Active Kinase Profiling, Genetic and Pharmacological Data Define Mtor

Home , Lapatinib, ROR1

ORIGINAL ARTICLE Active kinase proﬁling, genetic and pharmacological data deﬁne mTOR as an important common target in triple-negative breast cancer

JC Montero1, A Esparı´s-Ogando1, MF Re-Louhau1, S Seoane1, M Abad2, R Calero3, A Ocan˜a3,4 and A Pandiella1

Triple-negative breast cancer (TNBC) is an aggressive form of breast cancer. Despite response to chemotherapy, relapses are frequent and resistance to available treatments is often seen in the metastatic setting. Therefore, identification of new therapeutic targets is required. With this aim, we have profiled the activation status of 44 receptor tyrosine kinases (RTKs) and their major signaling pathways in patient-derived TNBC tumors. Frequent co-activation of several RTKs as well as the extracellular signal- regulated protein kinases 1 and 2 (Erk1/2) and mammalian target of rapamycin (mTOR) routes was found. Pharmacologic targeting of the activated kinases indicated that agents that attack the mTOR route are more potent and efficient antitumoral treatments than agents targeting RTKs. mTOR signals through two multiprotein complexes, mTORC1 and mTORC2. We used a genetic approach to explore the contribution of each of the two mTOR branches to the regulation of cell number of TNBC cells. RNA interference experiments indicated that mTORC1 predominated over mTORC2 in the control of TNBC cell proliferation. Moreover, RNA interference of mTOR had a superior antiproliferative action than separately acting on mTORC1 or mTORC2. To analyze the relevance of mTOR targeting in vivo, we used mice with TNBC. Treatment of these mice with BEZ235, a drug that targets mTOR, slowed tumor growth. Mechanistically, BEZ235 delayed cell cycle progression without affecting cell viability. Our results show that TNBCs are particularly sensitive to inhibition of the mTOR pathway, and indicate that mTOR targeting may be a more efficient anti-TNBC therapy than exclusively acting on the mTORC1 branch of the pathway. This is relevant as most mTOR inhibitors used in the clinic act on mTORC1. Collectively with the fact that BEZ235 synergized with drugs commonly used in the treatment of TNBC, our data support the clinical development of agents that act on mTOR as a therapy for this disease.

Oncogene (2014) 33, 148–156; doi:10.1038/onc.2012.572; published online 17 December 2012 Keywords: mTOR; PI3K; triple-negative breast cancer

INTRODUCTION bases to test drugs targeting some of these receptors in TNBC. Triple-negative breast cancer (TNBC) accounts for 15% of all breast Unfortunately, these clinical studies targeting individual RTKs have cancers. It is defined at an immunohistochemical level as the lack shown limited efficacy.13,14 A circumstance that may explain this of detectable expression of hormone receptors, and no human limited effect is concomitant expression of activated forms of epidermal growth factor receptor 2 (HER2) gene amplification.1–3 various RTKs.11 Therefore, the targeting of an individual RTK may Therapy of TNBC is based on chemotherapy with taxanes, be ineffective if other proto-oncogenic RTKs are functional.11 To vinorelbine or platinum compounds, which are likely to be overcome this situation, it would be desirable to identify the active effective in this type of tumors because of their rapid proliferation RTKs in a particular tumor to design drug combinations that target rates and frequent derangements in DNA repair mechanisms.4 the different active RTKs. Another strategy could be the use of Unfortunately, relapses are frequent in the early stage, and drugs that attack signaling pathways shared by different RTKs. The resistance to the chemotherapeutic agents is often seen in the pharmacological intervention on these convergent signaling metastatic setting.4–6 These facts, together with the relatively poor nodes could be effective and likely less toxic than using a knowledge of the molecular alterations present in TNBC, have multikinase RTK inhibitory approach.15 Importantly, recent deep- stimulated much research into this field to identify aberrant sequencing studies in TNBC have identified frequent clonal signaling networks that may be pharmacologically attacked in this mutations in PI3K and PTEN, two genes that participate in aggressive type of breast cancer.4 signaling downstream of RTKs.16 Receptor tyrosine kinases (RTKs) play a key role in cell With the aim of developing novel therapies in TNBC, we have proliferation, and their deregulation has been linked to the profiled the activation status of 44 different RTKs and their major pathogenesis of various tumors,7,8 including TNBC.3,9–11 These and signaling pathways in a set of tumors from patients with TNBC. other expression studies, together with the clinical success of We observed co-activation of various RTKs in single tumors, targeting RTKs in other breast tumors,12 have established the together with frequent activation of the PI3K/Akt/mTOR and the

1Instituto de Biologıá Molecular y Celular del Cańcer, CSIC-Universidad de Salamanca, Salamanca, Spain; 2Pathology Department, Hospital Clıńico Universitario de Salamanca, Salamanca, Spain; 3Servicio de OncologıáMe´dica, Complejo Hospitalario Universitario de Albacete, and AECC Unit, Albacete, Spain and 4Division of Medical Oncology and Hematology, Princess Margaret Hospital, University of Toronto, Toronto, Ontario, Canada. Correspondence: Dr A Pandiella, Instituto de Biologıá Molecular y Celular del Cańcer-CIC, Campus Miguel de Unamuno, Avenida Univ. Coimbra, 37007, Salamanca, Spain. E-mail: [email protected] Received 7 May 2012; revised 25 October 2012; accepted 26 October 2012; published online 17 December 2012 mTOR targeting in breast cancer JC Montero et al 149 Ras/Raf/Erk1/2 signaling routes. Genetic as well as pharmacologi- To explore the relevance of these frequently activated RTKs and cal studies showed that TNBC cells were particularly sensitive to signaling pathways in TNBC, their level of activation was explored inhibition of the mammalian target of rapamycin (mTOR) pathway, in seven TNBC cell lines. EGFR was expressed in all cell lines, and opening the possibility for the incorporation of agents that target was tyrosine phosphorylated in five of them (Figure 1b). HER2 was this route for the treatment of TNBC. detectable in five cell lines, and in those with higher levels (HCC1187 and HCC3153) resting tyrosine phosphorylation was evident. PDGFRb was expressed in HS578T and HBL100, and was tyrosine phosphorylated. HER4 was undetectable in the cell lines RESULTS analyzed. Akt was phosphorylated at serine 473 in all the cell lines, Profiling of active kinases defines mTOR as a relevant target in except in MDA-MB231. Detectable levels of dually phosphorylated TNBC Erk1/2 were observed in HS578T, HBL100 and BT549. MDA-MB231 Although expression analyses are useful in describing the and HCC70 cell lines showed higher phosphorylated Erk1/2 levels. presence of kinases or their signaling intermediates, a more To define the importance of the activated RTKs or their signal accurate assessment of their function is offered by evaluation of transduction routes in TNBC, we explored the action of several their activation status.7 The latter can be explored by analyzing inhibitory compounds, selected on the basis of being already tyrosine phosphorylation of RTKs, or phosphorylation of certain approved or under clinical development in other tumor types. activating residues in signaling proteins.7 The activation status of Lapatinib was used as an EGFR/HER2 dual inhibitor;17 dasatinib as 44 RTKs as well as of major RTK signaling pathways were analyzed a PDGFRb inhibitor;18 sorafenib19 and selumetinib20 as Raf and in 26 TNBC tumors using antibody arrays. The clinical MEK1 inhibitors, respectively; and BEZ235 as a PI3K/mTOR dual characteristics of this patient cohort are summarized in inhibitor.21 Dose-response experiments were performed in four Supplementary Table 1. The most frequently activated kinases TNBC cell lines, and estimation of the amount of cells analyzed were the epidermal growth factor receptor (EGFR; 75% of by an MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium patients), HER2 (33% of patients), HER4 (25% of patients), bromide)-based metabolization assay. As shown in Figure 1c, platelet-derived growth factor receptor-b (PDGFRb; 71% of lapatinib, sorafenib and selumetinib had a rather modest effect on patients), Akt (88% of patients) and extracellular signal-regulated the MTT metabolization of the distinct TNBC cell lines tested. protein kinases 1 and 2 (Erk1/2; 53% of patients; Figure 1a). Dasatinib presented a mild inhibitory effect on these cell lines. The Concomitant activation of several RTKs in a single tumor was most effective and potent agent of those tested was BEZ235, detected (see, for example, inset in Figure 1a, see also which reached half-maximal inhibitory concentration values in the Supplementary Table 2). nanomolar range. Western blotting demonstrated the efficiency of

100 100 RTK array

80 12 80 3

60 1. pEGFR 2. pHER2 3. pPDGFR 60

40 40

20 20 % of positive tumors

0 0 S6 Axl Src Lck Mer ALK Flt-3 TrkC TrkB TrkA Tie-1 Tie-2 Stat1 Stat3 c-KIT IRS-1 C-Ret A-Abl HER4 HER3 HER2 ROR2 ROR1 EGFR MuSK IGF-IR c-MET Erk1/2 EphA6 MSP R EphB6 EphB4 EphB3 EphB2 EphB1 EphA7 EphA4 EphA3 EphA2 EphA1 Zap-70 FGF R4 FGF R3 FGF R2 FGF R1 M-CSF R Akt T308 Insulin R Akt S473 PDGF R VEGF R2 VEGF R1 VEGF R3 PDGF R TYRO3/Dtk

Sorafenib 120 MDA-MB231 Dasatinib 120 HBL100 100 Lapatinib 100 BEZ235 80 80 pEGFR Selumetinib 60 60 EGFR 40 40

pHER2 (% control) 20 (% control) 20 HER2 0 0 MTT Metabolization 10 100 1000 10000 MTT Metabolization 10 100 1000 10000 pHER4 Concentration, nM Concentration, nM HER4 120 HS578T BT549 pPDGFR 120 100 100 PDGFR 80 80 pAkt 60 60 Akt 40 40 (% control) 20 (% control) 20 pErk1/2 0 0 Erk1/2 MTT Metabolization 10 100 1000 10000 MTT Metabolization 10 100 1000 10000 Concentration, nM Concentration, nM Figure 1. (a) Expression of activated forms of RTKs and signaling proteins in tumor samples. The inset shows an array of active RTKs from one TNBC tumor. (b) Expression and activation of EGFR, HER2, HER4, PDGFRb, AKT and Erk1/2 in seven TNBC cell lines. Activated forms of the RTKs were analyzed by immunoprecipitation with anti-receptor antibodies, followed by anti-PY blot. Detection of activated Akt and Erk1/2 was performed by western using antibodies that recognize residues phosphorylated in these proteins, which indicate active forms. (c) Effect of different kinase inhibitors on the MTT metabolization of four different TNBC cell lines. Cells were plated in 24-well dishes, and treated with the indicated concentrations of the drugs for 48 h. MTT uptake and metabolization was measured as described in the ‘Materials and methods’ section. Results are plotted as the mean±s.d.

& 2014 Macmillan Publishers Limited Oncogene (2014) 148 – 156 mTOR targeting in breast cancer JC Montero et al 150 these compounds in inhibiting the phosphorylation status of their a predominant role of mTORC1 over mTORC2 on this biological targets or downstream signaling intermediates (Supplementary property. Figure 1). mTOR inhibition provokes cell cycle arrest of TNBC cells To conﬁrm that BEZ235 acted on the mTOR route and to establish mTORC1 and mTORC2 control the proliferation of TNBC cells a dose that would efﬁciently block mTOR, we analyzed the Given the substantial inhibitory action of BEZ235 on the phosphorylation status of pS6 and p4E-BP1 as readouts of proliferation of TNBC cells, and as this inhibitor targets mTOR, mTORC1 activity, as well as pAkt as readout of mTORC2. BEZ235 we decided to explore the importance of this pathway in TNBC. decreased the resting phosphorylation status of pS6 and p4E-BP1 The mTOR kinase acts in concert with other proteins that in a dose-dependent manner (Figure 3a). In MDA-MB231 cells, 22 constitute two multiprotein complexes, mTORC1 and mTORC2. which did not present resting pAkt, treatment with BEZ235 mTORC1 includes mTOR, Raptor, PRAS40, mLST8/GbL and Deptor, provoked its phosphorylation at concentrations of BEZ235 whereas mTORC2 includes mTOR, Rictor, mLST8/GbL, Sin1, Protor between 10 and 50 nM. At higher doses of BEZ235 the pAkt signal and Deptor. decreased, disappearing at 250 nM (Figure 3a). In HBL100 cells, To investigate the role of mTOR and its two branches in the doses of BEZ235 of 10 and 25 nM caused a small but reproducible proliferation of TNBC, we performed knockdown of Raptor and increase in pAkt with respect to the resting levels. At higher doses, Rictor as targets to decrease mTORC1 and mTORC2 signaling, as BEZ235 decreased pAkt levels with complete disappearance of the well as knockdown of mTOR as readout of the overall importance signal at 250 nM. As this latter concentration of BEZ235 completely of this pathway in TNBC. For these experiments we used HBL100 blocked phosphorylation of mTORC1 and mTORC2 readout and MDA-MB231 cell lines selected on the bases of their proteins, we selected that dose for the following studies on the differential pAkt levels (Figure 1b). Western blotting of cell mechanism of action of the drug. It has been previously reported extracts from MDA-MB231 and HBL100 cells demonstrated the that treatment of HER2-overexpressing breast cancer cells with knockdown of these different components of the mTOR route BEZ235 results in activation of mitogen-activated protein kinase (Figure 2a). mTOR knockdown exerted the largest inhibitory effect signaling.23 However, in three different TNBC cell lines we did not on the MTT metabolization values in both cell lines (72% inhibition observe such an effect (Supplementary Figure 3). in MDA-MB231 and 75% of inhibition in HBL100 with respect to We explored whether the inhibitory action of the compound on the proliferation values of control cells, Figure 2b). Knockdown of TNBC cells was due to inhibition of proliferation, stimulation of cell Raptor had a higher inhibitory effect than the knockdown of Rictor death or both. Treatment with BEZ235 increased the number of in both cell lines. In the MDA-MB231 cell line the effect of the cells in the G0/G1 phases of the cell cycle (increase in G0/G1 of Raptor knockdown was very similar to the effect obtained by 32% and 11.5% in MDA-MB231 and HBL100, respectively, mTOR knockdown, indicating that mTORC1 is critical in channel- Figure 3b). BEZ235 did not increase Annexin V-positive cells, ing the cell growth signals transmitted through the mTOR suggestive of failure of this drug to cause death (Figure 3c). pathway in this cell line. Similar data were generated using a distinct set of short hairpin RNAs (shRNAs; Supplementary

Figure 2). Together, these data conﬁrm the relevant role of mTOR BEZ235 BEZ235 in sustaining cell proliferation in these TNBC cell lines, and indicate 0 10 25 50 100250 (nM) 0 10 25 50 100 250 (nM) 473 pS -Akt pS473-Akt * * Akt Akt pS6 pS6 S6 S6 p4E-BP1 p4E-BP1 Raptor Raptor 4E-BP1 4E-BP1 Rictor Rictor Calnexin Calnexin mTOR mTOR MDA-MB231 HBL100 GAPDH GAPDH MDA-MB231 HBL100 MDA-MB231 HBL100 MDA-MB231 HBL100 100 100

MDA-MB231 HBL100 80 80 1.4 1.2 G0/G1 )

sh-Control ) 60 60 Annexin V + sh-Control S

570 1.2 sh-Rictor 570 sh-Rictor Annexin V - 1.0 G2/M sh-Raptor 40 40 1.0 sh-Raptor sh-mTOR 0.8 sh-mTOR Cell percentage 0.8 Cell percentage 20 20 0.6 0.6 0 0 00250 250 00250 250 0.4 0.4 [BEZ235], nM [BEZ235], nM 0.2 0.2 Figure 3. (a) Effect of BEZ235 on mTORC1 and mTORC2 readout MTT Metabolization (A 0 MTT Metabolization (A 0 proteins. Cells were treated with the indicated doses of BEZ235 for 0 1 2 3 0 1 2 3 24 h, and cell lysates analyzed by western blot using antibodies to Days Days the indicated proteins. Calnexin was used as a loading control. The Figure 2. (a) Knockdown of different mTORC components in HBL100 asterisk indicates a non-specific band. (b) Effect of BEZ235 on the cell and MDA-MB231. The amount of the different proteins was analyzed cycle of MDA-MB231 and HBL100. Cells were treated with BEZ235 for by western blot. (b) Effect of the knockdown of mTOR, Raptor and 24 h, fixed and permeabilized and stained with propidium iodide Rictor on the proliferation of HBL100 and MDA-MB231. Cells were before analysis in a cytometer. The histograms represent the infected with viruses that included shRNAs for mTOR, Raptor or percentage of cells in the different phases of the cell cycle. (c) Effect Rictor, selected with puromycin for 48 h, and MTT metabolization of BEZ235 on apoptosis. Cells were treated with BEZ235 for 48 h, and was measured at the indicated times. Results are plotted as the then stained with Annexin V/FITC and propidium iodide. Stained mean±s.d. of an experiment that was repeated twice with similar cells were analyzed by flow cytometry. The percentage of Annexin results. V-positive or Annexin V-negative cells is shown.

Oncogene (2014) 148 – 156 & 2014 Macmillan Publishers Limited mTOR targeting in breast cancer JC Montero et al 151 To define the action of BEZ235 on the cell cycle we used Antitumoral action of BEZ235 in mice with epithelially deleted treatments that block cell cycle progression, followed by release Brca1 and p53 from these blockers in the absence or presence of BEZ235. To investigate the in vivo effect of BEZ235 on TNBC tumors we Nocodazole blocked cells in the M phase of the cell cycle, as used a conditional mouse model in which deletion of Brca1 and indicated by the accumulation of cyclin B or by phosphorylation of p53 can be created by crossing to animals with K14/cre, which the mitotic checkpoint protein BUBR1 (Figure 4a). After nocoda- eliminates the Brca1 and p53 genes in epithelia, including breast zole washout, cells were released from mitosis and entered into epithelia.25 Female mice from this strain develop mammary G1. The biochemical (for example, BUBR1 dephosphorylation, tumors with characteristics of TNBC.25 We confirmed by western Figure 4a) and cytometric profile (Figure 4b) at 3 h after release blot that HER2 as well as progesterone and estrogen receptors from nocodazole indicated that treatment with BEZ235 did not were undetectable or very low in these tumors (Figure 5a). We also substantially affect exit of the TNBC cells from mitosis. In failed to detect dually phosphorylated Erk1/2. In contrast, EGFR, untreated samples, a substantial number of cells transited from PDGFRb, pS6 and pAkt were expressed in these tumors (Figures 5a G1 to S and G2/M, especially at 16 h after release (Figure 4b). and b). Western analyses of the tumors from these mice showed However, in BEZ235-treated cells, the 2N population represented a that treatment with BEZ235 inhibited pAkt and pS6 levels substantial proportion (Figure 4b), suggesting that the drug (Figure 5b). BEZ235 significantly delayed tumor growth (70% slowed movement of the cells through G1 and S. The wide G1/S inhibition of tumor growth after 21 days of treatment, P ¼ 0.022, region in the propidium iodide staining also indicated that in Figure 5c). HBL100 cells BEZ235 could not fully block G1 to S progression (Figure 4b, at T16, T18 and T21 h after release). In HBL100 cells, increases in cyclin D1, cyclin D3, cyclin A or cyclin B, which mark BEZ235 synergizes with standard of care drugs progression of cells through G1, S and G2/M, were inhibited by As most antitumoral therapies are based on drug combinations, BEZ235 (Figure 4a). Western blotting of pS6 and pAkt confirmed we explored whether combining BEZ235 with drugs used in the that the drug inhibited their phosphorylation along the time therapy of TNBC augmented their effect. To this end, three TNBC course of the experiment (Figure 4a). Analogous experiments cell lines were treated with different doses of BEZ235, taxotere and carboplatin. Double and triple combinations were performed, performed in MDA-MB231 generated data also compatible with 26 the concept that BEZ235 slowed movement of cells through the and the results were analyzed by the Chou–Talalay algorithm. cell cycle (Supplementary Figure 4). Synchronization of HBL100 at Double and triple combinations of BEZ235 with these drugs were late G1 by a double thymidine block24 confirmed that BEZ235 found to be synergistic in HBL100 and HS578T cells (Figure 6a and slowed movement of cells at the G1/S transition (Supplementary Supplementary Figure 6A). In MDA-MB231, these double and triple Figure 5). combinations were additive. As BEZ235 affected cell cycle progression, it is possible that simultaneous administration of BEZ235 with chemotherapy might alter the efficacy of the latter. To test such a possibility, experiments in which BEZ235 and chemotherapy were added simultaneously or sequentially were performed. As shown in Figure 6b, sequential addition of BEZ235 Time after release, hours and chemotherapeutic agents did not show antitumoral benefit as 0 3 6 12161821 compared with the agents added together. Analogous results HBL100 Cyclin D1 Control were also observed in the HS578T cell line (Supplementary Control BEZ235 Cyclin D1 BEZ235 Figure 6B). T=0 h The action of the double combinations of BEZ235 with Cyclin D3 Control chemotherapy on cell cycle and cell viability was explored. Cyclin D3 BEZ235 T=3 h Addition of carboplatin to MDA-MB231 and HBL100 cells caused accumulation of cells in the G2/M phases, and strongly decreased Cyclin A Control T=6 h the presence of cells in the G1 phase (Figure 6c). Concomitant Cyclin A BEZ235 addition of BEZ235 and carboplatin augmented the proportion of Cyclin B Control T=12 h cells in the S phase, in line with a potential effect of BEZ235 in slowing progression along G1 and S phases of the cell cycle. Cyclin B BEZ235 T=16 h Taxotere had a profound effect on cell cycle profiles in both MDA- Cyclin E Control MB231 and HBL100, and addition of BEZ235 did not appreciably

Cyclin E BEZ235 T=18 h alter such profile. Annexin V staining, used to assess the effect of pBUBR1 the drugs on apoptotic cell death, indicated that a combination of Control BUBR1 BEZ235 with carboplatin slightly increased the amount of Annexin T=21 h pBUBR1 BEZ235 V-positive cells in HBL100 (Figure 6d). In contrast, the level of cell BUBR1 2N 4N 2N 4N death assessed by this technique did not change when BEZ235 pAkt Control * was combined with taxotere. pAkt BEZ235 * Combinations of BEZ235 with taxotere or carboplatin were also pS6 Control analyzed in vivo. For these experiments, MDA-MB231 cells were injected in nude mice, and when tumors reached 200 mm3, mice pS6 BEZ235 were treated with the different drugs, either alone or in Calnexin Control combination. Treatment with BEZ235 delayed tumor growth, Calnexin BEZ235 consistent with the data obtained in the Brca1 and p53 mice. The HBL100 combination of BEZ235 with carboplatin or taxotere had a higher antitumoral effect than individual treatments (Figure 6e). Figure 4. Effect of BEZ235 on the cell cycle. HBL100 cells were synchronized in mitosis by treatment with nocodazole for 14 h. Nocodazole was then washed out, and cells were placed in growth DISCUSSION media in the absence or presence of BEZ235 (250 nM). Cells were harvested at the indicated times, and analyzed biochemically (a) and The dismal prognosis of TNBC, as well as the lack of new treatment cytometrically (b). The asterisks indicate nonspecific bands. options to fight the disease, has stimulated research in this

& 2014 Macmillan Publishers Limited Oncogene (2014) 148 – 156 mTOR targeting in breast cancer JC Montero et al 152

ER WB: -ER

PR WB: -PR IP and WB: pAkt WB: -pAkt EGFR -EGFR pS6 IP and WB: WB: -pS6 HER2 -HER2 IP and WB: GAPDH WB: -GAPDH PDGFR -PDGFR --- + + + BEZ235 pErk1/2 WB: pErk1/2

Erk1/2 WB: -Erk1/2

12 *p=0.022 Control 10 BEZ235

2 Fold change in tumor volume

0 0 7 14 21 Days of treatment Figure 5. (a) Expression of different proteins in tumors from two mice with epithelially deleted Brca1/p53.(b) Effect of BEZ235 on the activation status of Akt and S6. Tumors from mice treated or not with BEZ235 were excised, homogenized and cell extracts analyzed by western blot for the expression of activated forms of Akt and S6. GAPDH blots were performed as loading controls. (c) Effect of treatment with BEZ235 on tumor growth in mice with epithelially deleted Brca1/p53. Mice were randomly assigned to two groups and treated with BEZ235 following the schedule detailed under ‘Materials and methods’.

pathology with the purpose of identifying novel treatments. In line more important to sustain proliferation than those sent through with this, and given the importance of kinases in breast and other mTORC2. This conclusion raises the possibility of using drugs that tumors, we investigated the level of activation of 44 different RTKs inhibit mTORC1, such as rapamycin or analogs, for the therapy of as well as several signaling intermediates in tumors from patients TNBC. However, several circumstances should be considered. First, with TNBC. As the tumor samples and the cell lines analyzed our data indicate that attack to the mTORC1 pathway by rapalogs frequently had co-activation of various RTKs, it is not unexpected causes an increase in pAkt, indicative of upregulation of mTORC2, that blocking only one of them may have a rather marginal effect as already reported in other cellular models.30,31 As our data with on cell proliferation. These data may explain the relatively poor the knockdown experiments indicate that mTORC2 also facilitates clinical activity of agents that target these receptors,14 and open proliferation of TNBC cells, the upregulation of mTORC2 by the important question of how to design efficient therapies to rapalogs could reduce effectiveness of these compounds. This overcome escape to antitumoral therapies because of proliferation concept, already reported in other cellular models and confirmed signals originating at different RTKs. Several strategies can be by us in TNBC cells, should be taken into account by clinicians contemplated, such as the concomitant administration of drugs evaluating rapalogs in different oncologic disorders. Furthermore, that target different RTKs or the use of wide-spectrum multikinase as activation of PI3K/Akt has been linked to resistance to therapies inhibitors. Although some preclinical studies support these used in other breast cancer subtypes,32,33 upregulation of pAkt by strategies,11,27 some clinical studies with the multikinase rapalogs may contribute to selection of tumoral clones resistant to inhibitor sunitinib have shown disappointing results.28 Another standard therapies used in the treatment of TNBC. Moreover, it is strategy could be the use of agents that target common signaling important to mention that in the human samples, phosphorylation routes that funnel proliferative signals originated at distinct of S6, a surrogate marker of mTORC1 activity, was not as RTKs.9,29 In support of the latter are the results obtained with frequently detected as in cell lines. Therefore, strategies to the dual PI3K/mTOR inhibitor BEZ235. This drug was more potent target the mTOR signaling network should consider acting on and effective than agents that target active RTKs in TNBC, mTOR, rather than just on the mTORC1 branch of the suggesting that targeting PI3K/mTOR could be important in TNBC. pathway.31,34 This conclusion is highly relevant as most mTOR This idea was evaluated using a genetic approach. Knockdown of inhibitors already approved or in clinical trials target mTORC1. mTOR profoundly compromised the proliferation of the two TNBC The mechanistic studies on the antitumoral effect of BEZ235 not cell lines tested. The knockdown of Raptor had a major inhibitory only provide insights into how the drug acts, but may also have effect than Rictor knockdown on the proliferation of TNBC cells. therapeutic relevance. Of note, low doses of BEZ235 reduced pS6 Together, those results confirm the important role of the mTOR but increased pAkt. As mTORC1 may inhibit mTORC2 signaling, it kinase routes in the control of proliferation of TNBC cells, and also is therefore possible that inhibition of mTORC1 at low BEZ235 suggest that signals channeled through the mTORC1 route are concentrations may explain the increase in pAkt observed when

Oncogene (2014) 148 – 156 & 2014 Macmillan Publishers Limited mTOR targeting in breast cancer JC Montero et al 153 > 2 antagonism

MDA-MB231 1 additivity

synergy 0 > 2 antagonism

HBL100 1 additivity

synergy 0 BEZ235 (nM) 555 57.5 7.5 7.5 7.555557.5 7.5 7.5 7.5 TXT (nM)0.1 0.5 0.1 0.50.1 0.1 0.5 0.5 0.1 0.1 0.5 0.5 0.1 0.1 0.5 0.5 Carboplatin (M) 100 250 100 250 100 250 100 250 100 250 100 250 100 250 100 250

HBL100 MDA-MB231 120 HBL100 100 MDA-MB231 Control 80 BEZ235 60 TXT 40 (% Control) Carb 20 MTT Metabolization 0 BEZ235+TXT

BEZ235+Carb

2N 4N 2N 4N

HBL100 MDA-MB231 1400 Control 100 TXT 1200 Carb 80 ) 3 1000 BEZ235 60 BEZ235+TXT 800 BEZ235+Carb 40 600 Cell Percentage 20 400

0 Tumor volume (mm - - - + + +- - -+++ BEZ235 200 - + - - + --+--+- TXT --+--+ - - + - - + Carb 0 0 3 7 10 14 17 21 24 Annexin V + Annexin V - Days of treatment Figure 6. (a) Effect of the combination of BEZ235 and standard of care drugs. Cell lines were treated with the indicated concentrations of the drugs, either in double or triple combinations, and their MTT values were measured. Combination indexes for the different drug combinations were obtained using the CalcuSyn program and plotted. (b) Effect of sequential versus concomitant administration of the drug combinations on MTT metabolization rates. Cells were plated, and 24 h later treated with the individual drugs (BEZ235, 7.5 nM; carboplatin, 100 mM;or taxotere, 0.5 nM), the first drug in the case of sequential combinations, or the drug combinations (B þ C, or B þ T) in the case of concomitant administration. For the sequential treatments, the second drug was added 15 h after the first drug. MTT metabolization was analyzed in all cases at 72 h. The results represent the mean±s.d. of quadruplicates of an experiment that was repeated twice. (c, d) Action of the drug combinations on the cell cycle (c) and apoptosis (d). Drugs were added for 48 h (BEZ235, 7.5 nM; carboplatin, 250 mM; or taxotere, 0.5 nM), and analyzed as described above. (e) In vivo antitumoral action of the drug combinations. Mice were injected with MDA-MB231 cells, and when tumors reached 200 mm3 mice were treated with the different drugs either alone or in combination, with the schedules indicated under ‘Materials and methods’. Data represent the mean±s.d. Statistical analyses were performed on the measurements obtained at the end of the experiment (day 24 of treatment) using the SPSS software v19, with the following results: Control vs BEZ235: mean volume ¼ 1.12 vs 0.53 cm3, difference ¼ 0.59 cm3, 95% confidence interval ¼ 0.77–0.41, P ¼ 0.000; Control vs TXT: mean volume ¼ 1.12 vs 0.50 cm3, difference ¼ 0.62 cm3, 95% confidence interval ¼ 0.81–0.44, P ¼ 0.000; Control vs carboplatin: mean volume ¼ 1.12 vs 0.84 cm3, difference ¼ 0.28 cm3, 95% confidence interval ¼ 0.47–0.083, P ¼ 0.009; BEZ235 vs BEZ235 þ TXT: mean volume ¼ 0.53 vs 0.27 cm3, difference ¼ 0.25 cm3, 95% confidence interval ¼ 0.37–0.13, P ¼ 0.000; BEZ235 vs BEZ235 þ carboplatin: mean volume ¼ 0.53 vs 0.32 cm3, difference ¼ 0.21 cm3, 95% confidence interval ¼ 0.34–0.079, P ¼ 0.004; TXT vs BEZ235 þ TXT: mean volume ¼ 0.50 vs 0.27 cm3, difference ¼ 0.22 cm3, 95% confidence interval ¼ 0.34–0.10 P ¼ 0.001; carboplatin vs BEZ235 þ carboplatin: mean volume ¼ 0.84 vs 0.32 cm3, difference ¼ 0.52 cm3, 95% confidence interval ¼ 0.68–0.36 P ¼ 0.000. using these low BEZ235 concentrations.30 These findings also important, as reports in different cell types,36,37 including a indicate the existence of distinct thresholds for mTORC1 and heterogeneous group of breast cancer cell lines,38 showed that mTORC2 inhibition by BEZ235, as formerly reported in ovarian the drug caused activation of caspases and PARP (poly (ADP- cancer cells.35 The antitumoral effect of BEZ235 in TNBC was ribose) polymerase) cleavage, indicative of apoptotic cell death.39 mainly caused by slowing cell cycle progression, without any However, BEZ235 failed to increase the number of Annexin evident effect on apoptosis. The latter conclusion is particularly V-positive cells in TNBC cell lines. Interestingly, although BEZ235

& 2014 Macmillan Publishers Limited Oncogene (2014) 148 – 156 mTOR targeting in breast cancer JC Montero et al 154 decreased tumor growth in mice, it did not cause shrinkage of the Biosciences. The anti-ER and anti-PR were from Master Diagno´stica tumors, in line with a cytostatic rather than a cytotoxic effect of (Granada, Spain). The anti-Rictor was from Bethyl Laboratories (Montgomery, the drug also in vivo. The blockade-release synchronization TX, USA). Horseradish peroxidase conjugates of anti-rabbit and anti-mouse experiments indicated that BEZ235 slows cell cycle progression. immunoglobulin G were from Bio-Rad Laboratories (Hercules, CA, USA). The The results obtained indicated that progression of cells through anti-calnexin antibody was from Stressgen Biotechnologies Corporation (British Columbia, Canada). The anti-pAkt (Serine 473) antibody has been G1 and S phases of the cell cycle was profoundly compromised by described previously.40 BEZ235. Biochemical experiments suggested that increases in the expression of various cyclins involved in cell cycle progression was inhibited in cells treated with BEZ235. Whether this is because of Cell culture and infection with lentivirus the role of mTOR on protein synthesis will require further All cell lines were cultured at 37 1C in a humidified atmosphere in the investigation. These mechanistic studies suggest that eradication presence of 5% CO2 and 95% air. Cells were grown in Dulbecco’s modified of the tumor will require combination of the effectiveness of Eagle’s medium (DMEM) or in RPMI medium containing high glucose BEZ235 as a cytostatic agent with drugs that provoke cell death. concentration (4500 mg/l) and antibiotics (penicillin at 100 mU/ml and streptomycin at 100 mg/ml) and supplemented with 10% fetal bovine serum. Importantly, BEZ235 synergized with other drugs used in the Cell lines were provided by Drs J Losada and A Balmain (originally from Dr treatment of TNBC, and those combinations, may cause cell death. JW Gray’s Laboratory, who in turn obtained them from the ATCC or from In fact, in HBL100 cells the combination of carboplatin with collection development in the laboratories of Drs S Ethier and A Gazdar to BEZ235 augmented the apoptotic action of the former. In addition avoid errors occurring when obtained through ‘second-hand’ sources).41 to this proapoptotic potentiation, this drug combination also The lentiviral vectors containing shRNA for Raptor, Rictor and mTOR42 affected cell cycle profiles when compared with the individual were obtained from Addgene (Cambridge, MA, USA). The following treatments. The fact that the drug combination increased the sequences were used: mTOR_shRNA_1 sense: proportion of cells in the S phase indicates that passage through 0 this cell cycle phase is compromised by the drug combination, 5 -CCGGCCGCATTGTCTCTATCAAGTTCTTCCTGTCAAACTTGATAGAGACAA TGCGGTTTTTG-30. and might contribute to the increased antitumoral action of this mTOR_shRNA_1 antisense: drug combination when compared with the individual treatments. 50-AATTCAAAAACCGCATTGTCTCTATCAAGTTTGACAGGAAGAACTTGATA The mechanism of potentiation of the action of taxotere by GAGACAATGCGG-30. BEZ235 is less clear. The combination of both agents did not cause Raptor_shRNA_1 sense: substantial apoptosis. Taxotere provoked profound changes of the 50-CCGGGGCTAGTCTGTTTCGAAATTTCTTCCTGTCAAAATTTCGAAACAGA cell cycle profiles in both MDA-MB231 and HBL100 cells, and these CTAGCCTTTTTG-30. profiles were unaffected by addition of BEZ235. We suspect that Raptor_shRNA_1 antisense: 50-AATTCAAAAAGGCTAGTCTGTTTCGAAATTTTGACAGGAAGAAATTTCGA the profound derangement of the cell cycle profile caused by 0 taxotere impedes adequate interpretation of a potential additive AACAGACTAGCC-3 . Rictor_shRNA_1 sense: action of the BEZ235 þ taxotere combination on the cell cycle. 50-CGGGCAGCCTTGAACTGTTTAACTTCCTgTCATTAAACAGTTCAAGGCTG Importantly, the in vivo experiments confirmed the major CTTTTTG-30. antitumoral action of the combined treatments when compared Rictor_shRNA_1 antisense: with the individual treatments, opening the possibility of 50-AATTCAAAAAGCAGCCTTGAACTGTTTAATGACAGGAAGTTAAACAGTTC evaluating such combinations for the therapy of TNBCs. AAGGCTGC-30. In summary, we show that co-activation of several RTKs is mTOR_shRNA_2 sense: frequent in TNBC, as it is activation of downstream signaling 50-CCGGTTCAGCGTCCCTACCTTCTTCTCTCGAGAGAAGAAGGTAGGGACG 0 pathways, especially the PI3K/Akt/mTOR route. Results obtained CTGATTTTTG-3 . using in vitro and in vivo preclinical models indicate that targeting mTOR_shRNA_2 antisense: 50-AATTCAAAAATCAGCGTCCCTACCTTCTTCTCTCGAGAGAAGAAGGTAGG of mTOR is an effective antitumoral treatment in TNBC. The GACGCTGAA-30. strategy of targeting this key pathway used by several proto- Raptor_shRNA_2 sense: oncogenic RTKs may not only result in higher clinical activity, but 50-CCGGAGGGCCCTGCTACTCGCTTTTCTCGAGAAAAGCGAGTAGCAGGG could also represent a less toxic alternative to the use of a cocktail CCCTTTTTTG-30. of RTK inhibitors. Finding a common target may in addition Raptor_shRNA_2 antisense: facilitate the management of TNBC, especially considering the 50-AATTCAAAAAAGGGCCCTGCTACTCGCTTTTCTCGAGAAAAGCGAGTAGC genomic diversity of this disease. These conclusions, together with AGGGCCC-30. Rictor_shRNA_2 sense: the fact that BEZ235 synergized with other drugs used in the 0 treatment of TNBC, represent a starting point for the clinical 5 -CCGGTACTTGTGAAGAATCGTATCTTCTCGAGAAGATACGATTCTTCACA AGTTTTTTG-30. development of mTOR inhibitors for the therapy of this disease. Rictor_shRNA_2 antisense: 50-AATTCAAAAAACTTGTGAAGAATCGTATCTTCTCGAGAAGATACGATTC MATERIALS AND METHODS TTCACAAGTA-30. 30 Reagents and antibodies Preparation of lentiviral vectors was performed as described previously. Cell culture media and fetal bovine serum were purchased from Invitrogen (Gaithersburg, MD, USA). Protein A-Sepharose was from GE Healthcare Life Antibody arrays, immunoprecipitation and western blotting Sciences (Piscataway, NJ, USA). Puromycin and MTT were from Sigma We used two different commercial arrays, the human phospho-RTK array Chemical (Madrid, Spain). Annexin V/fluorescein isothiocyanate (FITC) was kit (R&D Systems, Abingdon, UK) and the PathScan RTK Signaling Antibody purchased from BD Biosciences (San Jose, CA, USA). The polyethylenimide Array Kit (Cell Signaling). Quantitation of the different RTKs in the human reagent was from Polysciences, Inc. (Warrington, PA, USA). BEZ235, phospho-RTK array kit was performed using the ImageJ 1.44 software lapatinib, sorafenib and selumetinib were from LC Laboratories (Woburn, (National Institute of Health, Bethesda, MD, USA), and quantitation of the MA, USA). Other generic chemicals were purchased from Sigma-Aldrich different cell signaling intermediates was performed using the Odyssey (St Louis, MO, USA), Roche Biochemicals (Hoffmann, Germany) or Merck V3.0 program (LI-COR, Lincoln, NE, USA). The procedure for immunopre- (Darmstadt, Germany). The anti-GAPDH, anti-cyclin E, anti-Wee1, anti-pErk1/ cipitation and western blotting has been previously detailed.43 2, anti-Erk2, anti-pTyr, anti-EGFR, anti-HER2, anti-HER4 and anti-PDGFRb antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The anti-mTOR, anti-Raptor, anti-pS6 (S240/244), anti-S6, anti-p4E-BP1 Cell proliferation, cell cycle and apoptosis assays (T37/46), anti-4E-BP1, anti-Akt and anti-pCDK1(Y15) antibodies were from Cell Cells were plated in 24-well plates at 10 000–20 000 cells/well and cultured Signaling Technologies (Beverly, MA, USA). The anti-cyclin A, anti-cyclin B, overnight in DMEM or RPMI þ 10% fetal bovine serum. The next day, anti-cyclin D1, anti-cyclin D3 and anti-BUBR1 were purchased from BD medium was replaced with DMEM or RPMI containing the different drugs.

Oncogene (2014) 148 – 156 & 2014 Macmillan Publishers Limited mTOR targeting in breast cancer JC Montero et al 155 Cell proliferation was analyzed by an MTT-based assay as described amplification of HER2 were performed on paraffin-embedded tissue previously.44 Unless otherwise indicated, the results are presented as the following international guidelines.47 Scoring criteria include the intensity mean±s.d. of quadruplicates of a representative experiment that was and pattern of membrane staining, and ranged from 0 to 3, considering repeated at least three times. To determine whether the combinations of positive a score of 3.48 Hormonal receptor status was analyzed by BEZ235 with taxotere or carboplatin were synergistic, additive or immunohistochemical staining with anti-ER and anti-PR antibodies (Master antagonistic, we used the CalcuSyn v2.0 software program (Biosoft, Diagno´stica), and using the Bond Polymer Refine Detection Kit (Bond Max Ferguson, MO, USA)45 as described previously.46 Results are plotted as the Autostainer, Leica Microsystems, Buffalo Grove, IL, USA). Staining was mean±s.d. values of quadruplicates from two experiments. scored as described previously.49 To analyze the effect of BEZ235 on the different phases of cell cycle, cells were synchronized in G1/S or in mitosis. Mitotic cells were obtained by 24 treatment with nocodazole (275 nM during 14 h). The mitotic cells were CONFLICT OF INTEREST collected by mechanical shake-off, plated and treated with BEZ235 for the The authors declare no conflict of interest. indicated times. To enrich cells in G1/S, cells were treated with thymidine 24 (2 mM during 14 h) as described previously. Then, the cells were treated with BEZ235 at different times. DNA content and cell cycle analyses were ACKNOWLEDGEMENTS performed by using a FACScalibur flow cytometer and the CellQuest software (BD Biosciences). This work was supported by grants from the Ministry of Science and Innovation of For apoptosis analyses, cells were incubated with trypsin-EDTA, washed Spain (BFU2009–07728/BMC to AP, and PS09/00868 to AE-O). JCM and SS were twice with cold phosphate-buffered saline and resuspended in binding supported by the Cancer Center Network Program from the ISCIII (RD06/0020/0041). The Cancer Research Institute, and the work carried out at AP and AE-O laboratories, buffer (10 mM HEPES free acid, pH 7.4, 140 mM NaCl and 2.5 mM CaCl2). Cells were incubated for 15 min in the dark with Annexin V/FITC (BD receive support from the European Community through the regional development Biosciences) and propidium iodide staining solution. funding program (FEDER), and from the Fundacio´ n Ramoń Areces. AO receives research support from the FISCAM (Grant PI2010/017) and from the ISCIII (Grant PS09/02144). We thank Dr J Jonkers (Division of Molecular Pathology, The In vivo studies Netherlands Cancer Institute, Amsterdam, The Netherlands) for the generous gift of Mice were manipulated at the animal facility following legal guidelines. the Brca1/p53 conditional mouse model. The Brca1/p53 conditional mice were generously provided by Dr J Jonkers (Division of Molecular Pathology, The Netherlands Cancer Institute, 25 Amsterdam, The Netherlands). This model has a somatic insertion of REFERENCES Brca1 and p53 conditional alleles that, when crossed to K14/cre mice, create mice with deletions of these genes in the epithelia, including breast 1 Cleator S, Heller W, Coombes RC. Triple-negative breast cancer: therapeutic epithelia. The mice developed breast and skin tumors after 5 months. options. Lancet Oncol 2007; 8: 235–244. When the tumors were measurable, the mice were randomized and 2 Hudis CA, Gianni L. Triple-negative breast cancer: an unmet medical need. assigned into two groups (with equal average tumor volumes (50 mm3) Oncologist 2011; 16(Suppl 1): 1–11. before initiation of treatments): vehicle (n ¼ 5) and BEZ235 (n ¼ 4). When 3 Nielsen TO, Hsu FD, Jensen K, Cheang M, Karaca G, Hu Z et al. Immunohisto- the tumors reached a volume of 150 mm3, the mice were treated daily and chemical and clinical characterization of the basal-like subtype of invasive breast orally with 30 mg/kg BEZ235 dissolved in 10% N-Methyl-2-pyrrolidone carcinoma. Clin Cancer Res 2004; 10: 5367–5374. (Sigma-Aldrich) and 90% PEG300. Analyses of diameters and volumes of 4 Carey L, Winer E, Viale G, Cameron D, Gianni L. Triple-negative breast cancer: the tumors were calculated as reported previously.46 For biochemical disease entity or title of convenience? Nat Rev Clin Oncol 2010; 7: 683–692. 5 Metzger-Filho O, Tutt A, de Azambuja E, Saini KS, Viale G, Loi S et al. Dissecting the analyses, tumor samples were obtained after killing of the animals by CO2 inhalation, and 8 h after being treated with BEZ235, and immediately heterogeneity of triple-negative breast cancer. J Clin Oncol 2012; 30: 1879–1887. frozen in liquid nitrogen. The tumors were processed as described 6 Dent R, Trudeau M, Pritchard KI, Hanna WM, Kahn HK, Sawka CA et al. Triple- previously.40 negative breast cancer: clinical features and patterns of recurrence. Clin Cancer Female BALBc nu/nu mice (7 weeks old) were obtained from Charles Res 2007; 13: 4429–4434. River Laboratories (Wilmington, MA, USA). A total of 5 Â 106 MDA-MB231 7 Lemmon MA, Schlessinger J. Cell signaling by receptor tyrosine kinases. Cell 2010; cells in 75 ml of DMEM þ 10% fetal bovine serum and 75 ml of Matrigel (BD 141: 1117–1134. Biosciences) were injected into the mammary fat pad. When tumors 8 Hunter T. Tyrosine phosphorylation: thirty years and counting. Curr Opin Cell Biol reached 200 mm3, animals were randomized into 6 groups (n ¼ 5 animals 2009; 21: 140–146. for each condition, each bearing two tumors), and treatments were 9 Duncan JS, Whittle MC, Nakamura K, Abell AN, Midland AA, Zawistowski JS et al. initiated (BEZ235 (30 mg/kg daily, p.o.), Docetaxel Hospira (15 mg/kg, Dynamic reprogramming of the kinome in response to targeted MEK inhibition in weekly, intraperitoneally) and carboplatin Teva (20 mg/kg, each 5 days, triple-negative breast cancer. Cell 2012; 149: 307–321. intraperitoneally)). 10 Hochgrafe F, Zhang L, O’Toole SA, Browne BC, Pinese M, Porta Cubas A et al. Tyrosine phosphorylation profiling reveals the signaling network characteristics of basal breast cancer cells. Cancer Res 2010; 70: 9391–9401. Statistical analyses 11 Sun T, Aceto N, Meerbrey KL, Kessler JD, Zhou C, Migliaccio I et al. Activation of Comparisons of continuous variables between two groups for the mice multiple proto-oncogenic tyrosine kinases in breast cancer via loss of the PTPN12 tumor model experiments were performed using a two-sided Student’s phosphatase. Cell 2011; 144: 703–718. t-test. At least two independent experiments were performed for the 12 Higgins MJ, Baselga J. Targeted therapies for breast cancer. J Clin Invest 2011; 121: in vivo studies. Differences were considered to be statistically significant 3797–3803. when P-values were o0.05. Statistical data are presented as the 13 Baselga J, Stemmer S, Pego A, Chan A, Goeminne J-C, Graas M-P et al. mean±s.d. All data were analyzed using the statistical software SPSS Cetuximab þ cisplatin in estrogen receptor-negative, progesterone receptor- 19.0 (SPSS Inc., Chicago, IL, USA). negative, HER2-negative (triple-negative) metastatic breast cancer: results of the randomized phase II BALI-1 Trial. San Antonio Breast Cancer Conference 2010, (abstract PD01-). Patient samples and immunohistochemical staining 14 Cristofanilli M, Morandi P, Krishnamurthy S, Reuben JM, Lee BN, Francis D et al. The TNBC samples of patients were obtained from the University Hospital Imatinib mesylate (Gleevec) in advanced breast cancer-expressing C-Kit or of Salamanca, and the studies with human samples followed institutional PDGFR-beta: clinical activity and biological correlations. Ann Oncol 2008; 19: Ethics guidelines. Fresh tissue samples from the primary tumors were 1713–1719. embedded in OCT (Tissue-Tek, Aj Alphen aan den Rijn, The Netherlands), 15 Ocana A, Pandiella A, Siu LL, Tannock IF. Preclinical development of molecular- and then in isopentane. Samples were placed at À 80 1C. Selection of the targeted agents for cancer. Nat Rev Clin Oncol 2011; 8: 200–209. tumoral tissue was performed after excision of part of the tumor and 16 Shah SP, Roth A, Goya R, Oloumi A, Ha G, Zhao Y et al. The clonal and mutational staining with hematoxylin/eosin. Processing of the frozen samples was evolution spectrum of primary triple-negative breast cancers. Nature 2012; 486: performed as described previously.40 395–399. Immunohistochemical staining (DAKO HercepTest Kit, Dako, Dako 17 Xia W, Mullin RJ, Keith BR, Liu LH, Ma H, Rusnak DW et al. Anti-tumor activity of Diagno´sticos, S.A. San Just Desvern (Barcelona), Spain) and fluorescence GW572016: a dual tyrosine kinase inhibitor blocks EGF activation of EGFR/erbB2 in situ hybridization (FISH; DAKO HER2 FISH pharmDx Kit) to measure and downstream Erk1/2 and AKT pathways. Oncogene 2002; 21: 6255–6263.

& 2014 Macmillan Publishers Limited Oncogene (2014) 148 – 156 mTOR targeting in breast cancer JC Montero et al 156 18 Chen Z, Lee FY, Bhalla KN, Wu J. Potent inhibition of platelet-derived growth 34 Janes MR, Limon JJ, So L, Chen J, Lim RJ, Chavez MA et al. Effective and selective factor-induced responses in vascular smooth muscle cells by BMS-354825 targeting of leukemia cells using a TORC1/2 kinase inhibitor. Nat Med 2010; 16: (dasatinib). Mol Pharmacol 2006; 69: 1527–1533. 205–213. 19 Hilger RA, Kredke S, Hedley D, Moeller JG, Bauer RJ, Stellberg W et al. ERK1/2 35 Montero JC, Chen X, Ocana A, Pandiella A. Predominance of mTORC1 over phosphorylation: a biomarker analysis within a phase I study with the new Raf mTORC2 in the regulation of proliferation of ovarian cancer cells: therapeutic kinase inhibitor BAY43-9006. Int J Clin Pharmacol Ther 2002; 40: 567–568. implications. Mol Cancer Ther 2012; 11: 1342–1352. 20 Yeh TC, Marsh V, Bernat BA, Ballard J, Colwell H, Evans RJ et al. Biological 36 Baumann P, Mandl-Weber S, Oduncu F, Schmidmaier R. The novel orally bioa- characterization of ARRY-142886 (AZD6244), a potent, highly selective mitogen- vailable inhibitor of phosphoinositol-3-kinase and mammalian target of rapa- activated protein kinase kinase 1/2 inhibitor. Clin Cancer Res 2007; 13: 1576–1583. mycin, NVP-BEZ235, inhibits growth and proliferation in multiple myeloma. 21 Maira SM, Stauffer F, Brueggen J, Furet P, Schnell C, Fritsch C et al. Identification Exp Cell Res 2009; 315: 485–497. and characterization of NVP-BEZ235, a new orally available dual phosphatidyli- 37 McMillin DW, Ooi M, Delmore J, Negri J, Hayden P, Mitsiades N et al. Antimyeloma nositol 3-kinase/mammalian target of rapamycin inhibitor with potent in vivo activity of the orally bioavailable dual phosphatidylinositol 3-kinase/mammalian antitumor activity. Mol Cancer Ther 2008; 7: 1851–1863. target of rapamycin inhibitor NVP-BEZ235. Cancer Res 2009; 69: 5835–5842. 22 Laplante M, Sabatini DM. mTOR signaling at a glance. J Cell Sci 2009; 122: 38 Chiarini F, Grimaldi C, Ricci F, Tazzari PL, Evangelisti C, Ognibene A et al. Activity of 3589–3594. the novel dual phosphatidylinositol 3-kinase/mammalian target of rapamycin 23 Serra V, Scaltriti M, Prudkin L, Eichhorn PJ, Ibrahim YH, Chandarlapaty S et al. PI3K inhibitor NVP-BEZ235 against T-cell acute lymphoblastic leukemia. Cancer Res inhibition results in enhanced HER signaling and acquired ERK dependency in 2010; 70: 8097–8107. HER2-overexpressing breast cancer. Oncogene 2011; 30: 2547–2557. 39 Serra V, Markman B, Scaltriti M, Eichhorn PJ, Valero V, Guzman M et al. 24 Diaz-Rodriguez E, Pandiella A. Multisite phosphorylation of Erk5 in mitosis. J Cell NVP-BEZ235, a dual PI3K/mTOR inhibitor, prevents PI3K signaling and inhibits Sci 2010; 123: 3146–3156. the growth of cancer cells with activating PI3K mutations. Cancer Res 2008; 68: 25 Liu X, Holstege H, van der Gulden H, Treur-Mulder M, Zevenhoven J, Velds A et al. 8022–8030. Somatic loss of BRCA1 and p53 in mice induces mammary tumors with features of 40 Montero JC, Seoane S, Ocana A, Pandiella A. P-Rex1 participates in Neuregulin- human BRCA1-mutated basal-like breast cancer. Proc Natl Acad Sci USA 2007; 104: ErbB signal transduction and its expression correlates with patient outcome in 12111–12116. breast cancer. Oncogene 2011; 30: 1059–1071. 26 Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: the 41 Neve RM, Chin K, Fridlyand J, Yeh J, Baehner FL, Fevr T et al. A collection of breast combined effects of multiple drugs or enzyme inhibitors. Adv Enzyme Regul 1984; cancer cell lines for the study of functionally distinct cancer subtypes. Cancer Cell 22: 27–55. 2006; 10: 515–527. 27 Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO et al. MET 42 Sarbassov DD, Guertin DA, Ali SM, Sabatini DM. Phosphorylation and regulation of amplification leads to gefitinib resistance in lung cancer by activating ERBB3 Akt/PKB by the rictor-mTOR complex. Science 2005; 307: 1098–1101. signaling. Science 2007; 316: 1039–1043. 43 Diaz-Rodriguez E, Cabrera N, Esparis-Ogando A, Montero JC, Pandiella A. Cleavage 28 Barrios CH, Liu MC, Lee SC, Vanlemmens L, Ferrero JM, Tabei T et al. Phase III of the TrkA neurotrophin receptor by multiple metalloproteases generates sig- randomized trial of sunitinib versus capecitabine in patients with previously nalling-competent truncated forms. Eur J Neurosci 1999; 11: 1421–1430. treated HER2-negative advanced breast cancer. Breast Cancer Res Treat 2010; 121: 44 Yuste L, Montero JC, Esparis-Ogando A, Pandiella A. Activation of ErbB2 by 121–131. overexpression or by transmembrane neuregulin results in differential signaling 29 Hoeflich KP, O’Brien C, Boyd Z, Cavet G, Guerrero S, Jung K et al. In vivo antitumor and sensitivity to herceptin. Cancer Res 2005; 65: 6801–6810. activity of MEK and phosphatidylinositol 3-kinase inhibitors in basal-like breast 45 Chou TC, Motzer RJ, Tong Y, Bosl GJ. Computerized quantitation of synergism and cancer models. Clin Cancer Res 2009; 15: 4649–4664. antagonism of taxol, topotecan, and cisplatin against human teratocarcinoma cell 30 Sarbassov DD, Ali SM, Sengupta S, Sheen JH, Hsu PP, Bagley AF et al. Prolonged growth: a rational approach to clinical protocol design. J Natl Cancer Inst 1994; 86: rapamycin treatment inhibits mTORC2 assembly and Akt/PKB. Mol Cell 2006; 22: 1517–1524. 159–168. 46 Seoane S, Montero JC, Ocana A, Pandiella A. Effect of multikinase inhibitors on 31 Rodrik-Outmezguine VS, Chandarlapaty S, Pagano NC, Poulikakos PI, Scaltriti M, caspase-independent cell death and DNA damage in HER2-overexpressing breast Moskatel E et al. mTOR kinase inhibition causes feedback-dependent biphasic cancer cells. J Natl Cancer Inst 2010; 102: 1432–1446. regulation of AKT signaling. Cancer Discov 2011; 1: 248–259. 47 Wolff AC, Hammond ME, Schwartz JN, Hagerty KL, Allred DC, Cote RJ et al. 32 Eichhorn PJ, Gili M, Scaltriti M, Serra V, Guzman M, Nijkamp W et al. Phosphati- American Society of Clinical Oncology/College of American Pathologists guideline dylinositol 3-kinase hyperactivation results in lapatinib resistance that is reversed recommendations for human epidermal growth factor receptor 2 testing in breast by the mTOR/phosphatidylinositol 3-kinase inhibitor NVP-BEZ235. Cancer Res cancer. J Clin Oncol 2007; 25: 118–145. 2008; 68: 9221–9230. 48 de Alava E, Ocana A, Abad M, Montero JC, Esparis-Ogando A, Rodriguez CA et al. 33 Miller TW, Hennessy BT, Gonzalez-Angulo AM, Fox EM, Mills GB, Chen H et al. Neuregulin expression modulates clinical response to trastuzumab in patients Hyperactivation of phosphatidylinositol-3 kinase promotes escape from hormone with metastatic breast cancer. J Clin Oncol 2007; 25: 2656–2663. dependence in estrogen receptor-positive human breast cancer. J Clin Invest 49 Allred DC, Harvey JM, Berardo M, Clark GM. Prognostic and predictive factors in 2010; 120: 2406–2413. breast cancer by immunohistochemical analysis. Mod Pathol 1998; 11: 155–168.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2014) 148 – 156 & 2014 Macmillan Publishers Limited