Targeting Aberrant p70S6K Activation for Estrogen Receptor–Negative Breast Cancer Prevention

To reduce the burden of breast cancer, the fundamental strategy lies in prevention and early detection, which should be based on a better understanding of the molecular mechanisms driving early-stage breast diseases or lesions that progress to cancer. The mechanistic insight could also help to identify the appropriate patient population for personalized cancer intervention. Over the past two decades, extensive clinical trials testing the efficacy of selective estrogen receptor modulators (SERM, such as tamoxifen) or aromatase inhibitors on breast cancer prevention in high-risk populations have demonstrated that antiestrogen could reduce the development of estrogen receptor–positive (ER⁺) breast tumor by approximately 50% in various age groups (1–4). However, no benefit was observed for ER⁻ breast cancer in these trials; and so far, there are no effective agents available for the prevention of ER⁻ breast cancer. ER⁻ breast cancer is associated with HER2 (also named ErbB2) overexpressing (HER2⁺) luminal-type, EGFR-overexpressing (EGFR⁺) basal-type, and other signaling events dysregulating ER (5, 6). Low toxicity targeted agents for some of these aberrant signaling are available that may represent important opportunities to prevent ER⁻ breast cancer in women having tamoxifen-nonresponding (Tam-NR) atypia. Along this line, various targeted agents are being tested in clinical trials or approved by the FDA for the treatment of ER⁻ breast cancer, and the potential of some of these agents for prevention of ER⁻ breast cancer is also under clinical testing (7–9).

Aberrantly activated kinase pathways are ideal targets that can be specifically targeted with kinase inhibitors for the treatment of cancer and other diseases (10, 11). Several small-molecule kinase inhibitors, including lapatinib and saracatinib, have been shown to effectively prevent the development of HER2⁺/ER⁻ mammary tumors in animal models (12–14). Previously, using gene expression array, Speers and colleagues identified increased levels of multiple kinases in ER⁻ breast cancer, some of which regulate cell cycle, modulate immune response, or cluster at the ribosomal protein S6 pathway and MAPK pathway (15). These kinases were noted to confer advantages in progression to ER⁻ breast cancer. In particular, elevated S6 pathway signature is strongly associated with poor patient prognosis (15).

Aberrations at multiple levels may lead to the 70-kDa ribosomal protein S6 kinase (p70S6K) pathway activation, contributing to the development of breast cancer (16, 17). In the current study, our reverse-phase protein array (RPPA) study of breast tissue biopsy from women with high-grade Tam-NR atypia revealed high levels of Akt/p70S6K activation, and we confirmed that Akt/p70S6K signaling pathway was activated in early-stage breast disease. Therefore, we hypothesized that Akt/p70S6K activation in mammary atypia and other early lesions could drive initiation and progression of ER⁻ breast cancer, whereas agents targeting Akt/p70S6K may prevent ER⁻ and Tam-NR breast cancer in high-risk women. We investigated the effect of targeting this pathway with an Akt/p70S6K dual inhibitor LY2780301 in p70S6K-activated, ER⁻ human mammary epithelial cell (hMEC) model and the MMTV-neu mouse model of ER⁻ mammary tumors having p70S6K activation to develop targeted prevention strategies for ER⁻ breast cancer. We found that LY2780301 blocked the phosphorylation of ribosomal protein S6 and PRAS40, downstream of p70S6K and Akt, respectively, in hMECs and resulted in a drastic decrease in cell proliferation, although LY2780301 did not apparently reverse the acinar morphology of p70S6K-activated hMECs. Nevertheless, a brief treatment of LY2780301 in MMTV-neu mice with atypical hyperplasia (ADH) and mammary intraepithelial neoplasia (MIN) lesions was sufficient to suppress S6 phosphorylation in these lesions and led to a decreased cell proliferation in early-stage mammary lesions in vivo, suggesting that targeting p70S6K could deter abnormal proliferation of ER⁻ early mammary lesions.

Gene set enrichment analysis (GSEA) was performed using microarray data from GSE16873 (18) to find differential activated signaling pathways between 12 paired normal tissue and ADH lesions. GSEA software and MSigDB, including the C6 oncogenic signatures, are available from http://www.broadinstitute.org/gsea/index.jsp (19). GSE7882 (20), which contains 54 nonpaired samples [7 normal, 4 ADH, and 43 ductal carcinoma in situ (DCIS)], was used to validate GSEA analysis of GSE 16873. A common set of 28 genes was picked to display repression of AKT target genes in both datasets.

MCF10A cell line was obtained from ATCC and cultured in MCF10A medium, and the modified variants were obtained, generated, and cultured as described previously (21).

Three dimensional (3D) culture assay was performed following the protocol as described previously (22). Assay medium (DMEM/F12 supplemented with 2% donor horse serum, 10 μg/mL insulin, 1 ng/mL cholera toxin, 100 μg/mL hydrocortisone, 50 U/mL penicillin, and 50 μg/mL streptomycin) containing 5 ng/mL EGF and 2% growth factor–reduced Matrigel (BD Biosciences) was replaced every 3 days. Acini size was quantitated using ImageJ. At the end of 3D culture, cells were collected using Cell Recovery Solution (BD Biosciences) to remove Matrigel.

Cultures were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, washed 3 times with PBS/glycine buffer (130 mmol/L, 7 mmol/L, Na₂HPO₄, 3.5 mmol/L NaH₂PO₄, and 100 mmol/L glycine), and incubated with IF buffer (130 mmol/L, 7 mmol/L Na₂HPO₄, 3.5 mmol/L NaH₂PO₄, 7.7 mmol/L NaN₃, 0.1% BSA, 0.2% Triton X-100, and 0.05% Tween 20) for 1 hour. Antibodies were diluted in IF buffer and incubated overnight. The cultures were incubated with Alexa fluorophore–conjugated rabbit or mouse secondary antibodies (Life Technologies), counterstained with 4′, 6-diamidino-2-phenylindole (DAPI) and mounted with anti-fade solution (Life Technologies). Confocal microscopy was done using Zeiss LSM 710 laser scanning confocal microscope (Carl Zeiss AG).

All Western blotting were performed at the same time in both 10A.Vec and 10A.B2 cell lines and under all drug concentrations, for 2D and 3D, respectively. Total cell lysate was collected with RIPA buffer [50 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1% NP-40, 0.25% sodium deoxycholate, and protease inhibitor cocktail and phosphatase inhibitor cocktail]. Whole-cell lysate was obtained by sonication followed by centrifugation. Protein concentration was measured by BCA Protein Assay Kit (Pierce). Equal amounts of cell lysates were subjected to electrophoresis using SDS-PAGE and transferred to polyvinyldifluoride membrane (Bio-Rad). Membranes were blocked with 5% milk (in 1× PBST) for 30 minutes, followed by primary antibody incubation overnight at 4°C. After three washes with PBST (5 minutes each), membranes were incubated with secondary antibody (5% milk in PBST) for 60 minutes, and signal was detected by ECL (Amersham) following the manufacturer's instructions.

Cells were plated at 1,000 cells per well in triplicate in 96-well cell culture plates. At each time point, 20 μL of 5 mg/mL MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) in PBS (pH 7.5) was added into 100 μL of culture media and incubated for 1 hour at 37°C in the dark. After 1 hour of incubation, medium with MTT was removed and 100 μL of DMSO was added into each well. The absorbance was determined at 570 nm and 620 nm with a microtiter plate reader (BioTek).

Cell culture was labeled with bromodeoxyuridine (BrdU) at 25 μg/mL for 1 hour, then collected and fixed with 4 mL 70% ice-cold ethanol on ice for 20 minutes. Cell pellet was resuspended in 2 mL 4 mol/L HCl and incubated at room temperature for 1 hour. Cell pellet was collected and first washed with 0.1 mol/L Borax, then with 0.1% Brij-35/PBS. Cells were stained in 100 μL 0.1% Brij-35/PBS with anti–BrdU-APC–conjugated antibody (BioLegend #339807, 1:100) for 30 minutes at room temperature, washed with 0.1% Brij-35/PBS, and then stained with propidium iodide (PI) staining buffer (1 mg/mL BSA, 10 μg/mL RNase A, 10 μg/mL propidium iodide) for 30 minutes at room temperature. Cell pellet was washed again with 0.1% Brij-35/PBS and resuspended in 0.5 mL wash buffer. Analyze immediately with flow cytometry (Becton Dickinson).

Human samples for RPPA analysis were collected at Duke University School of Medicine (Durham, NC). A total of 22 random periareolar fine-needle aspiration (RPFNA) samples were obtained from 18 women (4 bilaterally, 14 unilaterally) who had taken tamoxifen for 6 to 12 months. Fifteen samples were classified as Tam-S and 7 samples were classified as Tam-NR based on their Masood cytology index pre- and post- tamoxifen. All studies were conducted with approval from the Institutional Review Board for Clinical Investigations at Duke University (protocol number Pro00011258, principal investigator, Victoria Seewaldt, MD).

RPPA for cell lysates was performed in MDACC Functional Proteomics core facility as described previously (23). Briefly, cellular proteins were denatured by 1% SDS, serially diluted, and spotted on nitrocellulose-coated slides. Each slide was probed with a validated primary antibody plus a biotin-conjugated secondary antibody. The signal obtained was amplified using a Dako Cytomation–catalyzed system and visualized by DAB colorimetric reaction. The slides were analyzed using customized Microvigene software (VigeneTech Inc.). Each dilution curve was fitted with a logistic model (“SuperCurve Fitting” developed at MDACC) and normalized by median polish. The data clustering was performed using Cluster 3.0 (centered by gene; followed by hierarchically clustering by gene and array using complete linkage) and TreeView.

Animal experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee of the UT MD Anderson Cancer Center (Houston, TX). Female MMTV-neu mice at 28 weeks of age were treated with either vehicle (n = 3, 0.5% hydroxypropyl methylcellulose with tween-80) or LY2780301 at 40 mg/kg (n = 5) by oral gavage once daily for 2 weeks. At the end of the treatment, mice were euthanized and the fourth pair of normal looking mammary fat pads (MFP) was isolated. For histologic analyses, nonserial sections throughout the MFPs were analyzed by a pathologist, and the scores were compiled and analyzed by another investigator.

IHC staining was performed as described previously (Lu and colleagues, 2009). Negative control slides without primary antibodies and positive control slides of tissues were included in each staining. IHC staining and statistics were performed in a blind manner. The pathologist who performed IHC staining and scoring was blinded to the hypothesis to be tested.

Statistics were performed using log-rank test, χ² test, or Student t test where applicable. Statistical analysis was performed using SPSS for Windows (16.0; SPSS, Inc.) and GraphPad Prism (Prism 5.0; GraphPad Software Inc.) packages. A P value <0.05 was considered significant. Error bars indicate the SD in all the figures (*, P < 0.05; **, P < 0.01; ***, P < 0.001 by two-tailed t test).

Aiming at developing prevention strategies for tamoxifen-resistant and ER⁻ breast cancer, we decide to identify druggable targets in tamoxifen-resistant early breast lesions. We have collected atypia needle biopsies from tamoxifen-sensitive (Tam-S) patients versus Tam-NR patients in a previously described cohort of women who are at high risk for breast cancer and have underwent Tam prevention treatment (13). Specifically, 18 high-risk women with greater than 20% lifetime risk of developing breast cancer were given tamoxifen (20 mg, orally once daily) for cancer prevention. After 6 to 12 months of treatment, samples from women who experienced a disappearance of atypia or did not progress to develop atypical lesions were classified as Tam-S and those who had persistent atypical lesions or developed atypical lesions were classified as Tam-NR. A total of 22 RPFNA samples were obtained from 18 women (4 bilaterally, 14 unilaterally) given tamoxifen. Fifteen samples were classified as Tam-S and 7 samples were classified as Tam-NR based on their Masood cytology index pre- and post-tamoxifen. To identify dysregulated signaling events as potential therapeutic targets in Tam-NR early lesions, we performed RPPA to compare expression levels of various proteins in Tam-S versus Tam-NR lesions. RPPA was performed in duplicate on Tam-S and Tam-NR RPFNA samples. Among the significantly altered signaling events, hyperphosphorylation of the glycogen synthase kinase 3 (GSK-3) and p70S6K, downstream of Akt and mTOR, respectively, suggested that Akt/p70S6K pathway is a major dysregulated signaling event in Tam-NR atypia compared with Tam-S atypia (Fig. 1A, highlighted in red).

Next, we performed GSEA of microarray data of paired normal tissues and atypia lesions from 12 patients (GSE16873) (18). We found that genes repressed by Akt are significantly (P < 0.001) decreased in ADH (top) and DCIS (bottom) compared with normal samples, indicating that Akt activity was increased as early as in ADH stage compared with normal tissue (Fig. 1B). Our further analysis of another dataset (GSE7882), which also had normal, ADH, and DCIS samples, validated the findings from GSE16873 (Fig. 1C; ref. 20). In both datasets, Akt-suppressed genes were significantly decreased in DCIS samples, while the gene expression profile in ADH samples appeared to be in between normal and DCIS samples (Fig. 1C).

To examine the function of Akt/p70S6K pathway in early transformation of ER⁻ hMECs, we stably overexpressed p70S6K in the nontumorigenic ER⁻ hMEC line MCF-10A (10A.S6K) along with vector control cells (10A.Vec; Fig. 1D). When cultured on reconstituted basement membrane (e.g., Matrigel), nontransformed hMECs (such as 10A.Vec cells) form 3D, polarized, and growth-arrested acini-like structure that recapitulate certain features of the mammary gland architecture in vivo (24). hMECs in 3D culture provide structurally and physiologically relevant interactions important for studying the morphogenesis of glandular epithelium and for modeling the biological activities of cancer genes in MECs (25) and are amenable for genetic manipulation and biochemical analysis (26). We grew 10A.S6K cells and 10A.Vec in 3D culture and examined their acini structures. Clearly, in 3D culture, 10A.Vec MECs form rounded acini with hollow lumen, whereas 10A.S6K cells formed noninvasive, disorganized, large atypical acini that are distinctively different from the round-shaped normal acini of the 10A.Vec cells (Fig. 1E, left). Furthermore, we treated the 10A.S6K cells and the 10A.Vec cells under 3D culture with LY2780301, a small-molecule dual inhibitor that acts as a selective and reversible ATP competitor for p70S6 kinase and Akt (Eli Lilly and Company). Remarkably, the abnormal acinar growth of 10A.S6K cells was effectively blocked by the inhibitor LY2780301 in 3D culture (Fig. 1E, right). These findings suggested that activation of p70S6K in early-stage breast disease contributes to the initiation and progression of breast cancer.

ErbB2 overexpression was found in >60% early-stage breast cancer in patients and could lead to p70S6K activation via PI3K/Akt pathway (21, 27). To test the general effect of targeting aberrant p70S6K activation in p70S6K-activated mammary epithelial cells of early transformation frequently seen in patients, we employed ErbB2-expressing vector transfected MCF-10A hMECs (10A.B2) that have constitutive p70S6K activation due to the overexpression of ErbB2 (21). We treated the 10A.B2 cells and the corresponding 10A.Vec control cells under 3D culture conditions with LY2780301. LY2780301 treatment efficiently reduced the levels of p-S6 in 10A.B2 cells in a concentration-dependent manner under 3D culture; however, it only mildly inhibited p-PRAS40 (downstream target of Akt) at higher concentrations (1–2 μmol/L; Fig. 2A). In 3D culture, 10A.B2 cells form noninvasive disorganized “grape-like” acinar structures with filled lumen due to increased proliferation and reduced apoptosis, which are distinctively different from the round-shaped acini of 10A.Vec cells (Fig. 2B, left). The 10A.B2 acinar structures mimic DCIS in patients and can be used for testing therapeutics (21, 22). LY2780301 (1 μmol/L) treatment significantly inhibited the growth of 10A.B2 disorganized acini and 10A.Vec acini, although the “DCIS-like” morphology in treated 10A.B2 cells was not completely reversed (Fig. 2B). Consistently, a proliferation marker, PCNA level was more dramatically reduced in 10A.B2 cells with LY2780301 treatment compared with 10A.Vec cells under the same conditions (Fig. 2C). Furthermore, 10A.B2 cells undergoing LY2780301 treatment had reduced Ki-67 staining, indicating a consistent decrease in cell proliferation, whereas they showed no significant change in cleaved caspase-3 staining, indicating no induction of apoptosis (Fig. 2D and E).

To further determine how LY2780301 takes effect on cell proliferation, we also assessed cell growth in 2D cell culture of 10A.Vec and 10A.B2 cells treated with LY2780301 for up to 6 days. Again, LY2780301 significantly inhibited proliferation of the 10A.B2 cells (P < 0.001) but had a less degree of inhibition on 10A.Vec cells when compared with vehicle-treated cells (Fig. 3A). We also detected the level of PCNA in these cells treated with LY2780301. Indeed, we found a dramatic decrease in PCNA levels in treated 10A.B2 cells in 2D culture conditions as well, while the effect was much milder in 10A.Vec cells under the same treatment (Fig. 3B). Moreover, flow cytometry detection of BrdU incorporation (DNA synthesis marker) and PI (marks DNA content) staining revealed a major dose-dependent cell-cycle stall in G₀–G₁ phase in 10A.B2 cells treated with LY2780301, resulting in dramatic decreases in actively proliferating cells (S-phase and G₂–M phase; Fig. 3C), whereas it showed much milder changes in cell-cycle distribution in 10A.Vec cells under the same treatment (Fig. 3C).

To gain systemic insights into the mechanisms of LY2780301 treatment-induced functional changes, especially, cell-cycle blockade, we performed RPPA analysis to detect global protein level changes in 10A.Vec and 10A.B2 cells growing in 2D and 3D cultures without or with LY2780301 treatment. RPPA data showed that LY2780301 indeed inhibited p70S6K effectively, leading to reduced phosphorylation on S6 at multiple sites (Fig. 4A). Notably, the level of Cyclin B1 was significantly decreased in 10A.B2 cells treated with LY2780301 under both 2D and 3D conditions, while Cyclin B1 level was not significantly reduced in 10A.Vec cells treated by LY2780301 (Fig. 4A). The decrease in Cyclin B1 is likely resulted from lowered protein synthesis due to p70S6K inhibition (28). The reduced cyclin B1 from RPPA result was also validated by Western blotting (Fig. 4B), and this reduction could led to the dramatic decrease in proliferating cells in G₂–M phase (Fig. 3C).

RPPA data also showed that Akt phosphorylation at both Ser473 and Thr308 sites was dramatically increased by LY2780301 treatment compared with vehicle control, in both culture conditions (Fig. 4A). This increase was confirmed by Western blotting, where a visible increase in p-Akt T308 and S473 was observed in 10A.B2 cells treated with as low as 0.1 to 0.2 μmol/L LY2780301 (Fig. 4C). The increased Akt phosphorylation could account for no induction of apoptosis by LY2780301 treatment. In addition, phospho-p44/42 MAPK (Erk1/2) (Thr202/Tyr204) was another consistently elevated phospho-protein by LY2780301 treatment (Fig. 4A), which was also validated by Western blotting when 10A.Vec or 10A.B2 cells were treated by >1 μmol/L of the inhibitor (Fig. 4D).

To suppress LY2780301-induced Akt phosphorylation, we used the PI3Kα/δ inhibitor GDC-0941 (29). Combination of GDC-0941 with LY2780301 effectively blocked S6 phosphorylation and hampered Akt activation (Fig. 4E). Furthermore, in 3D culture, GDC-0941 treatment reduced Ki-67⁺ proliferating cells and reversed the “grape-like” phenotype of 10A.B2 cells back to round acini similar to that from the 10A.Vec cells (Fig. 4F). Notably, combination treatment further reduced the sizes of the acini compared with those under single treatment (Fig. 4F). These data indicate that LY2780301-induced Akt phosphorylation is PI3K dependent.

As LY2780301 effectively inhibited the abnormal proliferation of the partially transformed 10A.B2 MECs in both 2D and 3D culture conditions, we reasoned that LY2780301 may also reduce the proliferation of hyperplasia mammary lesions in vivo. The MMTV-neu mouse is a neu (the rat homolog of the human ErbB2/HER2 gene)-driven ER⁻ mammary tumor mouse model (30), which develops hyperplastic lesions at about 10 to 18 weeks and palpable mammary tumors at approximately 32 weeks (13). Therefore, we treated 28-week-old mice (before development of palpable mammary tumors) with LY2780301 briefly for 2 weeks and examined whether LY2780301 treatment could inhibit S6K activity in the mammary gland, and whether the brief treatment could impact on mammary cell proliferation, atypia/DCIS progression, and tissue microenvironment. Specifically, 5 MMTV-neu mice at 28 weeks of age were treated with 40 mg/kg LY2780301 per day [comparable with 200 mg daily dose in human (31, 32), and it is ∼40% of recommended dose for cancer treatment clinical trials (33)] for 2 weeks, and control age-matched MMTV-neu mice were treated with vehicle only. No obvious drug-related toxicity (including, but not limited to, weight loss, rough coat, hypoactivity) was observed during the course of treatment (data not shown). Two MFPs from each mouse were collected at the end of the 2-week treatment for various analyses. Western blotting showed that LY2780301 decreased the p185 neu protein level in the treated MFPs, similar to what we observed in 3D 10A.B2 cell culture (Fig. 5A). Importantly, LY2780301 effectively inhibited the phosphorylation of ribosomal protein S6. We observed a clearly increased phosphorylation of Akt and Erk1/2 (Fig. 5A). Consistent with our findings in 2D and 3D culture (Figs. 2C and 3B), we detected a decrease in PCNA, which suggested that LY2780301 could suppress cell proliferation in the MFP of MMTV-neu mice in vivo (Fig. 5A). In addition, we also performed H&E and IHC staining in the MFP tissue samples. There was no apparent difference in the morphology of MFP, probably due to the short treatment time. The level of S6 phosphorylation in the treated mice was dramatically decreased (Fig. 5B, top), suggesting that LY2780301 could effectively inhibit its target in vivo, even at relatively low dose with short time treatment. Moreover, staining of the proliferation marker Ki-67 in the vehicle-treated group showed an average approximately 40% Ki-67⁺ cells (index; ref. 13), while the LY2780301-treated mice all had Ki-67 indices below 5% (Fig. 5B). These results are consistent with reduced PCNA (Fig. 5A) and confirmed that a brief LY2780301 treatment could inhibit abnormal cell proliferation of MMTV-neu mice in vivo.

Multiple clinical trials have suggested that the benefit for chemoprevention of breast cancer is limited to ER⁺ subtype, and there is an urgent need to develop effective strategies for the prevention of ER⁻ breast cancer. Although a few small-molecule kinase inhibitors, such as lapatinib and saracatinib, have been shown to have preventive effect on the development of ErbB2⁺/ER⁻ breast cancer in animal models, it remains imperative to identify the common molecular alterations that drive ER⁻ early lesions and target them for prevention.

Kinases can activate a wide spectrum of downstream effectors through phosphorylation of their substrates. Dysregulation of kinases in many cancer types has placed them at the central nodes of cancer cell signaling networks, especially in the early stages of malignant transformation (34, 35). As kinases are readily “druggable,” numerous efforts have led to successful development of inhibitors to various cancer-promoting kinases; some of these inhibitors have shown remarkable clinical efficacy in cancer treatment and have been approved by the FDA (10). However, these kinase inhibitors have not been keenly considered for cancer prevention even in high-risk population. One possible limitation is the toxicity due to high doses of kinase inhibitors used for cancer treatment, which could result in certain side effects, including anemia, folliculitis, and endocrine-related abnormalities, etc (36, 37). Here, we employed a unique approach to use low-dose, highly specific, and low-toxicity kinase inhibitors to target signaling pathways that play critical roles in promoting early breast lesion progression to ER⁻ breast cancer, as a novel and readily applicable strategy for prevention of ER⁻ breast cancer. We reason that no/low toxicity associated with these highly specific inhibitors could encourage the adoption of this prevention strategy.

Dysregulation of the PI3K/Akt/mTOR/p70S6K pathway is one of the most common signaling events found in multiple types of cancer, including breast cancer (38). Activation of this pathway regulates downstream protein synthesis and cell cycle and confers proliferation advantages that promote malignant transformation and tumorigenesis (39, 40). ErbB2 overexpression, which was reported in over 60% cases of early-stage breast lesions, could lead to constitutive activation of the PI3K/Akt/mTOR/p70S6K pathway (27, 41). Neu-driven PI3K/Akt pathway activation was also reported in other ER⁻ breast cancer mouse models (42). Our bioinformatics analysis in precancerous patient samples indicated that activation of this pathway occurred at very early stages of ER⁻ breast cancer development (Fig. 1), which may play a critical role in promoting the progression to advanced stages of ER⁻ breast cancer. Multiple small-molecule kinase inhibitors are being developed and tested for their efficacy on inhibiting excessive PI3K/Akt/mTOR/p70S6K signaling. In this study, we focused on Akt/p70S6K inhibition because of their central role in regulating protein translation and cell proliferation. LY2780301 is a small-molecule dual inhibitor that conveniently targets both p70S6 kinase and Akt activation. This agent has been already tested in phase I clinical trial, which showed favorable profile of tolerance (33), and phase Ib/II trial is ongoing to evaluate the tolerance and efficacy of LY2780301 in combination with paclitaxel in locally advanced or metastatic breast cancer (ClinicalTrials.gov NCT01980277). Using the clinically applicable dual kinase inhibitor LY2780301, we were able to inhibit p70S6K activity as indicated by the significantly reduced phosphorylation of downstream ribosomal protein S6 in ER⁻ MECs, which resulted in decreased level of Cyclin B1 (Fig. 4A and B), inhibited proliferating cells in G₂–M phase (Fig. 3C), and a drastic decrease in cell proliferation (Figs. 2–4). LY2780301 did not have a significant effect on reversing disorganized acinar growth of p70S6K-activated hMECs in 3D culture, which may be due to activation of Akt and Erk1/2 signaling in treated cells (Fig. 4). Nevertheless, a brief treatment of LY2780301 in the ER⁻ mammary tumor model of MMTV-neu mice before development of palpable tumors was sufficient to suppress S6 phosphorylation and to decrease cell proliferation in the MFP in vivo. Normal Akt/p70S6K activity is essential for normal cellular function as well. Although this Akt/p70S6K dual inhibitor seems to be well tolerated in mouse models, frequent adverse events were reported in the first-in-human phase I trial of LY2780301 given to patients at much higher doses (33); the long-term safety of this inhibitor at lower doses and other drugs with similar mechanism of action should be further tested in the future.

In addition, it is notable that ErbB2 activates multiple downstream signaling events, including the p70S6K pathway, therefore targeting more upstream signal, for example, PI3K may be more effective in the prevention of ER⁻ breast cancer. Indeed, the PI3Kα/δ inhibitor GDC-0941 effectively inhibited abnormal acini growth and reversed the disorganized acini structure by reducing cell proliferation and inducing apoptosis (Fig. 4E and F). However, acquired resistance such as genetic modifications and signaling bypass may emerge over time; therefore, long-term follow-up on the effect of the LY2780301 and PI3K inhibitor combination treatment in vivo would provide crucial preclinical indication of potential resistance and its underlying mechanisms. Our data warrant further testing of the efficacy of targeting p70S6K upstream signal, for example, PI3K and/or EGFR/HER2, on breast cancer prevention. These efforts, along with other innovative strategies (12–14, 43), will empower the development of effective agents for the general prevention of ER⁻ breast cancer.

No potential conflicts of interest were disclosed.

Conception and design: X. Wang, V.L. Seewaldt, D. Yu

Development of methodology: X. Wang, D. Yu

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): X. Wang, J. Wang, Q. Zhang, V.L. Seewaldt

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): X. Wang, J. Yao, J. Wang, S.W. Brady, B. Arun

Writing, review, and/or revision of the manuscript: X. Wang, J. Yao, S.W. Brady, B. Arun, D. Yu

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): X. Wang, D. Yu

Study supervision: D. Yu

The authors thank Dr. Yi Xiao for insightful comments on the manuscript and Eli Lilly & Company for providing LY2780301.

This work was supported by the Susan G. Komen Breast Cancer Foundation promise grant KG091020 (to D. Yu), RO1-CA184836 (to D. Yu), RO1-CA112567-06 (to D. Yu), R01CA208213 (to D. Yu), China Medical University Research Fund (to D. Yu), the Duncan Family Fund (to D. Yu), the MD Anderson Cancer Center Support GrantCA016672, and the Chinese Government Scholarship (to J. Wang).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.