Abrogating phosphorylation of eIF4B is required for EGFR and mTOR inhibitor synergy in triple-negative breast cancer

Abstract

Purpose

Triple negative breast cancer (TNBC) patients suffer from a highly malignant and aggressive disease. They have a high rate of relapse and often develop resistance to standard chemotherapy. Many TNBCs have elevated epidermal growth factor receptor (EGFR) but are resistant to EGFR inhibitors as monotherapy. In this study we sought to find a combination therapy that could sensitize TNBC to EGFR inhibitors.

Methods

Phospho-mass spectrometry was performed on the TNBC cell line, BT20, treated with 0.5 μM gefitinib. Immunoblotting measured protein levels and phosphorylation. Colony formation and growth assays analyzed the treatment on cell proliferation while MTT assays determined the synergistic effect of inhibitor combination. A dual luciferase reporter gene plasmid measured translation. All statistical analysis was done on CalucuSyn and GraphPad Prism using ANOVAs.

Results

Phospho-proteomics identified the mTOR pathway to be of interest in EGFR inhibitor resistance. In our studies, combining gefitinib and temsirolimus decreased cell growth and survival in a synergistic manner. Our data identified eIF4B, as a potentially key fragile point in EGFR and mTOR inhibitor synergy. Decreased eIF4B phosphorylation correlated with drops in growth, viability, clonogenic survival, and cap-dependent translation.

Conclusions

Taken together these data suggest EGFR and mTOR inhibitors abrogate growth, viability, and survival via disruption of eIF4B phosphorylation leading to decreased translation in TNBC cell lines. Further, including an mTOR inhibitor along with an EGFR inhibitor in TNBC with increased EGFR expression should be further explored. Additionally, translational regulation may play in important role in regulating EGFR and mTOR inhibitor synergy and warrants further investigation.

Introduction

While therapies to treat breast cancer continue to improve outcome, for one group of patients new treatments have not significantly changed survival. Women with Triple Negative Breast Cancer (TNBC) suffer from a highly malignant disease that is characterized by a lack of hormone receptor and HER2 overexpression and therefore, are not responsive to ER/ PR and HER2 targeted agents including tamoxifen and Herceptin. [1-3] New studies have demonstrated promising therapies in TNBC that target VEGF, mTOR, and PARP (NCT00827567, NCT00861705, NCT00528567) and may be of benefit when combined with cytotoxic chemotherapy.

Another potential targeted therapy for TNBC stems from the overexpression of the epidermal growth factor receptor (EGFR) observed both in vitro and up to 50% of patients. [1,3,4] In addition to breast cancer, EGFR is overexpressed in colon and non-small cell lung carcinoma where inhibitory antibodies and small molecule tyrosine kinase inhibitors are used efficaciously in the clinic.[5] Unfortunately, the use of the EGFR inhibitor cetuximab in TNBC has been ineffective.[6] One proposed mechanism for this intrinsic resistance to EGFR inhibitors in TNBC is crosstalk between EGFR and other signaling proteins.[4,7,5,2] Specifically, crosstalk between c-Met, c-Src, IGF-IR, HER2, and HER3 signaling with EGFR activation has been shown to abrogate the efficacy of monotherapy tyrosine kinase inhibitors and promote resistance to EGFR targeted therapies.[5]

Here we used a mass-spectrometry based phospho-proteomic technique to identify signaling proteins that remain phosphorylated after EGFR inhibition. We found that many components of the mTOR signaling pathway remained phosphorylated in the presence of the EGFR inhibitor, gefitinib. Based on these observations we investigated the combination of gefitinib with an mTOR inhibitor, temsirolimus. Our results suggested that gefitinib and temsirolimus in combination was synergistic in TNBC cell lines and decreased growth and colony formation through a non-MAPK and AKT mediated pathway. Instead our data suggested an important role for the translation initiation factor eIF4B in regulating gefitinib and temsirolimus synergy.

Materials and Methods

Cell Culture and Reagents

Gefitinib (Iressa) was provided by AstraZeneca (London, UK). Temsirolimus was purchased from LC Labs (Woburn, MA, USA). MDA-MB-231, MDA-MB-468, and BT20 cells were purchased from ATCC (Manassas, VA, USA). HEK293T cells were purchased from Life Technologies (Carlsbad, CA, USA). MDA-MB-231, MDA-MB-468, and HEK293T cells are grown in DMEM+10% FBS media (Dulbecco's modified Eagle's medium supplemented with 10% Fetal Bovine Serum). BT20 cells are grown in Eagle's + NEAA media (Eagle's MEM [Minimum Essential Medium] with 2 mM L-glutamine and Earle's Balanced Salt Solution adjusted to contain 1.5 g/L sodium bicarbonate, 0.1 mM non-essential amino acids, 1 mM sodium pyruvate, and 10% FBS). All other reagents were purchased from Thermo Fisher (Houston, TX, USA) or Sigma (St. Louis, MO, USA), unless indicated.

Phospho- Proteomics Analysis

BT20 cells were treated with 0.5 μM gefitinib or a DMSO vehicle control for 24 hours. Cells were collected in ice cold 100% EtOH and solubilized in 0.2 ml of Tris, 10 mM pH=7.5, LiF, 1 mM, Na₃VO₄, 0.1 mM, EDTA (Ethylenediaminetetraacetic acid) 1 mM and LiDS (Lithium Dodecyl Sulfate) 0.5%. Samples were filtered through 0.45 μm 13 mm GHP filters (Pall, Port Washington, NY, USA) and phosphopeptides were selected by incubation with 6 mg/sample TiO₂ beads (GL Sciences, Torrance, CA, USA, 5 μm). Eluted peptides were solubilized in 0.1% formic acid and analyzed by LC-MS/MS performed on a Thermo LTQ equipped with ETD (electron-disassociation transfer) (ThermoFisher Scientific, Watham, MA, USA). Samples were loaded on a peptide Captrap (Michrom, Auburn, CA, USA) trapping column and peptide separations were achieved using a linear gradient of 5% to 35% acetonitrile to elute from a Majic 0.1 mm x 150 mm AQ C18 column (Michrom). Tandem mass spectra were extracted by Proteome Discoverer (ThermoFisher Scientific) version 1.4.0.288. All MS/MS data were analyzed using Mascot (Matrix Science, London, UK; version 2.4.0) and X! Tandem (The GPM, thegpm.org; version CYCLONE (2010.12.01.1)). Additional details for the sample preparation and analysis parameters are available in the supplemental information.

Colony Formation Assays

Cells were cultured in triplicate in the presence of gefitinib and/or temsirolimus. BT20 cells (1 μM gefitinib and temsirolimus) and MDA-MB-231 and MDA-MB-468 (10 μM of each drug) were treated with the indicated compounds every other day for 10 days. Trypsinized cells were replated at 5,000 cells/ 35-mm dish (without treatment) for 7 days. Colonies were counted using Gelcount (Oxford Optromix; Abingdon, United Kingdom) and normalized to the untreated control. Experiments were done in triplicate and repeated at least three times. Graphs and statistics were done in GraphPad Prism (La Jolla, CA, USA) and analyzed using an ANOVA.

Cell Growth Assays

Cells were plated at 30,000 cells/ well of a 6 well plate in triplicate on Day 0. Treatment with gefitinib and/or temsirolimus began on Day 1 and continued every other day for 8 days. Day 1 untreated cells were counted using a hemocytometer. On Days 4 and 8 the respective plates were counted. Experiments were repeated at least three times. Graphs and statistics were done in GraphPad Prism using an ANOVA.

Cell Viability Assays

Cells were plated at 2,000 cells /well of a 96-well plate in triplicate. Cells were treated with log doses of gefitinib and/or temsirolimus. The MTS reagent was added per manufacture directions (Promega, Madison, WI, USA) and was read using a Dynex spectrophotometer after 72 hours. GraphPad Prism was used to generate non-linear inhibitory growth curves with top=1 and bottom=0. GI₅₀ values were generated by the program from the summary of at least three experiments performed in triplicate.

Statistical Analysis

ANOVAs and Student's T test were performed on GraphPad Prism software. Isobolograms were performed by determining the GI₅₀ values for gefitinib and temsirolimus using GraphPad Prism via standard MTS assay. Combinatorial Index (CI) values were calculated using the following equation: (IC₅₀ combination/IC₅₀ gefitinib) + (concentration of temsirolimus/IC₅₀ temsirolimus) by the computer program Calcusyn (Biosoft, Cambridge, United Kingdom).

Immunoblotting

Immunoblotting was performed as previously described.[8] Antibodies used in this study were purchased from Cell Signaling Technologies (Beverly, MA, USA): phospho-MAPK (Thr202/Tyr204), MAPK, phospho-P38 (Thr180/Tyr182), P38, phospho-AKT (Ser473), AKT, phospho-eIF4B (Ser422), eIF4B, eIF4A, eIF4A1, phosphoeIF4E (Ser209), eIF4E, phospho-eIF4G (Ser1108), eIF4G, eIF4H, phospho-P90RSK (Ser380), RSK1/2/3, phospho-P70S6K (Thr389), P70S6K, β-actin; secondary horseradish peroxidase-conjugated antibodies. Amersham ECL Prime Western Blotting Detection Reagent solution was purchased from GE Healthcare (Amersham, United Kingdom).

Bicistronic Luciferase Assay

Dual luciferase plasmid was purchased from Addgene (Cambridge, MA, USA). HEK293T cells were transfected with11510:pFR_HCV_xb. After 24 hours, media was removed and replaced with the indicated treatments. Cells were then harvested after an additional 24 hours and luciferase activity was measured using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, USA) and read using the BioTek Synergy 2 machine (Winooski, VT, USA) according to the manufacturers’ instructions. Relative luciferase units were plotted in GraphPad Prism.

Results

mTOR signaling remains activated in the presence of EGFR inhibitors in TNBC

Many TNBC cell lines express high levels of EGFR but have de novo resistance to EGFR inhibitors.[4] (See Online Resource 1 for demonstration of EGFR kinase inhibition in TNBC with gefitinib). To identify potential mechanisms of resistance in TNBC cells to EGFR inhibitors we treated the TNBC cell line BT20 with 0.5 μM gefitinib (a concentration previously shown to be selective for EGFR) or DMSO (a vehicle control) and performed phospho-mass spectrometry. Two hundred seventy-nine proteins whose phosphorylation levels did not significantly change in the presence of gefitinib were identified (Table 1, Online Resource 2). After pathway analysis using Ingenuity® Systems software (Redford, CA, USA) we observed that many of these phospho-proteins were involved in the mTOR pathway, particularly proteins critical to translation initiation (Table 1, Online Resource 2). The high number of proteins with unchanged phosphorylation after gefitinib treatment in BT20 cells led us to believe that the mTOR pathway might be contributing to EGFR inhibitor resistance. The EGFR and mTOR pathways have been intimately linked to cancer progression for many years.[9]'[7]

Table 1. mTOR signaling remains activated in the presence of EGFR inhibitors in TNBC.

Mass spectrometry on BT20 cells showed 279 proteins whose phosphorylation status did not significantly change in the presence of 0.5 μM gefitinib. Ingenuity® Pathways Analysis identified many components of the mTOR pathway and translation initiation factors to be of interest.[17,12] A sample of the mTOR associated proteins and their function are listed.

Proteins involved in translational control that remained phosphorylated after gefitinib treatment involved in mTOR pathway
Protein	mTOR Related Function
eIF1B	Enhance rate and accuracy of translation
eIF3B	Step 1 of Translation initiation 40S subunit binding
eIF3D	Step 1 of Translation initiation 40S subunit binding
eIF3G	Step 1 of Translation initiation 40S subunit binding
eIF3J	Step 1 of Translation initiation 40S subunit binding
eIF3C	Step 1 of Translation initiation 40S subunit binding
eIF2A	Step 2 Bind initiator tRNA and 40S subunit
eIF4G1	Step 3 of Translation initiation activation and binding of mRNA to 40S
Raptor	Major subunit of mTORC1
Rictor	Major subunit of mTORC2
GSK3A	Glycogen synthase kinase
4EBP1	Directly binds and activates eIF4E- rate limiting step in Translation initiation
FKHR (FOXO3)	Forkhead Transcription factor- glucose homeostasis, cell-cycle progression, and apoptosis

Inhibiting mTOR signaling sensitizes TNBC cells to EGFR inhibitors

The phospho-proteomic data suggest that the mTOR pathway may contribute to de novo resistance to EGFR inhibitors. We predict that dual inhibition of both the EGFR and mTOR pathways will inhibit the growth of TNBC cells. The drug we chose for mTOR inhibition is temsirolimus, an mTORC1 inhibitor approved for the treatment of renal cell carcinoma.[10] We identified three TNBC cell lines (BT20, MDA-MB-231, and MDA-MB-468) as being resistant to single agent gefitinib and temsirolimus treatment (GI₅₀ >5 μM). We performed growth assays and clonogenic survival assays to determine the effect of EGFR and mTOR inhibitor combination. Treatment with gefitinib (GEF) and temsirolimus (TEM) in combination significantly decreased cell growth in all three cell lines (Figure 1A). In all cell lines tested there was a significant decrease in cell growth between the no treatment (NT) group and the combination of gefitinib and temsirolimus (GEF+TEM) (p<0.01) and between GEF and GEF+TEM (p<0.05). Single agent temsirolimus treatment had a more pronounced effect decreasing cell growth than gefitinib but the decrease was not as robust or significant as the combination. In addition, this response to temsirolimus alone varied between the cell lines; however, in all three of the cell lines tested, the combination of EGFR and mTOR inhibitors was significantly decreased over gefitinib treatment alone.

Fig 1. Gefitinib and temsirolimus combination decreases TNBC cell growth and colony.

(a) Growth assays were done in BT20, MDA-MB-231, and MDA-MB-468 cells over 8 days in triplicate. Cells were treated every other day with 1 μM gefitinib (GEF), temsirolimus (TEM), or the combination (GEF+TEM) and counted on days 1, 4, and 8. Growth curves were generated by normalizing the number of cells to the Day 1 counts. The increase in growth is represented as a fold change. (b) Colony formation assays were plated in triplicate. Treatments were performed every other day at 1 μM for BT20 cells and 10 μM for MDA-MB-231 and MDA-MB-468 for two weeks. Cells were trypsinized, replated at low density, and allowed to grow for another week in normal growth media. Colony numbers were normalized to the NT (no treatment) group. Each experiment was repeated at least three times. p*<0.05 **<0.01 ***<0.001 ****<0.0001

To determine if the combination of EGFR and mTOR inhibition significantly decreased cell survival, we utilized clonogenic survival assays. These assays determine the ability of cells treated with the inhibitors to survive after drug washout. Inhibiting EGFR or mTOR singularly had a minimal effect on cell survival (red checkerboard and green vertical lines, respectively); however, the combination of GEF+ TEM (blue slanted lines) had a significant decrease in clonogenic survival in all three TNBC cell lines (Figure 1B). Specifically, when comparing NT (black solid) with the combination of GEF+TEM, there was a significant decrease in cell survival. These results suggest that the TNBC cells treated with GEF+TEM are sent down an irreversible path that will not allow them to recover from dual treatment, at least not at the same rate as with single treatments.

EGFR and mTOR inhibitors are synergistic in TNBC

We observed a decrease in cell growth and colony formation with treatment of gefitinib and temsirolimus in combination therefore, to determine if this drug combination is synergistic we performed MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assays. Using these assays we determined the growth inhibitory value at 50% (GI₅₀) for each drug alone (red and green lines) and in combination (blue lines) (Figure 2A). When the two drugs were combined at constant ratios another inhibitory dose response curve was generated and a clear shift in the GI₅₀ can be observed. These values were used to evaluate synergy of the two drug combination using the Chou-Talalay method.[11] Specifically, increasing concentrations of gefitinib and temsirolimus were added to the cells alone and in combination at constant ratios. The GI₅₀ was calculated from the fraction of viable cells using CalcuSyn software and a line was drawn on a graph connecting the individual GI₅₀s. The GI₅₀ was calculated for one drug in the presence of the other and plotted on the same graph (Figure 2B; triangles). Points falling under the line indicate synergy. Combinatorial index (CI) values were calculated as described and values less than 1.0 for all three TNBC cell lines indicate synergy between gefitinib and temsirolimus (BT20 CI=0.21, MDAMB-231 CI=0.28, MDA-MB-468 CI=0.56). Similar results were observed using additional EGFR and mTOR inhibitors, erlotinib and rapamycin (BT20 CI=0.30, MDA-MB-231 CI=0.22, MDA-MB-468 CI=0.19). Therefore, using multiple methods we were able to demonstrate a synergy between EGFR and mTOR inhibitors in TNBC cell lines.

Fig 2. Gefitinib and temsirolimus synergize in TNBC cells.

(a) The GI₅₀ curves for BT20 (top), MDA-MB-231 (middle), and MDA-MB-468 (bottom) cells were plotted in the presence of GEF, TEM or GEF+TEM using the raw absorbance data from the MTT assay. (b) Isobolograms were made by plotting the GI₅₀ of GEF on the Y axis and the GI₅₀ of TEM on the X axis. Triangles represent combination treatment points. Each experiment was performed a minimum of three independent times. GI₅₀ gefitinib: BT20=5.3 μM, MDA-MB-231>100 μM, MDA-MB-468=32.7 μM, GI₅₀ temsirolimus: BT20=9.8 μM, MDA-MB-231>100 μM, MDA-MB-468=61.05 μM

MAPK and AKT signaling remain phosphorylated in the presence of EGFR and mTOR inhibitors

To determine which proteins are responsible for the synergistic effect of gefitinib and temsirolimus in TNBC cell lines, we performed immunoblotting analysis for classical signaling proteins involved with EGFR and mTOR activation[7] (Figure 3). Specifically, phosphorylation of p44/42MAPK, p38MAPK, and AKT was assessed after treatment with gefitinib and/or temsirolimus over a 72 hour period with the 24 hour time point shown in the immunoblots. If phosphorylation of one of the tested proteins was responsible for the observed synergy, we would anticipate the combination treatment would abrogate its phosphorylation. Unexpectedly, phosphorylation of p42/44MAPK, AKT, and p38MAPK was not altered with the single or combination treatments (Figure 3, rows 1, 3, and 5). Therefore, these data demonstrate that classical EGFR and mTOR signaling pathways may not be involved in the synergic decrease in cell growth, survival, and viability observed with gefitinib and temsirolimus treatment in TNBC cells.

Fig 3. MAPK and AKT signaling remain activated in the presence of EGFR and mTOR inhibitors.

Cells were treated with 1 μM GEF and/or 1 μM TEM for 24 hours. Western blots were performed to detect each of the indicated proteins. Each blot was performed a minimum of three independent times with similar results. Blots are cropped for space and ease of explanation.

eIF4B phosphorylation is a fragile point in EGFR and mTOR signaling

One signaling pathway activated by mTOR that was suggested by the phospho-proteomic data to remain activated in the presence of EGFR inhibitors was the translation initiation pathway. Translation is a major downstream target of mTOR and is a highly regulated process that is often overactive in cancer.[12,13] Once activated, mTOR phosphorylates two effector proteins, P70S6K and 4E-BP1.[14,15] P70S6K controls the phosphorylation of eIF4B, a protein involved in eIF4A helicase efficiency, whereas, 4E-BP1 is a well-studied protein in the control of eIF4E, the rate limiting translation initiation factor.[12,16,15,17,18] The first step in translation initiation involves 4E-BP1 releasing eIF4E to complex with two other factors, eIF4G and eIF4A to form the eIF4F complex resulting in the recruitment of the 40S ribosome and the beginning of translation.[19,17] We analyzed the expression and phosphorylation of a number of translation initiation proteins anticipating their relevance to be observed as a decrease in phosphorylation in the combination treatment (Figure 4). Interestingly, eIF4B phosphorylation was completely abrogated as shown through immunoblotting with the combination gefitinib and temsirolimus treatment. This decrease was not observed with single agent treatment in the MDA-MB-231 and MDA-MB-468 (Figure 4, row 1). BT20 cells did show a decrease in phosphorylation of eIF4B with temsirolimus treatment. Further testing the role of eIF4B in cell viability compared to gefitinib and temsirolimus treatment we performed MTT cell viability assays with eIF4B siRNA knockdown (Online Resource 3). eIF4B knockdown significantly decreases MDA-MB-231 cell viability similar to gefitinib and temsirolimus combination. From these studies, our data suggest that dual inhibition of EGFR and mTOR pathways converge upon eIF4B to regulate translation initiation and may be a critical component of EGFR and mTOR inhibitor synergy in TNBC.

Fig 4. eIF4B phosphorylation is a fragile point in EGFR and mTOR signaling.

Cells were treated with 1 μM GEF and/or 1 μM TEM for 24 hours. Western blots were performed to detect each of the indicated proteins. Each blot was performed a minimum of three independent times with similar results. Blots are cropped for space and ease of explanation

P70S6K and P90RSK are responsible for phosphorylating eIF4B downstream of EGFR and mTOR signaling

P90RSK and P70S6K are two kinases identified as having the ability to phosphorylate eIF4B on Ser422.[20] To determine if these two kinases are involved in the phosphorylation of eIF4B downstream of EGFR and mTOR in TNBC, we treated the cell lines with gefitinib and/or temsirolimus for 24 hours and measured phosphorylation of P90RSK and P70S6K as surrogates for activation of each protein. We found that the phosphorylation of P90RSK was completely abrogated in BT20 cells and markedly reduced in MDA-MB-231 and MDA-MB-468 cells with the combination of gefitinib and temsirolimus while remaining activated with single agent treatment (Figure 5, row 1). P70S6K phosphorylation was abrogated with the addition of temsirolimus as a single agent or in combination with gefitinib, which was not surprising as mTOR inhibition is known to decrease P70S6K phosphorylation (Figure 5, row 3). Taken together, these data implicate P90RSK as a mediator of EGFR and mTOR signaling to translation in TNBC.

Fig 5. P70S6K and P90RSK phosphorylate eIF4B downstream of mTOR and EGFR signaling.

TNBC cells were treated with 1 μM GEF and/or 1 μM TEM. Western blots were performed to detect each of the indicated proteins. Each blot was performed a minimum of three independent times with similar results. Blots are cropped for space and ease of explanation

Inhibiting EGFR and mTOR blocks cap-dependent translation

There are two distinct modes of translation in eukaryotic cells: cap-dependent and cap-independent. The latter is also known as internal ribosome entry site (IRES) translation.[19] Eukaryotic cells are able to use cap-independent translation but mainly rely on the eIF proteins to mediate cap-dependent translation, comprising 95-97% of all translation.[14,21,19] It is believed that eIF4B is involved in cap-dependent translation and plays little to no role in cap-independent translation.[16,14,19] To examine the effects gefitinib and temsirolimus have on eIF4B and its role in translation we used a dual luciferase assay to measure both cap-dependent and cap-independent translation. We utilized a reporter plasmid (11510:pFR_HCV_xb) in which expression of firefly luciferase is regulated by cap-dependent translation and the expression of renilla luciferase is regulated by cap-independent translation from the same plasmid.[22] Both cap-dependent and cap-independent translation were measured simultaneously in a dual injector plate reader (Figure 6). Cells were treated with EGFR and/or mTOR inhibitors after transfection and lysates were prepared to measure luciferase activity. There was no significant change in cap-independent translation throughout the experiment and the relative light units generated as measures of renilla luciferase activity were therefore used as transfection controls for the assay (data not shown). The data are represented as cap-dependent/cap-independent normalized relative light units. Interestingly, there was no change in cap-dependent translation with gefitinib or temsirolimus (red checkerboard and green vertical lines, respectively) treatment compared to the NT control (black solid) (Figure 6). However, there was a significant decrease in cap-dependent translation with the combination of gefitinib and temsirolimus (blue slanted lines) when compared to NT, GEF, or TEM treated cells (Figure 6). Our results suggest the combination treatment of gefitinib and temsirolimus decreases cap-dependent translation while having a minimal effect on cap-independent translation.

Fig 6. Inhibiting EGFR and mTOR signaling blocks cap-dependent translation.

The plasmid 11510:pFR_HCV_xb was transfected into HEK293T cells to measure cap-dependent translation through firefly luciferase and cap-independent translation through renilla luciferase. Luciferase was measured using a dual-luciferase reported assay system from whole cell lysates. Values within each experiment were normalized to 1.0 as NT (no treatment). Relative light units from the firefly luciferase was plotted over relative light units from the renilla luciferase. Each experiment was performed in triplicate at least two times.

Discussion

While survival rates for breast cancer patients are rising, those with TNBC continue to have no suitable and successful first line chemotherapy other than standard cytotoxics. These patients do not express the cellular targets for common targeted therapies and only standard chemotherapy is available with limited success.[3,2] Through finding better drug combinations and understanding their mechanisms of action, physicians will be able to more effectively treat TNBC increasing quality of life and decreasing rates of relapse. EGFR and mTOR both are targets with approved inhibitors for different types of cancer and since TNBC often expresses high levels of EGFR, the receptor has potential to be exploited as a target.[2,23] Other researchers studying colon, prostate, and breast cancers have explored the use of mTOR inhibitors in combination with EGFR inhibitors.[24,4] Everolimus was shown to sensitize GEO colon, PC3 prostate, and MDA-MB-468 TNBC cells to gefitinib and cetuximab and decrease their growth in a dose-dependent manner. Using mouse xenografts the study showed a decrease in colon cancer tumor burden by 90% when the drugs were used in combination.[24] Additional studies have shown a synergistic combination of lapatinib and rapamycin in TNBC cell lines and decreased mouse xenograft tumor progression.[4] While EGFR inhibitors as monotherapy have not been effective due to intrinsic resistance, our results and those of others suggest that combinations of EGFR and mTOR inhibitors are synergistic and their combined potential as a targeted therapy in TNBC needs to be further studied.[4,25]

Similar to published studies, our work found that dual treatment with gefitinib and temsirolimus had a synergistic effect on decreasing TNBC cell viability and clonogenic survival. While the combination of EGFR and mTOR inhibitors is not novel in TNBC, the mechanism of synergy is not understood. Our studies found an EGFR and mTOR crosstalk mechanism involving the eukaryotic translation initiation factor 4B (eIF4B). Our data suggest that phosphorylation of eIF4B in TNBC may be a “fragile site” for EGFR and mTOR synergy such that inhibiting both pathways converge at abrogation of eIF4B phosphorylation. Data suggest that this decrease in eIF4B phosphorylation correlates with a decrease in cap-dependent translation, thereby implicating regulation of translation in the reduction of growth, viability, and survival in TNBC.

Aberrant translation has been implemented in cancer progression for a number of years. Cancer cells have a high demand for proteins needed to sustain increased proliferation and metabolic rate. The translation initiation factor eIF4E was shown to induce transformation in cells and has been considered an oncogene.[26] Most of the translation initiation factors have altered expression in different types of cancer [12] and ribosomal disorders are known to lead to increased cancer risk.[13,27] New research suggests that eIF4A is critically involved in BRAF and MEK inhibitor resistance in melanoma, colon, and thyroid cancer cell lines and eIF4F causes oncogenic translation.[28,29] It was previously discovered that the MAPK and mTOR/PI3K pathways converge on eIF4B [20] but to our knowledge this is the first evidence of EGFR and mTOR inhibitor synergy being regulated by eIF4B in TNBC. Taken together, these results have identified a new fragile point, eIF4B phosphorylation, and cap-dependent translation as a mediator of EGFR and mTOR crosstalk in TNBC.

Our results suggest that MAPK and AKT are not responsible for regulating the synergy as their activation remains unchanged following treatment with gefitinib and temsirolimus. Other studies in colon and lung cancer also have found that AKT and MAPK maintain activity after EGFR and mTOR inhibitor treatment. [7,24] While MAPK and AKT are often considered the main signaling proteins involved with EGFR and mTOR, respectively, our data suggest a more prominent role for P70S6K and P90RSK in TNBC resistant to EGFR inhibitors. P90RSK is phosphorylated by MAPK, p38MAPK, ERK5MAPK, PDK1, TIF1α, GSK3α, GSK3α, and fibroblast growth factor receptor-3 (FGFR3) and is known to be up-regulated in breast and prostate cancer. It is able to activate mTOR leading to increased P70S6K phosphorylation and can directly phosphorylate P70S6K.[30-32] Both kinases are involved in protein translation through their ability to phosphorylate eIF4B. P70S6K is also known to phosphorylate 4E-BP1 to release eIF4E, the rate-limiting step in translation initiation, allowing it to form the eIF4F complex initiating translation. Our data suggest that P70S6K regulation through EGFR and mTOR dual inhibition acts independent of 4E-BP1 as immunoblotting indicates no consistent change in 4E-BP1 phosphorylation after treatment in TNBC cells lines (Online Resource 4). Instead the gefitinib and temsirolimus combination abrogates eIF4B phosphorylation through P70S6K and P90RSK inhibition and decreases cap-dependent translation leading to the synergy we observe in TNBC cell lines. Our studies identified EGFR and mTOR inhibitors as a potential effective combination treatment for TNBC and the drug combination needs to be further explored. The effect translation has on cancer cells in regard to the mTOR and EGFR pathways is largely unexplored in TNBC and further implicates eIF4B as a protein of interest in understanding the gefitinib and temsirolimus synergy.

The data presented in this report represent the summary of results from three independently derived TNBC cell lines. In the Online Resource 5, the known mutations found in these cell lines that are present in the pathways studied in this report are identified. Each TNBC cell line contains a different compliment of mutations as does each breast tumor derived from an individual patient. We believe this in part explains some of the subtle difference in responses to the inhibitors in this report. For example, the BT20 cell line demonstrates sensitivity to temsirolimus with respect to growth and survival. This cell line also has a pronounced decrease in phosphorylation of eIF4B with temsirolimus alone, unlike the other cell lines. From the mutation table we note that BT20 cells have a PIK3CA mutation which may render this cell line more sensitive to mTOR inhibitors. Therefore, we believe that we have a good representative of TNBC breast cancer responding to a combination of inhibitors through a similar mechanism.