Recently the combination of the mammalian target of rapamycin (mTOR) inhibitor everolimus and the aromatase inhibitor exemestane has been shown to double the progression-free survival rate in advanced breast cancer. However, the effect of the interrelated pathways of hypoxia-inducible factor-1α (HIF-1α) and mTOR signaling, both of which are associated with a more aggressive breast cancer phenotype and endocrine resistance, on response in the neoadjuvant setting is unknown. We, therefore, have investigated the influence of these pathways with the aim of better defining those patients most likely to benefit from an endocrine-based therapy associated with/without mTOR inhibitors.
A total of 107 women with T2-4 N0-1 and estrogen receptor-positive breast cancer were randomly assigned to 6 months of primary letrozole (2.5 mg/daily) (LET) or LET plus oral “metronomic” cyclophosphamide (50mg/daily) (LET-CYC). Phospo-mTOR and HIF-1α were evaluated in tumor specimens collected before and after treatment using a tissue microarray format.
LET-based therapy induced a downregulation of phospho-mTOR and HIF-1α expression (P = .0001 and P < .004, respectively). The reduction of HIF-1α expression observed was positively correlated with phospho-mTOR reduction (P < .03); however, no treatment interaction between the two proteins was detected. HIF-1α expression was significantly modulated by the treatment (P < .004) with a reduction both in the LET arm (45%, n = 36/80) (P = .05) and LET-CYC arm (55%, n = 44/80) (P = .04). HIF-1α reduction showed a relationship with clinical response confined in LET arm only (P < .03).
In this neoadjuvant population, LET was able to modulate the phospho-mTOR and HIF-1α pathways and may define a subpopulation of nonresponders who may be most likely to benefit from mTOR inhibitors.
Aromatase inhibitors (AIs) are successfully used for the therapy of postmenopausal breast cancer (1) and are associated with increased patient survival compared with women with more aggressive estrogen receptor (ER)-negative tumors (2). Nevertheless, a significant number of patients treated with AIs do not respond or develop resistance. This has led to an intensive research effort into the mechanisms of endocrine sensitivity and resistance in breast cancer, which has revealed several growth factor pathways involved in bypassing the effects of endocrine treatment (3). One such pathway is PI3K/AKt/mTOR that critically influences cell growth, proliferation, and angiogenesis is also intimately involved in endocrine resistance as its activation leads to ligand-independent ER-α transcription through S6 kinase phosphorylation of Ser167 on the AF-1 domain (4,5). Nonetheless, the potential of interruption of PI3K/AKt/mTOR signaling has been demonstrated in preclinical models, where the mTOR inhibitor, everolimus with an AI led to abrogation of proliferation and induction of apoptosis and enhanced tumor regression (6,7). In the clinical trial context in ER-positive patients treated with primary systemic therapy, everolimus and letrozole (LET) resulted in an enhanced response rate than with an AI alone (8) and more recently, in the BOLERO-2 (NCT00863655) trial in advanced breast cancer, the addition of everolimus to exemestane more than doubled the progression-free survival compared with exemestane alone (9) (for a review of clinical trials, see Zagouri et al. (10)).
Our group along with others have previously shown that hypoxia-inducible factor-1α (HIF-1α) is a mediator of endocrine resistance in breast cancer (11,12). Studies have also shown that mTOR itself is involved in O(2)-sensitive signaling (13) and regulates HIF-1α expression in cancer (14,15). Since no conventional pathological or treatment variable is currently able to predict the degree of benefit, understanding the effects of AIs on the mTOR and HIF pathway may help define subgroups most likely to benefit from endocrine-based treatment and also define another group for alternative interventions.
To investigate this we have performed a biomarker study of these two target proteins in a phase II randomized LET-based clinical trial as primary systemic treatment in elderly breast cancer patients (16). In a previous explorative analysis, we also observed that LET plus/minus cyclophosphamide (CYC) was able to downregulate mTOR expression (17). Since there are such close molecular and signaling links between the ER, HIF, and mTOR signaling, we have now investigated the relationship of these interrelated pathways and endocrine response using LET therapy to test whether these pathways predicted response to treatment or were prognostic. The aim was to define a biomarker signature that can be used to determine those patients likely to respond to an AI and mTOR inhibitor combination.
Materials and Results
Patients and methods are previously fully described (16). The local ethical committee approved this substudy. Written informed consent was obtained from all patients before randomization (16). Histopathologic grade and immunohistochemistry are described elsewhere (16) and they were performed at the Pathology Unit of the Azienda Ospedaliera Istituti Ospitalieri of Cremona (Italy). Immunohistochemistry for HIF-1α, phospho-mTOR (Ser2448) rabbit mAb, and phospho-ER alpha (pERα) (Ser118) (16J4) mouse mAb stainings whose methodology is already described elsewhere (12), was performed on 5μ sections of tissue microarrays containing two 1-mm tumor cores (on tissue microarrays) taken from selected morphologically representative tumor regions of each paraffin-embedded breast tumor from both the initial diagnostic incisional biopsy and from tumor remaining at definitive surgery. The immunostaining for hypoxia-related markers and phospho-mTOR were quantified in carcinoma cells by semiquantitative scoring as previously described (12,17) blinded to patient outcome and whether the samples they examined was obtained from incisional biopsy or definitive surgery. Briefly, HIF-1α was scored 0 (no staining), 1 (weak staining) or 2 (strong staining) and phospho-mTOR were scored 0 (no staining), 1 (weak staining), 2 (moderate staining), and 3 (strong staining). Tumors were considered positive for each marker when any staining was present. Patient and tumor characteristics are described elsewhere (17). One hundred and seven patients had HIF-1α and phospho-mTOR assessed at baseline (51 randomized to receive LET and 56 to receive LET-CYC). A total of 97 patients (45 in LET and 52 in LET-CYC arm) received definitive surgery (16).
In total, 80 out of 107 patients showed HIF-1α expression at baseline, ranging (score 1+) from weak to strong (score 3+) and it was significantly positively correlated with baseline phospho-mTOR expression (P = .01). As previously reported LET-based therapy was associated with a significant reduction in phospho-mTOR expression across both LET (P = .001) and LET-CYC (P = .0001) arms of the trial (17) and comparatively HIF-1α expression was also significantly reduced by the treatment (P < .004). HIF-1α reduction occurred in both the LET arm (45%) (n = 36/80) (P = .05) and the LET-CYC arm (55%) (n = 44/80) (P = .04) and was positively correlated with phospho-mTOR reduction after treatment (P < .03) (Figure 1). However, there was no treatment interaction between HIF or mTOR in either arm of the study. Similarly there was no interaction with treatment when using the reduction of HIF-1α or mTOR. Disease response was observed in 37 out of 51 patients randomized in the LET arm (72.5%) and in 49 out of 56 patients randomized in the LET-CYC arm (87.5%). Dividing patients according to the treatment arm, the negative predictive effect of baseline HIF-1α on treatment response was evident in LET arm patients (P < .03) but not in the LET-CYC arm patients (P = .84) (Table 1).
* HIF-1α = hypoxia-inducible factor-1α; LET-CYC = letrozole-cyclophosphamide.
Although AIs are the most efficacious endocrine therapy for breast cancer patients, it is important to understand the mechanisms of de novo or acquired resistance to these compounds, to develop rational combination therapy. Several mechanisms have been hypothesized to be involved in the resistance of breast cancer cells to AI therapy (see review Johnston and Dowsett (18)). Recently our group showed, in the same series of patients examined in this study, that resistance to LET is associated significantly with baseline expression of HIF-1α together with p44/42MAPK (12). In the present paper, we explored the effect of LET treatment on mTOR and HIF-1α protein expression. Although we have previously reported that LET is able to reduce the HIF-1α expression, in this paper we provide important additional clinical information on the underlying molecular mechanisms in relation to mTOR expression. LET plus/minus CYC is able to significantly reduce the mTOR expression (17) and as mTOR is an upstream regulator of HIF-1α expression, the reduction in HIF-1α observed can be mediated by the reduction in mTOR expression. The significant direct relationship between mTOR and HIF-1 α expression at baseline condition and the significant correlation between treatment-induced reduction of mTOR and reduction in HIF-1α are consistent with this hypothesis.
Although mTOR and HIF-1α expression were significantly reduced by treatment, HIF-1α showed a negative predictive effect in LET-treated patients only. The chemoendocrine combination therapy could be responsible for a dampening of the HIF-1α response by modulation of vascular permeability and interstitial fluid pressure via vascular endothelial growth factor (19,20).
In conclusion, LET is able to modulate the expression of HIF-1α in breast cancers. The downregulation of intratumoral-estrogen levels induced by LET-inducing reduction in mTOR expression, could be an explanation of the observed downregulation of HIF-1α. LET modulation of phospho-mTOR and HIF-1α pathways may define a subpopulation of nonresponders who may be most likely to benefit from the combination with an mTOR inhibitor or an HIF-1α inhibitor.
This work is supported by the Victorian Breast Cancer Research Consortium (Australia), Fondazione Popolare di Cremona (Italy) and ARCO (Associazione per la Ricerca in Campo Oncologico) (042015) Onlus-Cremona (Italy).
- signal transduction
- aromatase inhibitors
- neoadjuvant therapy
- breast cancer
- frap1 protein, human
- estrogen receptor positive
- mtor inhibitors
- tissue microarray
- endocrine resistance
- transverse spin relaxation time
- progression-free survival