Abnormally activated OPN/integrin αVβ3/FAK signalling is responsible for EGFR-TKI resistance in EGFR mutant non-small-cell lung cancer

PC9, HCC827 and H1975 cell lines were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai, China). Caicun Zhou at Tongji University School of Medicine provided PC9GR cells. HCC827GR cells were derived from HCC827 cells by exposure to gefitinib, as previously described [24]. PC9 and PC9GR cells were routinely cultured in RPMI-1640 medium supplemented with 10% foetal bovine serum (Gibco, Carlsbad, CA, USA). HCC827, HCC827GR, and H1975 cells were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated FBS (Gibco). Cells were grown in a humidified incubator with 5% CO₂ at 37 °C.

Gefitinib was supplied by AstraZeneca (London, UK). Osimertinib and the FAK inhibitor VS-6063 were purchased from Selleck Chemicals (Houston, TX, USA).

Cells were plated in each well of a 96-well plate at a density of 3000 cells per well, grown overnight and then treated with varying drug concentrations for 72 h. The Cell Counting Kit-8 (CCK-8) assay kit (Boster, Wuhan, China) was used according to the manufacturer’s instructions to assess cell viability. Fluorescence at 630 nm and 450 nm was measured using a microplate reader after 1–2 h (Thermo, Waltham, MA, USA).

The protein profile was analysed using a human soluble receptor array kit, non-haematopoietic panel-ARY012 (R&D Inc. Minneapolis, MN, USA) according to the manufacturer’s protocols. PC9, HCC827, PC9GR and HCC827R cells were lysed with lysis buffer mixed with proteinase cocktail inhibitor (Roche, Branford, CT, USA). Cell lysates were pipetted up and down for resuspension on ice for 30 min and then centrifuged at 14,000 × g at 4 °C for 5 min. The protein lysates were collected, and the concentrations were determined by the bicinchoninic acid assay (BCA assay). The protein lysates (100–300 μg per membrane) were incubated overnight with nitrocellulose membranes containing 62 soluble receptors. The membranes were subsequently incubated first with a specific cocktail of biotinylated detection antibodies and later with the streptavidin–horseradish peroxidase solution. Signals were detected by using a chemifluorescence detection system (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s protocol. The relative density of specific protein expression was determined using Quantity One software.

We used Human Phospho-Kinase Array Kit-ARY003B (R&D) for the human receptor tyrosine kinase (RTK) assay. Cells were seeded into 6-well plates at a concentration of 3 × 10⁵ cells/well. After 24 h, the cells were harvested and lysed in RIPA buffer (Cell Signalling Technology, Danvers, MA, USA) containing a protease and phosphatase inhibitor cocktail (Sigma-Aldrich, Louis, MO, USA). The protein lysates were incubated with the array membrane, and the protein signal was visualized using chemifluorescent detection (Bio-Rad) according to the manufacturer’s protocol. The relative density of specific protein expression was determined using Quantity One software.

Antibodies against the following were obtained from Cell Signalling Technology: p-EGFR Y1068 (#2234), ERK (#9102), p-ERK T202/Y204 (#9101), AKT (#9272), p-AKT S473(#4060), FAK (#3285), p-FAK Y397 (#8556), SRC (#2108), p-SRC Y416 (#2101), and PARP (#9542). The β-actin antibody (CW0096M), GAPDH antibody (CW0100M), horseradish peroxidase (HRP)-conjugated anti-mouse antibody (CW0102), and HRP-conjugated anti-rabbit antibody (CW0103) were purchased from CoWin Biosciences (Beijing, China). Anti-EGFR (C-2), OPN (LFMb-14), -integrin αV (P2W7), and -integrin β3 (B-7) antibodies were purchased from Santa Cruz Biotechnology. The anti-integrin β1 (AF5379) antibody was purchased from Affinity Biotechnology. For immunoblotting, cells were harvested, washed in PBS, and lysed in RIPA buffer [50 mmol/L Tris–HCl (pH 8.0), 150 mmol/L sodium chloride, 5 mmol/L magnesium chloride, 1% Triton X-100; 0.5% sodium deoxycholate, 0.1% SDS, 40 mmol/L sodium fluoride, 1 mmol/L sodium orthovanadate, and complete protease inhibitors (Selleck Chemicals, Houston, TX, USA)]. Western Lightning ECL reagent (Thermo) was used for signal detection.

Transfection was carried out using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, NM, USA) according to the manufacturer's protocol. The target siRNA sequences used are listed in Additional file 1: Table S1. siRNAs were used at a concentration of 10 nmol/L (GenePharma, Shanghai, China). The efficacy of transfection was verified by qRT-PCR.

Cells were seeded in 6-well plates at a density of 5 × 10⁴ cells per well and treated with drugs at different concentrations or DMSO as a negative control. We analysed apoptosis and the cell cycle status of the cells by using Annexin V-FITC and propidium iodide (PI) (R&D) staining according to the manufacturer’s protocol.

The secreted OPN (sOPN) was measured using ELISA (MultiSciences, Hangzhou, China). Briefly, 50 μL of diluent was added to each well of the microplate. Then, another 50 μL standard, control, or sample was added to each well and incubated for 2 h at room temperature. After washing four times, the conjugate reagent was added to each well and incubated for 2 h at room temperature. After washing four times, 100 μL substrate solution was added to each well, and the reaction was stopped after 30 min by adding the stop solution. Absorbance at 450 nm was measured using a spectrophotometer (Thermo). PC9GR cells were exposed to medium containing 1% FBS for 12 h and then treated with the following inhibitors separately: the autophagy inhibitor 3-methyladenine (1 mM, Selleck Chemicals), the protein transport inhibitor brefeldin A (10 ng/ml, APExBIO, Houston, TX, USA), and the exosome secretion inhibitor 5-(N, N-dimethyl)- amiloride DMA (50 nM, APExBIO). The supernatant was collected from the cultured medium 24 h later.

Total RNA was extracted from cells using TRIzol reagent (Invitrogen) and reverse transcribed to cDNA using reverse transcription reagents (Takara Bio, Shiga, Japan) according to the manufacturer's protocol. The primer sequences used for reverse transcription quantitative polymerase chain reaction (qRT‐PCR) are listed in Additional file 1: Table S1. Quantitative RT-PCR was performed using SYBR Premix ExTaq™ (Takara) with an ABI StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA) according to the operator's manual. The expression values of genes were normalized to the internal control GAPDH.

Seven NSCLC tissue samples were collected from patients between 2017 and 2018 at the Respiratory Department of the First Affiliated Hospital of Soochow University. All participants provided written informed consent at the time of recruitment. All cases had clinically and pathologically confirmed diagnoses of NSCLC in accordance with the Revised International System for Staging Lung Cancer.

Immunohistochemical (IHC) analysis was conducted in our previous study. Briefly, sections were incubated with anti-ITGαV (EPR16800, 1:200 dilution; Abcam) overnight at 4 °C and then with biotinylated secondary antibodies. The reactions were developed using DAB Kit (BD Bioscience, San Jose, CA, USA), and the sections were counterstained with haematoxylin. The staining area was scored using the following scale: 0, 0–10% of tissue stained positive; 1, 10–20% stained positive; 2, 20–40% stained positive; 3, 40–70% stained positive; and 4, > 70% positive cells. The IHC score was generated from three different areas of the slides, and the average score was calculated for each sample.

Cultured cells were fixed with 4% paraformaldehyde for 15 min at room temperature, permeabilized with Triton (0.1% in TBS) for 30 min and blocked with 5% BSA in PBS for 1 h at room temperature. The cells were then incubated overnight at 4 °C with anti-ITGαV antibodies (EPR16800, 1:200 dilution; Abcam) followed by Alexa Fluor 488-conjugated anti-rabbit IgG (Beyotime, Shanghai, China) for 90 min. Finally, the samples were incubated in DAPI for 10 min (Life Technologies) for nuclear counterstaining. Images were acquired using a Leica SP8 confocal microscope with optimal settings for the fluorescent markers used.

Male athymic BALB/c nude mice were purchased from the Experimental Animal Center of Soochow University and bred under pathogen-free conditions. All experimental procedures were reviewed and approved in accordance with the guidelines for the care and use of laboratory animals, and informed written consent was obtained from Soochow University. To establish mouse xenograft models, the same amount of the indicated tumour cells was injected subcutaneously into both flanks of each mouse. Tumour volumes (mm³) were calculated as length × width²/2. When tumours reached ~ 200 mm³, gefitinib and VS-6063 were given by gavage at 12.5 mg/kg and 25 mg/kg daily, respectively, until the mice were sacrificed.

All results are presented as the mean standard deviation (SD). Two-way ANOVA was used to calculate the difference in IC50 for EGFR-TKIs among cells with various treatments. Statistical comparisons were determined with Student's t test, and P < 0.05 was regarded as significant. All statistical analyses were performed with GraphPad Prism 7.0 (GraphPad, San Diego, CA) and SPSS 17.0 software (SPSS, Chicago, IL).

To determine which specific protein contributes to acquired EGFR-TKI resistance in NSCLC, two EGFR mutant NSCLC cell lines, PC9 and HCC827 [25], were exposed to gefitinib at gradually increasing concentrations for more than 6 months to establish gefitinib-resistant (GR) cell lines (PC9GR and HCC827GR), as previously reported [24]. PC9GR and HCC827GR cells were highly insensitive to gefitinib compared to their respective parental cells (Additional file 2: Fig S1). Our previous work showed that the PC9GR cell line harbours EGFR 19del and T790M mutations; the HCC827GR cell line harbours the EGFR 19del mutation and overexpresses FGFR1 and AXL [26, 27]. To investigate whether there are other mechanisms responsible for acquired resistance to EGFR-TKIs in addition to variations such as the T790M mutation, MET amplification and overexpression of FGFR1 and AXL, we examined expression of soluble receptors and related proteins in PC9GR and HCC827GR cells by proteome profiler array analysis. We found that OPN was the only protein up-regulated simultaneously in the two EGFR TKI-resistant cell lines (fold change > 1.5) (Fig. 1a, Additional file 3: Table S2 and Additional file 4: Fig S2). Then, OPN up-regulation in PC9GR and HCC827GR cells was verified by qRT‐PCR and western blot assays (Fig. 1b, c). As OPN is considered to perform multiple functions as a canonical secretory protein [16], we detected expression of secreted OPN (sOPN) by ELISA. sOPN expression was apparently higher in PC9GR cells than in PC9 cells. In contrast, HCC827GR cells expressed less sOPN than HCC827 cells (Fig. 1d). In addition, we found OPN could be secreted via Golgi apparatus, secretory autophagy and exosome in PC9GR cells (Additional file 5: Fig. S3). These results suggest that OPN is more associated with EGFR-TKI resistance in PC9 cells. Therefore, further explorations were mainly performed using PC9 and PC9GR cells. To examine whether OPN is positively associated with acquired resistance to EGFR-TKIs, OPN mRNA and protein were significantly silenced by siRNA transient transfection in PC9GR cells (Fig. 1e). Secretory OPN was also down-regulated after si-OPN transfection (Fig. 1f). A cell viability (CCK8) assay demonstrated that silencing OPN significantly increased the gefitinib sensitivity of PC9GR cells (Fig. 1g). Furthermore, silencing OPN blocked the phosphorylation of AKT and ERK (Fig. 1h). Interestingly, OPN secretion was increased not only in gefitinib-resistant cells (PC9GR) but also in osimertinib-resistant cells (PC9GR/OR) (Additional file 6: Fig. S4A and B). These results suggest that OPN promotes acquired gefitinib resistance in NSCLC cells.

Given that OPNs are reported to bind to multiple integrins through Arg-Gly-Asp (RGD)-mediated integrin recognition sequences or alternative integrin recognition sequences to mediate cell–cell and cell-ECM interactions [28, 29], we hypothesize that integrins are involved in OPN-mediated acquired resistance to EGFR-TKIs. As shown in Fig. 2a, integrins αV and β3 were remarkably overexpressed in PC9GR cells compared to PC9 cells, which is consistent with the results of the proteome profiler array analysis (Additional file 3: Table S2 and Additional file 4: Fig S2). RNA data from the database TCGA showed that mRNA expression of ITGAV and ITGB3 correlated positively with SPP1 expression in lung adenocarcinoma and squamous patients (Additional file 7: Fig. S5). Furthermore, the sensitivity of PC9GR cells to gefitinib was increased with knockdown of integrin αV (Fig. 2b) or integrin β3 (Fig. 2c). Integrin αV expression was detected in 7 tumour tissues from lung adenocarcinoma patients before and after EGFR-TKI resistance acquisition (Additional file 8: Table S3). As shown in Fig. 2d, expression of integrin αV was increased in five patients with EGFR-TKI resistance (5/7 = 71.43%) but was not changed in patient 6 with MET amplification. These data demonstrate that integrin αVβ3 contributes to acquired EGFR-TKI resistance in NSCLC.

To identify kinases that are related to EGFR-TKI in EGFR mutant NSCLC, we compared the difference in activity of protein kinases between PC9 and PC9GR cells and between HCC827 and HCC827GR cells using kinome-wide array analysis (Additional file 9: Table S4 and Additional file 10: Fig S6). As shown in Fig. 3a, b, phosphorylated FAK (Tyr397) was activated in PC9GR cells in comparison with PC9 cells (Fig. 3a) but was reduced in HCC827GR cells compared with HCC827 cells (Fig. 3b). These results were further verified by western blot analysis (Fig. 3c). Consistent with FAK, its downstream factors, including SRC, AKT and ERK, were phosphorylated and activated in PC9GR cells. In addition, silencing integrins αV and β3 reduced the phosphorylation of FAK and its downstream factors in PC9GR cells (Fig. 3d). These results suggest that activation of the FAK signalling pathway mediates acquired EGFR-TKI resistance in NSCLC.

Because we found that OPN plays a role in increasing gefitinib resistance, it is natural to assume that integrins and FAK signalling are involved in this process. Expression of OPN (Fig. 4a, left) and secretion of sOPN (Fig. 4b, right) was up-regulated in PC9 cells treated with gefitinib. As illustrated in Fig. 4b, phosphorylation of EGFR and downstream AKT and ERK activities were significantly suppressed by gefitinib treatment at concentrations greater than 0.05 μM. In contrast, FAK phosphorylation was markedly enhanced in PC9 cells treated with 0.01 μM gefitinib (Fig. 4b). Moreover, upon 0.05 μM gefitinib stimulation, levels of p-FAK, as well as of integrins αV and β3, were gradually up-regulated with treatment time (Fig. 4c). Moreover, expression of p-AKT and p-ERK, which were inhibited by gefitinib-blocked p-EGFR, was partially restored after 72 h of gefitinib stimulation (Fig. 4c). As expected, we found that knockdown of OPN significantly prevented FAK phosphorylation as well as gefitinib-induced FAK phosphorylation in PC9 cells (Fig. 4d). FAK phosphorylation was also significantly activated by treatment with recombinant human OPN (rOPN), and in turn, this process was blocked by VS-6063, a specific inhibitor of FAK activity in PC9 cells (Fig. 4e). Furthermore, integrin αV was up-regulated by recombinant human OPN (rOPN) treatment, as determined by western blot (upper) and immunofluorescence (lower) assays (Fig. 4f). Finally, silencing integrin αV (Fig. 4g) or β3 (Fig. 4h) suppressed gefitinib-induced FAK phosphorylation, which mirrored the results of OPN knockdown. Given that OPN interacts mainly with various integrin αVs (particularly αVβ1 and αVβ3) [30, 31], it is necessary to determine whether integrin β1 is also involved in FAK signalling. We found that knockdown of integrinβ1 did not affect FAK activity after stimulation with gefitinib in PC9GR cells (Additional file 11: Fig. S7). These results indicate that OPN activates gefitinib-induced FAK signalling by up-regulating integrin αVβ3 expression in NSCLC.

To investigate whether the FAK signalling pathway contributes to acquired resistance to different EGFR-TKIs, we examined expression of p-FAK in PC9 cells treated with osimertinib at various concentrations and times. Similar to the results obtained with gefitinib treatment, p-FAK expression increased with increasing osimertinib concentration and time (Fig. 5a, b). Moreover, expression of integrin αvβ3 gradually increased with the extension of 0.05 μM osimertinib treatment time (Fig. 5b). Given that osimertinib is an oral, irreversible EGFR-TKI that is selective for both EGFR and T790M resistance mutations [32], we also evaluated p-FAK in T790M mutant cells (PC9GR and H1975) treated with osimertinib [25, 27]. As shown in Fig. 5c, d, p-FAK levels were elevated in PC9GR and H1975 cells treated with different concentrations of osimertinib. These results suggest that the FAK signalling pathway contributes to acquired resistance to osimertinib in NSCLC.

To verify whether the FAK signalling pathway is involved in acquired resistance to EGFR-TKIs, we explored the gefitinib sensitivity of PC9GR after treatment with a p-FAK inhibitor (VS-6063). As shown in Fig. 6a, when PC9GR cells were treated with VS-6063, their sensitivity to gefitinib was significantly increased. A cell proliferation assay indicated that gefitinib in combination with VS-6063 significantly inhibited PC9GR cell growth compared with each single agent (Fig. 6b). Moreover, flow cytometry results indicated that gefitinib in combination with VS-6063 significantly increased PC9GR cell apoptosis compared with each single agent (Fig. 6c). Then, we detected Cyclin D1 and cleaved PARP in PC9 GR cells treated with VS-6063 in the presence of gefitinib by western blot assay. When PC9GR cells were treated with gefitinib and VS-6063, we found that compared to single-agent treatment, Cyclin D1 and cleaved PARP levels were increased, accompanied by decreased phosphorylation of AKT and ERK (Fig. 6d). In the presence of gefitinib, silencing FAK also reduced phosphorylation of AKT and ERK (Fig. 6e). To verify our results, we assessed the effect of gefitinib in combination with VS-6063 in a mouse xenograft model. As shown in Fig. 7a–c, the effect of gefitinib combined with VS-6063 on inhibiting the growth of PC9GR cells was stronger than that of each single drug. Furthermore, we verified the inhibition of AKT and ERK phosphorylation by combination therapy by western blotting (Fig. 7d). The results indicated that inhibition of the FAK signalling pathway increased sensitivity to EGFR-TKIs in NSCLC. Taken together, our findings reveal that OPN contributes to acquired EGFR-TKI resistance by up-regulating expression of integrin αVβ3, which activates the downstream FAK/AKT and ERK signalling pathways to promote cell proliferation in NSCLC (Fig. 7i).