Loss of histone lysine methyltransferase EZH2 confers resistance to tyrosine kinase inhibitors in non-small cell lung cancer
Abstract
Tyrosine kinase inhibitor (TKI) treatment is the first-line therapy for non-small cell lung cancer (NSCLC) caused by activating mutations of epidermal growth factor receptor (EGFR). However, acquired resistance to EGFR-TKI occurs almost inevitably. Aberrant activation of proto-oncogene MET has been known to confer EGFR-TKI resistance; however, the mechanisms involved remains unclear. Recent evidence implicates epigenetic hetero- geneity as playing roles in cancer drug resistance, whereas links involving epigenetic heterogeneity and MET in NSCLC remain poorly understood. We found that expression of EZH2, a histone methyltransferase, was nega- tively correlated with MET activation and EGFR-TKI resistance in NSCLC cells and clinical samples, suggesting the potential for EZH2 to be used as a biomarker for EGFR-TKI sensitivity. Knockdown or inhibition of EZH2 up- regulated MET expression and phosphorylation, and elevated proliferation and EGFR-TKI resistance of cells in vitro. Meanwhile, inhibition of MET or PI3K/AKT enhanced EZH2 levels and restored sensitivity to EGFR-TKI. These findings indicate a “MET-AKT-EZH2” feedback loop regulating EGFR-TKI-resistance. Furthermore, com- bination therapy of PI3K/AKT inhibition and EGFR-TKI, which interrupts the loop, enhanced tumor-suppressive effects in an EGFR-TKI-resistant xenograft model, indicating a potential approach against drug resistance in NSCLC.
1. Introduction
Lung cancer is one of the most common cancers and the leading cause of cancer deaths worldwide [1]. Epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKI) are currently the most important targeted therapy for non-small cell lung cancer (NSCLC), which accounts for approximately 85% of lung cancer cases [2]. First-generation EGFR-TKIs, such as gefitinib and its derivative icotinib, as well as sec- ond- and third-generation EGFR-TKIs which have even broader thera- peutic applications, provide significant clinical benefits as the first-line therapy in NSCLC patients with activating EGFR mutations [3–5]. However, acquired resistance to EGFR-TKIs occurs almost inevitably [5].
In addition to mutations that alter drug binding, alternative survival pathways are an important molecular basis for EGFR-TKI resistance [6]. Activation of a receptor tyrosine kinase namely MET proto-oncogene (MET), either by its ligand hepatocyte growth factor (HGF) or via ligand-independent mechanisms such as MET amplification, has long been known to confer resistance to EGFR-TKIs [7,8]. In the presence of EGFR-TKI, the PI3K/AKT pathway can be persistently activated directly by MET or by MET-mediated EGFR-independent phosphorylation of ErbB3, providing a bypass for NSCLC cells to maintain downstream signal transmission and survival [9]. In addition, accumulating evidence implicates epigenetic heterogeneity in cancer drug resistance [10,11]. However, whether epigenetic heterogeneity influences MET expression to drive EGFR-TKI resistance remains unclear.
Histone methyltransferase EZH2 (Enhancer of zeste homolog 2) si- lences genes through di- or tri-methylation of lysine 27 on histone H3 (H3K27m2/3) [12]. Elevated levels of EZH2 have been reported to be associated with increased aggressiveness and poorer outcomes in lym- phoma and various solid tumors [13,14], and EZH2 inhibition sup- pressed progression of multiple tumors in vitro and in vivo [15]. However, a number of studies have also revealed a tumor-suppressive effect of EZH2, generating controversy over the exact role that EZH2 plays in cancer. For example, inactivating mutations of EZH2 predicted poor prognosis in myeloid malignancies [16], and prolonged EZH2 in- hibition in glioblastoma promoted tumorigenesis [17].
Although several EZH2 inhibitors are undergoing clinical trials for multiple malignancies including lung cancer [18], a trial of the EZH2 inhibitor GSK2816126 was recently terminated due to insufficient evi- dence of clinical efficacy, leading to concerns about the use of EZH2 inhibitors. Recent studies also revealed a potential link between EZH2 and drug-resistance. Low EZH2 expression is associated with shorter progression-free survival in patients with colorectal cancer who were treated with monoclonal antibodies targeting EGFR [19], and induces chemoresistance in acute myelocytic leukemia [20]; but little is known about whether and how EZH2 participates in coordinated response to TKI pressure in NSCLC.
Here, we demonstrated that EZH2 inhibition confers resistance to EGFR-TKI in NSCLC. Our findings using NSCLC cells and clinical samples demonstrated that EZH2 was negatively correlated with MET and EGFR- TKI resistance, indicating the potential use of EZH2 as a marker for EGFR-TKI sensitivity. We also discovered that a “MET-AKT-EZH2” feedback loop reinforced EGFR-TKI-resistance, which was disrupted by PI3K/AKT inhibition. Our results provide insight into the crosstalk be- tween MET and epigenetic heterogeneity, indicating that greater caution should be employed when using EZH2 inhibitors in lung cancer treat- ment, and suggest that simultaneously targeting PI3K/AKT and EGFR is a potential approach to overcoming EGFR-TKI resistance in NSCLC.
2. Materials and methods
2.1. Human tissue specimens
Paraffin-embedded lung tumor specimens of 102 patients before EGFR-TKI treatment (pre-TKI) (Supplementary Tables S1) and 11 after EGFR-TKI treatment (post-TKI) (Supplementary Table S2), and another 29 paired pre-TKI lung cancer and paratumor tissue specimens (Sup- plementary Table S3) were collected at Affiliated Tongji Hospital, Tongji Medical College (Wuhan, China). The study was approved by the Insti- tutional Review Board of Tongji Hospital; human sample collection procedures were in accordance with the established guidelines.
2.2. Cell culture, plasmids, and treatments
Human lung adenocarcinoma cell lines H1650, H1975, HCC827 and PC9 and gefitinib-resistant HCC827/PC9 were cultured in RPMI-1640 medium (Hyclone, South Logan, UT) containing 8% fetal bovine serum (PAN Biotech, Germany). The gefitinib-resistant PC9 cell lines were generated as previously reported [21,22]. Briefly, PC9 cells in culture were exposed to increasing concentrations of gefitinib. The exposure dose started at 0.01 μM and was progressively increased every 3–4 weeks, reaching a final gefitinib concentration of 3 μM. The resistant cell lines established in this manner were then maintained in drug-free medium; or with 0.1 μM gefitinib within 3 generations after thawing from storage in liquid nitrogen. The gefitinib-resistant HCC827 (HCC827GR5) cell line was kindly provided by Dr. Feng Zhu (HUST, Wuhan, China) and Dr. Pasi A. J¨anne (Harvard Medical School, Boston, MA). MET-overexpressing cell lines and MET/AKT knockdown cell lines were gifts from Dr. Feng Zhu. pCI-EZH2-Flag was a gift from Dr. Min Wu (Wuhan University, Wuhan, China). shEZH2 plasmids (targeting se- quences, Supplementary Table S4) were cloned into the pSUPER vector [23]. Gefitinib, GDC0941 and GSK343 were from Targetmol (Boston, MA); icotinib was from Betta Pharmaceuticals (Hangzhou, China);PHA665752 was from MedChemExpress (Monmouth Junction, NJ). For experiments on increasing doses of gefitinib, cells were continuously exposed to gefitinib for 48 h. For combination therapy in vitro, gefitinib and GDC0941/PHA665752/GSK343 were administrated to cells simul- taneously; cell viability and relative molecular changes were determined after 48 h of single or combination treatment.
2.3. Cell proliferation, MTT and soft agar colony formation assay
Cell number was counted using a hemocytometer. MTT assay was performed as previously described [24]. Soft agar colony formation
assay was performed as previously reported [25,26]. In brief, cells (8 × 103 per well) were suspended in soft agar cell culture mixture with or without different concentrations (0, 0.01, 0.1, 1 μM) of gefitinib, and plated on the bottom agar containing the corresponding concentrations of gefitinib. The cultures were maintained at 37 ◦C in a 5% CO2 incu- bator for 14 days, then colonies were counted.
2.4. Western blot, quantitative PCR (qPCR) and immunohistochemical study (IHC)
Western blot, qPCR and IHC were performed as we previously re- ported [27,28]. Bio-Rad Quantity One Software was used for quantifi- cation [29]. Antibodies and primers are shown in Supplementary Tables S5–S6.
2.5. Chromatin immunoprecipitation (ChIP) and co-immunoprecipitation (IP)
ChIP and IP were performed as previously described [30,31]. Anti- bodies and primers are shown in Supplementary Tables S5–S6.
2.6. Animal experiments
Mouse procedures were conducted in accordance with the Guidelines of the China Animal Welfare Legislation, as approved by the Committee on Ethics in the Care and Use of Laboratory Animals of College of Life Sciences, Wuhan University (Wuhan, China). 2 × 106 HCC827 or
HCC827GR cells were injected subcutaneously into the axillary region of 6-week-old male BALB/c nude mice (HFK Bioscience, Beijing, China) to establish TKI-resistant xenograft tumors as reported [32,33]. When tu- mors reached 200 mm3 in volume, mice were randomly divided into groups receiving different treatments. Gefitinib and GDC0941 were administrated via intraperitoneal injection 3 times a week at 25 mg/kg until sacrifice of the mice. Tumor volume was measured as we previ- ously described [24].
2.7. Statistical analysis
Results were expressed as the mean ± SEM and were analyzed using a two-tailed Student’s t-test and two-way analyses of variance. Pearson correlation coefficient analysis was performed using GraphPad Prism V5.0. Differences were considered statistically significant when P < 0.05. 3. Results 3.1. Low EZH2 expression is associated with gefitinib resistance in NSCLC cells To investigate the MET-mediated EGFR-TKI resistance, we per- formed experiments in EGFR-TKI-sensitive NSCLC cell lines, and matching resistant cells harboring aberrant MET activation. Gefitinib- sensitive HCC827 and PC9 are lung adenocarcinoma cell lines with an activating small in-frame deletion in exon 19 of EGFR (EGFR 19del). HCC827GR (HCC827GR5) is a cell line with MET amplification [21]. Fig. 1. Low EZH2 expression is associated with gefitinib resistance in NSCLC cells. (A) Experimental design to generate the gefitinib-resistant cell clones (GR clones). PC9 cells were treated with increasing concentrations (progressively increased every 3–4 weeks, from 0.01 to 3 μM) of gefitinib for 6 months. (B) Prolif- eration of paired parental and gefitinib-resistant cells in drug-free medium. (C) Colony formation of paired parental and gefitinib-resistant after 14-day gefitinib treatment. (D) Parental (black) and gefitinib-resistant (red) cells were treated with increasing concentrations of gefitinib for 48 h; cell viability was assessed using MTT assay. (E) Reduced EZH2 protein level was consistent in HCC827GR and PC9GR cells. (F) MET and AKT remained activated in HCC827GR/PC9GR cells under treatment with different gefitinib concentrations for 48 h, EZH2 significantly decreased in HCC827 and PC9 parental cells following gefitinib treatment. Data, mean ± SEM, *, P < 0.05, **, P < 0.01, ***, P < 0.001, ns, no significance. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.) The PC9GR (PC9GR10) cell line was selected from 16 gefitinib-resistant clones of PC9 cells, and was shown to exhibit enhanced MET expression and phosphorylation (Fig. 1A and Supplementary Fig. S1A). Consistent with previous reports, lung cancer cells went through “reversible EGFR-TKI tolerance” to finally acquire resistance to EGFR-TKI [34]. HCC827GR and PC9GR cells grew faster than the corresponding parental cells in either drug-free conditions or under EGFR-TKI pressure (Fig. 1B and C). Moreover, increased gefitinib IC50 observed in HCC827GR and PC9GR cells (Fig. 1D). These data confirmed the emergence of TKI resistance. EZH2, a histone methyltransferase with multiple roles in cancer, has yet to be related to the development of EGFR-TKI resistance in NSCLC cells. Interestingly, our results showed that EZH2 protein levels were significantly lower in HCC827GR and PC9GR cells (Fig. 1E), and a trend of negative correlation of gefitinib IC50 with EZH2 levels was observed in the 16 gefitinib-resistant clones (Supplementary Figs. S1A–B). Consistently, low EZH2 expression was associated with gefitinib resis- tance in 6 NSCLC cell lines (Fig. S1C and Supplementary Table S7). Since low EZH2 protein levels were observed in both HCC827GR and PC9GR cell lines (Fig. 1E) that were established via exposing the sen- sitive parental cells to gefitinib treatment, we questioned whether the down regulation of EZH2 was caused by EGFR-TKI pressure. Gefitinib treatment led to a decrease in EZH2 abundance without activating MET or AKT in gefitinib-sensitive HCC827 and PC9 cells (Fig. 1F). Recovery of EZH2 and phosphorylated EGFR was observed 8 days after removing gefitinib from the culture medium (Supplementary Fig. S1D). EGF, the ligand of EGFR, caused time dependent EZH2 elevation in HCC827 and H1650 (a lung adenocarcinoma cell line with EGFR 19del mutation) cells (Supplementary Fig. S1E). Immunofluorescent study also confirmed the association between EGFR phosphorylation and EZH2 increase (Supplementary Fig. S1F). Therefore, low EZH2 levels in gefitinib-resistant cells may be a response to EGFR-TKI pressure. Fig. 2. Knockdown of EZH2 elevates EGFR-TKI resistance through MET activation. (A–B) Knockdown of EZH2 increased gefitinib resistance and proliferation in HCC827 and PC9 cells. Stable cell lines were established by transfecting with shRNA-mock (mock) and shEZH2 (#1 and #2) and selection with puromycin. Cells were treated with either an increasing dose of gefitinib for 48 h (middle panel) or in drug-free medium for 4 days (right panel). (C) Colony formation of control (mock) and EZH2 knockdown cells (shEZH2 #2, hereafter, shEZH2) after 14 days of gefitinib treatment. (D) Knockdown of EZH2 increased p-MET/MET levels and decreased H3K27m3 in HCC827 and PC9. (E) Increasing doses of gefitinib caused increased p-MET/MET in EZH2 knockdown HCC827/PC9. Cells were cultured in 0.01 μM gefitinib for 1 day and 0.1 μM gefitinib for another 6 days. (F) HCC827 cells were treated with increasing gefitinib as follows: 0.01 μM gefitinib for 7 days, 0.1 μM for 12 days, 0.3 μM for 16 days. HCC827 without gefitinib treatment was harvested on day 0 as a control. (G) MET mRNA levels increased significantly in EZH2 knockdown HCC827 and PC9 cells. (H–J) ChIP assay showed EZH2 and H3K27m3 enrichment within 1500 bp upstream of the MET transcriptional start site (TSS). Data, means ± SEM, *, P < 0.05, **, P < 0.01, ***, P < 0.001, ns, no significance. Fig. 3. EZH2 inhibition enhances resistance to EGFR-TKI. (A) GSK343 treatment caused MET activation in a time- and dose-increasing treatment. (B–C) GSK343 enhanced gefitinib resistance. In combination treatment, both drugs were administrated at the same concentration as indicated. (D) GSK343 treatment increased p- MET and prevented gefitinib-induced inhibition of MET activation in HCC827GR and PC9GR cells. Data, means ± SEM. 3.2. Knockdown of EZH2 increases EGFR-TKI resistance via MET activation To further investigate the role of EZH2 in EGFR-TKI resistance, stable EZH2 knockdown HCC827 and PC9 cell lines were established. EZH2 knockdown in cells resulted in increased gefitinib IC50 and cell prolif- eration, and cells harboring shEZH2 #2 were used for further experi- ments (Fig. 2A and B). In contrast, EZH2 overexpression restored sensitivity to gefitinib in NSCLC cells (Supplementary Fig. S2). EZH2 knockdown HCC827 and PC9 cells formed more colonies when treated with 0.01 μM and 0.1 μM gefitinib, whereas 1 μM gefitinib caused drastic cytotoxicity (Fig. 2C). Since high gefitinib concentrations have been suggested to cause MET inhibition [35], further investigation was per- formed with gefitinib concentration below 1 μM. We hypothesized that loss of EZH2 may lead to alternative pathways that drive resistance to EGFR-TKI. Interestingly, we observed increased phosphorylation and expression of MET consistent with down- regulation of H3K27m3 in EZH2 knockdown HCC827 and PC9 cells (Fig. 2D). Meanwhile, increase of H3K27m3 and decrease of MET were found upon EZH2 overexpression (Supplementary Fig. S2). In EZH2 knockdown HCC827 cells, levels of MET and phosphorylated MET elevated following increasing doses of gefitinib; similar trend was also observed in gefitinib-treated EZH2 knockdown PC9 cells (Fig. 2E). These phenotypes were not observed in HCC827 and PC9 cells (Fig. 1F), sug- gesting that under EGFR-TKI pressure, cells with EZH2 attenuation may tend to survive by activating MET and related pathways. To determine whether MET activation is accompanied with EZH2 loss during the development of gefitinib resistance, we cultured HCC827 cells with increasing concentrations of gefitinib. We observed gradual down-regulation of EZH2 and H3K27m3, with a coincident in- crease in MET expression and phosphorylation after 19 days of treat- ment (Fig. 2F). In EZH2 knockdown HCC827 and PC9 cells cultured under drug-free conditions, increased mRNA levels of MET and HGF were observed (Fig. 2G and Supplementary Fig. S2E). Furthermore, ChIP assays demonstrated that EZH2 can bind to the MET promoter. Decreased EZH2 and H3K27m3 enrichment at the MET promoter region was observed in HCC827GR and EZH2 knockdown HCC827 cells, as well as in PC9GR and EZH2 knockdown PC9 cells, suggesting that down- regulation of EZH2 promoted MET activation via decreasing H3K27m3 levels and elevating MET transcription (Fig. 2H–J). Fig. 4. Low EZH2 is correlated with EGFR-TKI resistance in EGFR-mutant NSCLC patients. (A) Representative IHC images of EZH2 staining in TKI-sensitive (19del), -insensitive (G719X/L861Q/S768I) and -resistant (T790 M/20ins) human NSCLC samples. (B) IHC score of EZH2 in NSCLC samples harboring different mutant EGFR. (C) EZH2 was negatively correlated with p-MET in patients with 19exon del/L858R mutant EGFR. Immunostaining was classified as low (IHC score 0, 1 or 2) or high (IHC score 3, 4 or 5). (D) EZH2 level is higher in NSCLC patients who remained responsive to EGFR-TKI within 2 years after first treatment with TKI (Relapse-free, n = 26) than those who relapsed within 2 years (Relapse, n = 9). TKI therapeutic response data were collected in a 2-year follow-up investigation. (E) Patients with high EZH2 tend to benefit from TKI treatment based on the 2-year follow-up investigation. Immunostaining was classified as low (IHC score 0, 1 or 2) or high (IHC score 3, 4 or 5). (F) Compared to pre-TKI samples (n = 80) collected before TKI treatment, EZH2 dropped while p-MET was increased in post-TKI samples (n = 11) from patients who relapsed after TKI treatment. Among the 80 and 11 specimens, 3 were paired samples from 3 individual patients. (G) Representative IHC images of EZH2 and p-MET in unpaired treatment-naive (pre-TKI) and -relapsed (post-TKI) specimens derived from NSCLC patients. All immunostaining was scored blind. Data means ± SEM, *, P < 0.05, **, P < 0.01, ***, P < 0.001, ns, no significance. 3.3. EZH2 inhibition enhances resistance to EGFR-TKI via MET activation The important role of dysregulation of histone H3 methylation in the development of drug resistance has been reported [36]. Consequently, we sought to determine whether a specific EZH2 inhibitor, GSK343 that suppresses H3K27m3 and EZH2 levels [37], can enhance gefitinib resistance in NSCLC cells. We observed down-regulation of H3K27m3 and EZH2, as well as enhanced MET activation in HCC827 and PC9 cells treated with GSK343 (Fig. 3A). Both HCC827 and PC9 cells treated with GSK343 exhibited enhanced gefitinib-resistance (Fig. 3B); similar results were also obtained with HCC827GR and PC9GR cells (Fig. 3C). Furthermore, in HCC827GR and PC9GR cells, MET phosphorylation was mildly suppressed by gefitinib treatment, and upregulated in the presence of GSK343 (Fig. 3D). These data provided evidence that EZH2 inhibition promoted MET activation and EGFR-TKI resistance, regard- less of gefitinib pressure. 3.4. Low EZH2 is correlated with EGFR-TKI resistance in NSCLC patients To evaluate the association between the level of EZH2 and the therapeutic response to EGFR-TKI, we examined 102 human lung cancer specimens with different EGFR mutations that predicted to exhibit different sensitivity to EGFR-TKI treatment. For example, patients with common TKI-sensitizing activating EGFR mutations 19del and L858R, are highly responsive to gefitinib treatment [38]; uncommon EGFR mutations such as G719X, L861Q and S768I predict insensitivity to gefitinib treatment [39]; EGFR with T790 M mutation or an insertion mutation in exon 20 (20ins) are biomarkers for gefitinib-resistance [3]. The 102 samples collected before EGFR-TKI treatment (pre-TKI) were divided into three groups: sensitive (19del/L858R, n = 77), insensitive (G719X/L861Q/S768I, n = 18), and resistant (T790 M/20ins, n = 7). Immunohistochemical analysis revealed that EZH2 staining was strongest in the TKI-sensitive group (Fig. 4A and B), which also showed a negative correlation with p-MET (r = —0.403, P < 0.001, Fig. 4C) and a positive correlation with p-EGFR (r = 0.508, P < 0.001, Supplementary Fig. S3A). Lower EZH2 staining was observed in insensitive and resistant groups (Fig. 4A and B). We performed western blots on 29 additional sets of paired pre-TKI lung tumor and paratumor samples, which revealed an increased incidence of EZH2 expression in tumor samples with activating EGFR mutations (19del/L858R) compared to those with wild type EGFR (Supplementary Fig. S3B). Fig. 5. EGFR-TKI resistance is mediated by a MET-AKT-EZH2 feedback loop. (A) Dose- and time-course treatment with the MET inhibitor PHA665752 (PHA) on HCC827GR cells. HCC827GR cells were treated for 48 h in a dose-course experiment and 1 μM PHA was applied in a time-course experiment. (B) EZH2 levels were higher after 48-h combination treatment with PHA and gefitinib compared to those of the gefitinib-treated. (C) 48-h combination treatment of PHA and gefitinib showed a stronger effect in inhibiting proliferation of HCC827GR cells. (D) Knockdown of MET suppressed AKT activation and increased EZH2. (E) Overexpression of MET promoted AKT activation and suppressed EZH2. (F) Dose- and time-course treatment with the AKT inhibitor GDC0941 on HCC827GR cells. HCC827GR cells were treated for 48 h in a dose-course experiment, and 1 μM GDC0941 was applied in a time course experiment. (G) 48-h combination treatment of GDC0941 and gefitinib showed strong inhibition on the proliferation of HCC827GR cells. (H) 48-h combination treatment with GDC0941 and gefitinib showed higher EZH2 levels compared to those of the gefitinib-treated. (I) Knockdown of AKT suppressed EZH2 phosphorylation at S21 and decreased EZH2. (J) AKT binds with EZH2 in HCC827GR cells. Data, means ± SEM. Icotinib, a clinically equivalent derivative of gefitinib [40–42], is widely used in China as a first-line treatment for advanced NSCLC with activating EGFR mutations [43]. We were able to track the status of the 35 patients who were treated with icotinib to assess their EGFR-TKI therapeutic responses via a 2-year follow-up investigation. According to clinical definition for acquired resistance to EGFR-TKI in lung cancer [44], 26 patients were grouped as relapse-free and the remaining 9 as relapsed. Significantly higher EZH2 levels were observed in the relapse-free group (Fig. 4D), indicating the potential of EZH2 as a biomarker of EGFR-TKI sensitivity. Additionally, patients with high EZH2 and low MET phosphorylation were more likely to benefit from EGFR-TKI treatment (Fig. 4E). We then acquired 11 samples after TKI treatment (post-TKI), among which 3 samples had matched pre-TKI specimens. The post-TKI samples demonstrated overall low EZH2 and high MET phosphorylation, yet the trend was reversed in pre-TKI samples (Fig. 4F and G). This trend was also seen in the 3 pairs of matched pre- and post-TKI samples (Supple- mentary Fig. S3C). 3.5. EGFR-TKI resistance is mediated by MET-AKT-EZH2 feedback loop Our findings demonstrated that EZH2 regulates MET transcription, yet whether and how MET affects EZH2 remains unknown. Aberrant MET activation promotes the downstream PI3K/AKT pathway inducing EGFR-TKI resistance; meanwhile, AKT has been reported to phosphor- ylate EZH2 at serine21 (S21), resulting in decreased levels of EZH2 and H3K27m3 [45,46]. Hence, we hypothesized that AKT may play a role in the crosstalk between MET and EZH2, and carried out dose- and time-dependent treatment in HCC827GR cells, using either a MET in- hibitor (PHA665752 (PHA)) or a PI3K/AKT inhibitor (GDC0941). For prolonged drug administration, culture media were refreshed after 3 days to avoid loss of drug effects (Supplementary Fig. S4A). The inhib- itory effect of PHA and GDC0941 was confirmed by the reduction of p-MET and p-AKT respectively (Fig. 5). EZH2 and H3K27m3 levels were elevated in a dose- and time-dependent manner upon PHA treatment (Fig. 5A). Compared to gefitinib treatment, the combination of gefitinib plus PHA showed a stronger inhibitory effect on MET and AKT phos- phorylation in HCC827GR cells, accompanied with increased levels of EZH2 and H3K27m3 (Fig. 5B). The combination therapy also exhibited enhanced toxicity to HCC827GR cells as previously reported [21] (Fig. 5C). Similar results were obtained using PC9GR cells (Supple- mentary Fig. S4B). Moreover, knockdown of MET in HCC827GR cells decreased AKT phosphorylation and simultaneously increased EZH2 levels, whereas the opposite results were observed when MET was overexpressed in HCC827 cells (Fig. 5D and E). Similarly, GDC0941 treatment in a time- and dose-dependent manner increased EZH2 and H3K27m3 levels (Fig. 5F and Supplemen- tary Fig. S4C). Compared to single agent therapy, a combination of gefitinib and GDC0941 exhibited greater cytotoxicity for HCC827GR cells (Fig. 5G). EZH2 levels in HCC827GR cells receiving combination therapy were significantly higher compared to those treated with gefi- tinib alone; whereas they were lower in cells solely treated with GDC0941 (Fig. 5H), possibly due to gefitinib-induced suppression of EZH2. Presently, S21 is the only reported site on EZH2 that can be phos- phorylated by AKT, and phosphorylation of EZH2 at S21 results in decreased levels of EZH2 and H3K27m3 [45,46]. Indeed, knockdown of AKT suppressed EZH2 phosphorylation at S21 and increased EZH2 levels in HCC827GR cells (Fig. 5I); whereas, under HGF stimulation, which activates the MET-AKT pathway, reversed effects were observed (Supplementary Fig. S4D). In addition, overexpression of EZH2 pro- moted H3K27m3 and suppressed MET activation; compared to those transfected with wild-type EZH2, cells transfected with S21A mutated EZH2 (mimicking S21-unphosphorylated EZH2) [45] presented increased levels of EZH2 and H3K27m3 (Supplementary Figs. S4E and S4F). Moreover, enhanced binding of AKT and EZH2 was only observed in HCC827GR cells (Fig. 5J). These results (Figs. 2 and 5), together with previous reports that AKT-mediated phosphorylation suppressed EZH2 [45,46], suggested the existence of a “MET-AKT-EZH2” feedback loop: aberrant activation of the PI3K/AKT pathway downstream of MET may reinforce drug-resistance by inducing EZH2 deficiency in gefitinib-resistant cells, and EZH2 deficiency may further contribute to MET activation. Disruption of this feedback loop may thus restore EGFR-TKI sensitivity. 3.6. Combination therapy of gefitinib and GDC0941 suppresses HCC827GR-derived tumors To evaluate the change of EZH2 under TKI pressure in vivo, we mimicked the development of gefitinib resistance by treating HCC827- derived subcutaneous xenograft tumors with gefitinib. In the vehicle group, tumor volumes increased continuously and mice were sacrificed on Day 30. The tumor volumes in mice from the gefitinib group, how- ever, decreased over the first 30 days of treatment, then started to in- crease, suggestive of the development of TKI resistance (Fig. 6A). On Day 55, the mice from the gefitinib group were sacrificed, and xeno- grafts were collected for further analysis. Decreased EZH2 levels were found in tumors from the gefitinib group (Fig. 6B). Increased MET and AKT phosphorylation was evident in the gefitinib group compared to the control (Fig. 6B). We then sought to break the “MET-AKT-EZH2” feedback loop in our in vivo model. Currently, there are no EZH2 agonists, but several MET inhibitors have been identified with translational potential. MET inhi- bition combined with gefitinib has been shown to suppress TKI-resistant tumor in vivo [47,48]. A Phase lb/ll study against EGFR mutant plus MET-positive NSCLC also demonstrated the safety and efficacy of a MET inhibitor, INC280, in combination with gefitinib [49]. Therefore, we sought to explore a different approach by investigating the effect of AKT inhibition combined with gefitinib treatment in animal experiments. HCC827GR-derived xenograft tumors were established and treated with different therapies as indicated (Fig. 6C). Compared to single agent therapies, the combination of gefitinib and GDC0941 dramatically attenuated the growth of gefitinib-resistant tumors (Fig. 6C). Mice from the vehicle group, gefitinib group, GDC0941 group and combination therapy group were sacrificed on the 31st, 37th, 43rd, and 43rd day of each treatment, respectively. Xenografts were collected and analyzed. Similar to the results in HCC827GR cells (Fig. 5H), significantly increased levels of EZH2 and H3K27m3 were observed in the combi- nation group compared to those of the gefitinib group (Fig. 6D and E). Fig. 6. Combination therapy with gefitinib and GDC0941 suppresses HCC827GR-derived tumor in vivo. (A) HCC827 cells were used to develop NSCLC xe- nografts in nude mice. After the tumor reached 200 mm3, mice were randomly divided into vehicle group (n = 6) and gefitinib group (n = 7); vehicle and gefitinib (25 mg/kg) was given 3 times a week via intraperitoneal injection. (B) EZH2 and H3K27m3 was decreased in tumor samples from gefitinib group accompanied by the elevation of p-MET and MET. (C) HCC827GR cells were used to develop gefitinib-resistant NSCLC xenografts in nude mice. After the tumor reached 200 mm3, mice were randomly divided into 4 groups, respectively receiving vehicle (n = 6), gefitinib (n = 6, 25 mg/kg), GDC0941 (n = 6, 25 mg/kg) or gefitinib-GDC0941 combined therapy (combination, n = 6, 25 mg/kg each) 3 times a week via intraperitoneal injection. (D) Western blot examination of tumors from the 4 groups in (C). (E) Quantitative data of (D). Data, means ± SEM, *, P < 0.05, **, P < 0.01, ***, P < 0.001, ns, no significance. Fig. 7. A working model of EZH2 involvement in EGFR-TKI resistance in NSCLC cells. EGFR-TKI- resistant NSCLC cells display decreased EZH2 and aberrant activation of MET. Phosphorylation of MET leads to enhanced AKT activation. Activated AKT phosphorylates EZH2 at S21 and promotes ubiquitination-dependent degradation of EZH2, which results in decreased EZH2 enrichment at the promoter region of MET, thus activating MET tran- scription. The “MET-AKT-EZH2” feedback loop con- tributes to EGFR-TKI resistance in NSCLC cells and can be inhibited by PI3K/AKT inhibitors such as GDC0941, which increases EZH2 levels and restores EGFR-TKI sensitivity (Ub, ubiquitination; P, phos- phorylation; me, methylation). 4. Discussion EGFR-TKIs serve as one of the first-line therapies for NSCLC patients. However, the majority of patients inevitably develop EGFR-TKI resis- tance, through underlying mechanisms that are largely unclear [44]. In this study, we report that loss of EZH2 is associated with EGFR-TKI resistance. EZH2 expression was negatively correlated with EGFR-TKI resistance in NSCLC cell lines and clinical samples. Either knockdown or inhibition of EZH2 enhanced EGFR-TKI resistance, while the opposite effect was observed when EZH2 was overexpressed. Mechanistically, down regulation of EZH2 promoted MET activation, a well-known pathway involved in EGFR-TKI resistance [9]. Meanwhile, aberrant activation of the PI3K/AKT pathway downstream of MET suppressed EZH2 by AKT-mediated phosphorylation of EZH2. These data suggest the existence of a “MET-AKT-EZH2” feedback loop which reinforces EGFR-TKI-resistance (Fig. 7). There is discrepancy between epigenetic patterns in different drug- resistant cases. For example, chromatin repression via H3K9 and H3K27 methylation was reported to promote resistance to multiple TKIs in lung cancer cells [36]; EZH2 depletion in lung adenocarcinoma cells enhanced the response to platinum-based and VEGFR-2-targeted ther- apy [50]; and suppression of EZH2 resulted in acquired resistance to cytotoxic chemotherapy in acute myelocytic leukemia [20]. Moreover, the protein level and function of EZH2 can also be regulated by additional tumor promoters involved in drug resistance, such as CDK1 and AMPK [51,52]. Collectively, our data suggest that the “MET-A- KT-EZH2” loop is an important mechanism that contributes to the development of EGFR-TKI resistance. Consistent with previous findings [45,46], EZH2 was regulated by AKT-mediated phosphorylation at S21; however, we cannot rule out the possible existence of additional routes through which the MET-AKT pathway controls EZH2, such as modifi- cations at other amino acid sites; thus, future work is needed. EZH2 was previously regarded as an oncogene, but recent studies provided a more comprehensive understanding of its role in cancer. We demonstrated that NSCLC patients with high EZH2 were more likely to benefit from EGFR-TKI treatment. Consistent with our results, low EZH2 expression was associated with a shorter progression-free survival in colorectal cancer patients who received chemotherapy with anti-EGFR antibodies [19]. Moreover, EZH2-aberrant solid tumors may become less responsive to EZH2 inhibition through gain of MLL1-p300/CBP-mediated H3K27 acetylation, and further co-inhibition of EZH2 and MLL1 may activate an oncogenic MAPK pathway in some cancers [53], which is in agreement with our data that EZH2 knockdown promoted cell growth in HCC827 and PC9 cells. Currently, several EZH2 inhibitors are undergoing clinical trials for multiple malignancies including lung cancer [18]. These findings suggest two potential risks of using EZH2 inhibitors in clinical practice. First, EZH2 inhibition may lead to oncogenic reprogramming itself, limiting its therapeutic utility and suggesting the need to identify the subpopulation of patients who may benefit from EZH2 inhibition. Second, EZH2 inhibition can induce EGFR-TKI resistance in some cancers, including NSCLC, thus raising concerns about the compatibility of EZH2 inhibition with current first-line therapies. As a proof of concept, we interrupted the “MET-AKT-EZH2” loop in vitro and in vivo by inhibiting PI3K/AKT, which significantly enhanced EGFR-TKI sensitivity and therapeutic efficacy. Currently, there are in- hibitors for MET or PI3K/AKT pathway available or undergoing clinical evaluation [9,54]. Our data suggest the potential clinical efficacy of simultaneously targeting this feedback loop and EGFR, along with MET or PI3K/AKT inhibitors to combat EGFR-TKI resistance. In summary, the present study demonstrated a negative correlation between EZH2 and EGFR-TKI resistance, and the existence of a “MET- AKT-EZH2” feedback loop, suggesting that loss of EZH2 can lead to enhanced EGFR-TKI resistance. Our findings further provided insight into crosstalk between MET and epigenetic heterogeneity, indicating the need for caution before using EZH2 inhibitors in lung cancer treatment. These observations suggested the potential value of simultaneously targeting the “MET-AKT-EZH2” feedback loop and EGFR to restore EGFR-TKI sensitivity in clinical practice. Authors’ contributions CTQ, YCC, YKL, XMW, DY, QW, YXH performed the experiments. CTQ and KH designed the study. CTQ, YCC, YKL, RBP, XRL, LZ and KH analyzed the data. CTQ, YCC and KH wrote the manuscript. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments This work was supported by the Natural Science Foundation of China (31671195, 81703552 & 31871381), the Academic Frontier Youth Team Project of HUST, Integrated Innovative Team for Major Human Diseases Program of Tongji Medical College (HUST), the Fundamental Research Funds for the Central Universities (HUST: 2018JYCXJJ007). Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.canlet.2020.09.003. References [1] F. Bray, J. Ferlay, I. Soerjomataram, R.L. Siegel, L.A. Torre, A. Jemal, Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, CA, Cancer. J. Clin. 68 (2018) 394–424. [2] C.R. Chong, P.A. Janne, The quest to overcome resistance to EGFR-targeted therapies in cancer, Nat. Med. 19 (2013) 1389–1400. [3] J.G. Paez, P.A. Janne, J.C. Lee, S. Tracy, H. Greulich, S. Gabriel, P. Herman, F. J. Kaye, N. Lindeman, T.J. Boggon, K. Naoki, H. Sasaki, Y. Fujii, M.J. Eck, W. R. Sellers, B.E. Johnson, M. Meyerson, EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy, Science 304 (2004) 1497–1500. [4] Y.K. Shi, L. Wang, B.H. Han, W. Li, P. Yu, Y.P. Liu, C.M. Ding, X. Song, Z.Y. Ma, X. L. Ren, J.F. Feng, H.L. Zhang, G.Y. Chen, X.H. Han, N. Wu, C. Yao, Y. Song, S. C. Zhang, W. Song, X.Q. Liu, S.J. Zhao, Y.C. Lin, X.Q. Ye, K. Li, Y.Q. Shu, L.M. Ding, F.L. Tan, Y. Sun, First-line icotinib versus cisplatin/pemetrexed plus pemetrexed maintenance therapy for patients with advanced EGFR mutation-positive lung adenocarcinoma (CONVINCE): a phase 3, open-label, randomized study, Ann. Oncol. 28 (2017) 2443–2450. [5] C.W.S. Tong, W.K.K. Wu, H.H.F. Loong, W.C.S. Cho, K.K.W. To, Drug combination approach to overcome resistance to EGFR tyrosine kinase inhibitors in lung cancer, Canc. Lett. 405 (2017) 100–110. [6] D. Westover, J. Zugazagoitia, B.C. Cho, C.M. Lovly, L. Paz-Ares, Mechanisms of acquired resistance to first- and second-generation EGFR tyrosine kinase inhibitors, Ann. Oncol. 29 (2018) i10–i19. [7] B. Ko, T. He, S. Gadgeel, B. Halmos, MET/HGF pathway activation as a paradigm of resistance to targeted therapies, Ann. Transl. Med. 5 (2017) 4. [8] J. Bean, C. Brennan, J.Y. Shih, G. Riely, A. Viale, L. Wang, D. Chitale, N. Motoi, J. Szoke, S. Broderick, M. Balak, W.C. Chang, C.J. Yu, A. Gazdar, H. Pass, V. Rusch, W. Gerald, S.F. Huang, P.C. Yang, V. Miller, M. Ladanyi, C.H. Yang, W. Pao, MET amplification occurs with or without T790M mutations in EGFR mutant lung tumors with acquired resistance to gefitinib or erlotinib, Mol. Canc. Therapeut. 6 (2007) 3333s–3334s. [9] Q. Wang, S. Yang, K. Wang, S.Y. Sun, MET inhibitors for targeted therapy of EGFR TKI-resistant lung cancer, J. Hematol. Oncol. 12 (2019) 63. [10] R. Brown, E. Curry, L. Magnani, C.S. Wilhelm-Benartzi, J. Borley, Poised epigenetic states and acquired drug resistance in cancer, Nat. Rev. Canc. 14 (2014) 747–753. [11] H. Hammerlindl, H. Schaider, Tumor cell-intrinsic phenotypic plasticity facilitates adaptive cellular reprogramming driving acquired drug resistance, J. Cell. Commun. Signal. 12 (2018) 133–141. [12] D. Wan, C. Liu, Y. Sun, W. Wang, K. Huang, L. Zheng, MacroH2A1.1 cooperates with EZH2 to promote adipogenesis by regulating Wnt signaling, J. Mol. Cell Biol. 9 (2017) 325–337. [13] J.K. Lue, J.E. Amengual, Emerging EZH2 inhibitors and their application in lymphoma, Curr. Hematol. Malig. Rep. 13 (2018) 369–382. [14] S. Genta, M.C. Pirosa, A. Stathis, BET and EZH2 inhibitors: novel approaches for targeting cancer, Curr. Oncol. Rep. 21 (2019) 13. [15] K.H. Kim, C.W.M. Roberts, Targeting EZH2 in cancer, Nat. Med. 22 (2016) 128–134. [16] P. Guglielmelli, F. Biamonte, J. Score, C. Hidalgo-Curtis, F. Cervantes, M. Maffioli, T. Fanelli, T. Ernst, N. Winkelman, A.V. Jones, K. Zoi, A. Reiter, A. Duncombe, L. Villani, A. Bosi, G. Barosi, N.C.P. Cross, A.M. Vannucchi, EZH2 mutational status predicts poor survival in myelofibrosis, Blood 118 (2011) 5227–5234. [17] N.A. de Vries, D. Hulsman, W. Akhtar, J. de Jong, D.C. Miles, M. Blom, O. van Tellingen, J. Jonkers, M. van Lohuizen, Prolonged Ezh2 depletion in glioblastoma causes a robust switch in cell fate resulting in tumor progression, Cell Rep. 10 (2015) 383–397. [18] Y. Chen, X. Liu, Y. Li, C. Quan, L. Zheng, K. Huang, Lung cancer therapy targeting histone methylation: opportunities and challenges, Comput. Struct. Biotechnol. J. 16 (2018) 211–223. [19] I. Yamamoto, K. Nosho, S. Kanno, H. Igarashi, H. Kurihara, K. Ishigami, K. Ishiguro, K. Mitsuhashi, R. Maruyama, H. Koide, H. Okuda, T. Hasegawa, Y. Sukawa, K. Okita, I. Takemasa, H. Yamamoto, Y. Shinomura, H. Nakase, EZH2 expression is a prognostic biomarker in patients with colorectal cancer treated with anti-EGFR therapeutics, Oncotarget 8 (2017) 17810–17818. [20] S. Gollner, T. Oellerich, S. Agrawal-Singh, T. Schenk, H.U. Klein, C. Rohde, C. Pabst, T. Sauer, M. Lerdrup, S. Tavor, F. Stolzel, S. Herold, G. Ehninger, G. Kohler, K.T. Pan, H. Urlaub, H. Serve, M. Dugas, K. Spiekermann, B. Vick, I. Jeremias, W.E. Berdel, K. Hansen, A. Zelent, C. Wickenhauser, L.P. Muller, C. Thiede, C. Muller-Tidow, Loss of the histone methyltransferase EZH2 induces resistance to multiple drugs in acute myeloid leukemia, Nat. Med. 23 (2017) 69–78. [21] J.A. Engelman, K. Zejnullahu, T. Mitsudomi, Y. Song, C. Hyland, J.O. Park, N. Lindeman, C.M. Gale, X. Zhao, J. Christensen, T. Kosaka, A.J. Holmes, A. M. Rogers, F. Cappuzzo, T. Mok, C. Lee, B.E. Johnson, L.C. Cantley, P.A. Janne, MET amplification leads to gefitinib resistance in lung cancer by activating ERBB3 signaling, Science 316 (2007) 1039–1043. [22] S. Benavente, S. Huang, E.A. Armstrong, A. Chi, K.T. Hsu, D.L. Wheeler, P. M. Harari, Establishment and characterization of a model of acquired resistance to epidermal growth factor receptor targeting agents in human cancer cells, Clin. Canc. Res. 15 (2009) 1585–1592. [23] Y. Zhang, W. Xue, W. Zhang, Y. Yuan, X. Zhu, Q. Wang, Y. Wei, D. Yang, C. Yang, Y. Chen, Y. Sun, S. Wang, K. Huang, L. Zheng, Histone methyltransferase G9a protects against acute liver injury through GSTP1, Cell Death Differ. 27 (2020) 1243–1258. [24] X. Liu, Y. Zhou, X. Liu, A. Peng, H. Gong, L. Huang, K. Ji, R.B. Petersen, L. Zheng, K. Huang, MPHOSPH1: a potential therapeutic target for hepatocellular carcinoma, Cancer. Res. 74 (2014) 6623–6634. [25] X. Li, H. Li, S. Li, F. Zhu, D.J. Kim, H. Xie, Y. Li, J. Nadas, N. Oi, T.A. Zykova, D. H. Yu, M.-H. Lee, M.O. Kim, L. Wang, W. Ma, R.A. Lubet, A.M. Bode, Z. Dong, Z. Dong, Ceftriaxone, an FDA-approved cephalosporin antibiotic, suppresses lung cancer growth by targeting Aurora B, Carcinogenesis 33 (2012) 2548–2557. [26] S. Liu, Y. Sun, M. Jiang, Y. Li, Y. Tian, W. Xue, N. Ding, Y. Sun, C. Cheng, J. Li, X. Miao, X. Liu, L. Zheng, K. Huang, Glyceraldehyde-3-phosphate dehydrogenase promotes liver tumorigenesis by modulating phosphoglycerate dehydrogenase, Hepatology 66 (2017) 631–645. [27] X. Liu, Y. Li, L. Meng, X.Y. Liu, A. Peng, Y. Chen, C. Liu, H. Chen, S. Sun, X. Miao, Y. Zhang, L. Zheng, K. Huang, Reducing protein regulator of cytokinesis 1 as a prospective therapy for hepatocellular carcinoma, Cell Death Dis. 9 (2018) 534. [28] W. Wang, Q. Wang, D. Wan, Y. Sun, L. Wang, H. Chen, C. Liu, R.B. Petersen, J. Li, W. Xue, L. Zheng, K. Huang, Histone HIST1H1C/H1.2 regulates autophagy in the development of diabetic retinopathy, Autophagy 13 (2017) 941–954. [29] H. Chen, L. Wang, W. Wang, C. Cheng, Y. Zhang, Y. Zhou, C. Wang, X. Miao, J. Wang, C. Wang, J. Li, L. Zheng, K. Huang, ELABELA and an ELABELA fragment protect against AKI, J. Am. Soc. Nephrol. 28 (2017) 2694–2707. [30] H. Chen, Y.X. Huang, X.Q. Zhu, C. Liu, Y.M. Yuan, H. Su, C. Zhang, C.Y. Liu, M. R. Xiong, Y.N. Qu, P. Yun, L. Zheng, K. Huang, Histone demethylase UTX is a therapeutic target for diabetic kidney disease, J. Physiol. (London). 597 (2019) 1643–1660. [31] Y. Zhang, X. Guo, W. Yan, Y. Chen, M. Ke, C. Cheng, X. Zhu, W. Xue, Q. Zhou, L. Zheng, S. Wang, B. Wu, X. Liu, L. Ma, L. Huang, K. Huang, ANGPTL8 negatively regulates NF-kappaB activation by facilitating selective autophagic degradation of IKKgamma, Nat. Commun. 8 (2017) 2164. [32] Z. Zhang, J.C. Lee, L. Lin, V. Olivas, V. Au, T. LaFramboise, M. Abdel-Rahman, X. Wang, A.D. Levine, J.K. Rho, Y.J. Choi, C.M. Choi, S.W. Kim, S.J. Jang, Y. S. Park, W.S. Kim, D.H. Lee, J.S. Lee, V.A. Miller, M. Arcila, M. Ladanyi, P. Moonsamy, C. Sawyers, T.J. Boggon, P.C. Ma, C. Costa, M. Taron, R. Rosell, B. Halmos, T.G. Bivona, Activation of the AXL kinase causes resistance to EGFR- targeted therapy in lung cancer, Nat. Genet. 44 (2012) 852–860. [33] Y. Sun, Q. Wang, Y. Zhang, M. Geng, Y. Wei, Y. Liu, S. Liu, R.B. Petersen, J. Yue, K. Huang, L. Zheng, Multigenerational maternal obesity increases the incidence of HCC in offspring via miR-27a-3p, J. Hepatol. 73 (2020) 603–615. [34] K. Suda, H. Mizuuchi, Y. Maehara, T. Mitsudomi, Acquired resistance mechanisms to tyrosine kinase inhibitors in lung cancer with activating epidermal growth factor receptor mutation–diversity, ductility, and destiny, Cancer Metastasis Rev 31 (2012) 807–814. [35] D. Brehmer, Z. Greff, K. Godl, S. Blencke, A. Kurtenbach, M. Weber, S. Muller, B. Klebl, M. Cotten, G. Keri, J. Wissing, H. Daub, Cellular targets of gefitinib, Cancer. Res. 65 (2005) 379–382. [36] G.D. Guler, C.A. Tindell, R. Pitti, C. Wilson, K. Nichols, T. KaiWai Cheung, H.- J. Kim, M. Wongchenko, Y. Yan, B. Haley, T. Cuellar, J. Webster, N. Alag, G. Hegde, E. Jackson, T.L. Nance, P.G. Giresi, K.-B. Chen, J. Liu, S. Jhunjhunwala, J. Settleman, J.-P. Stephan, D. Arnott, M. Classon, Repression of stress-induced LINE-1 expression protects cancer cell subpopulations from lethal drug exposure, Canc. Cell 32 (2017) 221–237, e213. [37] M.Y. Ding, H. Zhang, Z. Li, C.L. Wang, J.M. Chen, L.Y. Shi, D.K. Xu, Y. Gao, The polycomb group protein enhancer of zeste 2 is a novel therapeutic target for cervical cancer, Clin. Exp. Pharmacol. Physiol. 42 (2015) 458–464. [38] T.J. Lynch, D.W. Bell, R. Sordella, S. Gurubhagavatula, R.A. Okimoto, B. W. Brannigan, P.L. Harris, S.M. Haserlat, J.G. Supko, F.G. Haluska, D.N. Louis, D. C. Christiani, J. Settleman, D.A. Haber, Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib, N. Engl. J. Med. 350 (2004) 2129–2139. [39] C.H. Chiu, C.T. Yang, J.Y. Shih, M.S. Huang, W.C. Su, R.S. Lai, C.C. Wang, S. H. Hsiao, Y.C. Lin, C.L. Ho, T.C. Hsia, M.F. Wu, C.L. Lai, K.Y. Lee, C.B. Lin, D. Yu- Wung Yeh, C.Y. Chuang, F.K. Chang, C.M. Tsai, R.P. Perng, J. Chih-Hsin Yang, Epidermal growth factor receptor tyrosine kinase inhibitor treatment response in advanced lung adenocarcinomas with G719X/L861Q/S768I mutations, J. Thorac. Oncol. 10 (2015) 793–799. [40] Y. Shi, L. Zhang, X. Liu, C. Zhou, L. Zhang, S. Zhang, D. Wang, Q. Li, S. Qin, C. Hu, Y. Zhang, J. Chen, Y. Cheng, J. Feng, H. Zhang, Y. Song, Y.-L. Wu, N. Xu, J. Zhou, R. Luo, C. Bai, Y. Jin, W. Liu, Z. Wei, F. Tan, Y. Wang, L. Ding, H. Dai, S. Jiao, J. Wang, L. Liang, W. Zhang, Y. Sun, Icotinib versus gefitinib in previously treated advanced non-small-cell lung cancer (ICOGEN): a randomised, double-blind phase 3 non-inferiority trial, Lancet Oncol. 14 (2013) 953–961. [41] J. Ni, L. Zhang, Evaluation of three small molecular drugs for targeted therapy to treat nonsmall cell lung cancer, Chin. Med. J. (Engl). 129 (2016) 332–340. [42] Y. Liu, Y. Zhang, G. Feng, Q. Niu, S. Xu, Y. Yan, S. Li, M. Jing, Comparison of effectiveness and adverse effects of gefitinib, erlotinib and icotinib among patients with non-small cell lung cancer: a network meta-analysis, Exp. Ther. Med. 14 (2017) 4017–4032. [43] Y. Shi, Y. Sun, C. Ding, Z. Wang, C. Wang, C. Bai, C. Bai, J. Feng, X. Liu, F. Li, Y. Yang, Y. Shu, M. Wu, J. He, Y. Zhang, S. Zhang, G. Chen, H. Luo, R. Luo, C. Zhou, Q. Pang, X. Hu, H. Zhao, Q. Zhao, A. Gu, Y. Ling, C. Huang, B. Han, S. Jiao, H. Jian, [China experts consensus on icotinib for non-small cell lung cancer treatment(2016 version)], Zhongguo Fei Ai Za Zhi 19 (2016) 489–494. [44] D. Jackman, W. Pao, G.J. Riely, J.A. Engelman, M.G. Kris, P.A. Janne, T. Lynch, B. E. Johnson, V.A. Miller, Clinical definition of acquired resistance to epidermal growth factor receptor tyrosine kinase inhibitors in non-small-cell lung cancer, J. Clin. Oncol. 28 (2010) 357–360. [45] T.L. Cha, B.P. Zhou, W. Xia, Y. Wu, C.C. Yang, C.T. Chen, B. Ping, A.P. Otte, M. C. Hung, Akt-mediated phosphorylation of EZH2 suppresses methylation of lysine 27 in histone H3, Science 310 (2005) 306–310. [46] M. Kaur, M.D. Cole, MYC acts via the PTEN tumor suppressor to elicit autoregulation and genome-wide gene repression by activation of the Ezh2 methyltransferase, Cancer. Res. 73 (2013) 695–705. [47] A.B. Turke, K. Zejnullahu, Y.-L. Wu, Y. Song, D. Dias-Santagata, E. Lifshits, L. Toschi, A. Rogers, T. Mok, L. Sequist, N.I. Lindeman, C. Murphy, S. Akhavanfard, B.Y. Yeap, Y. Xiao, M. Capelletti, A.J. Iafrate, C. Lee, J.G. Christensen, J. A. Engelman, P.A. J¨anne, Preexistence and clonal selection of MET amplification in EGFR mutant NSCLC, Canc. Cell 17 (2010) 77–88. [48] S. Nanjo, S. Arai, W. Wang, S. Takeuchi, T. Yamada, A. Hata, N. Katakami, Y. Okada, S. Yano, MET copy number gain is associated with gefitinib resistance in leptomeningeal carcinomatosis of EGFR-mutant lung cancer, Mol. Canc. Therapeut. 16 (2017) 506–515. [49] Y.-L. Wu, J.C.-H. Yang, D.-W. Kim, W.-C. Su, M.-J. Ahn, D.H. Lee, J. F. Vansteenkiste, L. Zhang, E. Felip, B. Peng, Y. Gong, S. Zhao, T. Amagasaki, M. Akimov, D.S.-W. Tan, Safety and efficacy of INC280 in combination with gefitinib (gef) in patients with EGFR-mutated (mut), MET-positive NSCLC: a single- arm phase lb/ll study, J. Clin. Oncol. 32 (2014), https://doi.org/10.1200/ jco.2014.32.15_suppl.8017, 8017-8017. [50] E. Riquelme, M. Suraokar, C. Behrens, H.Y. Lin, L. Girard, M.B. Nilsson, G. Simon, J. Wang, K.R. Coombes, J.J. Lee, W.K. Hong, J. Heymach, J.D. Minna, Wistuba II, VEGF/VEGFR-2 upregulates EZH2 expression in lung adenocarcinoma cells and EZH2 depletion enhances the response to platinum-based and VEGFR-2-targeted therapy, Clin. Canc. Res. 20 (2014) 3849–3861. [51] L. Wan, K. Xu, Y. Wei, J. Zhang, T. Han, C. Fry, Z. Zhang, Y.V. Wang, L. Huang, M. Yuan, W. Xia, W.C. Chang, W.C. Huang, C.L. Liu, Y.C. Chang, J. Liu, Y. Wu, V. X. Jin, X. Dai, J. Guo, J. Liu, S. Jiang, J. Li, J.M. Asara, M. Brown, M.C. Hung, W. Wei, Phosphorylation of EZH2 by AMPK suppresses PRC2 methyltransferase activity and oncogenic function, Mol. Cell. 69 (2018) 279–291, e275. [52] X. Liu, Y. Chen, Y. Li, R.B. Petersen, K. Huang, Targeting mitosis exit: a brake for cancer cell proliferation, Biochim. Biophys. Acta Rev. Canc 1871 (2019) 179–191. [53] X. Huang, J. Yan, M. Zhang, Y. Wang, Y. Chen, X. Fu, R. Wei, X.L. Zheng, Z. Liu, X. Zhang, H. Yang, B. Hao, Y.Y. Shen, Y. Su, X. Cong, M. Huang, M. Tan, J. Ding, M. Geng, Targeting epigenetic crosstalk as a therapeutic strategy for EZH2- aberrant solid tumors, Cell 175 (2018) 186–199. [54] C. Fumarola, M.A. Bonelli, P.G. Petronini, R.R. Alfieri, Targeting PI3K/AKT/mTOR pathway in non small cell lung cancer, Biochem. Pharmacol.Pictilisib 90 (2014) 197–207.