Triple-negative breast cancers (TNBC) account for approximately 15–25% of all breast cancer. They are characterized by the lack of expression of estrogen, progesterone receptors (ER/PgR) and human epidermal growth factor receptor 2 (HER2) and by an aggressive clinical course with higher rates of relapse and poor overall survival in metastatic disease (1–3). Treatment options for patients with TNBC are limited due to the absence of hormone receptors and HER2; therefore, cytotoxic chemotherapy currently remains the only available treatment (4).

Given these characteristics, TNBC is a challenge in clinical practice.

Several studies demonstrated that a subgroup of TNBC patients displayed a remarkable sensitivity to chemotherapeutic agents. Between 17 and 58% of TNBC patients have been shown to achieve pathological complete response (pCR) after anthracycline/platinum-based neoadjuvant chemotherapy and these patients had an excellent prognosis. On the contrary, those who failed to achieve a pCR had an exceptionally poor outcome (5,6). Over the past decades there have been several attempts to use genomic data in order to explain the highly variable responses to therapy and clinical outcome of this setting of patients. Recently, genomic analyses have provided additional insights, showing a wide heterogeneity of molecular characteristics of TNBCs by gene expression profile. On this basis, Lehmann et al (7) identified six different TNBC subtypes including basal-like (type 1 and 2), immunomodulatory, mesenchymal, mesenchymal stem like and luminal androgen receptor subtype demonstrating the heterogeneity of TNBC.

Ongoing research into the molecular and genetic mechanisms of TNBC tumorigenesis are helping to find out processes involved in local tumor progression and distant metastases.

One of the most important mechanisms involved in the control of neoplastic transformation is the PI3K/Akt/mTOR pathway. The aberrant activation of this cascade seems to be of great importance in breast cancer. In addition, a high activation level of the PI3K/Akt/mTOR pathway has been related to worse prognosis and resistance to conventional chemotherapy.

The mammalian target of rapamycin (mTOR) is an important serine/threonine protein kinase of the phosphoinositide 3-kinase (PI3K)-related kinase family, which functions as an environmental sensor and regulates organismal growth, cell physiology and homeostasis. mTOR is the catalytic subunit of two distinct complexes, mTOR complex 1 and mTOR complex 2 (mTORC1 and mTORC2), which consist of several additional regulatory proteins. The subunit composition of each mTORC dictates its substrate specificity. Main substrates of mTORC1 are S6 kinase 1 (S6KB1) and eIF4E-binding protein 1 (4E-BP1), both implicated in the regulation of mRNA and protein synthesis. mTORC2 controls several members of kinases including Akt and is thereby implicated in the regulation of cell survival, cell cycle progression and anabolism (8,9).

Previous in vitro studies showed that phosphorylated mTOR (p-mTOR), the activate form is closely related with the active status of mTOR and the cell proliferative capacity (10,11).

p-mTOR was found more frequently expressed in triple-negative than non-TNBC, suggesting that mTOR may play a crucial role in the molecular biology of TNBC. Besides, emerging preclinical evidence suggested that TNBC cells seem particularly sensitive to mTOR inhibitors, especially the androgen receptor-positive (AR⁺) TNBC cell lines, opening the possibility for the incorporation of target agents in treatments (12–15).

The aims of the present study were to assess the expression of the activated form p-mTOR in early TNBC and to evaluate possible correlations between immunohistochemical AR expression, clinicopathological parameters and disease outcome.

The PI3K/Akt/mTOR pathway regulates several cellular functions such as cell growth, survival and proliferation, characterizing tumorigenesis as well as tumor progression (18). In breast cancer, a high activation level of the PI3K/Akt/mTOR pathway has been related to resistance to conventional chemo and endocrine therapy. The recent BOLERO-2 trial comparing everolimus plus exemestane vs. placebo plus exemestane in women with resistance to no-steroidal aromatase inhibitors demonstrated a 6-month improvement in progression-free survival leading to the approval for the treatment of ER-positive metastatic breast cancer.

Previous in vitro studies showed that p-mTOR correlated with the activation of mTOR and an increase in proliferation (10,19). In the present study, the high expression of the active form of mTOR (p-mTOR) was present in almost one third of TNBC, suggesting that aberrant activation of the PI3K/Akt/mTOR may drive tumor proliferation in this subtype of breast cancer. Although there are no therapeutic evidence using mTOR inhibitors in TNBC, these data may open new therapeutic scenarios, also suggested by in vitro and in vivo assays, in which mTOR inhibitors demonstrated anti-tumor activity in TNBC respectively in cell lines and mouse xenograft models (18,20,21) indicating mTOR as the potential target.

Three reports have previously been published on the expression of p-mTOR in breast cancer; Bose et al (22) found increased levels of p-mTOR in high grade vs. low grade cancers. However, Zhou et al (13) found no relationship between p-mTOR and tumor grade. Recently Walsh et al (12) showed a high expression of p-mTOR in ~36% of triple-negative breast carcinomas; this result is consistent with the present study. Furthermore, they revealed that p-mTOR was significantly more frequently expressed in triple-negative than non-triple negative diseases, suggesting that inhibitors directed against this protein may be effective in at least some patients affected by this subtype of breast cancer.

In our results, 32.6% of cases were p-mTOR positive and its high expression was not related to the considered clinicopathological features neither to DFS and OS at the univariate and multivariate survival analysis. However, in patients who had small tumor size and early stage, p-mTOR positivity was significantly higher and that is consistent with a previous study (13). Of interest, the present investigation is the first clinical retrospective study showing the strong correlation between p-mTOR immunostaining and AR positivity in a subgroup of TNBC. It is consistent with other reports and it confirms microarray analysis recently conducted on TNBC (23).

A spectrum of somatic mutations have been discovered in TNBC; mutations in PIK3CA (10.2%), the gene that encodes the p110α catalytic subunit of phosphatidylinositol-3 kinase (PI3K) are the most common. Lehmann et al (23,24) observed that all AR-positive (AR⁺) TNBC cell lines contain the PIK3CA mutation (H1047R) and are highly sensitive to the PI3K/mTOR inhibitor NVP-BEZ235, suggesting that combination of AR antagonism and PI3K inhibition may have a synergistic effect on AR⁺ TNBC cell growth. Collectively, these findings are similar with other reports (1) and consistent with observations that hormonally responsive cancers, such as those expressing ER and AR are more likely to acquire PIK3CA mutations (24).

Up to date, in BC AR expression and relation with the PI3K pathway were studied in cell lines or xenograft models and the exact mechanism of action of AR in TNBC is still controversial (1,24–26). Due to the uncertain biological significance of the relationship between mTOR and AR, we defined the regulation pathway linking mTOR with AR by means of literature analysis and referring to Reactome (www.reactome.org) and KEGG databases (http://www.genome.jp/kegg/pathway.html).

The relationship between AR expression and the mTOR pathway could be explained by inherent biological data from literature analysis (Fig. 4). It should be noted that miR-21 and miR-34a, key elements of our pathway, were shown to be expressed in TNBC (20,26,27,28). In fact, there are many modifier pathways, specifically GTPase activating proteins TSC1 and TSC2 are negative regulators of mTORC1 and positive regulators of mTORC2 (29). These proteins act on Rheb GTPase hydrolyzing the GTP converting it to Rheb-GDP complex. Rheb is activated bound to GTP and turned off when bound with GDP, so respectively triggering or defusing mTORC1. Activated mTORC1, by phosphorylating some downstream effectors, gives rise to an increased protein translation and autophagy inhibition. In particular, it activates RPS6KB1 kinase that, by phosphorylating RPS6, induces cell growth and proliferation. Moreover, mTORC1 phosphorylates EIF4EBP1 that releases eIF4E so activating the translation of various mRNAs including HIF1A. This is the α subunit of a transcription factor that, in response to hypoxia, activates transcription of genes involved in regulation of erythropoiesis, angiogenesis, vascular tone, matrix metabolism, glucose metabolism, cell proliferation and survival, apoptosis (VEDF, IGF-2 and EPO) (30). Moreover, HIF1A induces DDIT4 transcription that, through the dissociation of 14-3-3 inhibitory protein from TSC2 (31), activates TSC2 that in turn inhibits mTORC1. eIF4E not only plays a key role in translation, but also in the nucleus where it promotes the export of specific mRNAs, as c-myc, Mdm2, NBN, ornithine decarboxylase (ODC1), and cyclin D1, that support proliferative and survival signalling pathways (32–34). MDM2, by ubiquitination, leads to p53 transcription factor degradation by the proteasome and, in turn, to miR-34 family repression, the latter is a direct transcriptional target of p53 (35) and the miR-34 silencing is frequent in several tumors, including breast cancer, and correlates with metastasis and poor survival (36). Since miR-34a/c can regulate AR (37), miR-34 repression causes lack of AR inhibition. AR promotes miR-21 transcription by some androgen responsive elements (AREs) in miR-21 promoter (in particular, three binding sites are known: ARE1 and ARE2/3 (38,39). The oncogene miR-21 indirectly suppresses p53 with a feedback loop mechanism (40) and downregulates some tumor suppressor genes, including PIAS3 (41), giving rise also to STAT3 upregulation. Furthermore, miR-21 blocks PTEN (42–44), the phosphatidylinositol-3,4,5-trisphosphate 3-phosphatase that dephosphorylates PIP3 to PIP2 negatively regulating AKT/mTORC1 signalling pathway. This pathway analyses demonstrates a correlation between activated mTORC1 and AR expression, of which the key elements are p53, miR-21 and miR-34. Therefore, low level of activated mTORC1 yields low AR expression.

It seems that some microRNAs are deregulated in TNBC. Particularly, miR-21 and miR-34a were significantly overexpressed in breast cancer, but miR-21 was significantly overexpressed in TNBC vs. non-TNBC, moreover, it seems to be associated to occurrence of lymph node metastases (27).

Other authors showed that miR-185 was strongly downregulated in TNBC tissues and its ectopic expression suppressed tumor proliferation, directly targeting DNMT1 and E2F6 (45). We can suggest that miR-185 acts also by suppressing AR. Another study showed miR-126 downregulation (46) and, according to our pattern, this should increase AKT/mTORC1 activation by targeting PI3K but, on the contrary, reducing mTORC1 by its target TSC2. Instead miR-145 downregulation (46) should ameliorate prognosis by TSC2 restoration and so mTORC1 reduction. It is interesting that miR-101 and miR-125a were associated to metastasis (46) instead we show that they should block mTORC1. The evidence that miR-31 downregulation causes an enhancement in metastasis (47) can be explained because it is an AKT inhibitor. The highly migratory and metastatic characteristics of TNBC with low miR-200 family expression (48) can be due to the inferred p53 low levels. The fact that miR-205 is downregulated but miR-200a/b/c is upregulated (46) cannot be realized by our pathway. Certainly their regulation is much more complex than that shown here and much remains to be clarified. However, this pathway can be useful also to plan new experiments.

All these complex results and analyses suggested that a high expression of the active form of mTOR (p-mTOR) and consequently an aberrant activation of the PI3K/Akt/mTOR pathway may drive tumor proliferation and that is true for almost one third of TNBC. Currently, TNBC is a subset of breast cancer with no available targeted therapies and hence adverse clinical outcome. Those findings could provide important information as to the potential opportunity for novel targeted and personalized treatment for these women but further translational investigations regarding the therapeutic efficacy of mTOR inhibitors in TNBC are required. Our analysis also confirms the biological association between mTOR activation and AR pathway, suggesting that may exist a subgroup of TNBC in which the combination of both AR antagonism and mTOR inhibition should have a synergistic effect on cell growth and tumor progression.