MBC accounts for less than 1% of all breast cancers and little is known on the epidemiologic, genetic and clinico-pathologic features of this malignancy [31–34]. Available data indicate that the prevalence of BRCA2 mutations in MBC ranges between 4% and 40% [32–34]. These differences are probably related to the genetic features of the families, with higher rates of BRCA2 mutations in men with a strong family history of cancer, particularly of those belonging to the BRCA2 spectrum.
In our series, the high prevalence of BRCA2 mutation (about 40%) correlated with the high number of relatives with breast cancer present in these families. In fact, excluding from the analysis the two apparently sporadic MBC, the frequency of BRCA2 mutations rises up to 50%. The mean age of breast cancer onset in males with and without BRCA2 mutations is still a controversial issue [16, 32, 35]. In the present study, we found a median age of diagnosis of 54.5 years for the BRCA2 mutation carriers (range 40–70 years) and of 53.8 years (range 38–62 years) for the w.t. type cases, supporting the findings reported by others  indicating that BRCA2 mutation carriers and noncarriers have a similar age at the time of diagnosis.
It has been reported that about 15–20% of MBC patients had a first-degree relative with breast carcinoma, supporting the notion that a positive family history of breast cancer is associated with increased risk of MBC [36, 37]. In our survey, all MBC patients carrying deleterious BRCA2 mutations had a family history of breast cancer. This was not due to a selection bias, since the eligibility criteria used for MBC did not require the presence of a positive family history. These findings are in keeping with those reported by others indicating that a positive history of breast cancer is more frequent in BRCA2 gene mutation carriers, with values ranging from 13 to 80% [36–38], although opposite results have been also reported . In our 8 families with a positive history of breast cancer, the presence of a deleterious BRCA2 mutation was not associated with a different occurrence of specific types of cancers as compared to BRCA2-negative families. A significantly higher frequency of breast cancers was observed in the first- and second-degree relatives of mutation carriers, whereas the prevalence of breast cancers in third degree relatives was higher in w.t. families.
Similarly to what reported in other studies , also the majority (75%) of our MBC were infiltrating ductal or intraductal carcinomas. One case was classified as intracystic atypic papilloma a relatively infrequent subtype accounting for about 2–5% of all MBC . The patient carrying this tumor was negative to the molecular screening, as those reported in other series , ruling out the possibility that BRCA1/2 mutations may confer a significantly increased risk of developing this peculiar MBC histotype.
Available data on the prevalence of BRCA1 and BRCA2 mutations in Italian MBC patients are limited. In a previous study , Ottini et al. reported one BRCA1 and three BRCA2 mutations including two mutations recurring in central Italy (BRCA1 3345delAG and BRCA2 6696delTC).
The three BRCA2 mutations identified in this study were however different from those we observed. Notably, the MBC cases from our series were characterized by the recurrence of a restricted set of BRCA2 mutations detected in families coming from a limited geographic area (North East of Italy), suggesting the possible existence of a founder effect. Indeed, it has been previously reported that all carriers of the IVS16-2A>G mutation shared a common haplotype, indicating a likely founder effect mainly confined to the Slovenian population . Consistently, our haplotype analysis carried out in six Italian families with the IVS16-2A>G BRCA2 mutation allowed the detection of a common haplotype shared by all mutation carriers. These findings, taken together, support the possible occurrence of a founder effect involving the 5' region flanking the BRCA2 gene in both North-East of Italy and the neighbouring Slovenia, suggesting that the IVS16-2A>G BRCA2 mutations probably recognize a common ancestral origin. The IVS16-2A>G mutation was previously reported by Krajc et al.  in three breast cancer-only Slovenian families, whereas it was not identified in three families with ovarian cancer, suggesting that the cancer phenotype associated with this mutation is limited to breast cancer. Conversely, our pedigree's analysis of six Italian families with the IVS16-2A>G mutation disclosed the presence of one family in which two ovarian carcinomas were present along with breast cancer, indicating thus that the risk conferred by this mutation is not restricted to breast cancer. Moreover, in the six families with the IVS16-2A>G mutation of our database, two additional MBC were detected (relatives of patients BR195 and BR312), further supporting the increased risk of developing breast cancer in males conferred by this mutation. Since the IVS16-2A>G mutation can result in either a partial or total skipping of exon 17 , the increased risk for breast cancer in males due to this mutation may be related to these effects on BRCA2 gene transcription. Studies are under way in our laboratory to elucidate this issue.
With regard to the 9106C>T mutation, the data obtained by haplotype evaluation coupled with the reconstruction of chromosome phasing suggest but do not conclusively prove the occurrence of a possible founder effect also for this mutation. In fact, all the mutation carriers of Family 25 showed the same haplotype for the markers D13S290-D13S260-D13S1698-D13S171-D13S1695; however, the three isolated cases showed a different allelic pattern for the markers D13S260 and D13S1698 located within the 3' region flanking the BRCA2 gene. These results may underlie the occurrence of a possible recombination event involving these two markers in the 3' region flanking the BRCA2 gene as also suggested by the finding that all mutation carriers of Family 25 and the three isolated cases shared the same allelic pattern 6-6 for the 5' markers D13S171 and D13S1695.
In support of this possibility, the analysis of 8 different intragenic BRCA2 gene polymorphisms demonstrated that all carriers of the 9106C>T mutation shared the same haplotype. It can be hypothesized that the common haplotype shared by all mutation carriers could be comprised within a region starting from the 3' end of the gene and extending to the 5' flanking microsatellite markers.
The analysis of larger series of families is however required to conclusively assess whether the 9106C>T mutation is associated with a founder effect in the North East of Italy, enlarging thus the number of founder mutations described for the BRCA2 gene [17, 40]. If this will be the case, it would be of interest to verify whether, similarly to what observed for the IVS16-2A>G BRCA2 mutation, also the 9106C>T mutation is associated with a founder effect shared by people from North-East Italy and Slovenia.