Breast cancer (BC) is the second most common cancer worldwide (1), and the first cause of cancer death among women in across all age groups (1). The etiology of BC is multifactorial, and both endogenous and environmental factors are implicated in its pathogenesis (2).

The risk of BC and/or ovarian cancer (OC) is increased in carriers of deleterious mutations in BRCA1 and BRCA2, these two high penetrance genes encode for proteins involved in DNA damage response and repair (3). Among endogenous factors, oxidative stress (OS) produces potentially mutagenic reactive oxygen species (ROS) and can play an important role in breast carcinogenesis (4); since breast tissue can be particularly exposed to OS due to estrogen metabolism and hormonal status (5-9). Considering that DNA damage may contribute to breast cancer development, an efficient repair of oxidative lesions is expected to protect mammary cells from neoplastic transformation. Among the genes involved in the response and repair of DNA damage, BRCA1 has been shown to have a decisive role, since it induces the gene expression of the antioxidant response and thus protecting cells from OS (10). In particular, BRCA1 stimulates the activity of key base excision repair (BER) enzymes, including 8-oxoguanine DNA glycosylase (OGG1), primarily by increasing transcription of BER enzymes (10). The BER pathway is responsible for the repair of many DNA nucleobases modified, indeed its action is essential for the maintenance of genetic integrity and stability (11) and in playing a crucial role in repair of DNA damage induced by ROS (12). BER glycosylases may provide a genome surveillance mechanism and may act as molecular sensors that induce apoptosis in response to extensive DNA damage through interaction of complex pathways (13-15). A deficiency and/or inactivity of the BER DNA glycosylase enzymes can induce deleterious outcomes in the cells driving the onset of various tumors (11). In particular, OGG1 (MIM 601982) and MUTYH (MIM 604933) remove DNA oxidative purine lesions and seem involved in the regulation of cell-cycle progression and cell division under OS (16,17). In fact, low activity and/or reduced expression of MUTYH and OGG1 enzymes may result in DNA repair impairment and failure to induce apoptosis in response to oxidative damage, resulting in survival of cells with oncogenic mutations (12,18,19). In previous studies the contribution to BC risk due to MUTYH seems to be limited or not relevant, although in most researches only specific mutations of MUTYH were considered, which had previously been characterized in families with gastrointestinal polyposis (20-22). Also the OGG1 contribution to BC risk has been significantly associated with the presence of some polymorphic nucleotide markers (SNPs) (21,23-25). However, these associations may have been influenced by the characteristics of the population examined or by the specificity of each germinal variant considered (21-24). To date, the contribution to BC or OC predisposition due to the reduced expression of these genes has not yet been considered in humans. Some doubts remain about the genetic susceptibility related to low penetrance genes and their expression, which could contribute to the onset of BC and/or OC (BC/OC) in people belonging to families with tumor phenotypes other than BC/OC (e.g. pancreas, thyroid and colon) or in patients with early onset of cancer and without family history (26). The patients with these characteristics and without germline mutations in high-penetrance BRCA1/2 genes might be a good model to clarify some aspects that contribute to the multifactorial etiology of BC and OC.

In light of the above, it can be assumed that genetic variability related to low expression in enzymes that protect cellular DNA from oxidative damage, causing genetic instability, may favor the onset of BC and OC.

In the present study we investigated MUTYH, OGG1 germline mutations and mRNA expression levels, in the peripheral blood mononuclear cell (PBMC) from patients with and without mutations and compared the gene expression with control individuals. In addition, to evaluate the relationship of these BER glycosylases (MUTYH, OGG1) and BRCA1 pathway, we investigated the possible mutual modulation. This exploratory study on correlation among mutational spectrum, gene expression and demographic characteristics, could improve the genetic diagnosis performing predictive testing of at-risk individuals belonging to families with reduced mRNA expression regardless of the mutation presence. The identification of carriers with reduced mRNA expression may be useful for improving clinical management of patients.

In this study we aimed to evaluate the involvement of two BER glycosylases (OGG1 and MUTYH) in the predisposition to breast and OC. We also investigated the correlation among BRCA1 and these BER genes expression, in the contest of the BC and OC predisposition. In this regard we evaluated a retrospective series, previously selected for genetic analysis, of the main breast and OC predisposing genes: BRCA1 and BRCA2 (27,28).

Since BRCA1 is involved in oxidative stress regulation and BER after oxidative damage (35), we evaluated germline mutations status and gene expression of the MUTHY and OGG1 associated to clinical characteristics of 59 BC/OC cases and to BRCA1 gene expression.

In this series we identified: 4 novel variants and one of them (in MUTYH) is predicted deleterious; one known deleterious mutation in OGG1; 6 VUS (3 in MUTYH and 3 in OGG1) (Table I). The OGG1 deleterious mutation, c.137 G>A, causing a substitution from basic to acidic amino acid (p.Arg46Gln), has never been reported in BC/OC patients. While, it was previously described as deleterious germline variant in non-polyposis hereditary colorectal cancer (HNPCC) with stable microsatellites (MSS) (33,36). In our study we found this mutation in a case (B58) that referred BC (age range, 40-45 years), BC direct transmission, bilateral ovariectomy and a brother died for leukemia.

It is interesting to note that the three novel missense mutations were found in three out of eight OC/BOC patients with OC family history (Table II); furthermore, one of these patients (B105), carrying a truncating BRCA1 mutation (E1373X), was also affected by BC. It is possible that, in this case, two defects simultaneously contributed to the onset and progression of the tumor in the breast tissue, although the truncating BRCA1 mutation doesn't affect gene expression levels in PBMCs, but its expression value falls within the BRCA1 median of the cases. The precise mechanisms that govern mutant allele penetrance depend on many factors, including personal and/or reproductive history, mutation location, and actually undefined genetic factors (‘modifier genes’) (26).

The BC, as extra-colic manifestation, occurred in 18% of female patients affected by MUTYH-associated polyposis (MAP) (20). To support these evidences, the MUTYH knock-out mice are prone to develop mammary tumors (37). We did not observe MUTYH mutations related to MAP, while we found 2 missenses and 1 untranslated variant (p.Pro18Leu, p.Gly25Asp, c.36+11C>T) in a patient (B66) affected by BC/TC, with family history for BC and polyposis (Table II). The Pro18Leu and Gly25Asp missense mutations were previously reported as occurring in the same MUTYH allele (38).

The presence of MUTYH rare variants resulted associated with reduced transcript expression of the same gene in carriers vs. non carriers (P=0.030), whereas the most frequent SNPs rs3219489 in MUTYH and rs1052133 in OGG1 did not affect mRNA expression. From this study a new assumption emerges that in BC/OC subjects, the MUTYH rare variants exert a gene pressure on the reduction of MUTYH expression, as already reported in MAP cases carrying MUTYH mutations (33).

The time lag between diagnosis of cancer and sampling, menopausal status, and cigarette smoking did not influence the median expression of BRCA1 and MUTYH whereas OGG1 expression displayed a significant rise (P=0.030) in the cases presenting a direct family history for tumor different than BC.

It is shown that OGG1 is involved in the acute and systemic inflammatory response that may favor carcinogenesis (39-41). OGG1 expression showed an increase in post-menopausal women suggesting that the physiological menopause-related decrease of estrogens may increase OGG1 expression in PBMC; this relation was already found in other female tissues (42).

These data, obtained from the mRNA analysis on PBMC, are very interesting and deserves further studies also on the MUTYH role in the sphere of immune functions, since the loss of this gene appears to be associated with immunosuppression and impairment of the inflammatory response (43). We observed that the expression of MUTYH and OGG1 showed significant correlation only in PBMC from cases (P=0.004) (Table V). This correlation has been verified in cell lines derived from various tissues, stressed with hydrogen peroxide (our laboratory data not shown), confirming that alterations in the redox balance and BER function are involved in the promotion and progression of cancer (12), although inter-individual differences in the oxidative stress regulation can explain a part of the variability in cancer susceptibility.

In this study the expression of OGG1 always correlated with that of BRCA1 both in cases and population controls without cancer family history (P=0.001, P=0.011 (Table V). This data agrees with evidences that BRCA1 have a role in the regulation of OS (35,44) and suggests a possible crosstalk between BRCA1 and OGG1. Consistent with these findings, Saha T. and colleagues (10) in 2010 found that BRCA1 over-expression increases the enzymatic activities related to the BER pathway while its under-expression decreases them. These finding prompt the thought that BRCA1 might exert its tumor suppressive function through oxidative stress regulation. In the context of the BC and OC predisposition BRCA1 showed low expression in cases, although it increased in controls with no family history of cancer, as also observed for BER genes. This feature has not yet been adequately explored and deserves further study. The aspects that relate BRCA1 and BER molecules, in response to ROS and carcinogenesis, are also supported by the fact that the BRCA1 Loss of Heterozygosity (LOH) induces an increase of 8-oxoG levels (35). Since, BRCA1 wild-type promotes 8-oxoG lesions repair via transcriptional regulation of BER. This mechanism was exploited in the therapeutic block of PARP in patients harboring BRCA1 mutations (35).

Wild-type BRCA1 expression suppresses basal and H₂O₂-induced ROS production in breast and ovarian cell models (10,44). Endogenous factors may play a role as well salient in promoting the effects of oxidative stress on breast and ovarian carcinogenesis. Indeed, some studies have shown a significantly higher level of oxidative DNA damage in normal breast tissues derived from cancer patients (45). Non-physiological OS can be decisive in the pathogenesis of cancer, in fact, it is known to induce phenotypic modifications of cancer cells through cross-talk with the surrounding stroma (46). For these reasons, individuals predisposed to BC/OC can undergo a high rate of mutations oxidative stress-related due to the deficiency on systems to repair DNA damage. Furthermore, basic metabolic changes could produce an increase in potentially reactive oxygen species, in tissues like breast and ovary, which already have a physiological exposure to oxidative stress due to specific hormonal metabolism (9) and or inflammation (39). The roles of BER go beyond maintaining DNA integrity, as they are also implicated in the metabolism regulation (47).

This study highlights that germline MUTYH/OGG1 transcript levels may reflect physiological and pathological changes that induce a different status in patients and in controls with and without cancer family history. This is the first study that evaluated germline mutations and expression of MUTYH and OGG1 genes in BC/OC patients in relationship to BRCA1, investigating their reciprocal modulation. From this exploratory study emerged interesting and significant correlations among these three genes related to cancer predisposition.

Relationship among mutational spectrum, gene expression and demographic characteristics, could improve the genetic diagnosis performing predictive testing of at-risk individuals belonging to families with reduced mRNA expression regardless of presence of mutation. Finally, an accurate evaluation of the reduced expression of MUTYH and OGG1 genes in PBMCs could represent a useful way to monitor primary prevention and clinical management of cancer-free patients.