Erlotinib Resistance in Lung Cancer Cells Mediated by Integrin B1/Src/Akt-Driven Bypass Signaling

Published OnlineFirst July 19, 2013; DOI: 10.1158/0008-5472.CAN-12-4502

Cancer Therapeutics, Targets, and Chemical Biology Research

Erlotinib Resistance in Lung Cancer Cells Mediated by Integrin b1/Src/Akt-Driven Bypass Signaling

Rina Kanda1, Akihiko Kawahara3, Kosuke Watari1, Yuichi Murakami1,6, Kahori Sonoda1, Masashi Maeda1, Hideaki Fujita4, Masayoshi Kage3, Hidetaka Uramoto5, Carlota Costa7, Michihiko Kuwano2, and Mayumi Ono1

Abstract EGF receptor (EGFR) kinase inhibitors, including gefitinib and erlotinib, exert potent therapeutic efficacy in non–small cell lung cancers harboring EGFR-activating mutations. However, most patients ultimately develop resistance to these drugs. Here, we report a novel mechanism of acquired resistance to EGFR tyrosine kinase inhibitors and the reversal of which could improve clinical outcomes. In erlotinib-resistant lung cancer cells harboring activating EGFR mutations that we established, there was increased expression of Src, integrin b1, a2, and a5 along with enhanced cell adhesion activity. Interestingly, RNAi-mediated silencing of integrin b1 restored erlotinib sensitivity and reduced activation of Src and Akt after erlotinib treatment. Furthermore, Src silencing inhibited Akt phosphorylation and cell growth, with this inhibitory effect further augmented by erlotinib treatment. Increased expression of integrin b1, a5, and/or a2 was also observed in refractory tumor samples from patients with lung cancer treated with erlotinib and/or gefitinib. Together, our findings identify the integrin b1/Src/Akt signaling pathway as a key mediator of acquired resistance to EGFR-targeted anticancer drugs. Cancer Res; 73(20); 6243–53. 2013 AACR.

Introduction acquired resistance to EGFR-TKIs include the T790M mutation fi Patients with non–small cell lung cancer (NSCLC) harboring in exon 20 of the EGFR-TKI domain (8, 9) and Met ampli ca- activating somatic mutations in the EGF receptor (EGFR) gene tion (10). Furthermore, the loss of PTEN and increased expres- show dramatic clinical responses. Of the somatic mutations, sion of mitogen-activated protein kinase (MAPK), ABCG2, b in-frame deletions in exon 19 (del E746-A750) and the L858R- insulin growth factor 1 receptor (IGF-IR), and TGF- have point mutation are most commonly observed in NSCLC (1). also been reported (7, 11, 12). In addition to these well- These EGFR mutations are closely associated with sensitivity characterized mechanisms, further elucidation of novel to EGFR tyrosine kinase inhibitors (TKI), erlotinib and gefiti- mechanisms for acquired drug resistance is essential for the nib. Both gefitinib and erlotinib have shown to improve development of personalized therapeutics and strategies to progression-free survival as compared with chemotherapy circumvent drug resistance. alone when given as first-line treatment for patients with In our laboratory, we have established various drug-resistant NSCLCs harboring activating EGFR mutations (2–6). cell lines from human lung cancer cell lines harboring acti- However, one of the serious problems encountered during vating EGFR mutations by selecting for resistance to EGFR- clinical treatment with EGFR-TKI is the appearance of drug- TKIs (13). Previously, we have reported the loss of PTEN resistant tumors (7). Well-characterized mechanisms for the expression with the loss of nuclear translocation of EGR-1, a transcription factor responsible for PTEN gene expression, in gefitinib-resistant clones (14–16). We also reported that either Authors' Affiliations: 1Department of Pharmaceutical Oncology and the complete or partial loss of the activated EGFR gene allele 2 Laboratory of Molecular Cancer Biology, Graduate School of Pharma- could also result in the acquisition of erlotinib resistance (17). ceutical Sciences, Kyushu University, Fukuoka; 3Department of Diagnostic Pathology, Kurume University Hospital, Kurume; 4Section of Functional In this present study, we further established erlotinib-resistant Morphology, Faculty of Pharmaceutical Science, Nagasaki International clones by step-wise selection following exposure to erlotinib University, Nagasaki; 5Second Department of Surgery, School of Medicine, University of Occupational and Environmental Health, Kitakyushu; 6St. and observed enhanced expression and activation of integrin Mary's Institute of Health Science, St. Mary's Hospital, Kurume, Japan; b1 and Src. Herein, we present a novel bypass mechanism and 7Pangaea Biotech, Dexeus University Institute, Barcelona, Spain through which the integrin b1/Src/Akt signaling may play a Note: Supplementary data for this article are available at Cancer Research pivotal role in the acquisition of erlotinib resistance in lung Online (http://cancerres.aacrjournals.org/). cancer cells. Corresponding Author: Mayumi Ono, Department of Pharmaceutical Oncology, Graduate School of Pharmaceutical Sciences, Kyushu Univer- Materials and Methods sity, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812-8582, Japan. Phone/Fax: 81-92-642-6296; Cell culture E-mail: [email protected] PC9 cells were kindly provided by Dr. Y Ichinose, National doi: 10.1158/0008-5472.CAN-12-4502 Hospital Organization, Kyushu Cancer Center (Fukuoka, 2013 American Association for Cancer Research. Japan) and 11–18 cells were provided by Dr. K Nakagawa,

www.aacrjournals.org 6243

Kanda et al.

Kinki University (Osaka, Japan). These cells and the resistant nating between 95C for 15 seconds and 60C for 1 minute. clones were cultured in RPMI medium supplemented with 10% The relative gene expression for each sample was determinedÀÁ FBS in an atmosphere of 5% CO2 (14, 18). Cells were routinely using the formula 2ðÞ¼DCt 2½CtðÞGAPDH Ct target , confirmed to be free of mycoplasma contamination using which reflected the target gene expression normalized to mycosensor QPCR Assay Kits (Agilent Technologies). Cells glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels. were transiently transfected with cDNA using Lipofectamine Each value represents the average from triplicate trials SD. LTX, PLUS reagent, and Opti-MEM medium (Invitrogen) according to the manufacturers' recommendations. Cells were Flow cytometry (FACS) transfected with siRNA duplexes using Lipofectamine RNAi- Cell suspension (106 cells/100 mL) in incubation buffer [0.5% MAX and Opti-MEM medium (Invitrogen) according to the bovine serum albumin (BSA) in PBS] was stained with anti- manufacturers' recommendations. Reagents are described in integrin b1 mAb or purified PE-Mouse IgG2a Control (Becton- Supplementary Material and Methods. Dickinson) for 1 hour at 4C. Flow cytometry was conducted using the FACSCalibur System (Becton-Dickinson). Cytotoxicity assays Exponentially growing cell suspensions were seeded into Tumor samples each well and the following day the indicated concentration of Informed consent for the use of tumor tissue specimens was drugs was added. After incubation for 72 hours, cytotoxicity obtained from patients according to the legal guidance (21, 22). was determined as described previously (17, 18). Each value Tumor samples, case 4 and case 10, are collected from Uni- represents the average from triplicate wells SD. versity Hospital, University of Occupational and Environmen- For colony formation assay, 5 102 cells were plated in 35 tal Health (Kitakyushu, Japan). Tumor samples, case RB4 and mm dishes and following day, cells were transfected with case RB11, are from EURTAC Study in Spain. The EURTAC was integrin b1 siRNA. After 48 hours, various concentrations of approved by the Institutional Review Board of each partici- erlotinib were added, followed by incubation for 7 days at 37C. pating center and written informed consent was obtained Colonies of more than 10 cells were counted after Giemsa from all patients. Mutation status of EGFR (delE746-A750 staining as described previously (19). Each value represents and L858R) and T790M was examined by direct sequencing the average of duplicate dishes. analysis. K-Ras mutations were investigated by PCR analysis (21–23). Western blot analysis Cells were rinsed with ice-cold PBS and lysed in Triton X-100 Immunohistochemical analysis buffer (50 mmol/L HEPES, 150 mmol/L NaCl, 50 mmol/L NaF, Expression of integrins and Src was examined by immuno- 1% Triton X-100, and 10% glycerol containing 5 mmol/L EDTA, histochemical (IHC) staining as previously described (23). Anti- 1 mmol/L phenylmethylsulfonyl fluoride, 10 mg/mL aprotinin, integrin b1 mAb was purchased from Abcam Inc., anti-integrin 10 mg/mL leupeptin, and 1 mmol/L sodium orthovanadate) a2 mAb was from Epitomics, anti-integrin a5 mAb was from and cell lysates were separated by SDS-PAGE and transferred Santa Cruz Biotechnology, and anti-Src was from Cell Signaling to Immobilon membranes (Millipore Corp.). For detection of Technology. The intensity of staining was scored using the active integrin b1, cells were plated on fibronectin- or collagen- following scale: no staining, 0; weak staining, þ1; moderate coated dishes and incubated with antiactive integrin b1 anti- staining, þ2; and strong staining, þ3, in more than 10% of body (HUTS-4) for 30 minutes at 37C. Cells were then rinsed cancer cells. with PBS, lysed in SDS sample buffer, and bound antibody was detected by Western blotting (20). Immunoprecipitation was Results described in Supplementary Material and Methods. Establishment of the erlotinib-resistant subclones from PC9 cells and 11-18 cells Cell adhesion assay To isolate erlotinib-resistant subclones from PC9 cells har- For cell adhesion assay, 96-well plates were coated with boring the activating EGFR delE746-A750 mutation and 11-18 collagen (10 mg/well, Trevigen) or fibronectin (5 mg/well, cells harboring activating EGFR L858R mutation, we cultured Trevigen) and cancer cells were seeded onto the extracellular PC9 cells and 11-18 cells in increasing, step-wise doses of matrix components, allowed to adhere for 30 minutes at 37C. erlotinib up to 15 mmol/L for approximately 6 months as Each value represents the average from triplicate trials SD. described previously (14, 17). We isolated drug-resistant cell lines, PC9/ER2 and 11-18/ER2, and further established drug- Quantitative real-time PCR resistant subclones, PC9/ER2-2 and PC9/ER2-3 from PC9/ER2 RNA was reverse transcribed from random hexamers using and 11-18/ER2-1 from 11-18/ER2 (Fig. 1A; Table 1). Growth avian myeloblastosis virus reverse transcriptase (Promega). In rates between each parental cell line and resistant subclones brief, 20 mL of PCR amplification reaction mixtures contained are similar (Table 1). Resistant subclones also manifested cDNA, primer pairs (Applied Biosystems), a dual-labeled increased resistance to gefitinib as compared with the fluorogenic probe (Applied Biosystems), and TaqMan Univer- parental cell lines (Table 1). In contrast, these 3 resistant sal PCR Master Mix (Applied Biosystems). The thermal cycle subclones exhibited similar sensitivities to picropodophyllin conditions included maintaining the reactions at 50C for 2 (an inhibitor of IGF-IR), SU11274 (an inhibitor of Met), and minutes and 95C for 10 minutes and then 40 cycles alter- cisplatin as compared with the parental cell lines (Table 1).

6244 Cancer Res; 73(20) October 15, 2013 Cancer Research

Integrin b1 Contributes to EGFR-TKI Resistance in NSCLC

Figure 1. Expression and activation of receptor tyrosine kinases and downstream signaling molecules in PC9, PC9/ER2-2, PC9/ER2-3, 11-18, and 11-18/ER2- 1 cells. A, the sensitivity to erlotinib was assessed by cell proliferation assays for 3 days. B, Western blots showing the expression levels of proteins and the loading control a-Tubulin. C, Western blots showing the effect of erlotinib on expression of proteins. Each cell line was exposed to various doses of erlotinib for 5 hours.

www.aacrjournals.org Cancer Res; 73(20) October 15, 2013 6245

Kanda et al.

Table 1. Comparison of drug sensitivity to various molecular-targeted drugs and cisplatin

Relative drug resistancea Doubling Cell lines Time (hour)b Erlotinib Geﬁtinib Picropodophyllin SU11274 Cisplatin PC9 21.6 1.0 (0.029) 1.0 (0.011) 1.0 (0.5) 1.0 (4.7) 1.0 (1.0) PC9/ER2-2 25.0 170 (4.9) 331 (3.9) 0.92 (0.4) 1.1 (5.0) 1.1 (1.1) PC9/ER2-3 24.3 199 (5.8) 310 (3.6) 0.77 (0.4) 1.2 (5.5) 1.2 (1.2) 11–18 25.2 1.0 (0.8) 1.0 (1.4) 1.0 (0.5) 1.0 (8.9) 1.0 (4.1) 11–18/ER2-1 24.9 10.1 (8.1) 12 (16.4) 1.1 (0.5) 1.0 (9.1) 1.2 (4.9)

a IC50 values (mmol/L) are calculated from logit regression lines from triplicate dishes and presented in parenthesis. The relative drug

resistance is deﬁned as the IC50 value divided by the value for each parental cell line. bDoubling time (hour) for each cell line is average of duplicate trials.

Neither alteration for secondary EGFR mutations (T790M, Integrins induce the activation of downstream regulatory T854A, L747S, and D761Y) nor hot spot activating mutation signaling pathways by adopting a high-affinity state to interact of PIK3CA was detected in resistant subclones (data not with extracellular matrices (25). Ligands of integrin a2b1 and shown). integrin a5b1 are collagen and fibronectin, respectively. As As compared with parental cell lines, all resistant subclones shown in Fig. 2D, increased activation of integrin b1 in PC9/ expressed relatively lower levels of phospho-HER2 (pHER2; Fig. ER2-2 and 11–18/ER2-1 cells as compared with parental cell 1B). In addition, 2 resistant subclones from PC9 expressed lines was observed in the presence of fibronectin or collagen lower levels of HER2, HER3, and pHER3. In contrast, the when a specific antibody (HUTS-4) recognizing ligand-occu- expression of phospho-Akt (pAkt) was moderately higher in pied integrin b1 was used. Furthermore, the adhesion of cells to the resistant subclones, although that of PTEN, c-Met, IGF1R, collagen- or fibronectin-coated plates in the resistant sub- Axl, and FGFR1-4 was also similar among cell lines (Fig. 1B, clones was significantly higher than that observed in parental data not shown). DelE746-A750 EGFR was similarly expressed cell lines (Fig. 2E). among both of the resistant subclones and the PC9 cell line (data not shown). However, L858R EGFR expression was less in Integrin b1 knockdown overcomes erlotinib resistance 11-18/ER2-1 cells because of the partial loss of mutated EGFR in resistant subclones gene allele, as described previously (17). The phosphorylation We next examined whether integrin b1 expression was of EGFR (pEGFR) and Erk1/2 was similarly inhibited by erlo- correlated with erlotinib resistance or constitutive Akt phos- tinib in a dose-dependent manner in parental cell lines as well phorylation. Treatment with integrin b1 siRNA almost as the resistant subclones. In contrast, the phosphorylation of completely suppressed the expression of integrin b1 protein Akt was observed even in the presence of erlotinib in the and mRNA (Fig. 3A and B) and there were no apparent changes resistant subclones (Fig. 1C). in the growth rates between each parental cell line and their resistant subclones upon exposure to integrin b1 siRNA for 5 Increased expression of b1, a2, and a5 integrins in the days (data not shown). In resistant subclones, integrin b1 erlotinib-resistant subclones siRNA-treatment restored sensitivity to erlotinib (Fig. 3C) and We further examined the expression levels of integrins in erlotinib treatment almost completely blocked the constitutive the resistant subclones by Western blot analysis. Surpris- phosphorylation of Akt when integrin b1 was silenced (Fig. 3D). ingly, all resistant subclones manifested higher levels of b1, We further examined whether integrins a2 and a5 were also a2, and a5 integrins than the parental cell lines. However, involved in the constitutive Akt phosphorylation in PC9/ER2-2 there were similar expression levels of avandb4 integrins cells (Supplementary Fig. S1B). The phosphorylation of Akt was between parental cell lines and the resistant subclones (Fig. moderately inhibited by erlotinib when integrin a2 or integrin 2A). PC9/ER2-2 and PC9/ER2-3 also exhibited higher expres- a5 was silenced in PC9/ER2-2 cells. Integrin b1 knockdown sion levels of integrin b1 mRNA than the PC9 cells, but 11- was also accompanied by the downregulation of integrin a5, 18/ER2-1 exhibited similar expression levels of integrin b1 suggesting a close regulatory link between integrin b1 and mRNA (Fig. 2B). Fluorescence-activated cell sorting (FACS) integrin a5. These results suggest that integrin b1, in concert analysis also revealed higher expression levels of membra- with integrin a2 and/or integrin a5, seems to play a key role in nous integrin b1 in PC9/ER2-2 and 11-18/ER2-1 as compared the acquisition of erlotinib resistance. with parental cell lines (Fig. 2C). Integrin b1isknowntobe We further established the stable cell lines PC9/ITG2 and dimerized with a integrin family proteins a2anda5 (24), PC9/ITG11 by transfecting PC9 cells with integrin b1 cDNA and immunoprecipitation Western blot analysis revealed (Supplementary Fig. S2A). There was no concomitant enhance- that a2ora5 integrins were coimmunoprecipitated with ment of the expression of integrin a2 or integrin a5 in either integrin b1 in PC9/ER2-2 and PC9/ER2-3 cells (Supplemen- cell line and no apparent changes in the growth rates of PC9/ tary Fig. S1A). Mock, PC9/ITG2, and PC9/ITG11 cells over the 5-day culture

6246 Cancer Res; 73(20) October 15, 2013 Cancer Research

Integrin b1 Contributes to EGFR-TKI Resistance in NSCLC

PC9 PC9/ER2-2 PC9/ER2-3 11-18 11-18/ER2-1 3 1.5 α2 α5 2 1 1 mRNA 1 mRNA β αv β 1 0.5 β1 Integrins β4 Integrin

Integrin 0 0 PC9 PC9/ PC9/ 11-18 11-18/ α -Tubulin ER2-2 ER2-3 ER2-1 Figure 2. Increased expression and activation of integrin b1 in erlotinib- C β β resistant subclones. A, Western Integrin 1 Integrin 1 160 160 blots showing the expression Control Control a levels of integrins and -tubulin as 120 PC9 120 11-18 a loading control. B, integrin b1 PC9/ER2-2 11-18/ER2-1 mRNA levels were determined by 80 80 qRT-PCR. , P < 0.05; , P < 0.01 40 Cell number 40 versus each parental cell line. C, Cell number flow cytometry of the cell surface 0 0 expression of integrin b1. D, 100 101 102 103 100 101 102 103 Western blots showing the Log fluorescence Log fluorescence expression level of active integrin b1 using the HUTS-4 antibody, D PC9 PC9/ER2-2 11-18 11-18/ER2-1 HUTS-4 HUTS-4 HUTS-4 HUTS-4 which recognizes ligand-occupied mIgG mIgG mIgG mIgG integrin b1. Plastic dishes were ECM Coat FN CL FN CL ECM Coat FN CL CLFN coated with fibronectin (FN) or Active Active collagen (CL). E, cell adhesion Integrin β1 Integrin β1 assay on plastic dishes coated with Integrin β1 Integrin β1 BSA, collagen, or fibronectin. α α-Tubulin -Tubulin , P < 0.05; , P < 0.01 versus each parental cell line. E 1.2 0.4 PC9 11-18 1 PC9/ER2-2 11-18/ER2-1 PC9/ER2-3 0.3 0.8 0.6 0.2 0.4 0.1 Adhesion activity 0.2 Adhesion activity 0 0 BSA Collagen Fibronectin BSA Collagen Fibronectin period (data not shown). These 2 stable cell lines exhibited pathway could be responsible for erlotinib resistance. Relatively slightly increased resistance to erlotinib (Supplementary Fig. higher expression levels of Src and pSrc (Y416) were observed in S2B). Furthermore, the phosphorylation of Akt was not sup- PC9/ER2-2 and 11-18/ER2-1 cells as compared with each paren- pressed by erlotinib in either the PC9/ITG2 or PC9/ITG11 cell tal cell line (Fig. 4A). In each resistant subclone from PC9 or 11- lines while Akt phosphorylation was suppressed in PC9/Mock 18 cells, the phosphorylation of Src was blocked by integrin b1 cells (Supplementary Fig. S2C). The pEGFR and Erk1/2 was knockdown and this reduction was augmented by erlotinib similarly suppressed by erlotinib in the PC9/Mock cells and the (Fig. 4B), suggesting a close correlation between integrin b1 corresponding integrin b1-overexpressing cell lines. Thus, the and EGFR with the activation of the Src/Akt pathway. overexpression of integrin b1 with integrin a2 and integrin a5 We next examined how Src activation could be involved in seems to be one of the mechanisms underlying acquired the integrin b1-driven activation of Akt and acquired erlotinib erlotinib resistance. resistance using Src siRNA and dasatinib—a Src kinase inhibitor. Treatment with Src siRNA almost completely suppressed Src knockdown and dasatinib treatment suppress Akt the expression of Src (Fig. 4C). Src siRNA treatment suppressed phosphorylation and cell survival in the resistant expression of Src and markedly suppressed Akt phosphoryla- subclones tion (Fig. 4D) and cell growth (Fig. 4E) in PC9/ER2-2 and 11-18/ Src protein, which interacts with PI3K, is often activated by ER2-1 cells. This suppression of Akt phosphorylation was EGFR or integrin b1 in association with extracellular matrices augmented by erlotinib. We further examined whether Akt (26). Thus, we next examined whether integrin b1 could directly phosphorylation was inhibited by erlotinib in PC9/ER2-2 cells induce the activation of the Src/Akt pathway and whether this treated with an inhibitor of Src, dasatinib. Dasatinib treatment

www.aacrjournals.org Cancer Res; 73(20) October 15, 2013 6247

Kanda et al.

A Day 1 Day 2 Day 3 Day 4 Day 5 Day 1 Day 3 Day 5 β1 si integrin –+ –+ –+ –+ –+ si integrin β1 –+ –+ –+

Integrin β1 Integrin β1

PC9 α-Tubulin 11-18 α-Tubulin Integrin β1 PC9/ ER2-2 α-Tubulin Integrin β1 ER2-1 β 11-18/ Integrin 1 α-Tubulin PC9/ ER2-3 α-Tubulin

B 0.02 0.02

0.01 0.01 ** ** ** * ** Integrin β 1 mRNA Integrin β 1 mRNA 0 0 si integrin β1 –+ –+ –+ si integrin β1 –+ –+ PC9 PC9/ PC9/ 11-18 11-18/ ER2-2 ER2-3 ER2-1 C PC9 PC9/ER2-2 PC9/ER2-3 si Control 100 100 si integrin β1 100

50 50 50 % of control % of control 0 0 % of control 0 0.1 10 0.1 10 0.1 10 Erlotinib (μmol/L) Erlotinib (μmol/L) Erlotinib (μmol/L)

11-18 11-18/ER2-1 si Control 100 100 si integrin β1

50 50 % of control % of control 0 0 0.1 10 0.1 10 Erlotinib (μmol/L) Erlotinib (μmol/L) D PC9 PC9/ER2-2 PC9/ER2-3 11-18 11-18/ER2-1 Erlotinib (1 μmol/L) – +– + – +– + – +– + –+–+ –+–+ si integrin β1 ––+++ ––+––++ ––+––+ + +

pEGFR EGFR Integrin β1

pAkt Akt

pErk1/2 Erk1/2 α−Tubulin

Figure 3. Effect of integrin b1 knockdown on drug sensitivity in erlotinib-resistant cell lines. A, Western blot showing integrin b1 levels after treatment with siRNA for the indicated number of days. B, integrin b1 mRNA expression was analyzed by qRT-PCR after 2 days of siRNA treatment. , P < 0.05; , P < 0.01 versus each parental cell line. C, the effect of integrin b1 knockdown on the sensitivity to erlotinib treatment for 7 days was assessed by the colony formation assay. D, the effect of integrin b1 knockdown on the pEGFR, Akt, and Erk1/2. The cells were treated with siRNA for 2 days followed by exposures to 1 mmol/L erlotinib for 5 hours.

at 0.1 mmol/L suppressed Akt phosphorylation and this effect to dasatinib than PC9 cells (Supplementary Fig. S3B), suggest- was augmented by erlotinib in PC9/ER2-2 cells (Supplemen- ing that Src activation in the resistant subclone may sensitize tary Fig. S3A). Moreover, PC9/ER2-2 cells were more sensitive those cells to the cytotoxic effects of dasatinib.

6248 Cancer Res; 73(20) October 15, 2013 Cancer Research

Integrin b1 Contributes to EGFR-TKI Resistance in NSCLC

A B PC9 PC9/ER2-2 11-18 11-18/ER2-1 Erlotinib(1μmol/L) – + –+ – +– + – +– + – +– + si integrin β1 –– ++ –– ++ –– ++ –– ++

11-18 pEGFR PC9 PC9/ER2-2 11-18/ER2-1 pSrc(Y416) EGFR β Src Integrin 1 α-Tubulin pSrc(Y416) Src α-Tubulin

C Day 2 Day 3 Day 5 Day 2 Day 3 Day 5 si Src –+ –+–+ si Src –+ –+–+

Src Src

PC9 α -Tubulin 11-18 α-Tubulin

Src Src PC9/ ER2-1 ER2-2 α-Tubulin 11-18/ α-Tubulin

D PC9 PC9/ER2-2 11-18 11-18/ER2-1 Erlotinib (1 μmol/L) – +–+ – +–+Erlotinib (1 μmol/L) – +–+ – +–+ si Src –– ++ –– ++ si Src –– ++ –– ++ pEGFR pEGFR EGFR EGFR Src Src pAkt pAkt Akt Akt α-Tubulin α-Tubulin

120 120 120 120 100 100 100 100 80 80 80 80 60 60 60 60 pAkt level 40 40 pAkt level 40 40 20 20 20 20 0 0 0 0 μ Erlotinib (1 mol/L) – +–+ – +– +Erlotinib (1 μmol/L) – +–+ – +–+ –– ++ ––+ + si Src si Src –– ++ –– ++

100 E 100 100 100 90 80 80 80 80 60 60 70 60 40 40 60 Cell growth Cell growth 50 20 20 40 40 Erlotinib (1 μmol/L) – +–+ – +–+Erlotinib (1 μmol/L) – +– + – +– + si Src –– ++ –– ++ si Src –– ++ –– ++

Figure 4. Effect of Src knockdown on Akt phosphorylation and the restoration of erlotinib sensitivity by Src knockdown. A, Western blots showing the expression of p-Src (Y416) and Src. B, the effect of integrin b1 knockdown on Src phosphorylation with or without erlotinib. Cells were treated for 2 days with integrin b1 siRNA, followed by further exposure to erlotinib for 5 hours. C, Western blots show the expression of Src protein after treatment with Src siRNA for the indicated number of days. Because Src and a-tubulin were detected on the same membrane, a-tubulin is pointed to with arrowheads. D, the cells were treated with or without Src siRNA for 2 days and exposed to erlotinib for another 5 hours. The quantitative analysis of the Western blots is shown and the values are normalized to pAkt levels in the absence of erlotinib and Src siRNA.E,theeffectsofSrcknockdownon proliferation of cells. Cells were treated for 2 days with Src siRNA, followed by treatment with erlotinib. The cell number in the absence of the drug and/or siRNA was set as 100%.

www.aacrjournals.org Cancer Res; 73(20) October 15, 2013 6249

Kanda et al.

β α α A Case 4 Integrin 1 Integrin 2 Integrin 5 Src (L858R)

Pre-gefitinib

1+ 1+ 0 1+

Post-gefitinib

3+ 3+ 0 3+

Integrin β1 Integrin α2 Integrin α5 Src B Case 10 (L858R)

Pre-gefitinib 0 3+ 0 3+

Figure 5. Immunohistochemical analysis of clinical tumor samples. Immunohistochemical analysis of Post-erlotinib integrins or Src expression in patients' refractory to geﬁtinib 3+ 3+ 3+ 2+ and/or erlotinib treatment. EGFR mutation status of each tumor C Case RB11 Integrin β1 Integrin α2 Integrin α5 Src sample was shown at the left side. (ΔE746- Magniﬁcation, 200. Tumor A750) samples were derived from Japan (A and B) and from Spain (C and D). Pre-erlotinib

1+ 1+ 0 2+

Post-erlotinib

3+ 2+ 0 3+

Integrin β1 Integrin α2 Integrin α5 Src D Case RB4 (ΔE746- A750, T790M)

Pre-gefitinib

2+ 1+ 2+ 2+

Post-gefitinib

3+ 3+ 2+ 3+

The expression of integrins in cancer cells from Table S1, 2 patients harbored L858R and 2 patients harbored refractory lung tumors delE746-A750. Figure 5(A–D) shows typical IHC images from We finally examined whether integrin expression could be the patients. Especially expression of integrin b1was observed in gefitinib- or erlotinib-refractory tumors. Four increased in the post-erlotinib and post-gefitinib tumor clinical samples were prepared from patients with lung samples with enhanced expression of integrin a2ora5. Src adenocarcinoma treated with gefitinib or/and erlotinib, but expression was also enhanced after treatment, but only post- subsequently became refractory to these EGFR-TKIs (Sup- erlotinib sample of case 10 shows lower expression of Src plementary Table S1; ref. 21). As shown in Supplementary than pretreatment sample. One sample, case RB4 harboring

6250 Cancer Res; 73(20) October 15, 2013 Cancer Research

Integrin b1 Contributes to EGFR-TKI Resistance in NSCLC

EGFR/EGFR family Inactive integrins EGFR/EGFR family Active integrins

β1 β1 β1 β1 β1 β1 α α α α α2 α5 2 2 5 5

Acquired resistance to Erlotinib Src

Mutant EGFR AKT Wild EGFR AKT EGFR family

Cell survival Cell survival

Figure 6. Our hypothetical model shows how erlotinib resistance is acquired in lung cancer cells. In erlotinib-sensitive cell lines, the proliferation and survival of human lung cancer cells harboring activating EGFR mutation depend upon the EGFR-driven PI3K/Akt pathway and this cell proliferation and survival is highly susceptible to erlotinib and other EGFR-TKIs. Erlotinib-resistant subclones express elevated levels of b1 and a2/a5 integrins together with Src, resulting in Akt activation. In erlotinib-resistant cell lines, activated integrin b1, a2ora5, and Src bypass pathway contributes to Akt signaling pathway.

T790M secondary mutation of EGFR, shows higher expres- lung cancer cells resistant to EGFR-TKIs. Our study showed sion of integrin b1, a2, and Src in post-gefitinib treatment. In that transfection with integrin b1 slightly induced erlotinib this case, integrin overexpression might not be directly resistance in PC9 cells (Supplementary Fig. S2) and in these correlated with acquired resistance (Fig. 5D). Expression of integrin b1-overexpressing cell lines, there was no enhanced integrins and Src in 4 clinical samples was summarized in expression of integrin a2ora5, suggesting that the coactiva- Supplementary Table S1. tion of b1/a2 and/or b1/a5 integrins is required for the acquisition of increased erlotinib resistance via the integrin Discussion b1 signaling pathway. In our present study, we established erlotinib-resistant sub- The interactions between cancer cells and the extracellular clones from PC9 cells harboring the activating del E746-A750 matrices often reduce the sensitivity to anticancer agents, EGFR mutation and 11-18 cells harboring the activating L858R including gefitinib (28–31) and the expression of integrin EGFR mutation. We observed the following characteristics in family proteins plays an important role in the malignant these resistant subclones: (i) Akt was constitutively activated progression of cancers (32–34). The cell proliferation rates and the phosphorylation of Akt was not suppressed by erlotinib were similar between parental cell lines and the resistant in resistant subclones; (ii) the expression of b1, a2, and a5 subclones, but the adhesion to collagen or fibronectin was integrins was upregulated along with Src activation; (iii) the much higher in the resistant subclones. Integrin b1 thus seems knockdown of integrin b1 restored cellular sensitivity to erlo- to be responsible for not only the acquisition of erlotinib tinib and reduced Akt phosphorylation; (iv) Src knockdown or resistance but also for cell adhesion activity. dasatinib treatment effectively inhibited cell survival and In this study, we observed constitutive phosphorylation of reduced Akt phosphorylation in the presence or absence of Akt even in the presence of erlotinib in resistant subclones. erlotinib; and (v) enhanced expression of integrin b1, a2, and/ This phosphorylation was markedly suppressed by erlotinib or a5 was also observed in refractory tumor samples from upon integrin b1 knockdown and cell sensitivity to erlotinib patients with lung cancer treated with gefitinib and/or erloti- was restored by integrin b1 knockdown. These data strongly nib. Together, these novel findings indicate that the acquisition suggest an essential role for integrin b1/a2 and/or integrin b1/ of erlotinib resistance is mediated by the activation of the a5 in the constitutive activation of Akt and in the drug integrin b1/Src/Akt signaling pathway (Fig. 6). resistance in the resistant subclones. Our study also sought Three major mechanisms including the alteration of onco- to understand how integrin b1 promotes Akt phosphorylation genic targets, the activation of downstream regulatory mole- and the acquisition of drug resistance. Of the potentially cules, and the activation of bypass effectors, are known to relevant molecules, Src is known to be a downstream regulator contribute to acquired resistance to molecularly targeted of EGFR and/or integrins and it plays an essential role in the drugs in cancer cells (7). This study suggests that the acqui- survival of lung cancer cells (35, 36). Src is also responsible for sition of erlotinib resistance is attributable to the activation of the acquisition of resistance to EGFR-targeted drugs (37, 38). an integrin b1-driven bypass signaling pathway. Furthermore, Consistent with these studies, we observed marked activation the study by Ju and colleagues (27) previously reported that of Src in PC9/ER2-2 and 11-18/ER2-1 cells as compared with transfection with integrin b1 cDNA protected PC9 cells from the parental counterparts. The blockade of Src signaling by Src apoptosis in response to gefitinib treatment. However, this siRNA resulted in the marked suppression of Akt phosphor- study did not evaluate how integrin b1 overexpression renders ylation, suggesting a close link between Src and Akt activation,

www.aacrjournals.org Cancer Res; 73(20) October 15, 2013 6251

Kanda et al.

as shown in a previous study (26). Furthermore, Src knock- pression of integrin b1/a2/a5, which resulted in the activation down also dramatically suppressed cell growth in PC9/ER2-2 of an integrin/Src/Akt pathway. This relationship between and 11-18/ER2-1 cells, even in the absence of erlotinib. Furthe- integrin expression and Src activity may be useful for the more, treatment with Src kinase inhibitor dasatinib also sup- development of both personalized therapeutics involving pressed Akt phosphorylation. Taken together, these data EGFR-TKIs and for circumventing drug resistance in patients suggest that the activation of the integrin b1/Src/Akt sig- with NSCLC. naling pathway may undergo compensatory activation during the acquisition of drug resistance (Fig. 6). On the other Disclosure of Potential Conflicts of Interest hand, clinical trials of Src inhibitor dasatinib in combination No potential conflicts of interest were disclosed. with erlotinib, was studied in lung cancer (39). We suggest that Src is one of the effective targets for lung cancer Authors' Contributions therapeutics. Conception and design: R. Kanda, M. Kage, M. Kuwano, M. Ono We also examined whether integrin b1, together with integ- Development of methodology: R. Kanda, M. Kuwano rin a2 or integrin a5, was expressed in lung cancer refractory Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Kanda, K. Watari, Y. Murakami, K. Sonoda, M. to gefitinib and/or erlotinib. These results showed increased Maeda, H. Fujita, H. Uramoto, C. Costa, M. Ono expression of b1, a2, and/or a5 integrins in 4 refractory tumors Analysis and interpretation of data (e.g., statistical analysis, biostatistics, fi computational analysis): R. Kanda, K. Watari, H. Fujita, M. Kuwano following treatment with ge tinib and/or erlotinib (Fig. 5 and Writing, review, and/or revision of the manuscript: R. Kanda, M. Kuwano, Supplementary Fig. S1). Enhanced expression of integrin b1 M. Ono and a2 were also observed in case RB4-harbored T790M Administrative, technical, or material support (i.e., reporting or orga- nizing data, constructing databases): A. Kawahara, M. Kage, M. Ono secondary EGFR mutation (Fig. 5D). It remains to be further Study supervision: M. Ono studied whether the enhanced expression of integrin b1, a2, and Src is closely correlated with T790M mutation. Recently, Grant Support integrin family targeted drugs are focused on cancer thera- This work was supported by the Third Term Comprehensive Control peutics. For example, anti-integrin a5b1 antibody (volocixi- Research for Cancer from the Ministry of Health, Labour and Welfare in Japan, and also by a Grant-in-Aid for challenging Exploratory Research from the Japan mab) is conducted on phase II clinical trials for several cancers Society for the Promotion of Science (JSPS; M. Kuwano). including NSCLC (40, 41). Further study should be also The costs of publication of this article were defrayed in part by the payment of required whether these integrin-targeted drugs could be useful page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. to overcome drug resistance in patients with NSCLC. In conclusion, the acquisition of erlotinib resistance in our Received December 14, 2012; revised May 21, 2013; accepted June 22, 2013; erlotinib-resistant subclones was attributable to the overex- published OnlineFirst July 19, 2013.

References 1. Sharma SV, Bell DW, Settleman J, Haber DA. Epidermal growth factor associated with a second mutation in the EGFR kinase domain. PLoS receptor mutations in lung cancer. Nat Rev Cancer 2007;7:169–81. Med 2005;2:e73. 2. Mitsudomi T, Yatabe Y. Mutations of the epidermal growth factor 10. Engelman JA, Zejnullahu K, Mitsudomi T, Song Y, Hyland C, Park JO, receptor gene and related genes as determinants of epidermal growth et al. MET amplification leads to gefitinib resistance in lung cancer by factor receptor tyrosine kinase inhibitors sensitivity in lung cancer. activating ERBB3 signaling. Science 2007;316:1039–43. Cancer Sci 2007;98:1817–24. 11. Ellis LM, Hicklin DJ. Resistance to targeted therapies: refining anti- 3. Inoue A, Suzuki T, Fukuhara T, Maemondo M, Kimura Y, Morikawa N, cancer therapy in the era of molecular oncology. Clin Cancer Res et al. Prospective phase II study of gefitinib for chemotherapy-naive 2009;15:7471–78. patients with advanced non-small-cell lung cancer with epidermal 12. Huang S, HolzelM,KnijnenburgT,SchlickerA,RoepmanP,€ growth factor receptor gene mutations. J Clin Oncol 2006;24:3340–6. McDermott U, et al. MED12 controls the response to multiple 4. Zhou Q, Zhang XC, Chen ZH, Yin XL, Yang JJ, Xu CR, et al. Relative cancer drugs through regulation of TGF-b receptor signaling. Cell abundance of EGFR mutations predicts benefit from gefitinib treat- 2012;151:937–50. ment for advanced non-small-cell lung cancer. J Clin Oncol 2011; 13. Ono M, Kuwano M. Molecular mechanisms of epidermal growth factor 29:3316–21. receptor (EGFR) activation and response to gefitinib and other EGFR- 5. Rosell R, Carcereny E, Gervais R, Vergnenegre A, Massuti B, Felip E, targeting drugs. Clin Cancer Res 2006;12:7242–51. et al. Erlotinib versus standard chemotherapy as first-line treatment for 14. Yamamoto C, Basaki Y, Kawahara A, Nakashima K, Kage M, Izumi H, European patients with advanced EGFR mutation-positive non-small- et al. Loss of PTEN expression by blocking nuclear translocation of cell lung cancer (EURTAC): a multicentre, open-label, randomised EGR1 in gefitinib-resistant lung cancer cells harboring epidermal phase 3 trial. Lancet Oncol 2012;13:239–46. growth factor receptor-activating mutations. Cancer Res 2010;70: 6. Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, 8715–25. Brannigan BW, et al. Activating mutations in the epidermal growth 15. Garofalo M, Romano G, Di Leva G, Nuovo G, Jeon YJ, Ngankeu A, et al. factor receptor underlying responsiveness of non-small-cell lung can- EGFR and MET receptor tyrosine kinase-altered microRNA expression cer to gefitinib. N Engl J Med 2004;350:2129–39. induces tumorigenesis and gefitinib resistance in lung cancers. Nat 7. Garraway LA, Janne PA. Circumventing cancer drug resistance in the Med 2011;18:74–82. era of personalized medicine. Cancer Discov 2012;2:214–216. 16. Okamoto K, Okamoto I, Hatashita E, Kuwata K, Yamaguchi H, Kita A, 8. Kobayashi S, Boggon TJ, Dayaram T, Janne€ PA, Kocher O, Meyerson et al. Overcoming erlotinib resistance in EGFR mutation-positive M, et al. EGFR mutation and resistance of non-small-cell lung cancer to non-small cell lung cancer cells by targeting survivin. Mol Cancer Ther gefitinib. N Engl J Med 2005;352:786–92. 2012;11:204–13. 9. Pao W, Miller VA, Politi KA, Riely GJ, Somwar R, Zakowski MF, et al. 17. Tabara K, Kanda R, Sonoda K, Kubo T, Murakami Y, Kawahara A, et al. Acquired resistance of lung adenocarcinomas to gefitinib or erlotinib is Loss of activating EGFR mutation gene contributes to acquired

6252 Cancer Res; 73(20) October 15, 2013 Cancer Research

Integrin b1 Contributes to EGFR-TKI Resistance in NSCLC

resistance to EGFR tyrosine kinase inhibitors in lung cancer cells. PLoS for extracellular matrix, TGFbeta and regulating microRNAs. Int J ONE 2012;7:e41017 Biochem Cell Biol 2010;42:25–30. 18. Ono M, Hirata A, Kometani T, Miyagawa M, Ueda S, Kinoshita H, et al. 31. Eke I, Deuse Y, Hehlgans S, Gurtner K, Krause M, Baumann M, et al. b1 Sensitivity to gefitinib (Iressa, ZD1839) in non-small cell lung cancer cell integrin/FAK/cortactin signaling is essential for human head and neck lines correlates with dependence on the epidermal growth factor (EGF) cancer resistance to radiotherapy. J Clin Invest 2012;122:1529–40. receptor/extracellular signal-regulated kinase 1/2 and EGF receptor/ 32. Natali PG, Nicotra MR, Bartolazzi A, Cavaliere R, Bigotti A. Integrin Akt pathway for proliferation. Mol Cancer Ther 2004;3:465–72. expression in cutaneous malignant melanoma: association of the 19. Koike K, Kawabe T, Tanaka T, Toh S, Uchiumi T, Wada M, et al. A alpha 3/beta 1 heterodimer with tumor progression.Int J Cancer canalicular multispecific organic anion transporter (cMOAT) antisense 1993;54:68–72. cDNA enhances drug sensitivity in human hepatic cancer cells. Cancer 33. Zutter MM, Santoro SA, Staatz WD, Tsung YL. Re-expression of the Res 1997;57:5475–9. alpha 2 beta 1 integrin abrogates the malignant phenotype of breast 20. Xu H, Zeng L, Peng H, Chen S, Jones J, Chew TL, et al. HMG-CoA carcinoma cells. Proc Natl Acad Sci U S A 1995;92:7411–5. reductase inhibitor simvastatin mitigates VEGF-induced "inside-out" 34. Mitra AK, Sawada K, Tiwari P, Mui K, Gwin K, Lengyel E. Ligand- signaling to extracellular matrix by preventing RhoA activation. Am J independent activation of c-Met by fibronectin and a5b1-integrin Physiol Renal Physiol 2006;291:F995–1004. regulates ovarian cancer invasion and metastasis. Oncogene 2011;30: 21. Uramoto H, Shimokawa H, Hanagiri T, Kuwano M, Ono M. Expression 1566–76. of selected gene for acquired drug resistance to EGFR-TKI in lung 35. Yamaguchi T, Yanagisawa K, Sugiyama R, Hosono Y, Shimada Y, adenocarcinoma. Lung Cancer 2011;73:361–5. Arima C, et al. NKX2-1/TITF1/TTF1-induced ROR1 is required to sustain 22. Onitsuka T, Uramoto H, Nose N, Takenoyama M, Hanagiri T, Sugio K, EGFR signaling in lung adenocarcinoma. Cancer Cell 2012;21:348–61. et al. Acquired resistance to gefitinib: the contribution of mechanisms 36. Song L, Morris M, Bagui T, Lee FY, Jove R, Haura EB. Dasatinib (BMS- other than the T790M, MET, and HGF status. Lung Cancer 2010; 354825) selectively induces apoptosis in lung cancer cells dependent 68:198–203. on epidermal growth factor receptor signaling for survival. Cancer Res 23. Kawahara A, Yamamoto C, Nakashima K, Azuma K, Hattori S, Kashi- 2010;66:5542–8. hara M, et al. Molecular diagnosis of activating EGFR mutations in non- 37. Yoshida T, Okamoto I, Okamoto W, Hatashita E, Yamada Y, Kuwata K, small cell lung cancer using mutation-specific antibodies for immu- et al. Effects of Src inhibitors on cell growth and epidermal growth nohistochemical analysis. Clin Cancer Res 2010;16:3163–70. factor receptor and MET signaling in gefitinib-resistant non-small cell 24. Barczyk M, Carracedo S, Gullberg D. Intgrins. Cell Tissue Res 2010; lung cancer cells with acquired MET amplification. Cancer Sci 339:269–80. 2010;101:167–72. 25. Ye F, Kim C, Ginsberg H. Reconstitution of integrin activation. Blood 38. Rexer BN, Ham AJ, Rinehart C, Hill S, Granja-Ingram Nde M, Gonzalez- 2012;119:26–33. Angulo AM, et al. Phosphoproteomic mass spectrometry profiling links 26. Thomas SM, Brugge JS. Cellular functions regulated by Src family Src family kinases to escape from HER2 tyrosine kinase inhibition. kinases. Annu Rev Cell Dev Biol 1997;13:513–609. Oncogene 2011;30:4163–74. 27. Ju L, Zhou C, Li W, Yan L. Integrin beta1 over-expression associates 39. Haura EB, Tanvetyanon T, Chiappori A, Williams C, Simon G, Antonia with resistance to tyrosine kinase inhibitor gefitinib in non-small cell S, et al. Phase I/II study of the Src inhibitor dasatinib in combination lung cancer. J Cell Biochem 2010;111:1565–74. with erlotinib in advanced non–small-cell lung cancer. J Clin Oncol 28. Rintoul RC, Sethi T. Extracellular matrix regulation of drug resistance in 2010;28:1387–94. small-cell lung cancer. Int J Radiat Biol 2007;83:733–41. 40. Millard M, Odds S, Neamati N. Integrin targeted therapeutics. Ther- 29. Giannelli G, Azzariti A, Fransvea E, Porcelli L, Antonaci S, Paradiso A. anostics 2011;1:154–88 Laminin-5 offsets the efficacy of gefitinib ('Iressa') in hepatocellular 41. Besse B, Tsao LC, Chao DT, Fang Y, Soria JC, Almokadem S, et al. carcinoma cells. Br J Cancer 2004;91:1964–9. Phase Ib safety and pharmacokinetic study of volociximab, an anti- 30. Helleman J, Jansen MP, Burger C, van der Burg ME, Berns EM. a5b1 integrin antibody, in combination with carboplatin and paclitaxel Integrated genomics of chemotherapy resistant ovarian cancer: a role in advanced non-small-cell lung cancer. Ann Oncol 2013;24:90–6.

www.aacrjournals.org Cancer Res; 73(20) October 15, 2013 6253

Erlotinib Resistance in Lung Cancer Cells Mediated by Integrin β 1/Src/Akt-Driven Bypass Signaling

Rina Kanda, Akihiko Kawahara, Kosuke Watari, et al.

Cancer Res 2013;73:6243-6253. Published OnlineFirst July 19, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-12-4502

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2013/11/12/0008-5472.CAN-12-4502.DC1

Cited articles This article cites 41 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/73/20/6243.full#ref-list-1

Citing articles This article has been cited by 11 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/73/20/6243.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/73/20/6243. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.