Abstract
Purpose: Epidermal growth factor receptor (EGFR) kinase inhibitors induce dramatic clinical responses in a subset of non-small cell lung cancer (NSCLC) patients with advanced disease, and such responses are correlated with the presence of somatic activating mutations within the EGFR kinase domain. Consequently, one of these inhibitors, erlotinib, has been Food and Drug Administration-approved as a second- or third-line treatment for chemotherapy-refractory advanced NSCLC. However, responses are typically relatively short-lived due to acquired drug resistance, prompting studies to determine whether first-line treatment with EGFR inhibitors could provide greater clinical benefit. NSCLC-derived cell lines have provided a powerful system for modeling EGFR mutation-correlated sensitivity to EGFR inhibitors and for modeling mechanisms of acquired drug resistance that are observed clinically.
Experimental Design: In a cell culture model of an erlotinib-sensitive EGFR-mutant NSCLC cell line, we tested the hypothesis that prior exposure to platinum agents, a standard component of NSCLC chemotherapy treatment, affects the subsequent response to erlotinib.
Results: Indeed, NSCLC cells initially selected for growth in cisplatin exhibit 5-fold reduced sensitivity to erlotinib, even after propagating the cisplatin-treated cells in the absence of cisplatin for several months. This lingering effect of cisplatin exposure appears to reflect changes in PTEN tumor suppressor activity and persistent EGFR-independent signaling through the phosphatidylinositol 3-kinase/AKT survival pathway.
Conclusions: These preclinical findings suggest that first-line chemotherapy treatment of EGFR-mutant NSCLCs may reduce the benefit of subsequent treatment with EGFR kinase inhibitors and should prompt further clinical investigation of these inhibitors as a first-line therapy in NSCLC.
- EGFR
- erlotinib
- resistance
Translational Relevance
We have tested the hypothesis that treatment of EGFR-mutant NSCLC with platinum-based chemotherapy drugs could influence the subsequent response to treatment with a selective kinase inhibitor. This was done in NSCLC cell culture models, where we showed that whereas the IC50 of erlotinib sensitivity in untreated versus cisplatin-treated cells is indistinguishable, longer-term clonogenic survival in the presence of erlotinib is substantially increased in the cisplatin-treated cells, associated with persistent PI3K signaling in the cisplatin-treated cells. This has important potential clinical implications when considering that erlotinib is approved for use in chemotherapy-refractory NSCLC, which often involves treatment with platinum agents. The findings suggest that NSCLC patients treated with first-line EGFR kinase inhibitors may exhibit longer times to disease progression than second- or later-line erlotinib-treated patients. This is particularly relevant in the context of several recent clinical studies that have shown impressive clinical response rates in EGFR-mutant NSCLC patients treated with first-line EGFR kinase inhibitors.
Non–small cell lung cancer (NSCLC) is the leading cause of cancer death worldwide. The prognosis for most patients with advanced NSCLC remains poor despite significant advances in medical oncology. Such patients typically experience modest clinical benefit from standard platinum-based chemotherapy treatments, associated with a limited increase in overall survival (1). The epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) erlotinib (Tarceva) yields a modest increase in survival when administered to unselected NSCLC patients following chemotherapy and was hence approved for this indication by the Food and Drug Administration in 2004 (2). However, recent studies have shown that a subset (10-20%) of NSCLC patients treated with EGFR TKIs experience striking clinical responses, which, in some cases, lead to durable remissions (3–5). Significantly, those responses are well correlated with the presence of a class of somatic activating mutations within the EGFR kinase domain (6–8), paving the way for recent genotype-based trials aimed at improving the overall response rate by preselecting patients that are more likely to respond to these agents in the first-line setting (9–11). Although none of the genotype-directed studies reported thus far have included a comparison arm in their design, initial results are promising, with response rates and durations being 2- to 3-fold better than those typically seen with standard chemotherapy (9–11).
Although such clinical studies are encouraging, and the concept of using a first-line treatment regimen that is targeted to a specific genetic lesion and is less toxic than conventional chemotherapy is appealing, there are significant considerations that need to be addressed before such an approach could be considered standard. Primarily, this strategy has not yet been compared with traditional chemotherapy in a randomized trial within a genotype-selected population, and consequently, its relative benefit has not yet been proven. Moreover, some have suggested that EGFR mutations are prognostic, not predictive factors for survival in the setting of EGFR-directed therapy, and therefore not optimal for therapeutic decision-making (12, 13).
To further investigate the potential benefit of first-line EGFR TKI therapy in NSCLC, we examined the effect of a platinum-based chemotherapy agent on subsequent sensitivity to EGFR kinase inhibitors using a cell culture-based preclinical model. NSCLC-derived cell lines have proven a reliable model of clinical response to EGFR kinase inhibitors. Thus, most tumor cell lines harboring activating EGFR kinase domain mutations exhibit substantially increased sensitivity to gefitinib and erlotinib (14), and continuous exposure of these cells to kinase inhibitors eventually yields drug-resistant clones that have acquired resistance through mechanisms that have been observed clinically in EGFR TKI-treated NSCLC patients (15, 16). To determine whether standard chemotherapy treatment of EGFR mutation-positive NSCLC can affect the subsequent responsiveness to second-line treatment, we examined an EGFR-mutant NSCLC-derived cell line. Exposing these cells to cisplatin substantially reduced their subsequent sensitivity to erlotinib via a mechanism that involves persistent activation of the phosphatidylinositol 3-kinase (PI3K)/AKT cell survival pathway. Moreover, the reduced sensitivity to erlotinib was observed even after cells had been maintained in cisplatin-free medium for several months. Our findings suggest that treatment with platinum-based agents may render EGFR-mutant NSCLC cells more resistant to EGFR kinase inhibitors than previously untreated cells. These findings provide further evidence to support a rationale for the use of EGFR kinase inhibitors as first-line therapy in NSCLC.
Materials and Methods
Cell line culture and pharmacologic inhibitors. PC9 and PC14 NSCLC-derived cells expressing the EGFR exon 19 deletion mutation (ΔE746-A750) were kindly provided by Dr. Kazuto Nisho (National Cancer Center Hospital, Tokyo) and maintained as described previously (17). Cisplatin-resistant PC9 cells were maintained in cisplatin containing RPMI 1640 with 10% fetal bovine serum, penicillin, and streptomycin (100 units/mL and 100 μg/mL, respectively). Erlotinib (from the MGH pharmacy) was resuspended in DMSO at a stock concentration of 10 mmol/L. Cisplatin was purchased from Calbiochem. LY 294002 was purchased from Cell Signaling Technology and resuspended in DMSO at a stock concentration of 50 mmol/L. All inhibitors were stored in small aliquots at −20°C.
Generation of cisplatin-resistant PC9 cells. PC9 cells at 70% confluency were exposed to 5 μmol/L cisplatin, equivalent to the empirically determined IC80 of cisplatin for PC9 cells in a 96-h survival assay. Fresh medium containing drug was added to the cells every 3 days until cells reached confluency (∼3 months). These cells were subjected to subsequent cell survival assays and found to be stably resistant to cisplatin when maintained in 5 μmol/L cisplatin. This pool of cells was used for further biochemical characterization and cell survival assays.
Antibody studies. The rabbit polyclonal antibodies directed against phospho-EGFR (Tyr1068) and EGFR were from Abcam and Santa Cruz Biotechnology, respectively. Rabbit polyclonal antibodies against phospho-p44/42 mitogen-activated protein kinase (Thr202/Tyr204), and phospho-AKT (Ser473) and antibodies directed against their nonphosphorylated counterparts were purchased from Cell Signaling Technology. The mouse monoclonal anti-c-K-Ras antibody was from Calbiochem. Secondary antibodies included horseradish peroxidase-conjugated anti-mouse and anti-rabbit antibodies and were also from Cell Signaling Technology.
Cell harvesting and protein analysis. Cells grown under the described conditions were lysed in a solution [1% NP-40, 20 mmol/L Tris-HCl (pH 7.5), 2 mmol/L EDTA, 137 mmol/L NaCl, 10% glycerol] containing the protease inhibitors aprotinin, leupeptin, and phenylmethylsulfonyl fluoride, and the phosphatase inhibitors NaF and Na3VO4. The lysate was cleared by centrifugation at 15,000 rpm for 20 min and the protein samples were quantified by BCA protein assay (Pierce). Total protein (15 μg) was resuspended in Laemmli sample buffer and the proteins were separated by electrophoresis on 10% SDS-polyacrylamide protein gels. Proteins were transferred to nitrocellulose membranes (0.45 μm Protran; Schleicher & Schuell) and the nonspecific protein binding sites were blocked by incubating filters in 5% nonfat dry milk resuspended in TBS-0.1% Tween. Filters were then incubated overnight with the appropriate primary antibody in TBS-0.1% Tween containing 5% bovine serum albumin. The next day, the filters were washed three times in TBS-0.1% Tween and incubated with the respective horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The filters were then washed three times in TBS-0.1% Tween and the specific protein bands were visualized by SuperSignal West Pico chemiluminescence (Pierce).
The in vivo Ras activation status was examined using the GST-Raf RBD pull-down assay. Cells grown under the described conditions were lysed in 500 μL MLB buffer consisting of 20 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 10 mmol/L MgCl2, 10% glycerol, 1% NP-40, and 0.25% sodium deoxycholate. The following inhibitors were added: 25 mmol/L NaF, 1 mmol/L sodium vanadate, 10 μg/mL aprotinin, and 10 μg/mL leupeptin. Lysates were transferred to prechilled 1.5 mL Eppendorf tubes, rocked at 4°C for 15 min, and cleared by centrifugation at 14,000 rpm for 5 min at 4°C. Each sample (50 μL) was normalized with BCA protein assay and used for loading control subsequently. A GST-Raf RBD-Sepharose bead slurry (5 μL; a stock concentration of 2 μg/μL) was added to the remainder of each lysate, and samples were incubated at 4°C on a rocker for 30 min. The beads were spun down at 4,000 rpm for 30 s at 4°C, the supernatant was aspirated, and beads were washed several times with MLB. Protein was extracted from beads in 10 μL of 2× SDS sample buffer. Samples were heated at ∼95°C for 5 min, resolved by 12% SDS-PAGE, transferred to polyvinylidene difluoride, subsequently processed as described above, and immunoblotted with K-Ras antibody.
Cell survival assays. Approximately 30,000 cells were plated in a well of a 12-well cluster dish. Twenty-four hours after plating, medium was removed from the wells and replaced with medium containing drugs. The experiments, unless otherwise stated, were done over 72 h and terminated by fixation of cells for 15 min at room temperature with 4% formaldehyde in PBS. Cells were then washed twice with PBS and stained with the fluorescent nucleic acid stain, Syto60 (700 nmol/L in PBS; excitation and emission wavelengths of 652 and 678 nm, respectively; Molecular Probes), for 15 min at room temperature. The dye was then removed and the cells were washed once with PBS. Quantitation of fluorescent signal intensity was carried out at 700 nm using an Odyssey Infrared Imager (Li-Cor Biosciences). Each experiment was done in quadruplicate and the results represent the average of the four values in comparison with untreated controls.
Clonogenic assays. To assess the effects of erlotinib on the respective cell lines, ∼105 cells corresponding to each cell line were plated in a 100 mm culture dish. The cell lines include PC9, cisplatin-resistant PC9, PTEN-infected cisplatin-resistant PC9, green fluorescent protein-infected cisplatin-resistant PC9, cisplatin-resistant PC9 (cultured in cisplatin-free medium), and PC14 cells. Erlotinib (2 μmol/L) was added the day after the cells were plated, and fresh medium with erlotinib was replaced every 3 days. These assays were done over 10 days, after which time the experiment was terminated by fixing the cells with 4% formaldehyde, and cells were stained with Syto60 for quantitation. Assays that examined the effects of LY294002 were done in a similar manner, except that LY294002 was added in addition to erlotinib during each drug and medium change. All experiments were done in triplicate.
Expression of PTEN in PC9 cells by lentiviral transduction. Recombinant green fluorescent protein-expressing (control) or PTEN-expressing lentiviral particles were generated in 293T cells by cotransfection of 1 μg pLenti-6V5-PTEN plus 0.9 μg pCMVΔ8.91 and 0.1 μg pMDG-VSV-G into two wells of a six-well plate. The day before infection, cisplatin-resistant PC9 cells were plated at a density of 1 × 105 per well of a six-well plate. The cells were then infected by adding 0.5 mL green fluorescent protein or PTEN-expressing lentiviruses per well in the presence of 8 μg/mL polybrene. Infection was enhanced by centrifugation of the cells at 1,200 × g for 1 h at 32°C. Twenty-four hours later, the medium was replaced with fresh complete growth medium. Lentiviral transduction was allowed to proceed for a further 48 h. Green fluorescent protein- or PTEN-expressing cells were selected by typsinizing the cells into 10 cm dishes and adding 10 μg/mL blasticidin in complete growth medium for 72 h. PTEN expression was verified by immunoblotting.
Results
EGFR-mutant NSCLC cells exhibit reduced sensitivity to erlotinib following prior treatment with cisplatin. To model the typical clinical experience of an EGFR mutation-positive NSCLC patient that involves first-line chemotherapy followed by second- or third-line EGFR TKI treatment, we used the PC9 human NSCLC-derived cell line. PC9 cells are derived from a previously untreated human NSCLC that harbors one of the recurrent in-frame deletions within EGFR exon 19 that gives rise to an activated kinase. These cells are exquisitely sensitive to EGFR TKIs, exhibiting an IC50 of ∼30 nmol/L in a 72-h cell viability assay (18). Thus, they appear to represent the subset of EGFR mutation-positive NSCLC patients that show a good clinical response to treatment with EGFR TKIs. We used this cell culture model to determine whether pretreatment with a platinum-based drug, a standard component of traditional first-line therapy for NSCLC, would affect the response to subsequent treatment with erlotinib. Treatment of PC9 cells with 5 μmol/L cisplatin for 3 days results in ∼80% cell death followed by the emergence of cells that are largely refractory to the cytotoxic actions of cisplatin and can be maintained in cisplatin indefinitely (Fig. 1A ). Thus, these cells appear to model a typical clinical response to cisplatin in EGFR mutation-positive NSCLC.
Sensitivity of PC9 and cisplatin-resistant PC9 cell lines to cisplatin and erlotinib. A, survival curves of PC9 and cisplatin-resistant PC9 (CR) cells following 72-h treatment with the indicated concentration of cisplatin. Points, percentage of surviving cells with respect to untreated cells. Average of four independently conducted experiments. B, survival curves of PC9 and PC9 cells following 72-h treatment with the indicated concentration of erlotinib. Points, percentage of surviving cells with respect to untreated cells. Average of four independently conducted experiments. Note that, at 2 μmol/L erlotinib (pharmacologic concentration), the curves corresponding to PC9 and PC9 (CR) cells are significantly different. C, clonogenic assays were done on PC9 and PC9 (CR) in the presence of 2 μmol/L erlotinib for 10 d. 1,000, 10,000, and 100,000 cells of each respective cell line were plated in triplicate, and the relative number of surviving cells was assessed by Syto60 staining. Average of three experiments. Bars, SD. D, representative cell culture plate from each cell line tested in C.
Next, we determined whether the cells that emerged following cisplatin exposure exhibit an altered sensitivity to erlotinib treatment relative to drug-naive PC9 cells. As shown in Fig. 1B, the IC50 of erlotinib sensitivity for PC9 cells and cisplatin-resistant PC9 derivatives is approximately equal (30 nmol/L) as measured in a 72-h viability assay. However, the IC90 value for erlotinib sensitivity was significantly increased in the cisplatin-resistant cells, suggesting the existence of cell heterogeneity within the cisplatin-resistant population with regard to erlotinib sensitivity. To explore this difference further, we did a similar comparison using a longer-term clonogenic survival assay. By plating varying numbers of PC9 or cisplatin-resistant PC9 cells (withdrawn from cisplatin) in the presence of 2 μmol/L erlotinib (the clinical concentration), we observed a 5-fold increase in the number of erlotinib-resistant clones that emerge from the cisplatin-resistant PC9 cells relative to the parental PC9 cells during 10 days of continuous erlotinib exposure (Fig. 1C and D). Together, these results suggest that pretreatment of EGFR-mutant NSCLC cells with cisplatin significantly reduces their sensitivity to subsequent erlotinib exposure.
EGFR-mutant NSCLC cells with primary cisplatin resistance show reduced erlotinib sensitivity. A significant fraction of advanced NSCLCs, including EGFR-mutant tumors, fail to respond to standard first-line chemotherapy, which frequently includes a platinum agent. We examined erlotinib sensitivity in the PC14 human NSCLC cell line, which harbors an EGFR exon 19 deletion and exhibits an IC80 for cisplatin sensitivity that is ∼10-fold higher than that of PC9 cells (Fig. 2A ). Thus, PC14 cells appear to model cisplatin-refractory NSCLC. As was seen with PC9 cells that had acquired cisplatin resistance, PC14 cells exhibit an IC50 for erlotinib that is similar to that of PC9 cells (Fig. 2B) but show substantially reduced erlotinib sensitivity as revealed in both the IC90 value (Fig. 2B) and in a longer-term clonogenic survival assay (Fig. 2C). These findings suggest that both primary and acquired cisplatin resistance in EGFR-mutant NSCLC cells are associated with reduced erlotinib sensitivity when assessed by clonogenic survival in the presence of drug.
Reduced erlotinib sensitivity in cisplatin-refractory EGFR-mutant NSCLC cells. A, survival curves of PC9 and PC14 cell lines treated with the indicated concentrations of cisplatin for 5 d. 50,000 cells were plated and treated with the indicated concentrations of cisplatin on the following day. Cells were re-fed with fresh medium containing the indicated concentrations of the drug every 2 d thereafter. Points, average value determined from four identically treated samples. Data are expressed as a percentage of surviving cells relative to untreated controls. Bars, SD. B, survival curves of PC9 and PC14 cell lines treated with the indicated concentrations of erlotinib for 5 d. 50,000 cells were plated and treated with the indicated concentrations of erlotinib on the following day. Cells were re-fed with fresh medium containing the indicated concentrations of the drug every 2 d thereafter. Points, average of four identically treated samples. Data are expressed as a percentage of surviving cells relative to untreated controls. Bars, SD. C, 1 × 105 PC9 or PC14 cells were plated in 10 cm dishes and treated with 2 μmol/L erlotinib for 19 d. Cells were re-fed every 3 d with medium containing drug. At the end of the treatment period, cells were fixed and stained with Giemsa. Note the substantially greater number of erlotinib-resistant clones that developed with the PC14 cell line.
Erlotinib fails to suppress AKT activation in cisplatin-resistant PC9 cells. To begin to address the molecular basis for the reduced erlotinib sensitivity seen in cisplatin-resistant NSCLC cells, we compared the signaling properties associated with mutationally activated EGFR in PC9 and cisplatin-resistant PC9 cells. We found that cisplatin-resistant cells exhibit significantly increased basal levels of EGFR phosphorylation, consistent with previous studies showing the activation of EGFR kinase activity by chemotherapy drugs (19). Similarly, these cells exhibit elevated levels of phospho-AKT and phospho-ERK, two key downstream effectors of EGFR-mediated cell survival. Interestingly, on treatment with erlotinib, cisplatin-resistant cells, like PC9 cells, rapidly experience suppressed phosphorylation of EGFR and ERK; however, they fail to show suppression of AKT phosphorylation, and phospho-AKT levels remain essentially unchanged even after exposure to 2 μmol/L erlotinib for 4 h (Fig. 3 ). These results suggest that in cisplatin-resistant PC9 cells, activation of AKT, a key survival protein, becomes uncoupled from EGFR kinase activity.
Persistent AKT activation in cisplatin-resistant PC9 cells. Cell lysates from PC9 and PC9 (CR) treated with increasing concentrations of erlotinib for 4 h were analyzed by SDS-PAGE followed by immunoblotting using antibodies directed against the indicated signal transduction proteins.
Cisplatin exposure has long-term consequences for subsequent erlotinib treatment. NSCLC patients who experience disease control with cisplatin typically undergo a period (a “drug holiday”) during which they are untreated, before initiating second- or third-line treatment with another agent, such as an EGFR TKI, at subsequent disease progression. Therefore, we wanted to determine whether withdrawal of cisplatin from cisplatin-resistant PC9 cells for several months would still affect subsequent sensitivity to erlotinib. Interestingly, after 4 months of culture in drug-free medium, the cisplatin-resistant cells retain their reduced sensitivity to erlotinib as measured in the clonogenic survival assay (Fig. 4A ). Furthermore, treatment of these cells with the chromatin-modifying agents azacytidine and trichostatin A failed to affect erlotinib sensitivity (Fig. 4A), suggesting that rapidly reversible epigenetic mechanisms, such as DNA methylation, are unlikely to account for the observed effects of cisplatin. These findings suggest that cisplatin treatment of erlotinib-sensitive NSCLC cells leads to reduced erlotinib sensitivity that is maintained even several months following cisplatin withdrawal.
Lingering effect of cisplatin exposure on erlotinib sensitivity. A, clonogenic assays were done in PC9, PC9 (CR), PC9 (CR maintained in cisplatin-free medium) and PC9 (CR treated with 1 μmol/L azacytidine and 30 nmol/L trichostatin A) cells. 100,000 cells corresponding to each cell line were plated and 2 μmol/L erlotinib was added the following day. Cells were maintained in erlotinib for the 10-d course of the experiment. Number of surviving cells as measured by Syto60 staining in the respective cell lines. Average of three independent experiments. Bars, SD. B, cell lysates from PC9, PC9 (CR), and PC9 (CR maintained in cisplatin-free medium) treated with increasing concentrations of erlotinib for 4 h were analyzed by SDS-PAGE followed by immunoblotting using antibodies directed against the indicated signal transduction proteins. C, 1,000 cells each of PC9, PC9 (CR), and PC9 (CR maintained in cisplatin-free medium) were plated on a 100 mm dish and grown in 10% fetal bovine serum. Cells were subjected to Syto60 staining after 1 wk to assess their respective proliferative rates. Note the increased staining seen in cisplatin-resistant cells following withdrawal of cisplatin.
Next, we examined the status of phospho-AKT in the cisplatin-resistant cells that had been removed from cisplatin for 4 months. In these cells, as was seen in the cisplatin-resistant cells immediately following cisplatin withdrawal, phospho-AKT levels are elevated, and significantly, they are not suppressed by erlotinib treatment (Fig. 4B). This observation suggests that persistent downstream activation of the AKT survival pathway may contribute to reduced erlotinib sensitivity as a long-term consequence of cisplatin exposure even several months following discontinuation of treatment. These findings are also consistent with previous studies that have shown persistent activation of the AKT pathway in cells resistant to EGFR TKIs (15, 16, 18, 20). Interestingly, the cisplatin-resistant cells cultured in cisplatin-free medium also exhibit a significantly increased proliferative rate (Fig. 4C), raising the possibility that cisplatin exposure enriches the population for tumor cells that grow more aggressively.
Reduced erlotinib sensitivity following cisplatin treatment is correlated with reduced PTEN function and can be partially overcome by reintroduction of PTEN. To pursue the mechanism underlying persistent AKT activation in the cisplatin-resistant PC9 cells, we examined the status of the PTEN tumor suppressor, a negative regulator of PI3K/AKT signaling that has been previously implicated in resistance to chemotherapy drugs, including cisplatin (21, 22), and is mutationally disrupted in a subset of NSCLCs (23). PC9 cells express readily detectable levels of PTEN protein, whereas another human NSCLC cell line (NCI-H1650) with an identical EGFR activating mutation, but which is ∼100-fold less sensitive to erlotinib (Supplementary Fig. S1), does not express PTEN due to an inactivating mutation (Fig. 5A ). Notably, NCI-H1650 cells exhibit relatively high levels of phospho-AKT compared with PC9 cells, suggesting that PTEN loss may contribute to the difference in erlotinib sensitivity between NIH-H1650 and PC9 cells. Significantly, in the cisplatin-resistant PC9 cells, PTEN levels are barely detectable (Fig. 5A), suggesting that PTEN attenuation may contribute to the persistent phospho-AKT levels seen in those cells on treatment with erlotinib (Fig. 4B).
Reduced erlotinib sensitivity following cisplatin treatment is correlated with loss of PTEN expression and can be partially overcome by reintroduction of PTEN. A, equivalent amounts of cell lysates from PC9, PC9 (CR), and 1650 cells were analyzed by SDS-PAGE followed by immunoblotting with PTEN and AKT antibodies. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) serves as loading control. B, cell lysates from PC9 cells treated cisplatin for the indicated number of days were subjected to SDS-PAGE and immunoblotted for PTEN expression (top). Cell lysates from PC9 (CR) maintained in cisplatin-free medium for increasing durations (1, 2, 3, and 4 mo) were similarly analyzed (bottom). C, cell lysates from green fluorescent protein and PTEN-expressing lentivirus-infected PC9 and PC9 (CR) were analyzed for PTEN expression by immunoblotting. D, 100,000 cells corresponding to PC9, PC9 (CR), and PC9 (CR PTEN-infected) lines were plated in a 100 mm dish for clonogenic assays. Erlotinib (2 μmol/L) was added the following day and cells were maintained in medium containing 2 μmol/L erlotinib for the 10-d assay. Number of surviving cells as measured by Syto60 staining in the respective cell lines. Average of three independent experiments. Bars, SD. Representative stained plate from each cell line (bottom).
To address the mechanism by which cisplatin affects PTEN, we examined the effects of cisplatin on PTEN expression and subcellular distribution. We observed that, within a few days following cisplatin treatment of PC9 cells, PTEN protein levels are significantly reduced (Fig. 5B, top). PTEN regulation also appears to involve altered subcellular distribution in response to cisplatin. Thus, within 2 h of cisplatin exposure, PC9 cells exhibit significant nuclear accumulation of PTEN (data not shown), a process that has been implicated in PTEN-mediated protection of the genome from cisplatin-mediated DNA damage (24). Notably, we observed that PTEN levels are gradually restored in the cisplatin-resistant PC9 cells during propagation in cisplatin-free medium (Fig. 5B, bottom). Overall, it appears that PTEN regulation in response to cisplatin exposure is complex and involves acute changes in protein distribution as well as longer-term changes in protein expression.
To directly address a potential role for reduced PTEN activity in the persistent phospho-AKT levels and reduced erlotinib sensitivity seen in the cisplatin-resistant PC9 cells, we restored high levels of PTEN in these cells via expression from a lentivirus. The virus-infected cells show a high level of PTEN expression relative to uninfected cells (Fig. 5C), although their growth properties were not detectably affected (data not shown). Significantly, in the clonogenic assay, the cells with restored PTEN levels exhibit significantly increased sensitivity to erlotinib albeit not to the same level as parental PC9 cells (Fig. 5D). Notably, basal levels of phospho-AKT remain somewhat elevated in the PTEN virus-infected cisplatin-resistant cells compared with parental PC9 cells, suggesting that AKT is activated in these cells by an additional upstream pathway. This elevation of phospho-AKT may also be influenced by the increased nuclear PTEN localization in the PTEN-infected PC9 cells (data not shown). The increased phospho-AKT may account for the fact that erlotinib sensitivity could not be completely restored in the cisplatin-resistant cells by expression of PTEN. Together, these findings suggest that the reduced PTEN function that arises during acquisition of resistance to cisplatin contributes to some but not all of the subsequently reduced sensitivity to erlotinib.
PI3K inhibition, together with restoration of PTEN function, reverses resistance to erlotinib. Our findings thus far have shown an important role for persistent AKT activation in the EGFR TKI sensitivity of cisplatin-resistant PC9 cells. Because PI3K is a major upstream activator of AKT, we examined the ability of a pharmacologic inhibitor of PI3K, LY294002, to modulate the erlotinib sensitivity of cisplatin-resistant PC9 cells. Indeed, we observed that the LY294002 restored a substantial level of erlotinib sensitivity in these cells (Fig. 6A ) and caused a corresponding reduction in phospho-AKT levels when combined with erlotinib (Fig. 6B). Furthermore, PI3K inhibition in the PTEN-infected cisplatin-resistant cells completely restores sensitivity to erlotinib in these cells, rendering them as sensitive to erlotinib as parental PC9 cells (Fig. 5A). The relevance of both a functional PTEN and PI3K inhibition in restoring erlotinib sensitivity is further supported by the observation that the cisplatin-resistant cells, which had been taken out of cisplatin for 4 months, also exhibit restored erlotinib sensitivity following treatment with LY294002 alone (Fig. 6C).
PI3K inhibition, together with restoration of PTEN function, reverses resistance to erlotinib. A, to determine if PI3K inhibition has any effect on reversal of resistance to erlotinib, clonogenic assays were done in triplicate on PC9, PC9 (CR), and PC9 (CR PTEN-infected) cells. 100,000 cells were plated and 2 μmol/L erlotinib ± 2.5 μmol/L LY 294002 was added the following day. The assay was carried out over 10 d and fresh medium containing 2 μmol/L erlotinib ± 2.5 μmol/L LY 294002 was replaced every 3 d. Number of surviving cells as measured by Syto60 staining in the respective cell lines. Average of three independent experiments. Bars, SD. Representative stained plate from each cell line (bottom). B, cell lysates were derived from PC9 (CR), PC9 (CR PTEN-infected), and PC9 (CR maintained in cisplatin-free medium) treated with either erlotinib (2 μmol/L), LY 294002 (2.5 μmol/L), or the combination of the two drugs for 4 h. These were subjected to SDS-PAGE analysis and immunoblotting with the indicated antibodies. C, a similar assay as described in A was done that included PC9 (CR) cells maintained in cisplatin-free medium for 4 mo. Number of surviving cells as measured by Syto60 staining in the respective cell lines. Average of three independent experiments. Bars, SD. Representative stained plate from each cell line.
Because the RAS pathway can activate AKT via PI3K, and K-RAS mutations are frequently detected in TKI-refractory NSCLC patients (25), we examined a potential role for K-RAS activation in the cisplatin-resistant cells. Interestingly, the levels of total as well as activated K-RAS are elevated in cisplatin-resistant PC9 relative to parental PC9 cells (Supplementary Fig. S2A). To determine if an activated RAS pathway is responsible for the persistently activated AKT pathway after erlotinib treatment, we also examined activated (GTP-bound) RAS levels in parental PC9 and cisplatin-resistant PC9 cells after 4 h of exposure to 1 μmol/L erlotinib (Supplementary Fig. S2B). Although phospho-AKT remained elevated, activated K-RAS levels were efficiently down-regulated by erlotinib. This result is consistent with the observation that phospho-ERK levels are attenuated in erlotinib-treated cisplatin-resistant PC9 cells as efficiently as in erlotinib-treated PC9 cells (Fig. 3). This suggests that, although the elevated AKT activation in these cells may, in part, be contributed by the RAS pathway, the resistance to erlotinib is unlikely to be a consequence of increased K-RAS activation in the presence of erlotinib. Together, these findings suggest that activation of the PI3K pathway is largely responsible for the reduced erlotinib sensitivity seen in cisplatin-resistant PC9 cells and that loss of PTEN expression and activation of an EGFR-independent pathway leading to PI3K activation both play an important role.
Discussion
Cultured NSCLC cell lines provide a very useful system for modeling the clinical response to EGFR TKIs as well as mechanisms of acquired drug resistance. Because EGFR TKIs are typically administered as second- or third-line treatment in NSCLC, subsequent to platinum-based chemotherapy, we extended the standard use of cell culture modeling to examine the effect of cisplatin on subsequent treatment with an EGFR TKI. To our knowledge, these represent the first studies to model this frequently administered NSCLC treatment sequence in cell culture. Our overall findings indicate that exposure to cisplatin can lead to significant resistance to subsequent TKI challenge in EGFR-mutant NSCLC. Interestingly, whereas cisplatin pretreatment does not detectably affect the IC50 of erlotinib in this model, the clonogenic growth potential of cisplatin-treated cells subsequently exposed to erlotinib is substantially increased. Extrapolating these cell culture findings to the clinical setting, this could indicate that, although cisplatin treatment may not affect the rate of clinical responses to subsequent EGFR TKI therapy, it could lead to a reduced time to disease progression in cases where EGFR TKI responses are observed due to a more rapid acquisition of TKI resistance. Indeed, EGFR mutation-positive NSCLC patients frequently respond to EGFR TKI therapy in the second- or third-line setting but will ultimately develop resistance to therapy (26). Thus far, clinical studies to compare the time to progression following erlotinib treatment in the first-line versus later-line settings have not been reported, yet our findings suggest that a differential effect may exist and, if clinically validated, could have an important effect on first-line therapeutic decision-making.
It is also interesting to consider the difference between drug responses in the PC9 and PC14 cell lines. The PC14 cells exhibit the same EGFR exon 19 deletion seen in PC9 cells but are more resistant to cisplatin initially. Thus, they may represent the subset of EGFR-mutant patients that fail to show an initial response to platinum therapy. Moreover, the fact that, like PC9 cells that have acquired cisplatin resistance, PC14 cells exhibit a strikingly increased clonogenic potential in the presence of erlotinib suggests that both primary and acquired cisplatin resistance among EGFR-mutant NSCLC may lead to a more rapid development of disease progression during treatment with EGFR TKIs.
The effects of cisplatin pretreatment on EGFR-mutant NSCLC cells in the PC9 model are associated with persistently activated PI3K/AKT signaling, a well-established tumor cell survival pathway that has been linked to drug resistance in a variety of contexts (26, 27). The persistent TKI-insensitive AKT activation observed in cisplatin-treated PC9 cells appears to involve at least two different mechanisms-reduced PTEN function and EGFR-independent AKT activation. Our results indicate that pharmacologic PI3K inhibition can efficiently suppress the alternatively activated AKT pathway, thereby rendering cells resensitized to EGFR TKIs when exposed to a combination of the PI3K inhibitor LY294002 and erlotinib. Several previous studies have similarly implicated PTEN and PI3K/AKT activity in cisplatin resistance in cancer cells (21, 22, 28, 29), consistent with the apparent role for PTEN and AKT in the NSCLC model described here.
The mechanism by which PTEN function is reduced in the cisplatin-treated PC9 cells appears to be complex and potentially involves changes in PTEN expression and subcellular localization, both of which have been previously implicated in PTEN regulation (30). Furthermore, the acute and longer-term effects of cisplatin treatment on these cells are distinct. Thus, within a few hours of cisplatin exposure, PTEN undergoes a gel mobility shift consistent with phosphorylation (data not shown) and exhibits a nuclear accumulation by immunofluorescence. However, over a period of days following treatment, the PTEN subcellular distribution is not obviously different from that seen in untreated cells, whereas PTEN protein levels are clearly reduced (data not shown). The initial response to cisplatin may reflect a cellular mechanism that engages survival signaling acutely following a stressful stimulus, whereas the longer-term response may reflect the selection of a fraction of cells that are capable of sustaining such survival signaling.
The PI3K/AKT pathway is among the critical effectors of oncogenic EGFR, and persistent EGFR-independent signaling through this pathway appears to contribute to erlotinib resistance in some NSCLCs (16). This presumably explains why cisplatin-resistant PC9 cells, in which this pathway persists in the presence of erlotinib, exhibit erlotinib-resistant clonogenic cell survival. Moreover, the fact that AKT activation is seen in cisplatin-resistant cells even several months following cisplatin withdrawal is likely to account for the longer-term consequences of cisplatin exposure to subsequent EGFR TKI treatment. However, it remains unclear as to why the effects of cisplatin on this pathway persist following cisplatin withdrawal. Although we found that treatment of cisplatin-resistant PC9 cells with two different chromatin modifying agents did not affect subsequent erlotinib sensitivity, it is not possible to exclude a role for epigenetic regulation of components of this pathway, particularly when considering the reversible nature of PTEN expression.
Although erlotinib is currently approved for clinical use in the second- or third-line setting, our cell culture findings highlight a potential disadvantage associated with subjecting EGFR mutation-positive NSCLC patients to platinum-based therapy before treatment with EGFR TKIs. Notably, it has been reported that NSCLC patients with reduced PTEN expression experience a shorter time to tumor progression and, consequently, a worse prognosis (31, 32). Thus, if these preclinical findings accurately model the experience of many NSCLC patients, they would suggest that patients who have received recent platinum therapy may be more resistant to TKIs and may also be intrinsically more resistant to subsequent lines of therapy.
It is also interesting to consider how first-line EGFR TKI therapy might affect subsequent platinum therapy. The T790M EGFR kinase domain mutation, along with MET gene amplification, appear to account for 50% to 70% of acquired EGFR TKI resistance in NSCLC patients, and both mechanisms can promote increased signaling via the PI3K/AKT pathway (16, 33, 34). Taken together with the fact that PI3K/AKT signaling reportedly affects cisplatin sensitivity, these EGFR TKI resistance mechanisms that arise in the context of first-line TKI treatment may similarly contribute to subsequent resistance to platinum-based agents.
In conclusion, we find that pretreatment of EGFR TKI-sensitive PC9 NSCLC cells with cisplatin leads to significant resistance to subsequent TKI treatment via a persistently activated PI3K/AKT pathway, in part contributed by cisplatin-induced reduction in PTEN function. The addition of a pharmacologic PI3K inhibitor to an EGFR TKI can partially resensitize these cells to TKIs, suggesting a potential therapeutic strategy for reversing or delaying resistance to TKIs. Importantly, our findings also suggest that these effects of cisplatin can persist for at least several months following drug withdrawal. These potential issues associated with cisplatin treatment before EGFR TKI administration should prompt further consideration of first line EGFR TKI use in EGFR mutation-positive NSCLC patients.
Disclosure of Potential Conflicts of Interest
J. Settleman has received commercial research support from AstraZeneca, Entermed, and Novartis and is on the speakers' bureau of Biogen.
Acknowledgments
We thank the members of the Settleman laboratory for helpful discussions during the course of this study.
Footnotes
-
Grant support: NIH grant RO1 CA115830 and V Foundation for Cancer Research award (J. Settleman).
-
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 Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).
- Accepted June 10, 2008.
- Received January 10, 2008.
- Revision received May 13, 2008.
References
Volume 14, Issue 21
