area ratio of 0.45 for G to A in cDNA relative to gDNA. The result indicates that transcripts with G (arrowhead) having the frameshift variant are twice less abundant than transcripts containing the missense variant (A). (B) Relative quantity (RQ) from the qRT-PCR experiment targeting the two main isoforms of NEIL1 using the housekeeping gene GAPDH as an endogenous control. Whiskers indicate 95% confidence limits. The index of family FAP104 shows reduced NEIL1 RNA expression compared to the controls. (C) Western blot of two healthy control individuals and the index of FAP104. GAPDH was used as a loading control. FAP104 index displays elevated NEIL1 protein levels compared to the controls (arrowhead). No truncated NEIL1 protein is seen.

Figure 4

Somatic mutational signature analysis of four NEIL1-associated tumors (A, B). (A) Heatmap indicating the relative contribution of SBS signatures (COSMICv2) to the mutational landscape of each tumor. Black arrowheads indicate NEIL1-deficiency associated signatures prominent in tumors from FAP104, whereas open arrowheads represent MSI-signatures present in a colorectal carcinoma from ARG046. (B) Heatmap of the relative contributions of ID signatures (COSMICv3.1, GRCh37) to the mutational profiles of the tumors. Subsequent discovery of the PMS2 alteration in the ARG046 case (C, D). (C) Immunohistochemical analysis of the MMR proteins reveals a selective loss of PMS2 in the tumor cells. Normal cells retaining the PMS2 expression are indicated with arrowheads. Scale bar represents 50 μm. (D) PMS2-MLPA analysis of blood DNA of the ARG046 case as well as a healthy control. Arrowheads indicate reduced emission peaks at exons 2-11. The average probe ratios of exons 2-11 (0.54 ± 0.03) are indicative of a heterozygous deletion.