Background
Icotinib has shown a favourable efficacy and appeared to be well tolerated in NSCLC patients harbouring EGFR mutations. Previous clinical trials reported that icotinib could prolong progression-free survival (PFS) and overall survival (OS) of patients with EGFR exon 19 deletion or L858R mutation [
1]. Unfortunately, partial patients with wild-type EGFR NSCLC could not benefit from the initial treatment with icotinib [
2], and other patients with EGFR mutant NSCLC only received short-term clinical benefits from icotinib treatment [
3], indicating that innate and acquired resistance limit broad application and long-term efficacy of icotinib. Thus, it is required to clarify the possible resistance mechanisms and explore corresponding treatment strategies.
Autophagy has been considered as a highly conserved and complex process whereby sequential formation of autophagosomes and autolysosomes, and component degradation [
4]. As an essential protein in autophagy activation, ATG7 appears to play a scenario-dependent role in the cross-talk between autophagy-dependent resistance and autophagic apoptosis. For instance, ATG7-dependent autophagy facilitated promoting erlortinib resistance and silencing ATG7 overcame resistance of breast cancer cells to erlortinib [
5]. On the contrary, ATG7 was also required for autophagic apoptosis under the treatment of erlortinib in lung cancer with EGFRT790M [
6]. Thus, the role and mechanisms of ATG7 in icotinib resistance need to be explored.
The JAK2/STAT3 pathway and PI3K/AKT/FOXM1 pathway have been identified as primary targets of EGFR-TKIs. Of note, activation of STAT3 and FOXM1 diminished the therapeutic advantages of EGFR-TKIs [
7,
8]. On the other hand, STAT3 and FOXM1 are also considered as upstream molecules in autophagy. They inhibited autophagy by preventing the formation of Beclin-1/VPS34 complex [
9], and rendered tumour cells resistant to chemotherapy by targeting AMPK/mTOR pathway-induced autophagy [
10]. Nevertheless, it is worth exploring whether and how STAT3 and FOXM1 affect icotinib resistance and autophagy.
Based on the above observations, we demonstrated the responses of EGFR-mutated NSCLC cells to icotinib, and explored the biological functions and action mechanisms of autophagy in icotinib-resistant cells in vitro and in vivo. Further study will be needed to verify the relationships between autophagy and EGFR-TKIs resistance in EGFR-mutated NSCLC tissues after EGFR-TKIs treatment.
Methods
Cell culture, treatments and reagents
Human cancer cell line PC-9 was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and genotyped for identity at China Center for Type Culture Collection (CCTCC, Wuhan, China). Gefitinib-resistant PC-9/GR cells were gifted by Prof. Shiyue Li and Dr. Ming Liu from Guangzhou Medical University (Guangzhou, China). H1975 cells were kindly provided by Dr. Xinling Ren from Air Force Military Medical University (Xi’an, China). These cell lines were all cultured with DMEM or RPMI 1640 medium (HyClone, Utah, USA) supplemented with 10% fetal bovine serum (FBS, BI, Israel) at 37C, with 5% CO2. To prepare for further studies, PC-9/GR and H1975 cells were exposed to 2 μM icotinib or DMSO for 4 weeks. Icotinib was provided by Betta Pharmaceuticals Co., Ltd. (Zhengjiang, China). 3-MA, bafilomycin A1 (BafA1), and chloroquine (CQ) were purchased from Sigma Aldrich (Merck KGaA, Darmstadt, Germany). Cryptotanshinon (CTN) was obtained from MCE (USA).
Cell viability assay
Cell viability was assessed with Cell Counting Kit (CCK)-8 (Dojindo Molecular Technologies, Kumamoto, Japan). Lung cancer cells were seeded into a 96-well plate at a density of 5 × 103/well, and treated with various concentrations of icotinib for up to 24 h, 48 h and 72 h. After treatment, 20ul of CCK-8 was added to the cells and the incubation was continued for 2 h. Cell viability was measured by the absorbance at 450 nm on an ELISA reader (Bio-Rad Laboratories, Hercules, CA, USA).
Cell apoptosis and clonogenic assays
Procedures of cell apoptosis and clonogenic assays were performed as described previously [
11]. Cancer cells were seeded into a 6-well plate of 3 × 105/well, and harvested by a set of treatment, trypsinization, centrifugation, wash, and staining with Annexin V-PE and 7-AAD. Apoptosis was detected by flow cytometry (Becton-Dickinson; BD Biosciences). For clonogenic assay, cells were seeded in 6-well plates at a density of 250–300 cells/well, and treated with the indicated treatment conditions. Cells were then fixed with 4% PFA (Servicebio Technology CO., LTD, Wuhan, China), stained with crystal violet (Solarbio, Beijing, China) and photographed using a camera (Canon, Tokyo, Japan).
Transmission Electron microscopy (TEM)
Cells and tissues were fixed with the electron microscope fixative (Servicebio Techology CO., LTD, Wuhan, China) overnight. After osmium acid fixation and dehydration, the samples were embedded to prepare ultrathin sections. The intracellular structures were detected using TEM HT7700 (Hitachi, Tokyo, Japan).
Autophagic flux experiments measurement
The mRFP-GFP-LC3 lentivirus vectors were used to monitor autophagy flux (GeneChem Biotechnology Co., Ltd., Shanghai, China). According to the manufacturer’s instructions, lung cancer cells were infected with mRFP-GFP-LC3 lentivirus. Following an incubation period of 8 h, cells were replaced with complete medium for 48 h. Puromycin was added into cells for 1 week, and puromycin resistant cells were survived. After treatment, fluorescence images were observed by laser confocal fluorescence microscopy IX83 (Olympus, Tokyo, Japan). In pictures, yellow dots represent autophagosomes and red dots represent autolysosomes, which were counted in at least 100 cells in each sample.
Short interfering RNA (siRNA) and plasmids transfection
SiRNAs targeting ATG3, ATG7, Beclin-1 and negative control were purchased from Santa Cruze Biotechnology (Dallas, Texas, USA). SiRNAs targeting FOXM1 (siFOXM1-1: 5′-CUCUUCUCCCUCAGAUAUA-3′, siFOXM1-2: 5′-GCCGGAACAUGACCAUCAA-3′ and control siRNA (si-NC 5′-TAAGGCTATGAAGAGATAC-3′) were synthesized by GenePharma (Shanghai, China). Wild-type STAT3 plasmids (pGV492-Flag-STAT3WT), constitutively active STAT3 plasmid (pGV492-Flag-STAT3-CA), dominant-negative STAT3 vector (pGV492-Flag-STAT3-Y705F) or the respective empty vectors plasmids were constructed by GeneChem. After exposed to DMSO or icotinib for the indicated time, cancer cells were seeded into a 6-well plate and transfected with siRNAs or plasmids for 72 h using RNAiMAX or Lipofectamine2000 (Invitrogen, MA, USA). Then cells were collected as indicated.
Immunoblotting, immunohistochemistry (IHC) and TUNEL staining
For immunoblotting, total protein in lung cancer cells and xenografted tumour tissue was harvested with RIPA lysis buffer (Beyotime, Shanghai, China) containing a protease/phosphatase inhibitor cocktail (MCE, USA). Following procedures were performed as previously described [
11]. For IHC, Paraffin-embedded sections of LUAD specimens and xenografts were subjected to heat-induced epitope retrieval and blocked with 3%H2O2 and goat serum at room temperature. After incubation with primary antibodies and secondary antibodies linked with DAB, images were captured and scored using the IRS (immunoreactive score, IRS) score system based on the intensity and percentage of positive staining [
12]. All slides were viewed and assessed by two independent pathologists. For TUNEL staining, apoptotic Cells in xenograft tissues were identified by the TdT-mediated dUTP labelled by fluorescein after retrieval and permeabilization, and the nucleus was stained with DAPI. Photograph acquisition was finished using a fluorescence microscope (Olympus, Tokyo, Japan).
Plasmid and shRNA Lentivirus infection
Wild-type STAT3 plasmid, dominant-negative STAT3 plasmid (pGV492-Flag-STAT3-Y705F), FOXM1 shRNA (shFOXM1) and scrambled shRNA (shNC) lentivirus were purchased from Hanheng Biotechnology (Hanheng Biotechnology Co., Ltd., Shanghai, China). Cells were plated into six-well plates and infected with plasmids and shRNA lentivirus according to the manufacturer’s instructions. Six hours later, cells were washed with PBS and cultured with complete medium in the incubator for 48 h.
Xenograft studies
All experiments involving animals were performed following the protocol of the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 8023, revised 1978), and approved by the Laboratory Animal Center of Xi’an Jiaotong University. 4-5 weeks old female nude mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China), and kept in specific pathogen-free (SPF) rooms at the Animal Center of Xi’an Jiaotong University for feeding and observations. Notably, to relieve the pain to these mice, a system containing isoflurane was used to introduce inhalation anaesthesia before the subcutaneous injections. The mice were sterilised with medical iodophors before and after subcutaneous injection. PC-9/GR cells (5 × 106) were inoculated into the subcutaneous tissue of the right flanks of nude mice, and growth was observed every three days. To explore the impacts of CQ on resistance to icotinib in vivo, the mice were randomly divided into four groups of four mice each with or without oral administration of icotinib (50 mg/kg), or subcutaneous injection of chloroquine (50 mg/kg), or the indicated combinations for 21 days. To explore the role of STAT3/FOXM1 in resistance to icotinib in vivo, seven groups were designed: (1) Ctrl, (2) icotinib(50 mg/kg), (3) icotinib(50 mg/kg) plus LV-STAT3WT, (4) icotinib(50 mg/kg) plus LV-STAT3-Y705F, (5) icotinib(50 mg/kg) plus shNC, (6) icotinib(50 mg/kg) plus shFOXM1, (7) icotinib(50 mg/kg) plus LV-STAT3-Y705F and shFOXM1. When tumour volume reached to 50-100 mm3 in Ctrl group, the mice received drug treatment. At the endpoint of treatment, nude mice were sacrificed. Tumour volumes were measured as (length × width2) × 0.5. Tumour weight and body weight were evaluated at the endpoint of treatment and every three days, respectively.
Patient-derived samples
Advanced lung adenocarcinoma patients detected for EGFR mutations at the 2nd Affiliated Hospital of Xi’an Jiaotong University from 2013 to 2018 were recruited in this study. Thirty-two biopsy specimens from patients were obtained before single-agent gefitinib or icotinib treatment, and the other thirty-one specimens were from patients who had developed radiographic progression of disease after continuous gefitinib or icotinib therapy. Biopsy specimens were obtained in the least invasive manners, including fine-needle aspiration (FNA) and core biopsy done with image guidance. The Ethics Committee of the 2nd Affiliated Hospital of Xi’an Jiaotong University approved this study. Before the study, all patients signed informed consent. The study was carried out in compliance with the Code of Ethics of the World Medical Association (Declaration of Helsinki Declaration of 2013).
Statistical analysis
Comparisons among multiple groups were analyzed using the one-way ANOVA, and Student t-test was conducted for data between two groups. The correlation between EGFR-TKIs resistance and gene expression was examined using the Chi-square test. The rank-based Spearman correlations were conducted to detect the association among p-STAT3 expression, FOXM1 expression, and ATG7 expression. The Kaplan-Meier method and log-rank test were used to evaluate survival. A Cox proportional hazard regression model was used to evaluate the predictive value of STAT3/FOXM1/ATG7 signalling on the therapeutic effect of icotinib. Each assay in this study was carried out in triplicate, and all data were shown as mean ± standard deviation (SD). SPSS 23.0 (SPSS, Inc., IL, USA) and GraphPad Prism 7.0 (GraphPad Software, Inc., CA, USA) were separately used to perform statistical analysis and generate the set of graphics. P < 0.05 was considered statistically significant.
Discussion
As documented, protective autophagy is regarded to deliver damaged organelles and misfolded proteins to the lysosome for degradation and readily produces metabolites to maintain cellular homeostasis that can result in tumorigenesis and resistance to treatment [
14]. Protective autophagy limited the therapeutic effects of EGFR-TKIs in different cancer cells. Experimentally altering specific EGFR-related genes and pathways caused autophagic cell death or growth [
15], suggesting the crucial role of autophagy in the sensitivity and resistance of cancer cells to EGFR-TKIs [
16]. Notably, STAT3/FOXM1/ATG7 signalling played a vital supportive role in EGFR-mutated NSCLC resistance to icotinib through inducing aberrant autophagy. Impressively, icotinib could upregulate ATG7 to induce protective autophagy in resistant NSCLC cells, but silencing Beclin-1 or blockade of the class III PI3K complex could not affect protective autophagy initiated by icotinib.
Notably, there is a discrepancy in the biological effects of autophagy induced by icotinib in various NSCLC cells. The pro-apoptotic effect of autophagy induced by icotinib has been reported in HCC827 cells, and it has also been shown that autophagy may not be essential for the resistance of A549 cells to icotinib [
17]. However, we found that autophagy could be revealed as a potential therapeutic target for reversing resistance to icotinib in PC-9/GR and H1975 cells. In addition, autophagy indeed relies on distinct mechanisms in different cancer types or tumour treatments. For instance, such a discrepancy of biological effects of Beclin-1 was common, especially in the regulation of autophagy. Beclin-1 was known for its ability to affect the sensitivity of cancer to targeted therapy such as gefitinib, osimertinib and sorafenib [
18‐
20], but was not involved in autophagy-dependent resistance to erlotinib in tongue squamous carcinoma [
21]. Similarly, our results displayed that autophagy-modulated resistance to icotinib was independent of Beclin-1. These results indicate a non-essential role for Beclin-1 in autophagy-mediated resistance to targeted therapy. On the other hand, the dual role of ATG7 in lung cancer is context-dependent, including pro-apoptotic and anti-apoptotic effects. For example, ATG7-dependent autophagy suppressed the progression of lung cancer driven by activation of oncogenic HRasV12 [
22]. In contrast to the tumour suppressive effect of ATG7 on tumour development [
23], blockage by ATG7 deficiency contributed to promoting the sensitivity to chemotherapy in resistant cells [
24]. Interestingly, we found that silencing ATG7 also accelerated lung tumour apoptosis induced by icotinib, suggesting the importance of ATG7 in autophagy-dependent resistance to icotinib.
Activation of STAT3 is a well-known event in resistance to therapy. For instance, activated STAT3 was believed to be necessary for paclitaxel-induced autophagy due to the addition of IL-6 [
25], but acted to inhibit docetaxel-regulated autophagy in castration-resistant prostate cancer cells [
26], leaving the puzzle of how STAT3 affecting autophagy unresolved. We proposed that activation of STAT3 was capable of elevating FOXM1 expression to promote ATG7-dependent autophagy, leading to resistance to icotinib treatment. Until now, FOXM1 was reported to increase the expression of LC3 and Beclin-1 by binding to their promoters [
27]. We indicated that ATG7 could be a functional target of FOXM1 in autophagy initiated by icotinib. Our findings established a connection between STAT3/FOXM1 signalling and ATG7-dependent autophagy to maintain icotinib resistance in EGFR-mutated NSCLC.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.