Epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) have been used to treat non–small cell lung cancer (NSCLC). However, the overall response rate to EGFR TKIs is limited, and the mechanisms mediating resistance to the drugs are poorly understood. Here, we report that insulin-like growth factor-I receptor (IGF-IR) activation interferes with the antitumor activity of erlotinib, an EGFR TKI. Treatment with erlotinib increased the levels of EGFR/IGF-IR heterodimer localized on cell membrane, activated IGF-IR and its downstream signaling mediators, and stimulated mammalian target of rapamycin (mTOR)–mediated de novo protein synthesis of EGFR and survivin in NSCLC cells. Inhibition of IGF-IR activation, suppression of mTOR-mediated protein synthesis, or knockdown of survivin expression abolished resistance to erlotinib and induced apoptosis in NSCLC cells in vitro and in vivo. Our data suggest that enhanced synthesis of survivin protein mediated by the IGFR/EGFR heterodimer counteracts the antitumor action of erlotinib, indicating the needs of integration of IGF-IR–targeted agents to the treatment regimens with EGFR TKI for patients with lung cancer. (Cancer Res 2006; 66(20): 10100-11)
- non–small cell lung cancer
The 5-year survival rate for lung cancer patients remains extremely poor (≤15; ref. 1), underscoring the need for more effective treatment strategies. Recently, new therapeutic approaches targeting signaling pathways involved in cell proliferation, apoptosis, angiogenesis, and metastasis have been investigated ( 2). Among the many potential target pathways, the epidermal growth factor (EGF) receptor (EGFR) signaling pathway has been studied most extensively because EGFR overexpression has been observed in a number of solid tumors, including 40% to 80% of non–small cell lung cancers (NSCLC; ref. 3). The EGFR signaling pathway activates the phosphatidylinositol 3-kinase (PI3K)/Akt and mitogen-activated protein kinase (MAPK) pathways, which play major roles in cell proliferation, survival, and transformation and in therapeutic resistance ( 4, 5). In addition, the EGFR pathway is implicated in angiogenesis, and cell invasion by its regulation of the expression and activity of matrix metalloproteinases ( 6, 7).
These findings indicate the therapeutic potential of inhibitors of EGFR tyrosine kinase activation. EGFR tyrosine kinase activity can be inhibited by antibodies against the extracellular domain of EGFR, such as cetuximab, or by small molecules that block the ATP binding site of the cytoplasmic domain, such as gefitinib (ZD1839, Iressa; AstraZeneca Pharmaceuticals, Macclesfield, United Kingdom) and erlotinib (Tarceva®; OSI Pharmaceuticals; Genentech, South San Francisco, CA). Both forms of EGFR inhibition have single-agent antitumor activity against previously treated NSCLC ( 3, 8– 10). Erlotinib exhibits an antiproliferative effect at nanomolar concentrations and has induced apoptosis and reversible cell cycle arrest at G1 ( 11). In vivo preclinical models have shown that erlotinib administration markedly reduces EGFR autophosphorylation and growth in human head and neck cancer xenografts (HN5 and A431 cells) in nude mice ( 11, 12). In addition, gefitinib, combined with standard chemotherapeutic agents and/or radiotherapy in preclinical studies, has inhibited EGFR activation, thus causing G1 cell cycle arrest and contributing to synergistic growth inhibition ( 13).
Despite a similar chemical structure, these two EGFR tyrosine kinase inhibitors (TKIs) have provided contrasting results in phase III clinical trials, in which only erlotinib showed significantly improved survival compared with placebo ( 14– 16). The response to gefitinib and erlotinib has been suggested to be associated with sex, smoking status, tumor histology, and somatic mutations of the EGFR ATP binding site ( 17, 18). Recent data have suggested that the insulin-like growth factor-1 receptor (IGF-IR) pathway is also implicated in the resistance of gefitinib and anti-EGFR monoclonal antibody ( 19, 20). However, to our knowledge, the mechanisms involved in the IGF-IR–mediated acquired resistance to erlotinib in NSCLC cells have not been completely defined. In this article, we report that erlotinib induce EGFR/IGF-IR heterodimerization on the cell membrane, transmitting a survival signal through IGF-IR and its downstream mediators PI3K/Akt and p44/42 MAPK to stimulate mammalian target of rapamycin (mTOR)–mediated synthesis of EGFR and antiapoptotic survivin proteins. Consequently, inactivation of IGF-IR, suppression of mTOR-mediated protein synthesis, or knockdown of survivin protein renders EGFR-overexpressing NSCLC cells sensitive to the erlotinib treatment.
Materials and Methods
Cells, reagents, and animals. The human NSCLC cell lines H596, H226B, H226Br, H460, H1299, A549, H358, H661, and H322 were from the American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 supplemented with 10% fetal bovine serum (FBS; Life Technologies, Gaithersburg, MD) in a humidified atmosphere with 5% CO2. IGF and EGF were from R&D Systems (Minneapolis, MN). Erlotinib were prepared as 10 mmol/L stock solution in DMSO and stored at −20°C. LY294002 (an inhibitor of PI3K), PD98059 (an inhibitor of the MEK1), rapamycin, and AG1024, a TKI of IGF-IR, were from Calbiochem-Novabiochem (Alexandria, New South Wales, Australia); these inhibitors were prepared as 20 mmol/L stock solutions in DMSO and also stored at −20°C. Adenoviral vectors expressing survivin (Ad-survivin; ref. 21) or dnIGF-IR/482 [adenovirus-expressing, dominant-negative IGF-IR (Ad-dnIGF-IR)], a soluble extracellular domain of IGF-IR with an engineered stop codon at amino acid residue 482 ( 22), and control adenoviral vector [adenovirus-expressing empty vector (Ad-EV)] were amplified as described elsewhere ( 23). We confirmed increases in the levels of IGF-IR protein by Western blot assay with an antibody for the α-subunit (anti-IGF-IRα N-20, Santa Cruz Biotechnology, Santa Cruz, CA) using medium from the cells that were infected with Ad-dnIGF-IR because IGF-IR/482 has been shown to produce and release the truncated α-subunit of IGF-IR into the medium ( 22). The effect of the combination of erlotinib and Ad-dnIGF-IR on established s.c. tumor nodules was studied in athymic nude mice (Harlan Sprague-Dawley, Indianapolis, IN) in a defined pathogen-free environment. Six-week-old female mice were used in this study; mice with necrotic tumors or tumors ≥1.5 cm in diameter were euthanized.
Cell proliferation assay. Cells were treated with erlotinib, rapamycin, LY294002, PD98059, AG1024, Ad-dnIGF-IR, Ad-EV, or their combinations in the absence or presence of 10% FBS, EGF (50 ng/mL), or IGF (50 ng/mL). For the experiments with the viruses, cells were infected with 5 and 10 particle forming unit (pfu) for Ad-dnIGF-IR or Ad-EV, in serum-free medium for 2 hours and then incubated for 3 days in RPMI medium supplemented with 10% FBS in the absence or presence of the indicated concentrations of erlotinib. Cell proliferation was measured with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. The drug concentrations required to inhibit cell growth by 50% were determined by interpolation from the dose-response curves. For defining the effect of the combined drug treatments, any potentiation was estimated by multiplying the percentage of cells remaining by each individual agent. The synergistic index was calculated as previously described ( 24). In the following equations, A and B are the effects of each individual agent, and AB is the effect of the combination. Subadditivity was defined as %AB / (%A × %B) < 0.9; additivity was defined as %AB / (%A × %B) = 0.9-1.0; and supra-additivity was defined as %AB / (%A × %B) > 1.0.
Clonogenic growth assay. The anchorage-dependent clonogenic growth assay was done by seeding NSCLC cell lines into six-well plates at low density (≈3 × 103 cells per well). Cells were either left uninfected or infected with 5 or 10 pfu/cell of Ad-dnIGF-IR or Ad-EV, incubated for 72 hours with different concentrations of erlotinib (0.1, 1.0, and 5.0 μmol/L), AG1024 (5 μmol/L), or combinations of the two drugs in serum-free RPMI medium in the absence or presence of IGF (50 ng/mL). Cells were replated in six-well plates and cultured in growth medium for 7 to 10 days, in a humidified atmosphere with 5% CO2, at 37°C, and then colonies were fixed with 0.1% Coomassie blue (Bio-Rad Laboratories, Hercules, CA) in 30% methanol and 10% acetic acid. We then counted the number of colonies with >50 cells. For the anchorage-independent clonogenic growth assay, ∼3 × 103 cells were suspended in 0.75 mL of 0.22% soft agar that was layered on top of 1 mL of 1% solidified agar in each well of 24-well plates. The plates were then incubated for 10 to 15 days in serum-free RPMI medium containing 0.1 or 1.0 μmol/L concentrations of erlotinib in the absence or presence of 10% FBS or IGF (50 ng/mL). The medium was changed daily during this period, at the end of which tumor cell colonies measuring at least 80 μm were counted under using a dissection microscope.
Cell cycle and apoptosis assays. For cell cycle and apoptosis assays, both adherent and nonadherent cells were harvested, pooled, and fixed with 1% paraformaldehyde and 70% ethanol. For the cell cycle analysis, we stained cells with 50 μg/mL propidium iodide and determined the percentage of cells in specific cell cycle phases (G1, S, and G2-M) by using a flow cytometer equipped with a 488 nm argon laser (Epics Profile II; Beckman Coulter, Miami, FL). Approximately 1 × 104 cells were evaluated for each sample. Apoptosis was assessed with a flow cytometry–based terminal deoxyribonucleotide transferase–mediated nick-end labeling (TUNEL) assay processed with an APO-bromodeoxyuridine (APO-BrdUrd) staining kit (Phoenix Flow Systems, San Diego, CA); this assay was modified as previously described ( 25). Cells treated with DMSO were used as a negative control, and for a positive control, we used the HL-60 leukemic cells treated with camptothecin provided with the kit.
Establishment of resistant cell line. The H460 cell cultures were continuously exposed to erlotinib (10 μmol/L) in routine culture medium that was replaced every day for 5 months. Initially, H460 cell numbers were substantially reduced, and for the next 2 months, the surviving cells were passaged approximately every 10 days with a seeding ratio of 1:2. Cell proliferation slowly increased to allow a passage every 7 days with a seeding ratio of 1:4 over the next 2 months. A stable growth rate was reached after a total of 5 months with routine maintenance of the H460/TKI-R cells involving passage every 4 days with a seeding ratio of 1:8 of the confluent cell number.
Subcellular fractionation. The following procedures were done at 4°C. Cells were scraped into PBS [10 mmol/L sodium phosphate (pH 7.4) and 150 mmol/L NaCl] and then collected by centrifugation. Cell pellets were resuspended with 1 mL of hypotonic buffer [10 mmol/L Tris-HCl (pH 7.5), 1 mmol/L MgCl2, 50 μg/mL leupeptin, 1 mmol/L phenylmethylsulfonyl fluoride, and 1 mmol/L Na3VO4]; 10 minutes later, the cells were transferred to a Dounce homogenizer and further disrupted by 25 strokes with a tight-fitting pestle. The homogenate was adjusted to the indicated NaCl concentration from a 5 mol/L stock solution, and nuclei were removed by centrifugation at 1,700 × g for 5 minutes. The postnuclear supernatant was centrifuged again at 10,000 rpm for 20 minutes to remove the mitochondrial fraction; the postmitochondrial supernatant was centrifuged at 45,000 × g for 60 minutes. The supernatant fraction, representing the cytosolic fraction, was adjusted to 1% NP40 from a 10% stock solution. The pellet, representing the plasma membrane fraction, was gently rinsed with 1 mL PBS and then resuspended in 1 mL hypotonic buffer containing 1% NP40.
Immunoblotting and coimmunoprecipitation. NSCLC cells (1 × 106 cells/100 mm3 dish) were either left uninfected or infected with Ad-EV (50 pfu/cell) or Ad-survivin (50 pfu/cell) and then left untreated or treated with various concentrations of erlotinib (0.1-10.0 μmol/L), AG1024 (5.0 μmol/L), LY294002 (10.0 μmol/L), PD98059 (10.0 μmol/L), rapamycin (1.0 μmol/L), or their combinations in growth medium that was changed daily. When growth factor stimulation was done, cells were cultured in serum-free medium for 1 day and then incubated in EGF (50 ng/mL) or IGF (50 ng/mL) for 15 minutes. For the small interfering RNA (siRNA) transfection, H460 cells in the logarithmic growth phase in six-well plates (5 × 105 cells per well) were transfected with 10 μL of 20 μmol/L survivin siRNA or control scrambled siRNA (Dharmacon Research, Lafayette, CO) using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA), according to the protocol of the manufacturer. After 24 hours of incubation in growth medium, erlotinib was added, and the cells were harvested after 3 days of incubation. Immunoprecipitations were done using 3 mg protein from the total cell lysates and 1 μg mouse monoclonal anti-EGFR antibody, mouse monoclonal anti-IGF-IR antibody (Oncogene Sciences, Uniondale, NY), or healthy preimmune serum anti-mouse for the negative control and by incubating overnight at 4°C. The immunocomplexes were precipitated with protein-G agarose (Pharmacia-LKB Biotechnology, Piscataway, NJ). The immunoprecipitates were resolved on 6% SDS-PAGE gels, followed by Western blotting as described elsewhere ( 25).
Metabolic labeling. Metabolic labeling was done with H460 and TKI-R cells (5 × 105 in six-well plates). Cells were washed in PBS and incubated in RPMI medium without methionine and cysteine (Sigma, St. Louis, MO) for 2 hours. Next, the medium was replaced with fresh medium containing methionine and cysteine, to final concentrations of 150 mg/L, and the cells were labeled with trans-35S (0.5 mCi; ICN, MP Biomedicals, Irvine, CA). The cells were then treated with 0.1% DMSO or erlotinib (10 μmol/L) for 1, 3, 6, 12, and 24 hours. At harvesting time, the cells were washed in ice-cold PBS and lysed in radioimmunoprecipitation assay buffer. Lysates containing equal amounts of protein (100 μg) were immunoprecipitated using 1 μg of antibody to detect EGFR or 1 μg of antibody to detect survivin (both from Santa Cruz Biotechnology) and 30 μL of 50% slurry of protein G agarose beads (Pharmacia-LKB Biotechnology, Piscataway, NJ). The immunoprecipitates were washed five times with lysis buffer, separated by SDS-PAGE, and analyzed autofluorographically. Cell extracts were also subjected to Western blot analysis for β-actin to ensure that equal amounts of protein had been used. Two independent experiments were done with similar results; representative results of one experiment are presented.
Northern blot analysis. H460 cells (1 × 106 in 10 mm3 plates) were treated with erlotinib (10 μmol/L) for different times (0, 24, 48, and 72 hours). The total cellular RNA was isolated by the application of TRIzol. For the Northern blotting, 10 μg of the total cellular RNA prepared from each sample was subjected to electrophoresis on a 1% agarose gel containing 2% formaldehyde and then stained with ethidium bromide, photographed, transferred to a Z probe membrane (Bio-Rad Laboratories), and hybridized to an [α-32P]dCTP-labeled EGFR cDNA probe.
In vivo model. The effect of the combination of erlotinib and Ad-dnIGF-IR on established s.c. tumor nodules was studied in athymic nude mice (Harlan Sprague-Dawley) in a defined pathogen-free environment. Briefly, 6-week-old female nude mice were irradiated with 350 rad from a cesium-137 source and then were injected s.c. with 1 × 107 H1299 cells in 100 μL of growth medium at a single dorsal site. The mice were randomly assigned to one of four treatment groups, with each group containing eight mice. Group 1 (control mice) received 1× PBS and Ad-EV, group 2 received erlotinib and Ad-EV, group 3 received 1× PBS and Ad-dnIGF-IR, and group 4 received erlotinib and Ad-dnIGF-IR. Tumor growth was quantified by measuring the tumors in three dimensions with calipers for a total of 35 days. After the tumor volumes reached ∼75 mm3 (considered day 0), the mice were treated with p.o. administered erlotinib (100 mg/kg of body weight) twice a day. We chose this dosage of erlotinib because it had had no notable effect on H1299 tumor growth in preliminary experiments (data not shown). On day 23, when tumor volumes reached ∼125 mm3, each mouse was given a single intratumoral injection of 2 × 109 particles of Ad-dnIGF-IR or Ad-EV in 100 μL of PBS. Mice with necrotic tumors or tumors ≥1.5 cm in diameter were euthanized immediately. The results were expressed as the mean tumor volume (n = 5) with 95% confidence intervals (95% CI). On day 35, all mice were sacrificed and tumor tissues were collected from the xenografts to determine whether the combination of erlotinib and Ad-dnIGF-IR induced apoptosis in vivo by Western blot and immunohistochemical analyses as previously described ( 23).
Statistical analyses. The data acquired from the MTT assay were analyzed using Student's t test. All means and 95% CIs from eight samples were calculated using Microsoft Excel software (version 5.0; Microsoft Corporation, Seattle, WA). Cell survival comparisons among groups and statistical significance of differences in tumor growth in the combination treatment group and in the single-agent treatment groups were analyzed by ANOVA for 2 × 2 factorial design. All means from triplicate to eight samples and 95% CIs were calculated using SAS software (release 8.02; SAS Institute, Cary, NC). In all statistical analyses, two-sided P values of <0.05 were considered statistically significant.
Differential apoptotic responses of NSCLC cells after treatment with erlotinib. To test the effects of EGFR TKIs on NSCLC cell proliferation, a subset of NSCLC cell lines (H460, H1299, H661, H596, H226B, H226Br, A549, H322, and H358) was treated with erlotinib in regular growth medium containing 10% FBS or in serum-free medium containing EGF. A MTT assay revealed that erlotinib has different levels of antiproliferative activities, depending on cell lines ( Fig. 1A ): Compared with the other NSCLC cell lines, H322 and H358 cells were more sensitive to the erlotinib treatment (P < 0.001). Approximately 1 μmol/L erlotinib significantly inhibited proliferation of H322 and H358 cells after 72 hours of treatment. In contrast, the drug concentrations required to inhibit cell growth by 50% for H460, H1299, H661, H596, H226B, H226Br, and A549 cells were 10 to 20 times higher than those needed to inhibit H322 and H358 cells. Consistent with the results from the MTT assay, erlotinib only slightly affected the anchorage-dependent and anchorage-independent colony-forming abilities of H460 and H1299 cells (data not shown) at concentrations <1 μmol/L, a concentration that significantly inhibited those abilities of H322 (P < 0.001; Fig. 1B and C; P < 0.001) and H358 (data not shown) cells.
We next asked whether the ability of 1 μmol/L erlotinib to inhibit H322 and H358 cell proliferation was due to decreased cell cycle progression and/or increased apoptosis. Flow cytometric analyses of propidium iodide–stained H460, H1299, and H322 cells revealed that treatment with 1 μmol/L erlotinib for 3 days resulted in no marked change in the cell cycle distribution (data not shown). However, fluorescence-activated cell sorter analysis followed by terminal deoxynucleotidyl transferase–mediated dUTP-biotin nick-end labeling (TUNEL) staining revealed induction of apoptosis in 30.4% of the H322 cells treated with 1 μmol/L erlotinib. In contrast, treatment with up to 10 μmol/L erlotinib did not detectably increase the apoptotic population of H1299 and H460 cells ( Fig. 1D). In agreement with these findings, the protein levels of the active form of caspase-3 (Ac-caspase-3) and the cleaved form of poly(ADP-ribose) polymerase (PARP; cl-PARP, the 89 kDa fragment) increased in the H322 cells treated with >0.1 μmol/L erlotinib but not in the H460 cells, even after treatment with 10 μmol/L erlotinib ( Fig. 1E). H358 cells also responded with apoptosis to similar concentrations of erlotinib (data not shown). Together, these findings indicated the presence of mechanisms that affect the response of NSCLC cells to erlotinib-mediated apoptosis.
Role of IGF-IR signaling pathways in the development of resistance to erlotinib treatment in NSCLC cells. To investigate the mechanisms involved in the sensitivity of NSCLC cells to erlotinib, we first tested whether erlotinib successfully blocks activation of EGFR and its downstream mediators in NSCLC cell lines. Figure 2A shows that 0.1 to 1.0 μmol/L erlotinib suppressed the levels of phosphorylated EGFR (pEGFR), phosphorylated Akt (pAkt), and phosphorylated p44/42 MAPK (pp44/42 MAPK) in H460 and H1299 cells (cell lines that are weakly sensitive to erlotinib) and H322 and H358 cells (cell lines that are very sensitive to the drug). However, comparable induction of pAkt and p44/42 MAPK was evident in H460 and H1299 cells after treatment with >5.0 μmol/L erlotinib, doses that induce apoptosis in most of the H322 and H358 cells. pAkt and p44/42 MAPK are located in the nodal points of growth factor–mediated cell survival signaling, and the IGF-IR pathway can modulate the action of the erbB family blocking agents in various cancer cells ( 25– 27). Hence, we tested whether IGF-IR was involved in increases in pAkt and p44/42 MAPK. Indeed, erlotinib concentrations >5.0 μmol/L induced phosphorylated IGF-IR (pIGF-IR) in H460 and H1299 cells but not in H322 and H358 cells.
We then studied the influence of IGF-IR signaling pathways on the response of NSCLC cells to erlotinib. H460 cells exhibited significantly decreased proliferation and anchorage-dependent and anchorage-independent colony-forming abilities when the cells were treated with erlotinib in serum-free medium compared with when they were treated in the presence of IGF-I. H322 cells, however, showed statistically significant sensitivity to erlotinib in all conditions ( Fig. 2B), suggesting that the induced activation of the IGF-IR signaling pathway allows NSCLC cells to survive and proliferate when the EGFR pathway is blocked by erlotinib treatment.
To test our hypothesis, we compare the effects of erlotinib, either single or in combination with AG1024, an IGF-IR TKI, on the proliferation, clonogenic survival ability, and apoptosis of H460 cells. AG1024 has shown significantly lower affinity for the insulin receptor than for the IGF-IR ( 26). Combined treatment with erlotinib and AG1024 synergistically enhanced the antiproliferative effects of erlotinib on H460 cells compared with single treatment with each drug when cultured in complete (FBS) or serum-free medium in the absence or presence of IGF (P < 0.001; Fig. 2C; Supplementary Table S1). Erlotinib also showed significantly enhanced antiproliferative properties in H460 cells infected with an Ad-dnIGF-IR compared with the control cells infected with Ad-EV (P < 0.001; Fig. 2D). Moreover, combined treatment with erlotinib and AG1024 ( Fig. 2E) or Ad-dnIGF-IR ( Fig. 2F) significantly suppressed the anchorage-dependent, colony-forming ability of H460 cells (P < 0.001; Supplementary Table S2). Furthermore, TUNEL staining and flow cytometric analysis revealed that ∼1% of control H460 cells, 1.2% of erlotinib-treated cells, and 26% (95% CI, 14.3-37.6%, P < 0.05) of AG1024-treated cells underwent apoptosis. In contrast, combined treatment with both erlotinib and AG1024 significantly enhanced TUNEL staining (77.1%; 95% CI, 64.4-89.8%; P < 0.001; Fig. 2G) and induced cleavage of the 113-kDa PARP to the 89-kDa fragment in parallel with the concomitant decreases in the levels of pIGF-IR, pAkt, and p44/42 MAPK ( Fig. 2H). These findings suggest that the IGF-IR pathway provides an alternative proliferation and/or survival mechanism for NSCLC cancer cells in which EGFR is blocked by erlotinib.
Evidence of increased heterodimerization and membrane localization of IGF-IR and EGFR in erlotinib-treated H460 cells. We investigated the mechanism underlying NSCLC cell resistance to erlotinib using in vitro model of the H460 cell line (H460/TKI-R) that had been continuously treated with erlotinib. Treatment with >10 μmol/L erlotinib decreased the number of H460 cells, but proliferation of the remaining cells gradually increased after 3 months of erlotinib treatment. Compared with the parent H460 cells, the H460 cells treated with 10 μmol/L erlotinib for 5 months (H460/TKI-R) had higher levels of pIGF-IR, pAkt, and pp44/42MAPK, with no detectable differences in the expression of IGF-IR, Akt, and p44/42MAPK ( Fig. 3A ). The dose-response curves in Fig. 3B revealed no detectable cytotoxicity from erlotinib treatment of the H460/TKI-R cells up to a concentration of 20 μmol/L, whereas inhibition of IGF-IR activation by the AG1024 treatment induced greater antiproliferative effects on the H460/TKI-R cells than it did on the H460 cells. Moreover, compared with a single agent, combined treatment with erlotinib and AG1024 significantly suppressed the anchorage-dependent, colony-forming ability of H460/TKI-R as well as H460 cells (P < 0.001; Fig. 3C). The effects of combined treatment with erlotinib and AG1024 on colony formation was greater in H460/TKI-R cells than in H460 cells, indicating the dependence of H460/TKI-R cells on the IGF-IR signaling pathway for maintaining cell proliferation and tumorigenic potential.
We investigated the mechanism of erlotinib-mediated activation of IGF-IR in H460 cells. One mechanism by which the growth factor receptor is activated in tumor cells is by receptor dimerization; another is by overwhelming negative regulatory mechanisms that suppress receptor activation. Recent studies have revealed the interaction between the ErbB receptor families and IGF-IR in several tumor models, including breast cancer and oral cancer cell lines ( 19, 28– 32). We, therefore, tested whether EGFR interacts with IGF-IR in H460/TKI-R cells by performing immunoprecipitation. EGFR immunoprecipitates from H460/TKI-R cells showed greater IGF-IR binding compared with that from the parental H460 cells ( Fig. 3D). Control immunoprecipitates using preimmune serum exhibited no immunoreactive band. The interaction between EGFR and IGF-IR was observed as early as 30 minutes after the erlotinib treatment ( Fig. 3E). In contrast, no detectable change was observed in the levels of EGFR-EGFR or EGFR-ErbB2 interaction in H460 cells treated with erlotinib for 3 days ( Fig. 3F, top). Similarly, IGF-IR immunoprecipitates from erlotinib-treated H460 cells showed greater levels of EGFR binding than untreated cells did ( Fig. 3F, bottom), whereas no detectable binding was observed when the IGF-IR immunoprecipitates were immunoblotted to ErbB2 or ErbB3. Increased EGFR/IGF-IR heterodimerization was also observed in H1299 cells treated with 10 μmol/L erlotinib ( Fig. 3G). In contrast, the EGFR and control (preimmune serum) immunoprecipitates from untreated or erlotinib-treated H322 cells exhibited no immunoreactive band. These results suggested that erlotinib induces physical contact between EGFR and IGF-IR, which is accumulative.
Erlotinib treatment–induced expression of survivin protein protects NSCLC cells from apoptosis. We next attempted to find evidence connecting erlotinib-induced activation of the IGF-IR with survival of NSCLC cells. Because the inhibitor of apoptosis proteins, including survivin and XIAP, decrease the sensitivity of tumor cells to chemotherapeutic drugs, thereby conferring resistance to apoptosis ( 21, 33), we tested the effects of erlotinib on the expression of these proteins in a subset of NSCLC cells with weak or great sensitivity to erlotinib. We found that expression of survivin but not XIAP markedly increased in H1299, H460, H661, A549, A596, and 226B cells ( Fig. 4A, top ) during the time IGF-IR was phosphorylated by the erlotinib treatment ( Fig. 4A, bottom). In contrast, H358 and H322 cells showed no detectable changes in the protein levels of survivin, XIAP, and pIGF-IR during the time EGFR was inactivated by erlotinib treatment. Erlotinib induced survivin expression in a time- and dose-dependent manner ( Fig. 4B). Interestingly, a similar but less pronounced increase in EGFR expression was observed in H1299, H460, H661, A549, H596 H226Br, and H226B, H460/TKI-R cells but not in H358 and H322 cells. Increases in the survivin and EGFR expression were also observed in H460/TKI-R cells ( Fig. 4C).
We then tested the response of H460 cells, in which survivin expression was abolished by siRNA transfection. Western blot analysis revealed an obvious increase in PARP cleavage in the 460 cells by the treatment with erlotinib ( Fig. 4D). Among H322 cells, in which survivin overexpression was induced by the infection with Ad-survivin, the PARP cleavage was substantially reduced after the erlotinib treatment ( Fig. 4E). These results indicated that increased expression of survivin protein protected NSCLC cells from the erlotinib-induced apoptosis.
mTOR pathway induces de novo protein synthesis of EGFR and survivin and protects NSCLC cells from apoptosis. We investigated the mechanisms of erlotinib-mediated increase in survivin and EGFR protein expression. According to Northern blot analysis, exposure of H460 cells to erlotinib resulted in no change in the mRNA levels of survivin ( Fig. 5A ) and EGFR (data not shown). We then determined the effects of erlotinib on the rates of survivin and EGFR protein synthesis. Metabolic labeling of the H460 cells with [35S]Met-Cys revealed that the rate of [35S]labeled survivin ( Fig. 5B) and EGFR (data not shown) synthesis was remarkably greater in the erlotinib-treated H460 and H460/TKI-R cells than in the untreated parental H460 cells. We then determined whether mTOR is involved in the erlotinib-induced protein synthesis of survivin and EGFR by determining the levels of phosphorylated 4E-BP1 and p70s6k, downstream mediators of mTOR ( 27, 34), in the H460 cells treated with erlotinib alone or in combination with rapamycin, an mTOR inhibitor, or AG1024 for 3 days. Western blot analysis revealed that erlotinib up-regulated the protein levels of p4E-BP1, pp70s6k in association with increases in survivin and EGFR expression, all of which were suppressed by treatment with AG1024 or rapamycin ( Fig. 5C). Moreover, the combined treatment with AG1024, LY294002 (PI3K inhibitor), PD98059 (MEK inhibitor), or rapamycin reduced the levels of membranous EGFR expression ( Fig. 5D) and of the EGFR and IGF-IR heterodimer ( Fig. 5E) induced by the 3 days treatment of erlotinib. These findings suggested that the increases in the levels of EGFR/IGF-IR heterodimer on cell membrane and protein expressions of survivin and EGFR were mediated at least in part through translation-dependent events mediated by IGF-IR signaling pathways. We then tested whether inhibitors of the IGF-IR and mTOR pathways sensitize the H460 cells to the erlotinib treatment. H460 cells treated with erlotinib and AG1024 ( Fig. 2G and H) or rapamycin ( Fig. 5F) showed an increase in the PARP cleavage, suggesting that suppression of IGF-IR and mTOR pathways could restore the apoptotic activities of erlotinib in NSCLC cells.
Antitumor efficacy of dual targeting of EGFR and IGF-IR signaling pathways in vivo. To determine whether the inhibition of IGF-IR signaling can enhance the antitumor activities of erlotinib in vivo, we tested the effects of erlotinib, Ad-dnIGF-IR, and their combination on the growth of H1299 NSCLC xenograft tumors established in athymic nude mice. The mice treated with erlotinib plus Ad-dnIGF-IR showed synergistically reduced tumor growth compared with the control mice or the mice treated with erlotinib or Ad-dnIGF-IR alone ( Fig. 6A ; Supplementary Table S3). At the end of the study, the mean tumor volume in combined treatment group was 23% (P < 0.001) of the mean volume in the control group. Thus, the combination of erlotinib and Ad-dnIGF-IR enhanced the antitumor effects on the growth of NSCLC cells in vivo.
We then determined the effects of erlotinib, Ad-dnIGF-IR, and their combination on the activation of the IGF-IR and EGFR, the expression of survivin and EGFR, and the induction of apoptosis in vivo. According to Western blot analysis of total protein extracts harvested from the H1299 xenograft tumor tissues, the levels of pEGFR were decreased by erlotinib. In addition, erlotinib treatment induced marked increases in the levels of pIGF-IR, EGFR, and survivin, all of which were effectively blocked by Ad-dnIGF-IR ( Fig. 6B). Combined treatment with erlotinib and Ad-dnIGF-IR also increased the levels of Ac-caspase-3, which is confirmed by the immunohistochemical staining of the H1299 xenograft tumor tissues ( Fig. 6C). Together, these findings suggested that the combined treatment with erlotinib and Ad-dnIGF-IR exert enhanced in vivo antitumor activities by decreased expression of survivin and EGFR and induction of apoptosis.
Several preclinical and clinical discoveries have associated EGFR TKIs with antitumor activities. However, the limited response rates of patients to EGFR TKIs, even in patients with high levels of EGFR ( 15, 16, 35), have been raising questions about the mechanisms leading to the EGFR TKI resistance. Although somatic mutations of the EGFR ATP binding site have been associated with the response to the EGFR TKIs in some cases ( 17, 18), increasing number of evidence have suggested that the presence of other pathways that mediate the resistance of cancer cells to EGFR TKI therapy ( 36, 37). In this article, we have shown, to our knowledge for the first time, that erlotinib induces survival of NSCLC cells by inducing heterodimerization of EGFR/IGF-IR, activating IGF-IR pathway and its downstream mediators Akt and p44/42 MAPK, and thus stimulating mTOR-mediated protein synthesis of survivin that plays a crucial role in the blocking apoptosis. We showed here that the blockade of the IGF-IR to mTOR signaling pathway was sufficient to suppress de novo survivin protein synthesis and to restore apoptotic activities of erlotinib in NSCLC cells in vitro and in vivo. Our data present clear evidence that crosstalk between the IGF-IR and the EGFR signaling pathways and consequential survivin expression are involved in the NSCLC cell resistance to erlotinib. The erbB2/Her2/neu, a known preferred coreceptor for the EGFR, has been suggested to play a role in inducing NSCLC cell survival against EGFR TKIs ( 28). However, we found that the erbB2/Her2/neu was inactivated by the EGFR TKIs in NSCLC cells (data not shown), consistent with previous reports ( 38). In addition, an interaction between IGF-IR and erbB2 was undetectable, regardless of erlotinib treatment. Therefore, erbB2/Her2/neu is not likely to have a role in inducing EGFR TKI resistance in NSCLC cells.
We investigated the detailed mechanism that mediates IGF-IR activation by erlotinib and the consequent development of drug resistance. Given that gefitinib-resistant DU145/TKI-R prostate cells have shown considerably higher basal levels of IGF-II mRNA than wild-type cells ( 19), erlotinib might have increased the expression of IGF and conferred resistance to the drug to NSCLC cells. However, in gefitinib-resistant breast cancer cell lines, the IGF II mRNA level did not differ from that of the original clone ( 19), indicating that autocrine/paracrine production is not entirely responsible for the sensitivity of the cell to EGFR TKIs. Perhaps our most striking finding was that in H460 and H1299 cells, erlotinib induced heterodimerization between IGF-IR and EGFR. The interaction between EGFR and IGF-IR also has been observed in cancer cells ( 29– 31). Given the considerable similarity between EGFR and IGF-IR in the sequence of their extracellular domain ( 39) and their reliance on EGF and IGF to achieve cell cycle progression and survival, it is plausible that erlotinib-mediated IGF-IR/EGFR heterodimerization can stimulate intracellular signaling components in a distinct pattern and allow NSCLC cells to resist the drug. Previous studies have shown the ability of erlotinib to induce EGFR mRNA and protein expression in the erlotinib-resistant biliary tract cancer cell line HuCCT1 but not in the susceptible A431 epidermoid cell line ( 40). In current study, we found that erlotinib induces mTOR-mediated de novo protein synthesis of survivin and EGFR with no detectable change in their mRNA levels, indicating diverse responses of different cancer cells to erlotinib. Our results may explain some apparently paradoxical findings in several clinical trials ( 41), in which up-regulation of pEGFR was observed after treatment of breast cancer patients with erlotinib. In another report, modifications of EGFR serum values during treatment of NSCLC seemed to reflect gefitinib activity; responding patients showed decreased serum levels of EGFR relative to those in patients with refractory disease ( 42). Our data also show the critical role of the induced survivin proteins in the development of resistance to erlotinib; (a) survivin expression was induced in NSCLC cell lines with weak sensitivity to the erlotinib treatment; (b) overexpression of survivin protected the sensitive NSCLC cells from erlotinib-induced apoptosis; and (c) knockdown survivin expression by siRNA provoked apoptosis in NSCLC cells with weak erlotinib sensitivity. mTOR has been known to regulate the translation of subsets of mRNA, many of which encode for proteins involved with driving cell growth, proliferation, and angiogenesis ( 43). Therefore, resistance and sensitivity to erlotinib in NSCLC may be determined at least in part by the ability of the cancer cells to stimulate mTOR-mediated synthesis of specific proteins that have key roles in cell proliferation and/or survival and thus to adapt to a stressful environment.
In summary, our findings provide definitive in vitro and in vivo evidence that erlotinib induces heterodimerization of the EGFR/IGF-IR and stimulates IGF-IR and downstream pathways, including PI3K/Akt, MEK/extracellular signal-regulated kinase (ERK), resulting in the mTOR-mediated increases in EGFR and survivin proteins. On the basis of these findings, it is plausible to suggest that increased EGFR proteins further enhance the interplay between the EGFR and IGFR on the cell membrane, resulting in a further amplification of IGF-IR to mTOR signaling. In addition, the increased survivin proteins seem to provide survival potential to the NSCLC cells against the erlotinib treatment ( Fig. 6D).
Our findings have direct effect to the treatment of NSCLC with erlotinib. The data showing no detectable interaction between EGFR/IGF-IR in NSCLC cell lines with low levels of IGF-IR expression suggests that the expression of IGF-IR is an important factor for the EGFR and IGF-IR complex and erlotinib sensitivity. Therefore, IGF-IR expression may serve as a predictor for erlotinib resistance in NSCLC. IGF-IR signaling pathway also plays a key role in the resistance to several therapeutic drugs ( 12, 27, 44– 47). Overexpression of IGF-IR has been observed in various human cancers, including lung cancer ( 48), and is associated with a poor prognosis ( 49). Importantly, our unpublished data showed that clinical samples from NSCLC patients revealed that the majority of EGFR-overexpressing samples showed correlative increases in IGF-IR protein levels compared with their paired normal counterparts from the same patients. With this prospect, IGF-IR–targeting combination treatment may be required when erlotinib is considered as a therapeutic agent for NSCLC patients. Alternatively, mTOR inhibitors could confer benefit to erlotinib-resistant patients. In light of this notion, we have shown that combined treatment with erlotinib and inhibitors of IGF-IR or mTOR suppressed survivin and EGFR expression, decreased proliferation of NSCLC cells, and induced apoptosis in NSCLC cells in vitro and in vivo. Further studies are warranted to validate whether the erlotinib combined with inhibitors of IGF-IR or mTOR inhibitors could enhance the objective response and survival rates in NSCLC patients.
Grant support: NIH grant R01 CA100816 (H-Y. Lee), American Cancer Society grant RSG-04-082-01-TBE 01 (H-Y. Lee), Department of Defense grants W81XWH-04-1-0142-01-VITAL and W81XWH-06-1-0303-BATTLE, and partly by The National Foundation for Cancer Research Fellow grants (W.K. Hong).
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.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
W.K. Hong is an American Cancer Society Clinical Research Professor. H-Y. Lee is a faculty member at the Graduate School of Biomedical Sciences.
- Received May 10, 2006.
- Revision received July 6, 2006.
- Accepted August 22, 2006.
- ©2006 American Association for Cancer Research.