Contribution of MUTYH Variants to Male Breast Cancer Risk: Results From a Multicenter Study in Italy

Inherited mutations in BRCA1, and, mainly, BRCA2 genes are associated with increased risk of male breast cancer (MBC). Mutations in PALB2 and CHEK2 genes may also increase MBC risk. Overall, these genes are functionally linked to DNA repair pathways, highlighting the central role of genome maintenance in MBC genetic predisposition. MUTYH is a DNA repair gene whose biallelic germline variants cause MUTYH-associated polyposis (MAP) syndrome. Monoallelic MUTYH variants have been reported in families with both colorectal and breast cancer and there is some evidence on increased breast cancer risk in women with monoallelic variants. In this study, we aimed to investigate whether MUTYH germline variants may contribute to MBC susceptibility. To this aim, we screened the entire coding region of MUTYH in 503 BRCA1/2 mutation negative MBC cases by multigene panel analysis. Moreover, we genotyped selected variants, including p.Tyr179Cys, p.Gly396Asp, p.Arg245His, p.Gly264Trpfs*7, and p.Gln338His, in a total of 560 MBC cases and 1,540 male controls. Biallelic MUTYH pathogenic variants (p.Tyr179Cys/p.Arg241Trp) were identified in one MBC patient with phenotypic manifestation of adenomatous polyposis. Monoallelic pathogenic variants were identified in 14 (2.5%) MBC patients, in particular, p.Tyr179Cys was detected in seven cases, p.Gly396Asp in five cases, p.Arg245His and p.Gly264Trpfs*7 in one case each. The majority of MBC cases with MUTYH pathogenic variants had family history of cancer including breast, colorectal, and gastric cancers. In the case-control study, an association between the variant p.Tyr179Cys and increased MBC risk emerged by multivariate analysis [odds ratio (OR) = 4.54; 95% confidence interval (CI): 1.17–17.58; p = 0.028]. Overall, our study suggests that MUTYH pathogenic variants may have a role in MBC and, in particular, the p.Tyr179Cys variant may be a low/moderate penetrance risk allele for MBC. Moreover, our results suggest that MBC may be part of the tumor spectrum associated with MAP syndrome, with implication in the clinical management of patients and their relatives. Large-scale collaborative studies are needed to validate these findings.

Inherited mutations in BRCA1, and, mainly, BRCA2 genes are associated with increased risk of male breast cancer (MBC). Mutations in PALB2 and CHEK2 genes may also increase MBC risk. Overall, these genes are functionally linked to DNA repair pathways, highlighting the central role of genome maintenance in MBC genetic predisposition. MUTYH is a DNA repair gene whose biallelic germline variants cause MUTYH-associated polyposis (MAP) syndrome. Monoallelic MUTYH variants have been reported in families with both colorectal and breast cancer and there is some evidence on increased breast cancer risk in women with monoallelic variants. In this study, we aimed to investigate whether MUTYH germline variants may contribute to MBC susceptibility. To this aim, we screened the entire coding region of MUTYH in 503 BRCA1/2 mutation negative MBC cases by multigene panel analysis. Moreover, we genotyped selected variants, including p.Tyr179Cys, p.Gly396Asp, p.Arg245His, p.Gly264Trpfs * 7, and p.Gln338His, in a total of 560 MBC cases and 1,540 male controls. Biallelic MUTYH pathogenic variants (p.Tyr179Cys/p.Arg241Trp) were identified in one MBC patient with phenotypic manifestation of adenomatous polyposis. Monoallelic pathogenic variants were identified in 14 (2.5%) MBC patients, in particular, p.Tyr179Cys was detected in seven cases, p.Gly396Asp in five cases, p.Arg245His and p.Gly264Trpfs * 7 in one case each. The majority of MBC cases with MUTYH pathogenic variants had family history of cancer including breast, colorectal, and gastric cancers. In the case-control study, an association between

INTRODUCTION
Male Breast Cancer (MBC) is a rare disease whose etiology appears to be associated with genetic factors. Inherited mutations in BRCA1 and, mainly, BRCA2, predispose to MBC and account for up to 13% of all cases in the Italian population (1). Even though there is evidence supporting an association between increased MBC risk and pathogenic variants in PALB2 and CHEK2 (2-4), these two genes are unlikely to account for a substantial fraction of MBC cases. Thus, additional genes that may contribute to MBC genetic susceptibility need to be investigated.
BRCA1, BRCA2, PALB2, and CHEK2 belong to or are functionally linked to the Homologous Recombination (HR) mechanism, one of the most important DNA Double-Strand Break (DSB) repair pathways, highlighting the central role of genome maintenance in MBC predisposition (5). Overall, the maintenance of genomic integrity is achieved by a coordinated interplay of different mechanisms of DNA repair, including Mismatch Repair (MMR), Nucleotide Excision Repair (NER) and Base Excision Repair (BER), in addition to DSB repair (6,7). While dysregulation of DSB repair is known to play a relevant role in breast cancer (BC) pathogenesis, the involvement of other DNA repair pathways in BC is much less established.
MUTYH encodes a DNA glycosylase involved in BER, preventing 8-oxo-G:A mispairs generated by oxidative damage (8). Oxidative DNA damage, including 8-oxoG, may be due to hormonal metabolism and may contribute to BC susceptibility (9,10). In this context, it is noteworthy that BRCA1 and BRCA2 are also involved in 8-oxoG repair (11), thus further supporting a possible role of BER and, more specifically, MUTYH in BC pathogenesis.
Overall, the association between MUTYH mutations and BC risk remains controversial, some studies have shown an increased BC risk among MUTYH mutation carriers, while others have not (22)(23)(24)(25)(26)(28)(29)(30). An increased risk of BC associated with biallelic and monoallelic variants of MUTYH has been reported in BRCA1/2 mutation negative individuals (21)(22)(23)26). A higher frequency of monoallelic MUTYH mutations in families with both breast and colorectal cancer has been also reported compared to general population (21). Recently, an increased BC risk has been also reported for women with the common p.Gln338His variant (31).
To date the possible association between MUTYH variants and MBC risk has not been investigated. MBC is recognized as being primarily a hormone-dependent malignancy and is widely accepted as an estrogen-driven disease specifically related to hyperestrogenism (32) thus, oxidative DNA damage, due to hormonal metabolism, may particularly contribute to BC susceptibility in men. In this context, impairment of MUTYH activity due to inactivating/pathogenic variants may contribute to increase MBC risk.
To assess if MUTYH germline variants may contribute to MBC susceptibility, we screened a large series of BRCA1/2 mutation negative MBC patients by sequencing the entire MUTYH coding region. Furthermore, to explore whether MUTYH variants were significantly associated with MBC risk, we performed a case-control study of selected MUTYH variants.

Study Population
A total of 560 BRCA1/2 mutation negative MBC cases and 1,540 male controls, enrolled in the frame of the ongoing Italian Multicenter Study on MBC (33), were included in the present study. For each MBC case, information on the main clinical-pathologic characteristics were collected as previously described (33,34). Controls were male individuals without personal history of cancer, enrolled under research or clinical protocols, or blood donors. All controls were recruited in the same geographical area of cases. For each study participant, samples of blood or DNA from peripheral blood leukocytes were collected. DNA from blood samples was extracted and quantified as previously described (35). The study was approved by Local Ethical Committee (Sapienza University of Rome, Prot. 669/17) and informed consent for using information and biological samples was obtained from all participants to the study.

MUTYH Gene Sequencing
A total of 503 MBC cases underwent next generation sequencing (NGS) of a custom panel of 50 cancer susceptibility genes including MUTYH. Briefly, paired-end libraries were prepared using the Nextera Rapid Capture Custom Enrichment kit (Illumina, San Diego, California, USA), pooled and loaded into a MiniSeq system (Illumina) for automated cluster generation, sequencing, and data analysis, including variant calling. Variant annotation and filtering was performed with Illumina Variant Studio Software version 2.2 against the human reference genome GRCh37. Variants were classified as pathogenic or likely pathogenic (collectively termed, pathogenic) according to the American College of Medical Genetics and Genomics (ACMG) recommendations (36). Briefly, variants were classified as pathogenic if they had a truncating, initiation codon or splice donor/acceptor effect or if pathogenicity was demonstrated by functional studies supportive of a damaging effect on the gene or gene product. All pathogenic variants were confirmed by double-stranded Sanger Sequencing (primer sequences are available upon request). Variants were named according to Human Genome Variation Society nomenclature (HGVS, hpp://www.hgvs.org).

Statistical Analysis
Chi-square test was performed in a case-case analysis in order to evaluate potential associations between pathogenic variants and specific clinical-pathologic characteristics.
The genotype frequency for each variant was evaluated in both series of cases and controls. The association between each variant and overall MBC risk was measured by the odds ratio (OR) and its corresponding 95% confidence interval (CI) by univariate logistic regression, and also by a multivariate analysis including adjustment for age, center and type of enrolment. A p-value <0.05 was considered statistically significant. All the analyses were performed using STATA version 13.1 statistical program.

Clinical-Pathologic Characteristics of MBC Cases
The study population consisted of 560 BRCA1/2 mutation negative MBC cases, enrolled in the frame of the ongoing Italian Multicenter Study on MBC. Overall, mean age at first BC diagnosis was 61.8 years (range 22-91 years); 91 cases (16.2%) reported first-degree family history of breast and/or ovarian cancer (BC/OC), 247 cases (44.1%) had first-degree family history of cancer and 101 cases (18%) had a personal history of cancer in addition to BC, mostly colorectal and prostate cancer. The majority of male breast tumors were invasive ductal carcinomas (85.9%), estrogen receptor positive (ER+, 94.2%), progesterone receptor positive (PR+ 88.4%), and HER2 negative (79.2%).

MUTYH Gene Sequencing in MBC Cases
The entire coding region of MUTYH was screened in 503 BRCA1/2 mutation negative MBC cases, by a custom multigene panel using NGS technologies. MUTYH variants detected are shown in Table 1. p.Tyr179Cys and p.Gly396Asp variants were the most frequently detected pathogenic variants and were identified in 1.6 and 1.0% of the MBC cases, respectively. The common variant p.Gln338His was identified in 41.7% of the MBC cases (Table 1).
Overall, pathogenic variants were identified in 15 (3.0%) MBC cases ( Table 2), 14 cases were carriers of monoallelic (heterozygous) pathogenic variants and one case was carrier of the biallelic p.Tyr179Cys/p.Arg241Trp (compound heterozygous) pathogenic variants. The majority of MBC cases with MUTYH pathogenic variants had family history of cancer including breast, colorectal, and gastric cancers ( Table 2). In particular, the biallelic MUTYH pathogenic variant carrier was a man diagnosed with BC at 51 years of age who developed colon cancer, with phenotypic manifestation of adenomatous polyposis, at early age (41 years) and had a first-degree relative affected by melanoma at young age (26 years). With the exception of this case, clinical features of the other MBC patients with MUTYH pathogenic variants did not suggest a MAP phenotype.
Overall, comparison of the clinical-pathologic characteristics between MUTYH pathogenic variant carriers and non-carriers did not show any statistically significant differences ( Table 3).
The distribution of genotype frequencies and the estimates for the association between each genotyped variant and overall MBC risk are summarized in Table 4. Significant differences in the distribution of genotypes between MBC cases and controls emerged for p.Tyr179Cys (rs34612342) variant. The analysis of the genotype-specific risks showed that men with heterozygous genotype for MUTYH p.Tyr179Cys variant were at increased BC risk both in the univariate (OR = 5.56; 95%CI:1.67-18.55; p = 0.005) and in the multivariate analysis (OR = 4.54; 95%CI:1.17-17.58; p = 0.028). No statistically significant differences in genotype distribution between case and controls emerged for the other variants analyzed.

DISCUSSION
In this study, we aimed to evaluate the contribution of MUTYH variants in MBC susceptibility. To this purpose, we obtained NGS data of the entire coding region of MUTYH from a large series of BRCA1/2 mutation negative MBC cases, from the ongoing Italian Multicenter Study on MBC, and further genotyped selected variants in a case-control study. To date, there is contrasting evidence on the impact of MUTYH pathogenic variants on risk of BC in women and, to the best of our knowledge, no study has been performed in MBC. By NGS, we identified 15 MBC patients (3.0%) with germline MUTYH pathogenic variants, including one biallelic and 14 monoallelic variant carriers. The MBC patient with biallelic MUTYH pathogenic variants was affected by colorectal cancer at early age with phenotypic manifestation of adenomatous polyposis. Thus, our results allowed a molecular diagnosis of MAP. To the best of our knowledge, to date, only another MBC case has been reported with MAP syndrome (23). Taking into account the rarity of both MBC and MAP, the occurrence of MBC in MAP patients may underline a possible common genetic pathway and suggest that MBC could be considered a MAP-related malignancy.
Overall, MUTYH monoallelic pathogenic variants, including p.Tyr179Cys, p.Gly396Asp, p.Arg245His, and p.Gly264Trpfs * 7, were found with a frequency of 2.8% in our MBC series. p.Tyr179Cys and p.Gly396Asp were the most frequently variants detected and were identified in 2.4% of the cases. Published data showed that these two variants are the most frequent pathogenic variants in populations of European origin and account for 50 to 90% of MUTYH pathogenic variants identified in MAP patients (13,14,37,38). The p.Arg245His variant was identified in a MBC patient with family history of breast and gastric cancers. This variant has been reported strongly associated with familial colorectal cancer (23,39), and has also been identified in patients with suspected Lynch Syndrome and in a patient with gastric cancer (23,40). The p.Gly264Trpfs * 7 variant was identified in a MBC patient, from North-East of Italy, where it occurs as a founder mutation accounting for about 15.0% of the MUTYH pathogenic variants identified in MAP patients (41). By contrast, this variant has been reported with lower frequency, ranging from 1.0 to 8.0%, in MAP patients from other populations of Caucasian ethnicity (23,(41)(42)(43)(44)(45)(46)(47).
To investigate whether MBC arising in MUTYH pathogenic variant carriers may be characterized by specific features, we compared clinical-pathologic characteristics between carriers and non-carriers. No statistically significant association emerged for any of the clinical features tested. However, the great majority of MBC patients with MUTYH pathogenic variants had family history of cancer, including, breast, colorectal, and gastric cancers. These findings, if confirmed by additional data, may be useful in decisions concerning clinical management of patients and their families.
To further investigate the role of MUTYH in MBC, we evaluated the risk of MBC associated with selected MUTYH variants previously proposed to be associated with increased cancer risk, including BC risk (21,27,31), by performing a case-control study. Among the pathogenic variants examined, the p.Tyr179Cys variant was associated with an increased MBC risk (OR = 4.54, 95%CI = 1.17-17.58). A higher frequency of p.Tyr179Cys has been reported in families with both breast and colorectal cancer compared to the general population (21), but an association between p.Tyr179Cys variant and increased BC risk has not been observed (25, 26,28,30). Our results, suggest that p.Tyr179Cys variant may be a low/moderate penetrance risk allele for BC in men. This variant, located at 8-oxo-G binding site, causes major structural protein changes and a reduction in functionality (48,49). Thus, oxidative DNA damage due to hormonal metabolism, like estrogen-induced 8-oxo-dG generation, may particularly contribute to MBC susceptibility, as BC in men is primarily a hormone-dependent tumor, specifically related to hyperestrogenism. Furthermore, it can be hypothesized that MBC, unencumbered by the many confounding factors that exist in female BC (i.e., reproductive factors and high frequency) might facilitate the identification of genetic factors and molecular mechanisms that may influence BC risk in general (50).
We also assessed whether the common p.Gln338His variant, reported to increase BC risk in women (31), was associated with MBC risk. We did not observe any significant differences in p.Gln338His genotypes distribution between MBC cases and controls inconsistent with a possible role of this variant in MBC risk. The other common variant, p.Val22Met, has not been reported to be associated with cancer risk (51)(52)(53) and was not examined in this study.
Overall, we observed that the majority of MBC patients with pathogenic MUTYH variants have first-degree family history of cancers. This raises the question of whether MUTYH variants, especially the Tyr179Cys variant, may be associated with MBC risk only, or with the risk of familial or multi-syndromic diseases, including MBC. Further clinical/phenotype assessments and detailed statistical analyses would be useful in future studies to answer this question.
In conclusion, our study suggests that MUTYH pathogenic variants may have a role in MBC, in particular, p.Tyr179Cys variant may be a low/moderate penetrance risk allele for MBC. Our findings also suggest that MBC may be part of the tumor spectrum associated with MAP syndrome, with implications in the clinical management of the patients and their relatives.
Although we have a large series of MBC cases, this study may be underpowered to detect smaller risk effects and largescale collaborative studies are needed to investigate any possible association with rarer variants and to have a more comprehensive examination and characterization of the link between MUTYH variants and MBC risk.

DATA AVAILABILITY STATEMENT
Datasets are available on request. The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

AUTHORS CONTRIBUTIONS
PiR drafted the manuscript, performed NGS and statistical analyses and interpreted the results. VS performed genotyping and statistical analyses, and interpreted the results. AB and IC performed genotyping analysis. VZ and VV performed NGS analysis. IZ, GM, SB AS, ST, MT, AR, LV, AC, DC, LC, AV, BB, JA, SM, MM, PaR, and DP recruited samples and collected clinicalpathologic data. PP contributed to study design, recruited samples and collected clinical pathologic data. LO conceived, designed and coordinated the study, and drafted the manuscript. All authors reviewed, edited, and approved the manuscript for publication.

FUNDING
This study was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC IG 16933) to LO.