The management of patients with breast carcinoma currently uses the following prognostic factors: Disease stage (which takes into account axillary lymph node involvement, tumour size and distant tumour dissemination), degree of differentiation (tumour grade), histological type, proliferation index and receptor status [progesterone receptor (PR), oestrogen receptor (ER) and receptor 2 of the human epidermal growth factor (HER2)] of the primary tumours (1–2). Among these features, the expression levels of hormone receptors appear to best predict the breast cancer response to different therapeutic strategies (2–3).

Triple-negative breast cancer (TNBC) has been classified as a breast cancer subgroup that is negative for ER, PR and HER2 expression. TNBC accounts for 15 to 20% of breast cancer cases (4). Despite a notably favourable rate of response to chemotherapy, TNBC patients present with a higher risk of relapse and a relatively poor outcome (5).

Mutations in the BRCA1 and BRCA2 tumour suppressor genes have been associated with breast cancer risk among families with strong recurrence of the disease, whereas no clear role of the BRCA1/2 mutations has emerged for the majority of breast cancers occurring sporadically in individuals with little or no family history (6–8). Overall, BRCA1/2 mutations have a prevalence of ~5% in the general population and ~25% in the families with a history of breast cancer (7–8). Among BRCA1 mutation carriers, TNBC represents the predominant breast cancer subtype (more than two-thirds of cases). BRCA1 germline mutations have been observed in up to one-third of TNBC patients (mainly among those with an age at diagnosis of <45 years) (9). One possible role for BRCA2 mutations in TNBC has been reported previously (10). Using data from BRCA mutation analysis in different TNBC cohorts, a low BRCA2 mutation frequency (5% average) has been indicated among this subset of patients (9–10).

In Sardinia, breast cancer is the principal malignancy that causes mortality, with an incidence rate comparable with that observed in other Western countries (standardized rate, 95 per 100,000 inhabitants per year), and the median age of onset for breast cancer among Sardinian females is 65 years (11). Among breast cancer patients from Sardinia, the contribution of BRCA1/2 mutations to the population incidence of such a disease has been extensively evaluated by our group in previous years (12–14). The Sardinian population is considered genetically homogeneous due to its high rate of inbreeding and the subsequent inheritance of numerous common genetic traits (15–16). This may be instrumental in further defining the association between the germline mutations in the two genes (BRCA1 and BRCA2) and the TNBC status. The geographical distribution of the BRCA mutations in the Sardinian breast cancer population has been demonstrated to be particularly heterogeneous; in the northern and middle areas of the island, BRCA2 mutations are the most common genetic variants (with a predominant founder mutation), while in the southern area, BRCA1 mutations are largely prevalent instead (14).

In the present study, the association between the occurrence of BRCA1/2 mutations and the TNBC status among breast cancer patients from Sardinia was investigated.

The germline DNA from 726 consecutive breast cancer patients, who provided informed consent and were included in the study, was investigated for mutations in the BRCA1 and BRCA2 genes, as previously described (14). Briefly, mutation screening was performed by DHPLC analysis; all PCR products presenting an abnormal denaturing profile in comparison to the normal controls were sequenced using an automated approach. Overall, deleterious BRCA1/2 mutations were detected in 21/726 (2.9%) breast cancer cases. In particular, BRCA1 mutations were detected in 7/21 (33.3%) patients, while BRCA2 mutations were identified in the majority of patients (14/21; 66.7%).

When the mutation status was evaluated according to the age of breast cancer onset, 11/285 (3.9%) patients with an age at diagnosis of ≤50 years were found to carry a BRCA1/2 mutation (Table I). No significant difference was observed between these patients and the remaining patients with a diagnosis age of >50 years [10/441 (2.3%) BRCA1/2 mutation-positive (Table I)]. Of the patients with a BRCA1/2 mutation, 11/21 (52.4%) were ≤50 years old at the time of diagnosis (Table I); no statistical correlation was detected.

Using Pearson’s χ² test, the occurrence of a BRCA1/2 mutation was evaluated for association with the expression levels of oestrogen, progesterone and HER2 receptors. The distribution of BRCA1/2 mutation carriers and non-carriers was similar in the different subsets of patients, according to such pathological parameters, and therefore no statistically significant correlation was observed (Table IIA–C). When the analysis was focused on patients with TNBC classification, the rate of BRCA1/2 mutations was significantly higher in the group of patients with a triple-negative primary tumour compared with those without such a feature [7/49 (14.3%) vs. 14/677 (2.1%), respectively; P=0.012; Table IID].

As listed in Table III, the presence of a triple-negative phenotype was strongly associated with the BRCA1 mutations at a highly significant level (P<0.001), whereas no association was found with the BRCA2 mutations (P=0.837). With respect to patient origin within Sardinia, the distribution of mutations was confirmed as significantly heterogeneous: 6/7 (85.7%) mutated cases in Southern Sardinia were represented by BRCA1 mutations, while 13/14 (92.9%) mutated cases in Middle-Northern Sardinia were constituted by BRCA2 mutations (P<0.001; Table III; Fig. 1). These discrepancies did not result from incorrect standard sequencing, as confirmed by an independent duplicate analysis. As a further indication of a strong correlation between TNBC status and the type of BRCA mutation, patients from the geographical area with prevalent BRCA1 mutations presented a frequency of triple-negative tumours much higher than that observed in cases from the area with prevalent BRCA2 mutations (11.8 vs. 4.3%, respectively; P=0.037).

In the present study, a low prevalence (3%) of patients with germline mutations in coding regions and splice boundaries of BRCA1 or BRCA2 was observed. Breast cancers carrying BRCA1/2 germline mutations often occur in younger women, and present a lack of ER/PR expression (mostly among BRCA1-positive tumours) (8,17–19). In our series, the age at diagnosis was younger in patients carrying BRCA1/2 mutations [11/21 (52%) <50 years vs. 10/21 (48%) >50 years] than in cases with no detectable mutation [274/705 (39%) <50 years vs. 431/705 (61%) >50 years] (Table I); however, such a difference was not statistically significant.

In Sardinia, the proportion of breast cancer patients with a diagnosis of TNBC was much lower than expected [7% instead of 15–20%, as reported in breast cancer worldwide (4)]. Among the TNBC cases, the rate of BRCA1/2 mutations was significantly higher (14%; P=0.012); such a significance was due to the occurrence of BRCA1 mutations (Table III). As a confirmation of this, patients from South Sardinia (the geographical area with the preponderance of BRCA1 mutations) presented a significantly higher TNBC frequency compared with that in breast cancer cases from Middle-North Sardinia (the area with prevalent BRCA2 mutations); ~12 vs. 4%, respectively (P=0.037).

In a population sharing a similar lifestyle and diet habits across the island, these results strongly indicate that a different ‘genetic background’ may indeed affect the phenotypic characteristics in the onset of a complex disease such as breast cancer. Similar data was reported by our group for melanoma and colorectal carcinoma (20–21); within the island, the geographical distribution of the genetic variants appears to be correlated to the specific large areas of Sardinia, which reflect its ancient history: The northern area, which is linguistically different from the rest of the island and delineated by the mountain chain crossing Sardinia, and the middle-southern area, which is the domain of pastoral culture and the land of the ancient Sardinian population. Together, these findings clearly indicate that mutation frequency for candidate cancer genes requires accurate evaluation in each geographical area within every single population.

Finally, it is noteworthy that the occurrence of TNBC was significantly associated with the BRCA1 mutation carrier status (P<0.001), regardless of the geographical origin of patients in the present series. It may therefore be hypothesized that the simultaneous lack of expression of ER/PR/HER2, mainly when associated with an early diagnosis age, is somehow predictive for the presence of BRCA1 germline mutations. The absence of an association between TNBC classification and BRCA2 mutations in the present study was consistent with data previously reported [reviewed in (4)].

Overall, the prevalence of BRCA1/2 mutations in TNBC cases remains limited and characterizes only a subset in this group of malignancies. The present data, along with those from the literature, support the hypothesis that additional breast cancer genes should be tested in such patients, in order to improve the subclassification of the heterogeneous TNBC disease from a genetic point of view. Further effort is required to reduce the fraction of patients with a ‘non-mutant’ TNBC status.