Nuclear Translocation of Epidermal Growth Factor Receptor

ORIGINAL ARTICLE

Molecular Mechanisms of Tyrosine Kinase Inhibitor Resistance Induced by Membranous/Cytoplasmic/ Nuclear Translocation of Epidermal Growth Factor Receptor Xuezhu Rong, MD, PhD,a Yuan Liang, MD, PhD,b Qiang Han, MD, PhD,a Yue Zhao, MD, PhD,a Guiyang Jiang, MD, PhD,a Xiupeng Zhang, MD, PhD,a Xuyong Lin, MD, PhD,a Yang Liu, MD, PhD,a Yong Zhang, MD, PhD,b Xu Han, MD, PhD,a Meiyu Zhang, MD, PhD,a Yuan Luo, MD, PhD,a Pengcheng Li, MD, MA,a Lai Wei, MD, MA,a Ting Yan, MD, MA,a Enhua Wang, MD, PhDa,* a

Department of Pathology, College of Basic Medical Sciences and First Affiliated Hospital of China Medical University, Shenyang, China b Department of Pathology, Cancer Hospital of China Medical University, Liaoning Cancer Hospital & Institute, Shenyang, China Received 16 January 2019; revised 9 June 2019; accepted 13 June 2019 Available online - 19 June 2019

ABSTRACT Introduction: The molecular mechanism underlying the induction of resistance to tyrosine kinase inhibitors (TKIs) via the membranous/cytoplasmic/nuclear translocation of EGFR has not yet been reported. Methods: We performed immunohistochemistry to detect the distribution of EGFR in lung adenocarcinoma specimens after TKI treatment and analyzed the relationship between different EGFR locations and patient survival duration. Mass spectrometry analysis and immunoprecipitation were performed to show the interaction of cytosolic EGFR with YY1 associated protein 1 (YAP) and salt inducible kinase 2 (SIK2). Dual-luciferase assays, immunoblotting, real-time polymerase chain reaction, and functional experiments were used to elucidate the role of EGFR cytoplasmic/nuclear translocation in Hippo pathway dysregulation. Results: Patients with advanced lung adenocarcinoma with membranous mutant EGFR (19del or 21 L858R) showed significantly longer progression-free survival than those with cytoplasmic mutant EGFR after gefitinib treatment. The concentration that inhibits 50% in PC-9 with cytoplasmic EGFR was higher than that in hunman non-small cell lung cancer 827 with membranous EGFR. During firstgeneration TKI resistance induction, membrane EGFR translocated to the cytoplasm/nucleus, accompanied by the Hippo pathway inhibition. Cytoplasmic EGFR and SIK2 interaction inhibited large tumor suppressor kinase 1 (LATS1) and macrophage stimulating 1 (MST1) interaction,

promoting YAP nuclear translocation. However, cells with osimertinib-induced resistance also showed EGFR translocation and lower phospho-EGF receptor but did not show Hippo pathway inhibition. Moreover, osimertinib and erlotinib could restore sensitivity to each other in resistant cells. Conclusions: Plasma/nuclear translocation of EGFR and inhibition of the Hippo pathway are some of the important mechanisms underlying the resistance induced by firstgeneration TKIs. Membrane/plasma translocation of EGFR induced by osimertinib may be another resistance phenomenon besides MNNG HOS transforming gene (c-MET) amplification, C797S mutation, and ERK pathway inhibition.
2019 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. Keywords: TKI resistance; EGFR translocation; Hippo pathway; YAP; SIK2

*Corresponding author. Disclosure: The authors declare no conflict of interest. Address for correspondence: Enhua Wang, MD, PhD, Department of Pathology, College of Basic Medical Sciences and the First Affiliated Hospital of China Medical University, No.77 Puhe Road, North New Area, Shenyang 110122, China. E-mail: [email protected] ª 2019 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. ISSN: 1556-0864 https://doi.org/10.1016/j.jtho.2019.06.014

Journal of Thoracic Oncology

Vol.

-

No.

-: ---

2 Rong et al

Introduction Of all the malignant tumors, lung cancer currently has the highest incidence and mortality, with 85% to 90% of these cancers being NSCLCs.1-3 In an Asian population, approximately one-third or more of the patients with NSCLC showed EGFR 19 del (54%) or 21 L858R (43%).4 Therefore, targeted therapy involving the use of EGFRspecific tyrosine kinase inhibitors (TKIs) gefitinib/IRESSA and erlotinib/Tarceva as the first-line drugs is generally recommended for the treatment of lung cancers with somatic mutations in EGFR. These targeted drugs have better objective response rate and progression-free survival (PFS) than chemotherapy.5,6 However, more than 50% patients are predicted to develop TKI resistance after the administration of these two drugs. More than half of the cases of resistance to the first- and second-generation TKIs have been reported to be caused by the EGFR mutation T790M; however, the mechanism of TKI resistance in the remaining 30% cases remains unclear.7 The thirdgeneration TKIs such as osimertinib/TAGRISSO not only overcome the resistance caused by the secondary T790M mutation induced by the first-generation of TKIs but also show reduced gastrointestinal and skin toxicity, compared with that observed after treatment with the first-generation TKIs.8,9 Moreover, osimertinib can cross the blood-brain barrier, thus providing a better curative effect in patients with NSCLC metastasis to the central nervous system compared with that of the first- and second-generation TKIs.10 In the study by Erbellein et al.,11 the treatment of the 19del population with thirdgeneration TKIs did not cause T790M mutation; however, seven of eight gefitinib-resistant populations and two of three afatinib-resistant populations showed the presence of the T790M mutation. In phase III FLAURA trial (NCT02296125), osimertinib showed a higher overall remission rate (osimertinib versus gefitinib and erlotinib: 80% versus 76%, respectively; p < 0.0001) and longer PFS (osimertinib versus gefitinib and erlotinib: 18.9 months versus 10.2 months, respectively; p < 0.0001) than that observed after treatment with the first-generation TKIs gefitinib and erlotinib.12 The phase III AURA3 trial (NCT02151981) recruiting patients harboring T790M mutation after treatment with firstand second-generation EGFR TKIs revealed significantly lower progression rate (osimertinib group versus platinum-pemetrexed group: 50% versus 79%, respectively) and significantly longer PFS (osimertinib group versus platinum-pemetrexed group: 10.1 months versus 4.4 months, p < 0.001) in the osimertinib group (279 patients) than that in the platinum-pemetrexed group (136 patients).13 Thus, as osimertinib showed better clinical efficacy and less adverse effects compared with

Journal of Thoracic Oncology

Vol.

-

No.

-

those with the standard of care in the first and second line, it has been approved both as a first-line drug for patients with common EGFR mutations and second-line drug for patients with T790M mutations. The main resistance mechanisms observed in patients treated with upfront osimertinib are the amplification of MNNG HOS transforming gene (c-Met), the mutation of EGFR-C797S, and the modulation of the ERK pathway.14,15 However, with a long-term treatment with osimertinib, the possibility of development of new drug resistance mechanisms remains to be elucidated. The EGFR family plays an important role in tumor formation and development. High EGFR expression can promote the occurrence and development of cancer and result in poor prognosis.16 In the past decade, several articles have reported that epidermal growth factor (EGF) can stimulate the translocation of membranous EGFR (mEGFR) into the cytoplasm, followed by its entry into the nucleus by binding to importin-b through its nuclear localization sequence.17,18 This suggests that EGFR localization changes between the cell membrane, cytoplasm, and nucleus. It has also been suggested that EGFR acts as a transcription factor in plasma/nucleus shuttle and plays an important role in tumor progression.19 Over the past 10 years, several studies have confirmed the important mechanism of nuclear EGFR (nEGFR) function. nEGFR promotes the activation of seven oncogenes (cyclin D1 [CCND1], inostitol-3phosphate synthase 1 [iNOS], MYB proto-oncogene like 2 [B-Myb], Aura Kinase A, mitochondrially encoded cytochrome c oxidase II [COX-2], DENN domain containing 4A [C-Myc], and BCRP) and promotes the development of tumors and the emergence of TKI resistance.20,21 Moreover, nEGFR can regulate RNA stability, promote protein translation, and ultimately induce resistance to radiotherapy.22-24 However, it is not clear whether mutant EGFR can enter the nucleus and whether its mechanism of nuclear translocation is the same as that of the wild-type. Its role after nuclear translocation also remains unclear. Recently, several studies have reported TKI resistance caused by inhibition of the Hippo pathway.25-28 The Hippo pathway was first discovered and studied in Drosophila melanogaster.29 Its main components include the upstream molecules (neurofibromin-2/NF2, Kidney and brain/KIBRA protein, etc.), central kinase complex (MST1/2-SAV1-LATS1/2-MOB), the main downstream effector molecules (YY1 associated protein 1 [YAP]), and the target genes in the nucleus (cellular communication network factor 2 [CTGF], cyclin E1 [CCNE], alkaline phosphatase, intestinal [IAP], etc.).30 When the Hippo pathway is activated, MST1/2 phosphorylates SAV1, which functions as a partner of MST1/2 in promoting LATS1 phosphorylation; this results in phosphorylation

---

2019

of YAP, which binds to the 14-3-3 protein and is eventually degraded.31 When the Hippo pathway is inhibited, YAP is not phosphorylated by the central kinase complex and it translocates into the nucleus, activates the target genes (CTGF, CCNE, etc.) with its co-activator (TEAD), and ultimately promotes the proliferation, invasion, and drug resistance of cancer cells.32,33 Although TKI resistance is known to be related to the inhibition of the Hippo pathway, the specific mechanism underlying Hippo pathway inhibition is not fully understood. In this study, we investigated whether (1) intracytoplasmic translocation of EGFR occurs in drug resistance induced by both first-generation TKIs and osimertinib; (2) EGFR translocation inhibited the Hippo pathway via some mechanism and ultimately leads to the nuclear translocation of YAP; (3) the resistance after TKI treatment is due to loss of its target, and if not, can the resistance induced by the first-generation TKIs be overcome by osimertinib; and (4) the first-generation TKIs can overcome the resistance caused by osimertinib. Such a detailed investigation into TKI resistance mechanism is essential to improve the prognosis evaluation and treatment strategy formulation for patients with lung cancer. Furthermore, it will also provide an experimental basis for preventing TKI resistance and developing new effective targeted drugs.

Material and Methods

TKIs Resistance Caused by mEGFR Translocation

3

Immunohistochemistry Staining Immunohistochemistry assays were performed as described previously.34 Briefly, tissue sections were incubated with mutant EGFR rabbit monoclonal antibody (EGFR E746-A750del, #2085, 1:100; EGFR L858R, #3197, 1:100; Cell Signaling Technology (CST) Inc., Danvers, Massachusetts). Positive EGFR expression was determined by the intensity of membrane and/or cytoplasm staining of tumor cells and the proportion of EGFR-positive cells in the tissue sections.35 The specimens were divided into four grades (0, 1þ, 2þ, 3þ) and were scored as follows: 0 (no staining), 1þ (light yellow staining without clear granular staining or yellow staining with clear granular staining area <10%), 2þ (yellow staining with clear granular staining area 10% or brown staining with clear granular staining area <10%), or 3þ (brown staining with clear granular staining area 10%).

Cell Lines HCC827 (EGFR 19del), PC-9 (EGFR 19del), H1975 (EGFR 21L858R mutation and EGFR T790M mutation), and HEK293 were obtained from Shanghai Cell Bank (Shanghai, China), and cultured according to the instructions of the American Type Culture Collection/ ATCC. All cell lines were authenticated by short tandem repeat DNA profiling.

Specimen Collection

Generation of TKI-Resistant Cell Lines

We randomly selected 21 puncture biopsy specimens from patients with advanced lung adenocarcinoma and EGFR mutation (19del or 21 L858R) confirmed by amplification-refractory mutation system (ARMS) and willing to receive TKI treatment. Informed consent was obtained from all the patients for analysis of their specimens. The average age of the patients was 57 years. According to the WHO 2015 lung cancer histopathologic diagnostic criteria, all 21 cases were alveolar adenocarcinoma. According to the eighty edition of the American Joint Committee on Cancer TNM staging criteria published by the International Association for the Study of Lung Cancer in 2017, all 21 cases were stage IV cancers. These cases were not treated with chemotherapy or TKIs before a definite pathologic diagnosis was determined. We detected EGFR location in the 21 lung adenocarcinoma specimens by immunohistochemistry. The patients were then treated with gefitinib (Astra Zeneca, Inc., London, United Kingdom; 250 mg/d) and followed up from the date of TKI treatment to the date of progression or death (this period was called PFS). All 21 patients were well balanced for Eastern Cooperative Oncology Group performance score (grade 1-2) and did not show central nervous system metastasis.

HCC827 cell line (EGFR localized on the membrane) was selected for inducing TKI resistance using gefitinib and erlotinib (CST Inc.) and osimertinib (AstraZeneca). For resistance induction using first-generation TKIs, 3000 cells were inoculated into 96-well plates (cells were maintained in Roswell Park Memorial Institute –1640 with 10% fetal bovine serum and were cultured at 37 C in a humid incubator with 5% CO2). Gefitinib or erlotinib with gradient concentration (2, 4, 8, 16, 32, 64, 128, and 256 nmol/L; diluted according to the manufacturer’s instructions) were added the next day, and a 0-nmol/L control group was also established. After 24 hours, the concentration that inhibits 50% (IC50) of the two drugs was determined using the MTT assay (gefitinib: 14 nmol/L, erlotinib: 20 nmol/L). Then, the hunman non-small cell lung cancer 827 (HCC827) cells were treated with the above-mentioned concentrations of gefitinib and erlotinib. These TKI-containing media were replaced with normal media after 24 hours. HCC827 acquired gefitinib and erlotinib resistance following a stepwise increasing dose of these TKIs. Every dose was repeated twice until the final concentration of 180 nmol/L, and the resultant cell lines were named HCC827/GR and HCC827/ER. HCC827/GR and HCC827/ER cells

4 Rong et al

were found to grow stably in 140-nmol/L TKIs. A similar procedure was followed for resistance induction using osimertinib. Osimertinib with concentration of 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, 6, 15, 20, 25, and 30 ng/mL was added, and a 0-ng/mL control group was set up. Osimertinib, at a concentration two times IC50, was added to the flask, and the normal medium was replaced after 6 hours. When the cells resumed their logarithmic growth phase, the same concentration of drugs was added; that is, every dose was repeated twice, followed by four times the IC50 concentration, until the final concentration of 107 ng/mL or 180 nmol/L; the resultant cell line was termed HCC827/OR. The induction time was 2 months.

Matrigel Invasion, Colony Formation, and MTT Assays For colony formation assays, cells were seeded in three 6-cm cell culture dishes (1000 cells per dish) and incubated for 12 days (cells were maintained in Roswell Park Memorial Institute –1640 with 10% fetal bovine serum and were cultured at 37 C in a humid incubator with 5% CO2). The plates were then washed with phosphate-buffered saline, stained with Giemsa, and the colonies were counted. Matrigel invasion and MTT assays were performed as described previously.36

Journal of Thoracic Oncology

Vol.

-

No.

-

(Cambridge, Massachusetts) and those against p-ERK 1/2 (sc-81492), CyclinE (sc-377100, 1:200), and CTGF (sc-34772, 1:200) were from Santa Cruz Biotechnology (Santa Cruz, California). Secondary antibodies were purchased from CST Inc. Expression was quantified using densitometry and ImageJ software. Protein complexes obtained by immunoprecipitation were detected and analyzed by Hoogen Biotechnology Company (Shanghai, China) for mass spectrometry.

Transfection and Interference pCMV6-Entry (#PS100001), YAP1 (#RC231269), SIK2 (#RC221327), and MST1 (#RC229352) were purchased from Origene Technologies Inc. (Rockville, Maryland). Flag-LATS1 (#18971) and pGL3b_8xGTIICluciferase (#34615) were obtained from Addgene (Cambridge, Massachusetts). siEGFR (sc-29301), siYAP (sc-38637), siSIK2 (sc-44364), siLATS1 (sc-35797), siMST1 (sc-39570), shYAP (sc-38637-SH), and shSIK2 (sc-44264-SH) were purchased from Santa Cruz Biotechnology, Inc. pRL-TK (#E2241) was purchased from Promega (Madison, Wisconsin). Lipofectamine 3000 (Invitrogen, Carlsbad, California) was used for transfection and interference. G418 (#A1720, Sigma, St. Louis, Missouri) was used for selecting stably transfected cell lines.

Cell Apoptosis Assay HCC827, HCC827/GR, and HCC827/ER cells were seeded in 6-cm cell culture dishes and incubated for 2 days. Then, the cells in each tube were suspended in 500 mL binding buffer, and 10 mL Annexin V-FITC/APC and 10 mL propidium iodide were added sequentially (4A Biotech, Beijing, China).

Western Blot, Immunoprecipitation, and Mass Spectrometry The western blot (WB) and immunoprecipitation assays were performed as described previously.37 The primary antibodies against EGFR (#4267S, 1:500), phospho-EGF receptor (p-EGFR) (#3777S, 1:500), EGFR E746-A750del (#2085S, 1:500/WB, 1:100/IP), EGFR L858R (#3197S, 1:500), YAP (#14074, 1:500), p-YAP (#13008, 1:500), MST1 (#14946, 1:500), p-MST1 (#49332, 1:500), LATS1 (#3477, 1:500), p-LATS1 (#9157, 1:500), SIK2 (#6919, 1:500/WB, 1:100/IP), pStat5 (#4322, 1:500), p-Gab1 (#12745, 1:500), p-Shc (#2434, 1:500), p-PLCg1 (#2821, 1:500), SAV1 (#1330, 1:500/WB, 1:100/IP), ATP1A1 (#23565, 1:500), Histone H3 (#4499, 1:500), GAPDH (#5174, 1:10000), and phospho-p44/42 MAPK (Erk1/2, #8544, 1:500) were purchased from CST Inc. Primary antibodies against LaminB1 (#ab16048, 1:1000/IB) and tubulin (#ab52866, 1:1000/IB) were purchased from Abcam

Immunofluorescence Staining HCC827, H1975, PC-9, and TKI-resistant cell lines were fixed, permeabilized, and incubated with primary antibodies (SIK2, sc-393139, 1:50, Santa Cruz Biotechnology, Inc; EGFR E746-A750del, 1:100, YAP 1:100) and fluorescein isothiocyanate–conjugated or tetramethylrhodamine isothiocyanate–conjugated secondary antibodies (Zsbio Store, Beijing, China). Nuclei were stained with 40 ,6-diamidino-2-phenylindole, and cells were observed under a confocal microscope (Carl Zeiss, Thornwood, New York).

Dual-Luciferase Assays Cells were transfected with pGL3b_8xGTIIC (with the activity of TEAD) in 24-cm cell culture dishes and collected after 48 hours. Dual-luciferase reporter gene detection system was used to detect the YAP/TAZ activity, and cells transfected with pRL-TK were used as control.

Transplantation of Tumor Cells Into Nude Mice Four-week-old female BALB/c nude mice were purchased from Charles River (Beijing, China). The mice were treated according to experimental animal ethics guidelines issued by China Medical University; the study was approved by the Institutional Animal Research

---

2019

Committee of China Medical University. The 4-week-old female nude mice (16 g to 20 g) were randomly selected and subcutaneously inoculated with 1  107/0.2 mL HCC827, HCC827/ER, HCC827/ER-shYAP, or HCC827/ ER-shSIK2 cells. When the tumor volume reached approximately 0.1 cm3 (tumor volume (cm3) ¼ a2  b/2), the mice were randomly divided into two groups, with five mice in each group. Erlotinib suspension (60 mg/kg), prepared using 150-mg erlotinib tablets (Tarceva, Roche, Basel, Switzerland) and sodium carboxymethyl cellulose, was administered intragastrically to the experimental group, whereas sodium carboxymethyl cellulose of the same volume was administered intragastrically to the control group.38 The changes in tumor volume were measured and recorded every 3 days.

RNA Extraction and Reverse Transcription Quantitative Polymerase Chain Reaction The RNA extraction and reverse transcription quantitative polymerase chain reaction (RT-qPCR) assays were performed as described previously.37 Quantitative real-time PCR was performed using SYBR Green PCR master mix (Applied Biosystems, Inc., Foster City, California) in a total volume of 20 mL on the 7900 HT Fast Real-Time PCR System (Applied Biosystems, Inc.) as follows: 95 C for 30 seconds, 40 cycles at 95 C for 5 seconds, and 60 C for 30 seconds. A dissociation step was performed to generate a melting curve to confirm amplification specificity. The relative levels of gene expression were represented as D Ct¼Ctgene-Ctreference, and the fold change of the gene expression was calculated by the 2-DDCt method. Experiments were performed in triplicate. The relative transcript levels of genes were normalized to GAPDH mRNA levels. The primers used were: BCRP: 5ʹ-CAGGTGGAGGCAAATCTTCGT-3ʹ (forward) 5ʹ-ACCCTGTTAATCCGTTCGTTTT-3ʹ (reverse) B-Myb: 5ʹ-CCGGAGCAGAGGGATAGCA-3ʹ (forward) 5ʹ-CAGTGCGGTTAGGGAAGTGG-3ʹ (reverse) CCND1: 5ʹ-GCTGCGAAGTGGAAACCATC-3ʹ (forward) 5ʹ-CCTCCTTCTGCACACATTTGAA-3ʹ (reverse) C-Myc: 5ʹ-GGCTCCTGGCAAAAGGTCA-30 (forward) 5ʹ-CTGCGTAGTTGTGCTGATGT-3ʹ (reverse) iNOS: 5ʹ-TTCAGTATCACAACCTCAGCAAG-3ʹ (forward) 5ʹ-TGGACCTGCAAGTTAAAATCCC-3ʹ (reverse) GAPDH: 50 -GGAGCGAGATCCCTCCAAAAT-30 50 -GGCTGTTGTCATACTTCTCATGG-30

DNA Sequencing HCC827/ER and HCC827/OR cell lines were washed three times with phosphate-buffered saline and centrifuged at 200 rpm. The supernatant was discarded and the collected cells were delivered to Hoogen

TKIs Resistance Caused by mEGFR Translocation

5

Biotechnology Company (Shanghai, China) for DNA sequencing. The primer sequence was as follows: EGFR900: 50 -TGAAGCTGTCCAACGAATG-30 (forward) 50 -AGGGCGTTCTCCTTTCTCCAG-30 (reverse)

Fluorescence In Situ Hybridization HCC827, HCC827/ER, and HCC827/OR cell lines were fixed and sent to the Department of Pathology, the First Affiliated Hospital of China Medical University for fluorescence in situ hybridization (FISH) analysis. Currently, there is no uniform standard for c-MET FISH result interpretation; therefore, the following criteria were used for the result interpretation: (1) Ratio greater than or equal to 1.8 was considered as a positive result of c-MET signal connection in clusters, indicating the amplification of the c-MET gene in the sample. Ratio greater than or equal to 5.0 was considered to indicate high amplification; a ratio greater than 2.2 but less than 5.0, moderate amplification, and ratio greater than or equal to 1.8 but less than or equal to 2.2, low amplification. (2) Ratio less than 1.8 was considered as a negative result, indicating that the c-MET gene in the sample has not been amplified. We counted 30 cells, recorded the number of c-MET (red) signals and CSP7 (green) signals per cell, and calculated the ratios (ratio ¼ number of red signals per 30 cells / number of green signals per 30 cells).

Statistical Analysis SPSS 22.0 software (SPSS, Chicago, Illinois) was used for all analyses. Correlations between EGFR location and PFS of patients were assessed using the chi-square test. The gray value of WB was detected (Image Lab) and compared using the Student’s t-test; p value less than 0.05 (*: Fig. 1; Fig. 6), p value less than 0.01 (**: Fig. 2), and p value less than 0.001 (***: Figs. 2-4; Fig. 6) indicated significance.

Results Patients With Membrane Localization of EGFR Have a Longer PFS Than Those With Only Cytoplasmic Localization of EGFR During Gefitinib Treatment We screened the puncture biopsy specimens of advanced lung adenocarcinoma with EGFR mutation (19del or 21L858R mutation) by ARMS and detected EGFR location using specific antibodies against mutant EGFR. As reported previously, we found that mutant EGFR localized to the cell membrane and the cytoplasm.35 To verify whether this different localization affects the sensitivity of patients to TKIs, we selected 21 patients who received standard TKI treatment with their consent. Immunohistochemistry confirmed that all 21

6 Rong et al

specimens were mutant EGFR-positive (17 cases: þþþ [strong positive], 4 cases: þþ [positive]). Immunohistochemistry results are consistent with the ARMS results with respect to the type of EGFR mutations (19del: 13 cases, 21 L858R: 8 cases). Of 21 cases, 13 (19del: 8 cases, 21 L858R: 5 cases) and 8 (19del: 5 cases, 21 L858R: 3 cases) showed EGFR localization to the membrane and cytoplasm, respectively. There was no connection between EGFR location and mutant type (p > 0.05). All patients were administered gefitinib (250 mg/d) orally and were checked every 3 months. Therapeutic observation showed that after gefitinib treatment, the minimum and maximum PFS of patients with cytoplasmic EGFR was 32 and 374 days, respectively (Fig. 1A), with an average of 202.5 ± 96 days. However, the minimum and maximum PFS of patients with membranous EGFR was 271 and 681 days, respectively (Fig. 1B), with an average of 411 ± 107.6 days, which is significantly longer than that of patients with cytoplasmic EGFR (Figs. 1C and D) (p < 0.05). These results suggest that cytoplasmic localization of EGFR may be involved in the shortening of effective gefitinib therapy duration. Therefore, we speculated that the different localization of EGFR in cancer cells may be associated with TKI resistance.

Induction of Resistance Using First-Generation TKIs in Lung Cancer Cells is Accompanied With EGFR Membranous/Cytoplasmic/Nuclear Translocation To determine whether the mEGFR/cytoplasmic mutant EGFR (cEGFR) is related to TKI resistance, we first selected different cell lines with EGFR gene mutation and mutant EGFR protein expressed in membrane or cytoplasm. Laser confocal analysis using EGFR (E746A750del) antibody revealed the membranous and cytoplasmic localization of mutant EGFR protein in HCC827 and PC-9 cells, respectively; whereas the use of 21 L858R mutant antibody showed the cytoplasmic localization of mutant EGFR protein in H1975 cells (Fig. 2A). Because both HCC827 and PC-9 contain EGFR 19del mutation, we tested the sensitivity of these cell lines to gefitinib and erlotinib. The results showed that IC50 of gefitinib in HCC827 cells (20.7 nmol/L) was significantly lower than that in PC-9 (40.1 nmol/L, p < 0.01); IC50 of erlotinib in HCC827 cells (22 nmol/L) was also significantly lower than that in PC-9 (64.69 nmol/L, p < 0.001). Therefore, it is further suggested that the difference in localization of mutant EGFR in membrane/ cytoplasm affects TKI sensitivity of cells (Fig. 2B). To investigate whether mutant EGFR expression is accompanied with its membranous/cytoplasmic/nuclear translocation during the induction of TKI resistance, we induced TKI resistance in the HCC827 cell line. The

Journal of Thoracic Oncology

Vol.

-

No.

-

initial concentration of gefitinib was 14 nmol/L and erlotinib was 20 nmol/L, whereas the final concentration for both was 180 nmol/L. The induction time for gefitinib was 6 months and erlotinib was 7 months. Colony formation assay, MTT assay, Matrigel invasion assay, and cell apoptosis assay were performed to investigate the differences between TKI-resistant cell lines and the HCC827 cell line. The two TKI-resistant cell lines showed significantly enhanced proliferation, invasion, and antiapoptotic ability (Supplementary Figs. 1A–D), and the IC50 of TKI-resistant cells were higher than that of HCC827 cells (p < 0.001) (Fig. 2F, Supplementary Fig. 2B). We observed changes in localization of mEGFR during the induction of TKI resistance. With the increase in TKI concentration, mEGFR gradually translocated into the cytoplasm (gefitinib: 120 nmol/L for 4 weeks; erlotinib: 120 nmol/L for 5 weeks). With further increase in TKI concentration, the cEGFR started to enter the nucleus (gefitinib: 180 nmol/L for 7 weeks; erlotinib: 180 nmol/L for 10 weeks) (Fig. 2C, Supplementary Fig. 2A). Furthermore, membrane-cytoplasm isolation and cytoplasm-nucleus isolation confirmed that mEGFR expression was lower whereas cEGFR and nuclear EGFR (nEGFR) expression was higher in TKI-resistant cells, compared to the corresponding values in HCC827 cells (Fig. 2D). RT-qPCR was used for comparing HCC827 and TKI-resistant cells with respect to the transcriptional levels of nuclear oncogenes. It was found that the transcription levels of BCRP, b-Myb, c-Myc, and CCND1 increased simultaneously in HCC827/GR when EGFR entered the nucleus (Fig. 2E). However, in HCC827/ER, only the transcription level of BCRP was increased (Supplementary Fig. 2C).

First-Generation TKIs Induce Resistance via cEGFR Translocation and Inhibition of Hippo Pathway Although it has been reported that inhibition of the Hippo pathway is related to TKI resistance, the precise mechanism remains to be explored. Immunoblotting was applied to detect the difference in the expression of Hippo pathway–related proteins between the HCC827 and TKI-resistant cells. The expression of p-MST1, p-LATS1, and p-YAP1 was downregulated, whereas the expression of YAP1 and nuclear YAP1 was upregulated (Figs. 3A, C, and D). Notably, the levels of cytosolic and nEGFR phosphorylation decreased in TKI-resistant cells, compared to those in HCC827 (Fig. 3B). Dual luciferase assay revealed that the activity of co-activator TEAD in nucleus increased after drug resistance (Fig. 3E) (p < 0.001). Immunoblotting analysis revealed that the expression of nuclear target genes CTGF and CCNE of the Hippo pathway was also upregulated (Figs. 3F and G) in the HCC827/GR and HCC827/ER cells, compared to that

---

2019

TKIs Resistance Caused by mEGFR Translocation

7

8 Rong et al

in the HCC827 cells, indicating that TKI resistance induced by EGFR translocation may be closely related to the inhibition of the Hippo pathway. In particular, TKI treatment in TKI-resistant cell lines continues to inhibit the phosphorylation of EGFR and its downstream proteins (Figs. 3F and G). Collectively, these results reveal that TKIs continue to inhibit EGFR phosphorylation, suggesting that TKI resistance at this time is not caused by the decrease of EGFR dephosphorylation ability of TKIs, but may be related to Hippo pathway inhibition via EGFR translocation.

Interaction of EGFR and SIK2 in Cytoplasm Inhibits the Hippo Pathway We found that during the induction of TKI resistance in HCC827, EGFR first translocated from the cell membrane into the cytoplasm, and then the Hippo pathway was inhibited. Therefore, to identify the proteins that may interact with cytoplasmic EGFR in TKI-resistant cells, we used EGFR (E746-A750del) antibody for immunoprecipitation, and the obtained proteins were detected by mass spectrometry. We found an interaction between EGFR and SIK2 (Supplementary Fig. 3A). Using laser confocal microscopy, we further confirmed that in HCC827, EGFR was mainly localized in the cell membrane, whereas SIK2 was mainly localized in the cytoplasm. However, in HCC827/GR and HCC827/ER, EGFR and SIK2 co-localized in the cytoplasm (Fig. 4A). Immunocoprecipitation confirmed that EGFR and SIK2 interacted with each other (Fig. 4B). Therefore, to confirm the relationship between EGFR, SIK2, and the Hippo pathway, we first transfected SIK2 in HEK293 cell line, which shows high transfection efficiency, and found that SIK2 could reduce the binding of LATS1 to MST1 (Supplementary Fig. 3B). Therefore, we transfected HCC827, HCC827/ER, and HCC827/GR cells, and compared these cells lines. The results revealed that the binding of SIK2 to SAV1 increased but the binding of LATS1 to MST1 decreased in TKI-resistant cells compared to HCC827 cells (Fig. 4C, Supplementary Fig. 3C). Furthermore, after interfering with EGFR expression in TKI-resistant cells, SIK2 binding to SAV1 decreased but LATS1 binding to MST1 increased (Fig. 4D, Supplementary Figs. 3D and E). Moreover, the IC50 in SIK2-transfected TKI-resistant cells was higher and that in TKI-resistant cells after interference with SIK2 was lower than the IC50 of HCC827/GR and

Journal of Thoracic Oncology

Vol.

-

No.

-

HCC827/ER cells. The efficiency of SIK2 transfection and silencing were detected by immunoblotting analysis (p < 0.001) (Figs. 4E–H, Supplementary Fig. 3F). Notably, the binding ability of SIK2 to SAV1 weakened and that of LATS1 to MST1 increased only by EGFR silencing in TKIresistant cells, whereas the binding of SIK2 to SAV1 increased and LATS1 to MST1 weakened by SIK2 transfection in TKI-resistant cells. Moreover, the binding ability of SIK2 to SAV1 did not increase and the binding ability of LATS1 to MST1 did not decrease in EGFRsilenced and SIK2-transfected cell line compared to the SIK2-transfected cell line (Fig. 4I, Supplementary Figs. 3G and H). These results suggest that cEGFR can enhance the binding of SIK2 to SAV1 by interacting with SIK2 (possibly combined with SIK2) and inhibit the Hippo pathway. Next, we validated the effect of cEGFR on the Hippo pathway by immunoblotting analysis. After EGFR silencing in TKI-resistant cells, phosphorylation of LATS1 and YAP molecules of the Hippo pathway increased, and the expression levels of nuclear target genes CTGF and CCNE decreased, but the phosphorylation of MST and expression of YAP did not change significantly (Supplementary Figs. 4A and C). Similar results were observed after SIK2 silencing in the two TKI-resistant cells (Supplementary Figs. 4B and D). Thus, it is suggested that cEGFR regulates the Hippo pathway mainly by regulating LATS1 phosphorylation and that the inhibitory effects of cEGFR and SIK2 on the Hippo pathway are similar after inducing TKI resistance.

EGFR Transport Into the Nucleus Occurs via Binding to YAP Our experiments revealed that EGFR and YAP entered the nucleus simultaneously when inducing TKI resistance in cells (HCC827/GR: 24 weeks; HCC827/ER: 28 weeks). The presence of YAP was also detected by mass spectrometry after immunoprecipitation with EGFR antibody (E746-A750del) (Supplementary Fig. 2A). We also observed that both EGFR and YAP co-localized in the cytoplasm and nucleus in TKI-resistant cells (Fig. 5A) and that the two molecules interacted with each other (Fig. 5B). After YAP silencing in TKI-resistant cells, it was found that levels of nEGFR decreased but cEGFR increased (Figs. 5C and E), whereas the reverse result was obtained after YAP transfection in TKI-resistant cells (Figs. 5D and F). When LATS1 and MST1 were transfected into the TKI-resistant cells to activate the Hippo

Figure 1. Difference in localization of EGFR leads to significant differences in progression-free survival (PFS) after gefitinib treatment. (A) In patients with cytoplasmic localization of EGFR, PFS was observed for a minimum of 32 days and a maximum of 374 days (original magnification 400; amplification-refractory mutation system [ARMS] method proves that there are two separated peaks). (B) In patients with membranous EGFR localization, PFS was observed for a minimum of 271 days and a maximum of 681 days (original magnification 400; ARMS method proves that there are two separated peaks). (C, D) The average PFS of patients with membranous EGFR localization (411 ± 107.6 days) was significantly longer than that of patients with mainly cytoplasmic EGFR localization (202.5 ± 96 days; *p < 0.05).

---

2019

pathway, it was found that nuclear translocation of both YAP and EGFR decreased. After the co-transfection of LATS1 and MST1, the levels of nuclear YAP and nEGFR

TKIs Resistance Caused by mEGFR Translocation

9

decreased further (Supplementary Figs. 5A and B). Moreover, when LATS1 and MST1 silencing was introduced in TKI-resistant cells to inhibit the Hippo pathway,

10 Rong et al

the opposite results were obtained (Supplementary Figs. 5C and D). When SIK2 silencing was introduced in TKI-resistant cells, the Hippo pathway was activated and nuclear translocation of YAP and EGFR decreased (Supplementary Figs. 5E and F).

Partial Restoration of Erlotinib Sensitivity by Knockdown of YAP or SIK2 in HCC827/ER Because the expression of YAP in HCC827/ER cells was higher than that in HCC827/GR cells (Fig. 3A), we selected HCC827/ER cells to establish shYAP stable cell lines (ER-shYAP) and shSIK2 stable cell lines (ERshSIK2). Because the rate of tumorigenesis of cell lines (HCC827, HCC827/ER, HCC827/ER-shYAP, and HCC827/ER-shSIK2) used in vivo is different, it is impossible to observe the effect of the erlotinib on each cell line simultaneously or at the same tumor size after subcutaneous inoculation; therefore, we used the grouping observation method. Female nude mice were subcutaneously inoculated with HCC827, HCC827/ER, HCC827/ER-shYAP, or HCC827/ER-shSIK2 cells, and tumor volume changes were measured every 3 days after inoculation in each group for a period of 15 days. When the tumor volume reached 0.1 cm3, the mice were randomly divided into two groups and subjected to erlotinib treatment. In nude mice inoculated with HCC827 cells, the volume and weight of tumors decreased with the prolongation of treatment time (0.225 ± 0.014 g versus 0.040 ± 0.013 g, p < 0.001) (Figs. 6A and B). In nude mice inoculated with HCC827/ER cells, the time of reaching 0.1 cm3 was significantly shortened (14 days) (p < 0.05). Compared with the control group, the experimental groups showed a slow increase in tumor volumes with increase in treatment time and decreased weight of tumors, but the difference was not statistically significant (0.176 ± 0.058 g versus 0.112 ± 0.0092 g, p ¼ 0.06) (Figs. 6C and D). These results indicate that HCC827/ ER cells with cytoplasmic translocation of EGFR are resistant to erlotinib. Therefore, to prove that the TKI

Journal of Thoracic Oncology

Vol.

-

No.

-

resistance after EGFR translocation was due to the binding of EGFR with SIK2 and the inhibition of YAP phosphorylation, we established stable YAP and SIK2 knockdown HCC827/ER cell lines (HCC827/ER-shYAP and HCC827/ER-shSIK2, respectively). These cells were then subcutaneously inoculated in nude mice and the sensitivity to erlotinib was observed. For both HCC827/ER-shYAP and HCC827/ER-shSIK2 cells, the volume of the tumors decreased slowly with increased treatment time (Figs. 6E and F) in the experimental group, compared to the control group. Moreover, the weights of the tumors in the experimental groups were significantly lower than those in the corresponding control groups (HCC827/ER-shYAP: 0.283 ± 0.076 g versus 0.134 ± 0.051 g, p < 0.05; HCC827/ER-shSIK2: 0.3108 ± 0.0269 g versus 0.1412 ± 0.0124 g, p < 0.001) (Figs. 6G and H). The above experimental results were also confirmed via in vitro experiments. Compared to the control HCC827/ER and HCC827/GR cells, YAP-knockdown HCC827/ER and HCC827/GR cells showed decreased cell proliferation, invasiveness, and IC50. In contrast, YAP-transfected HCC827/ER and HCC827/GR cells showed increased proliferation, invasiveness, and IC50 (Supplementary Figs. 6A–J). Furthermore, compared to the control HCC827/ER and HCC827/GR cells, SIK2knockdown HCC827/ER and HCC827/GR cells showed decreased cellular proliferation and invasiveness. In contrast, the SIK2-transfected HCC827/ER and HCC827/ GR cells showed increased proliferation and invasiveness (Supplementary Figs. 7A–F, Supplementary Table 1). Collectively, these results suggest that YAP or SIK2 knockdown in HCC827/ER cell line restores the erlotinib sensitivity of TKI-resistant cells to some extent.

Osimertinib-Induced Resistance Was Also Expressed as cEGFR Translocation but Did Not Inhibit Hippo Pathway We first used osimertinib to detect the drug-sensitive concentrations for the HCC827 and PC-9 cell lines, which

Figure 2. Induction of gefitinib resistance in HCC827 cell line and the membranous/cytoplasmic/nuclear translocation of EGFR. (A) Laser confocal microscopy images (original magnification 400) showing mainly the cell membrane localization of mutant EGFR in HCC827 (EGFR E746-A750del) cell line and mainly the cytoplasmic localization of mutant EGFR in PC-9 (EGFR E746-A750del) and H1975 (EGFR 21L858R) cell lines. (B) The tyrosine kinase inhibitor (TKI) sensitivity of HCC827 and PC-9 cell lines as tested by gefitinib and erlotinib treatment. TKI sensitivity of HCC827 (concentration that inhibits 50% [IC50]: gefitinib 20.7 nmol/L, erlotinib 22 nmol/L) was higher than that of PC-9 (gefitinib 40.1 nmol/L, p < 0.01; erlotinib 64.69 nmol/L, p < 0.001). (C) Laser confocal microscopy images (original magnification 400) showing that increase in gefitinib concentration was associated with the gradual translocation of EGFR (E746-A750del) from cell membrane to cytoplasm and finally into the nucleus. (D) Results of membranous-cytoplasmic protein isolation and nuclear-cytoplasmic protein isolation. The expression of mutant EGFR on the cell membrane decreased, whereas that in the cytoplasm and nucleus increased in HCC827/GR and HCC827/ER cells compared to HCC827 cells. ATP1A1, Tublin, and LaminB1 were used as membrane, cytoplasm, and nucleus loading controls, respectively. (E) Reverse transcriptase quantitative polymerase chain reaction results showing that the transcriptional level of nuclear oncogenes increased significantly (***p < 0.001) when EGFR translocated into the nucleus in HCC827/GR cells compared to that in HCC827 cells. (F) Results of the MTT assay showing that IC50 in HCC827/GR cells was significantly higher than that in HCC827 cells (***p < 0.001). DAPI, 40 ,6-diamidino-2-phenylindole; HCC827, hunman non-small cell lung cancer 827.

---

2019

TKIs Resistance Caused by mEGFR Translocation

11

Figure 3. Inhibition of the Hippo pathway after induction of resistance using first-generation tyrosine kinase inhibitors (TKIs). (A) Western blot (WB) showing that the phosphorylation of the Hippo pathway-related factors MST1, LATS1, and YAP decreased but the expression of total YAP increased in HCC827/GR and HCC827/ER cells compared to HCC827 cells. (B) WB showing the levels of cytosolic and nuclear EGFR phosphorylation decreased in TKI-resistant cells, compared to those in HCC827. (C,D) Laser confocal microscopy (original magnification 400) and nuclear-cytoplasmic protein isolation images showing that nuclear YAP increased in HCC827/GR and HCC827/ER cells compared to HCC827 cells. (E) Results of the dualluciferase assay confirming the increase of TEAD activity in HCC827/GR and HCC827/ER cells (*** p < 0.001). (F,G) WB results confirming that the phosphorylation level of EGFR and its major downstream proteins remained at a low level after the successful induction of TKI resistance and the expression of YAP nuclear target genes CTGF and CCNE was higher in HCC827/ GR and HCC827/ER cells than that in HCC827 cells. DAPI, 40 ,6-diamidino-2-phenylindole; HCC827, hunman non-small cell lung cancer 827.

12 Rong et al

harbor the EGFR 19del mutation. The IC50 of HCC827 with membrane-localized EGFR was also significantly lower than that of the PC-9 with plasma-localized EGFR (HCC827 versus PC-9: 5.590 ± 0.299 ng/mL versus 8.842 þ 0.405 ng/mL, p < 0.01) (Supplementary

Journal of Thoracic Oncology

Vol.

-

No.

-

Fig. 8A). It is suggested that the different localization of EGFR (membrane/cytoplasm) is related to sensitivity to osimertinib. Next, we induced osimertinib resistance in HCC827 using the high-dose shock method. The established drug-resistant cell line, HCC827/OR, showed

---

2019

increased IC50 (HCC827 versus HCC827/OR: 6.658 ± 1.060 ng/mL versus 34.20 ± 2.123 ng/mL, p < 0.001), colonization (HCC827 versus HCC827/OR: 218.3 ± 17.85 versus 463.0 ± 34.24, p < 0.01), invasion (HCC827 versus HCC827/OR: 258.3 ± 21.87 versus 844.0 ± 33.56, p < 0.001), and proliferation (Supplementary Figs. 8B– D). Confocal laser microscopy revealed that EGFR on the cell membrane had translocated into the cytoplasm, similar to that observed in cells resistant to firstgeneration TKIs. Membrane/plasma separation experiments confirmed the decreased expression of mEGFR, increased expression of cEGFR, and decreased levels of pEGFR in HCC827/OR cells (Supplementary Figs. 8E and F). Using luciferase reporter gene, we confirmed that the activity of the Hippo pathway did not change significantly after resistance induced by osimertinib (p ¼ 0.0561.7) (Supplementary Fig. 8G). DNA sequencing was performed to detect the changes in EGFR gene in the TKI-induced resistant cells. T790M (HCC827/ER) and C797S (HCC827/OR) mutations were not found (Supplementary Figs. 9A and B). FISH and WB revealed neither c-MET amplification (Ratio for HCC827 ¼ 0.84 [negative]; for HCC827/ER ¼ 0.73 [negative]; for HCC827/OR ¼ 0.77 [negative]) nor change of ERK protein in HCC827/OR cells (Supplementary Figs. 9C and D). Next, we treated HCC827/ER cells with osimertinib and found that the sensitivity of HCC827/ER cells to osimertinib was restored compared with that to erlotinib (IC50: erlotinib versus osimertinib: 96.46 ± 3.812 nmol/ L versus 15.87 ± 2.617 nmol/L, p < 0.001) (Supplementary Fig. 9E). Similarly, the sensitivity of HCC827/OR cells treated with erlotinib was restored to some extent (IC50: osimertinib versus erlotinib: 92.07 ± 7.684 nmol/L versus 20.58 ± 0.916 nmol/L; p < 0.001) (Supplementary Fig. 9F). These results suggest that the roles and mechanism of cytoplasmic translocation of EGFR caused by first-generation TKIs and osimertinib were different.

Discussion Although several studies have focused on the mechanisms underlying TKI resistance and the involvement of

TKIs Resistance Caused by mEGFR Translocation

13

cEGFR, nEGFR, and the Hippo pathway has been shown, the precise link between TKI resistance and Hippo pathway inhibition is not yet clear. By observing the therapeutic effect of gefitinib, we found that the PFS of patients with membranous localization of mutant EGFR was significantly longer than that of patients with cytoplasmic localization mutant EGFR. Therefore, we speculated that the patients with same EGFR mutations but different subcellular localizations may have different sensitivity to TKIs. In vitro, when first-generation TKIs (gefitinib and erlotinib) and a third-generation TKI (osimertinib) were used to induce TKI resistance, the IC50 in cells with mutant mEGFR (HCC827 cells) was lower than that in cells with mutant cEGFR (PC-9 cells). We found that the emergence of TKI resistance (increased IC50, invasion, and anti-apoptotic ability) was associated with the translocation of mEGFR into the cytoplasm/nucleus. Membranous-cytoplasmic protein isolation and nuclearcytoplasmic protein isolation also confirmed that the expression of mEGFR decreased and the expression of cEGFR increased. In particular, increased nEGFR expression was also observed in cells with resistance induced by first-generation TKIs, and this result was consistent with those observed in clinical cases. Although TKIs have been reported to inhibit the Hippo pathway and lead to TKI resistance, the specific molecular regulatory mechanism remains unclear.39 Here, we revealed the mechanism underlying the inhibition of the Hippo pathway by first-generation TKIs (Supplementary Fig. 10). We found that cEGFR showed enhanced binding ability to SAV1 via interaction with SIK2 and that it inhibited the interaction between LATS1 and MST1, reduced YAP phosphorylation, increased the nuclear translocation of YAP, and ultimately inhibited the Hippo pathway. Using EGFR- or SIK2-knockdown TKI-resistant cells, we proved that cEGFR affected pLATS1 and p-YAP. SIK2 or YAP knockdown in TKIresistant cells partially restored sensitivity to TKIs both in vivo and in vitro. These results suggest that the cytoplasmic/nuclear translocation of EGFR after TKI treatment promotes cell proliferation and invasion by inhibiting the Hippo pathway, which reduces or

Figure 4. Inhibition of the Hippo pathway by cytoplasmic EGFR (cEGFR) via SIK2. (A) Laser confocal microscopy (original magnification 400) results showing the cell membrane and cytoplasmic localization of EGFR and SIK2, respectively, in hunman non-small cell lung cancer 827 (HCC827) cells. However, in homologous tyrosine kinase inhibitor (TKI)–resistant cells (HCC827/ER and HCC827/GR), EGFR was translocated into the cytoplasm and co-localized with SIK2 (without Triton punching). (B) Immunocoprecipitation results showing the interaction between EGFR and SIK2. (C) Immunoprecipitation (quantitative) results showing that SIK2 binding to SAV1 increased and LATS1 binding to MST1 decreased in HCC827/ER cells compared to HCC827 cells. (D) After EGFR interference, SIK2 binding to SAV1 decreased and LATS1 binding to MST1 increased in HCC827/ER cells compared to HCC827 cells. (E-H) MTT assay results indicating that SIK2 transfection in TKI-resistant cells increased their concentration that inhibits 50% (IC50) (p < 0.001), whereas SIK2 knockdown decreased their IC50 (p < 0.001). (I) Immunocoprecipitation results for HCC827/ER cells showing that EGFR silencing decreased the binding of SIK2 to SAV1 but increased the binding of LATS1 to MST1; SIK2 transfection increased the binding of SIK2 to SAV1 but decreased the binding of LATS1 to MST1. After simultaneous EGFR silencing and SIK2 transfection, the binding of SIK2 to SAV1 did not increase and the binding of LATS1 to MST1 did not decrease compared to that in the SIK2-transfected group.

14 Rong et al

Journal of Thoracic Oncology

Vol.

-

No.

-

Figure 5. EGFR translocation into the nucleus by interacting with YAP. (A) Laser confocal microscopy images (original magnification 400) showing that EGFR and YAP mainly co-localized in the nucleus of TKI-resistant cells. (B) Immunocoprecipitation (quantitative) analysis showing increased binding of EGFR and YAP in tyrosine kinase inhibitor (TKI)–resistant cell lines. (C, E) The results of nuclear-cytoplasmic protein isolation and laser confocal microscopy (original magnification 400) showing that nuclear EGFR decreased after YAP silencing in TKI-resistant cells. (D, F) Nuclearcytoplasmic protein isolation and laser confocal microscopy results showing that YAP transfection increased nEGFR levels in TKI-resistant cells.

---

2019

TKIs Resistance Caused by mEGFR Translocation

15

Figure 6. Partial restoration of erlotinib sensitivity after stable knockdown of YAP or SIK2 in hunman non-small cell lung cancer 827 (HCC827)/ER cells. (A, B) HCC827 cells were subcutaneously inoculated in nude mice. When erlotinib was administered to the experimental group, the volume and weight of the tumors were significantly reduced (*** p < 0.001). (C, D) When mice were inoculated with HCC827/ER cells, erlotinib treatment resulted in minor decrease in tumor volume and weight (not significant [n.s.]: p > 0.05). (E–H) When nude mice were inoculated with HCC827/ER-shYAP or HCC827/ER-shSIK2 cells, the volume and weight of the tumors decreased significantly in the erlotinib-treated group, compared with the control group (*p < 0.05; ***p < 0.001).

16 Rong et al

counteracts the biological effects caused by the inhibition of EGFR phosphorylation by TKI inhibitors, thus exhibiting TKI resistance. The nuclear translocation of EGFR is related to TKI resistance; however, the precise mechanism underlying the entry of EGFR into the nucleus remains unclear. In the present study, we found that the nuclear translocation of cEGFR coincided with that of YAP. Simultaneously, we found that both EGFR and YAP were co-localized in the cytoplasm and nucleus, and that these two molecules interacted with each other. Furthermore, in TKI-resistant cells, YAP knockdown resulted in decreased nEGFR and increased cEGFR, whereas the opposite result was observed after YAP transfection. Importantly, the transcriptional activity of the downstream target genes (BCRP, b-Myb, c-Myc, and CCND1) of EGFR increased after nuclear translocation. These results suggest that the mutant EGFR is transported into the nucleus by binding with YAP. nEGFR initiates the transcription of its downstream oncogenes while YAP initiates the transcriptional activities of CTGF and CCNE, ultimately leading to the development of TKI resistance. Another important finding is that although TKI resistance is associated with translocation of EGFR, it is not associated with EGFR phosphorylation. After the cytoplasmic/nuclear translocation of EGFR during induction of TKI resistance, the phosphorylation of EGFR decreased compared with that observed before inducing TKI resistance, and the phosphorylation level of the proteins downstream of pEGFR (p-Gab1, p-Shc, p-PLCr1, and p-Stat5) also decreased after the successful induction of TKI resistance. These results indicate that TKIs continue to inhibit EGFR phosphorylation after the cells have gained TKI resistance. This also confirms our inference that the observed drug resistance does not involve the absence of drug function but involves an alteration in its biological effects due to other reasons. Notably, the osimertinib sensitivity of PC-9 cell line, which shows cytoplasmic localization of EGFR, was significantly lower than that of HCC827 cell line, which shows membrane localization of EGFR. Similarly, the membrane/plasma translocation of EGFR occurred during the induction of resistance by osimertinib; however, the translocation of EGFR into cytoplasm did not play a role in inhibiting the activity of the Hippo pathway. This is different from the mechanism underlying the resistance induced by the first-generation TKIs. Further studies are warranted to understand the role of EGFR translocation or an accompanying phenomenon in drug resistance. Nevertheless, we confirmed that osimertinibinduced resistance did not involve c-MET amplification, mutation of EGFR-C797S, and changes in ERK protein

Journal of Thoracic Oncology

Vol.

-

No.

-

levels, thereby indicating that cytoplasmic EGFR may cause osimertinib resistance via other mechanisms. Finally, we found that HCC827/ER cells were sensitive to osimertinib, whereas HCC827/OR cells were sensitive to erlotinib. These results indicate that the T790M mutation-lacking cells resistant to firstgeneration TKIs were sensitive to osimertinib and that osimertinib-resistant cells without C797S mutation were sensitive to erlotinib. This may be because of the different mechanisms underlying their drug resistance and because the apparent drug resistance was not due to the loss of TKI action (irrespective of the drug used to induce resistance, the pEGFR levels remained low). However, this result is inconsistent with the clinical manifestations in patients. At present, there is only one clinical case supporting our conclusion.40 However, whether the patients that are osimertinib-resistant and C797S-negative are sensitive to first-generation TKIs has not been reported. We believe that this phenomenon is of great interest. This is possibly because in vitro experiments only used two kinds of drug-resistant cell lines in this article (HCC827/ER and HCC827/OR), and neither of them showed T790M and C797S mutations and c-Met amplification. Whether this result is due to the specificity of the cell line is unclear. Therefore, more cell lines resistant to first- and third-generation TKIs should be used so that the results may be more rigorously verified. Moreover, in clinical trials, we should also collect more cases (lung adenocarcinoma) of secondary drug resistance to first- or third-generation TKIs showing the absence of T790M and C797S mutations and c-Met amplification, observe the sensitivity of these patients after the application of the third- or firstgeneration TKIs, and check whether membranous/cytoplasmic/nuclear translocation of mutant EGFR is detected. However, this study does not clarify the mechanisms that induce EGFR translocation to the cytoplasm/nucleus after TKI treatment. It also does not explain the reason for absence of Hippo pathway inhibition by osimertinibinduced cytoplasmic translocation of EGFR. Furthermore, it remains unclear whether (1) cytoplasmic translocation is a concomitant phenomenon or another mechanism for TKI resistance; (2) the nuclear translocation of mutant EGFR protein is different from that of wild-type EGFR; and (3) the nuclear translocation is related to importin-beta, via its tripartite nuclear localization sequence (645RRRHIVRKRTLRR657).41 Our study is a preliminary study showing that firstgeneration TKIs can restore the sensitivity in osimertinib-resistant cells at the cytological level; however, further studies are essential to elucidate the detailed mechanisms underlying this phenomenon and to ensure clinical application.

---

2019

Acknowledgments The study was supported by grants from the National Natural Science Foundation of China (grant numbers 81871887, 81572854, 81401885, and 81772489) and China Postdoctoral Science Foundation (grant number 2018M641737). The authors thank Prof. Yong Zhang and Yuan Liang for collection of specimens from patients with lung adenocarcinoma treated with tyrosine kinase inhibitors and for their generous help.

Supplementary Data Note: To access the supplementary material accompanying this article, visit the online version of the Journal of Thoracic Oncology at www.jto.org and at https://doi. org/10.1016/j.jtho.2019.06.014

References 1. Heydt C, Michels S, Thress KS, Bergner S, Wolf J, Buettner R. Novel approaches against epidermal growth factor receptor tyrosine kinase inhibitor resistance. Oncotarget. 2018;9:15418–15434. 2. Novello S, Barlesi F, Califano R, et al. Metastatic non– small-cell lung cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2016;27:v1–v27. 3. Ribaudo G, Zanforlin E, Zagotto G. Overcoming resistance in non–small-lung cancer: a practical lesson for the medicinal chemist. Arch Pharm (Weinheim). 2018;351:e1800037. 4. Shigematsu H, Lin L, Takahashi T, et al. Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J Natl Cancer Inst. 2005;97:339–346. 5. Lin JJ, Cardarella S, Lydon CA, et al. Five-year survival in EGFR-mutant metastatic lung adenocarcinoma treated with EGFR-TKIs. J Thorac Oncol. 2016;11:556– 565. 6. Greenhalgh J, Dwan K, Boland A, et al. First-line treatment of advanced epidermal growth factor receptor (EGFR) mutation positive non-squamous non–small cell lung cancer. Cochrane Database Syst Rev. 2016:CD010383. 7. Majem M, Remon J. Tumor heterogeneity: evolution through space and time in EGFR mutant non–small lung cancer patients. Transl Lung Cancer Res. 2013;2:226–237. 8. Remon J, Steuer CE, Ramaligam SS, Felip E. Osimertinib and other third-generation EGFR TKI in EGFR-mutant NSCLC patients. Ann Oncol. 2018;29:i20–i27. 9. Finlay MRV, Anderton M, Ashton S, et al. Discovery of a potent and selective EGFR inhibitor (AZD9291) of both sensitizing and T790M resistance mutations that spares the wild type form of the receptor. J Med Chem. 2014;57:8249–8267. 10. Ballard P, Yates JW, Yang Z, et al. Preclinical comparison of osimertinib with other EGFR-TKIs in EGFR mutant NSCLC brain metastasis models, and early evidence of clinical brain metastasis activity. Clin Cancer Res. 2016;22:5130–5140.

TKIs Resistance Caused by mEGFR Translocation

17

11. Eberlein CA, Stetson D, Markovets AA, et al. Acquired resistance to the mutant-selective EGFR inhibitor AZD9291 is associated with increased dependence on RAS signaling in preclinical models. Cancer Res. 2015;75:2489–2500. 12. Cho BC, Chewaskulyong B, Lee KH, et al. Osimertinib versus standard of care EGFR TKI as first-line treatment in patients with EGFRm advanced NSCLC: FLAURA Asian Subset. J Thorac Oncol. 2019;14:99–106. 13. Mok TS, Wu Y-L, Ahn -MJ, et al. Osimertinib or platinumpemetrexed in EGFR T790M-positive lung cancer. N Engl J Med. 2017;376:629–640. 14. Wang S, Tsui ST, Liu C, Song Y, Liu D. EGFR C797S mutation mediates resistance to third-generation inhibitors in T790M-positive non–small cell lung cancer. J Hematol Oncol. 2016;9:59. 15. Ku BM, Choi MK, Sun JM, et al. Acquired resistance to AZD9291 as an upfront treatment is dependent on ERK signaling in a preclinical model. PLoS One. 2018;13:e0194730. 16. Dittamann K, Mayer C, Czemmel S, Huber SM, Rodemann HP. New roles for nuclear EGFR in regulating the stability and translation of mRNAs associated with vEGF signaling. PLoS One. 2017;12:e0189087. 17. Bazzani L, Donninis Finetti F, Christofori G, Ziche M. PGE2/EP3/SRC signaling induces EGFR nuclear translocation and growth through EGFR ligands release in lung adenocarcinoma cells. Oncotarget. 2017;8:31270–31287. 18. Brand TM, Iida M, Corrigan KL, et al. The receptor tyrosine kinase AXL mediates nuclear translocation of the epidermal growth factor receptor. Sci Signal. 2017;10(460):pii:eaag1064. 19. Wang SC, Hung MC. Cytoplasmic/nuclear shuttling and tumor progression. Ann N Y Acad Sci. 2005;1059:11–15. 20. Porcelli L, Giovannetti E, Assaraf Y, et al. The EGFR pathway regulates BCRP expression in NSCLC cells: role of erlotinib. Curr Drug Targets. 2014;15:1322– 1330. 21. Brand TM, Idia M, Luthar N, Wleklinski MJ, Starr MM, Wheeler DL. Mapping c-terminal transactivation domains of the nuclear HER family receptor tyrosine kinase HER3. PLoS One. 2013;8:e71518. 22. Dittmann K, Mayer C, Paasch A, et al. Nuclear EGFR renders cells radio-resistant by binding mRNA species and triggering a metabolic switch to increase lactate production. Radiother Oncol. 2015;116:431–437. 23. Dittmann K, Mayer C, Czemmel S, Huber SM, Rodemann HP. New roles of nuclear EGFR in regulating the stability and translation of mRNAs associated with vEGF signaling. PLoS One. 2017;12:e0189087. 24. Wei H, Zhu Z, Lu L. Inhibition of EGFR nuclear shuttling decreases irradiation resistance in HeLa cells. Folia Histochem Cytobiol. 2017;55:43–51. 25. Lee BS, Park DI, Lee DH, et al. Hippo effector YAP directly regulates the expression of PD-L1 transcripts in EGFR-TKI-resistant lung adenocarcinoma. Biochem Biophys Res Commun. 2017;491:493–499. 26. McGowan M, Kleinberg L, Halvorsen AR, Helland A, Brustugun OT. NSCLC depend upon YAP expression and nuclear localization after acquiring resistance to EGFR inhibitors. Genes Cancer. 2017;8:497–504.

18 Rong et al 27. Hsu PC, You B, Yang YL, et al. YAP promotes erlotinib resistance in human non–small cell lung cancer cells. Oncotarget. 2016;7:51922–51933. 28. Lee JE, Park HS, Lee D, et al. Hippo pathway effector YAP inhibition restores the sensitivity of EGFR-TKI in lung adenocarcinoma having primary or acquired EGFR-TKI resistance. Biochem Biophys Res Commun. 2016;474:154–160. 29. Moon S, Park SY, Park HW. Regulation of the Hippo pathway in cancer biology. Cell Mol Life Sci. 2018;75:2303–2319. 30. Taha Z, Janse van Rensburg H, Yang X. The Hippo pathway: immunity and cancer. Cancers (Basel). 2018;10:E94. 31. Jang JW, Kim MK, Bae SC. Reciprocal regulation of YAP /TAZ by the Hippo pathway and the small GTPase pathway. Small GTPases. 2018;20:1–9. 32. Lee TF, Tseng YC, Nguyen PA, Li YC, Ho CC, Wu CW. Enhanced YAP expression leads to EGFR TKI resistance in lung adenocarcinomas. Sci Rep. 2018;8:271. 33. Ghiso E, Migliore C, Ciciriello V, et al. YAP-dependent AXL overexpression mediates resistance to EGFR inhibitors in NSCLC. Neoplasia. 2017;19:1012–1021. 34. Lin XY, Zhang XP, Wu JH, Qiu XS, Wang EH. Expression of LATS1 contributes to good prognosis and can negatively regulate YAP oncoprotein in non–small-cell lung-cancer. Tumor Biol. 2014;35:6435–6443. 35. Jiang G, Fan C, Zhang X, et al. Ascertaining an appropriate diagnostic algorithm using EGFR

Journal of Thoracic Oncology

36.

37.

38.

39.

40.

41.

Vol.

-

No.

-

mutation-specific antibodies to detect EGFR status in non–small-lung-cancer. PLoS One. 2013;8:e59183. Han Q, Lin X, Zhang X, et al. WWC3 regulates the Wnt and Hippo pathways via dishevelled proteins and large tumor suppressor 1, to suppress lung cancer invasion and metastasis. J Pathol. 2017;242:435–447. Imajo M, Miyatake K, Iimura A, Miyamoto A, Nishida E. A molecular mechanism that links Hippo signaling to the inhibition of Wnt/b-catenin signaling. EMBO J. 2012;31:1109–1122. Wang P, An F, Zhuang X, et al. Chronopharmacology and mechanism of antitumor effect of Erlotinib in Lewis tumor-bearing mice. PLoS One. 2014;9:e101720. Xu W, Wei Y, Wu S, et al. Up-regulation of the Hippo pathway effector TAZ renders lung adenocarcinoma cells harboring EGFR-T790M mutation resistant to gefitinib. Cell Biosci. 2015;5:7. Chen J, Soudy H. Osimertinib in the treatment of leptomeningeal disease in T790M-negative epidermal growth factor receptor-mutated non–small cell lung cancer: a case report. Chin Clin. 2019 [Epub ahead of print]. Wang YN, Wang H, Yamaguchi H, Lee HJ, Lee HH, Hung MC. COPI-mediated retrograde trafficking from the Golgi to the ER regulates EGFR nuclear transport. Biochem Biophys Res Commun. 2010;399: 498–504.