Blocking autophagy improves the anti-tumor activity of afatinib in lung adenocarcinoma with activating EGFR mutations in vitro and in vivo

Afatinib, a second-generation tyrosine kinase inhibitor (TKI), has been approved for the treatment of advanced EGFR-mutant non-small cell lung cancer (NSCLC). However, afatinib’s clinical application is still hampered by acquired resistance. Recently, autophagy is considered as an important mechanism of resistance to TKI. Herein, we investigated the autophagy induction as well as its influence on anti-lung adenocarcinoma activity of afatinib in two activating EGFR-mutants H1975 and H1650 cells. First, Growth inhibition and caspase-dependent apoptosis were observed in afatinib-treated H1975 and H1650 cells. Then we confirmed afatinib-induced autophagy in H1975 and H1650 cells. Importantly, autophagy inhibition using chloroquine (CQ) and 3-MA enhanced the cytotoxicity of afatinib, elucidating the cytoprotective role of autophagy in lung adenocarcinoma therapy with afatinib. Further study suggested that Akt/mTOR and Erk signaling pathways were involved in afatinib-induced autophagy, and reactive oxygen species (ROS) acted as an intracellular transducer regulating both autophagy and apoptosis in afatinib-treated H1975 and H1650 cells. Moreover, the in vivo experiment in xenograft model using H1975 cell line confirmed the enhanced anti-lung adenocarcinoma efficacy of afatinib when combined with autophagy inhibitor CQ. Thus, blocking autophagy may be a promising strategy to overcome resistance and increase sensitivity to afatinib in lung adenocarcinoma harboring activating EGFR mutations.

Despite improved clinical efficacy in the treatment for advanced NSCLC with activating EGFR mutations, afatinib's clinical application is still hampered by acquired resistance and adverse events 11 . It is suggested that combination therapy may be useful in terms of overcoming resistance and improving tolerability to afatinib 14 . Autophagy is an intracellular catabolic process that maintains cellular energetic balance through the degradation of proteins and organelles in lysosomes 15,16 , and can be upregulated by environmental stimuli including chemotherapeutic agents, oxidative stress and nutrient starvation [17][18][19][20] . Although autophagy has a role as a double-edged sword in disease and health 21 , many studies show that it serves a cytoprotective function particularly in cancer treatment [22][23][24] . Recently, autophagy is considered as an important mechanism of resistance to TKI including afatinib 25 . It has been proved that autophagy is involved in the induction of erlotinib resistance in tongue squamous cell carcinoma 26 , and blocking autophagic flux in pancreatic cancer cell lines sensitizes EGFR-TKI-induced non-apoptotic cell death 27 . However, the role of autophagy in advanced EGFR-mutant NSCLC therapy using afatinib has not been clearly elucidated.
In this study, two activating EGFR-mutants H1975 (L858R/T790M) and H1650 (exon 19 deletion) cells were selected to study the activities of cytotoxicity and autophagy induction of afatinib. We found that afatinib not only induced growth inhibition and caspase-dependent apoptosis, but also activated autophagic response in H1975 and H1650 cells, and autophagy played a survival role in this process. The mechanism by which afatinib initiated autophagy was also explored. In addition, in tumor xenograft model using H1975 cell line performed in nude mice, the combination treatment of afatinib and CQ significantly reduced the tumor volume and tumor weight. Taken together, autophagy plays a cytoprotective role in afatinib therapy and the inhibition of autophagy shows an increased sensitivity of lung adenocarcinoma cells harboring activating EGFR mutations to afatinib.

Afatinib induces growth inhibition and apoptosis in H1975 and H1650 cells. Afatinib is a anili-
no-quinazoline derivative that can covalently bind to Cys 773 of EGFR, Cys 805 of HER2 and Cys803 of ErbB4, and its structure is shown in Fig. 1A. H1650 and H1975 cells were dealt with afatinib (0-20 μM) for indicated time, then the MTT assay was used to detect the cell viability. We can see from the Fig. 1B and C that H1975 and H1650 cells viability could be suppressed by afatinib dose-and time-dependently. Meanwhile, the results of western blot illustrated that the expression of cleaved-PARP increased in afatinib-treated cells ( Fig. 1D and E), indicating that the apoptosis was induced by afatinib in H1650 and H1975 cells. These data show that afatinib can induce growth inhibition and apoptosis in lung adenocarcinoma cells with activating EGFR mutations through a dose-and time-dependent way.

Afatinib-triggered apoptosis is partially caspase 3-dependent in H1975 and H1650 cells.
To further determine whether apoptosis induced by afatinib in H1975 and H1650 cells was correlated to the activation of caspase 3, we used casepase 3 activity assay kit to detect the casepase 3 activity of cells which were treated with different concentrations of afatinib for 24 h. From Fig. 2A, we knew that afatinib dramatically increased casepase 3 activity in a dose-dependent manner in both cells. In addition, a pan-caspase inhibitor benzyloxycarbonyl Val-Ala-Asp (O-methyl)-fluoro-methyl ketone (z-VAD-fmk) was employed to test its effect on cell viability and casepase 3 activity. The results showed that 20 μM of z-VAD-fmk could significantly decrease the cytotoxicity of afatinib in both cell lines (Fig. 2B). Moreover, when cells were co-treated with afatinib and z-VAD-fmk for 48 h, both caspase 3 activity and the level of cleaved-PARP were down-regulated significantly as compared to that of afatinib single treatment ( Fig. 2C and D).
These results suggest that afatinib-induced apoptosis in lung adenocarcinoma cells partially depends on caspase 3 activation.
Afatinib induces autophagy in H1975 and H1650 cells. It has been reported that TKI treatment can induce autophagy in many cancer cells including tongue cancer, pancreatic cancer and chronic myeloid leukaemia (CML), which is now believed to be the main reason for drug resistance [26][27][28] . We then determined whether autophagy could be induced by afatinib in H1975 and H1650 cells. First, transmission electron microscopy (TEM) analysis was applied to detect the formation of autophagic vacuoles in H1975 and 1650 cells after exposure to afatinib for 24 h. Many double-membrane-enclosed autophagosomes were observed in afatinib-treated cells, while there were little autophagosomes in the negative control (Fig. 3A). Second, Cyto-ID ® green dye autophagy detection kit was used to identify the level of LC3-II, which is a protein on the membrane of autophagosomes. After incubated with10 μM afatinib for 24 h, both H1975 and H1650 cells had notable green fluorescence when compared with the untreated cells which displayed no specific fluorescence. Rapamycin (50 nM) treated cells served as the positive control and also had great green fluorescence (Fig. 3B). At last, we determined the conversion of LC3 in the cells treated with different concentrations of afatinib using western blot. The shift of LC3 from the LC3-I to LC3-II is related to the formation of autophagy. From Fig. 3C and D, we found the production of LC3-II increased dose-and time-dependently responding to afatinib treatment.
All of these results strongly support the idea that autophagy can be induced by the afatinib in H1975and H1650 cells.
Afatinib triggers autophagic flux in H1975 and H1650 cells. In order to deeply illustrate autophagy triggered by afatinib in H1975 and H1650 cells, we investigated the autophagic flux in the course of afatinib treatment. Chloroquine(CQ), a lysosome inhibitor, can result in the accumulation of autophagosomes and up-regulation of LC3-II by inhibiting the fusion of autophagosomes and lysosomes. Western blot results showed that the aggregation of LC3-II was time-dependent upon afatinib treatment combined with or without CQ. Meanwhile, the LC3-II concentration in the absence of CQ were lower than that in the presence of CQ, notably in 12 h, 24 h, 48 h after afatinib treatment (Fig. 4A), which implied that some LC3 was blocked by CQ to be delivered to lysosomes for degradation. In addition, through confocal fluorescence microscopy (Cyto-ID ® and Lyso-Tracker staining), we observed three steps of autophagic flux in afatinib-incubated H1975 and H1650 cells including the production and aggregation of autophagosomes at 12 h (green fluorescence), autophagosome-lysosome fusion at 24 h (yellow fluorescence) and decomposition of autophagosomes by lysosomes at 48 h (red fluorescence) ( Fig. 4B and C).
The results strongly support that afatinib induces the autophagic flux in H1975 and H1650 cells.

Afatinib-induced cytotoxicity and apoptosis can be improved by autophagy inhibition in H1975 and H1650 cells.
Most researches indicate that autophagy serves a cytoprotective role particularly in cancer treatment, and autophagy inhibition can enhance therapy-induced cell death 29,30 . To determine the role of autophagy caused by afatinib in lung adenocarcinoma cells, two autophagy inhibitor CQ and 3-MA were applied to co-treat cancer cells, and we analyzed their influence on afatinib-induced growth inhibition and apoptosis. It is well known that 3-MA is a PI3K inhibitor that blocks autophagosome aggregation and the transformation of LC3-I to LC3-II, while CQ increases LC3-II level through aggregation of autophagosome. Figure 5A showed that both 3-MA and CQ intensified afatinib-induced cell death in H1975 and H1650 cells. Meanwhile, western blot analysis showed that afatinib-induced autophagy was successfully inhibited by CQ and 3-MA (Fig. 5B). Thus, we got the conclusion that autophagy suppression could increase afatinib-induced cytotoxicity. We further studied the influence of autophagy on afatinib-induced apoptosis by detecting the changes of the cleaved-PARP level and caspase 3 activity with or without CQ or 3-MA. Western blot analysis revealed that the cleaved-PARP concentrations upon co-treatment with CQ or 3-MA were higher than that upon afatinib alone (Fig. 5C). Moreover, afatinib combined with CQ or 3-MA could increase the activity of caspase 3 significantly in both cells. (Fig. 5D). (B and C) H1975 and H1650 cells were dealt with various concentrations of afatinib for indicated time, then MTT assay was used to measure cell viability. All these data were represented as mean ± SD (*P < 0.05, **P < 0.01, ***P < 0.001 versus control). (D) H1975 and H1650 cells were incubated with various concentrations of afatinib for 24 h, then the levels of PARP and cleaved-PARP were measured by western blot analysis. These blots were cropped with Photoshop CS6. (E) H1975 and H1650 cells were treated with 10 μM afatinib for different time, then western blot analysis was performed to assess the levels of PARP and cleaved-PARP. These blots were cropped with Photoshop CS6.
All these data illustrate that autophagy has a protective function against the afatinib-triggered cytotoxicity and apoptosis, and autophagy inhibition can improve the anti-cancer effect of afatinib in lung adenocarcinoma cells with activating EGFR mutations.
The afatinib-induced autophagy is related to the Akt/mTOR and Erk signling pathway. It has been reported that the Akt/mTOR signaling pathway is one of the major signaling pathways acting as a negative regulator of autophagy 31 . To uncover the underlying molecular mechanism of afatinib-induced autophagy, we next investigated the activation status of the key members in Akt/mTOR pathway using western blot analysis. Figure 6A and C indicated that the phosphorylation of key members including Akt, mTOR and its two downstream substrates, p70S6K and eukaryotic initiation factor 4EBP-1 decreased dose-and time-dependently in H1975 and H1650 cells after treatment with afatinib. mTOR can be phosphorylated by its upstream regulator phosphorylated (p)-AKT-serine(S)473 to form p-mTOR-S2448, which negatively regulates autophagy through the phosphorylation of its substrates p70S6K and 4EBP-1 32 . The reduced phosphorylation levels of Akt, mTOR and its substrates indicated that afatinib triggered autophagy via Akt/mTOR signaling pathway.
Extracellular signal-regulated kinase (Erk1/2) is another important pathway involved in autophagy which can regulate the expression of autophagy and lysosomal related genes 33, 34 . Then we investigated the level of phosphorylated Erk1/2 using western blot analysis. A time-and dose-dependent up-regulation of Erk1/2 phosphorylation was observed in H1975 and H1650 cells treated with various concentrations of afatinib for 24 h or 10 μM of afatinib for 3 h, 6 h, 12 h and 24 h (Fig. 6B and D).
These experiment results suggest that the autophagy in afatinib-treated lung adenocarcinoma cells is related to Akt/mTOR and Erk signaling pathway.
The afatinib-triggered autophagy and growth inhibition are related to intracellular ROS. Many studies prove that reactive oxygen species (ROS) accumulation induced by various stresses serves as a critical intracellular signal transducer sustaining autophagy 35 . To investigate whether ROS was involved in afatinib-induced autophagy, we first detected intracellular ROS accumulation in H1975 and H1650 cells after exposure to 10 μM of afatinib with or without antioxidant N-acetyl-cysteine (NAC). As shown in Fig. 7A, afatinib-treated cells displayed significant ROS accumulation as compared with non-treated cells, and NAC could down-regulate oxidative stress. In addition, MTT assay showed that scavenging ROS by NAC could rescue H1975 and H1650 cells from afatinib-induced growth inhibition (Fig. 7B). Second, western blot assay showed that NAC could decrease afatinib-induced LC3-II and cleaved-PARP accumulations significantly in both cells, suggesting that removing ROS resulted in autophagy inhibition as well as apoptosis down-regulation ( Fig. 7C and D).
Furthermore, ROS detection kit Mito SOX TM and autophagy detection kit Cyto-ID ® were used to investigate the generation process of intracellular ROS and autophagy. Figure 7E showed that afatinib-treated cells exhibited obvious ROS-specific red fluorescence and autophagosome-specific green fluorescence compared with non-treated cells, whereas cells co-treated with NAC displayed almost no red and green fluorescence. Moreover, as shown in Fig. 7F, after exposure to 10 μM of afatinib for 18 h both cells displayed enhanced red fluorescence indicating ROS generation, then both cells had clear green fluorescence representing autophagy activation in 24 h of afatinib incubation. Additionally, Fig. 6E and F showed that both Akt/mTOR and Erk singal pathways were reversed after scavenging afatinib-induced ROS by NAC.
The results reveal that ROS acts as an intracellular transducer regulating both autophagy and apoptosis in afatinib-treated H1975 and H1650 cells.  Fig. 8A and B, combination treatment with afatinib and CQ have more remarkable anti-tumor effect than treatment with afatinib alone, and the average values of tumor volume decreased by 427 mm 3 . Moreover, the combination treatment also significantly reduced tumor weight (Fig. 8C).
In brief, these data prove that the suppression of autophagy can improve the anti-tumor effect of afatinib in vivo.

Discussion
NSCLC, especially lung adenocarcinoma, is the commonest fatal malignancy worldwide. The advent of first-generation reversible EGFR-TKIs improved the approaches for NSCLC diagnosis and treatment 36 . Patients with sensitizing EGFR mutations show a better prognosis and sensitivity to EGFR-TKIs than conventional chemotherapy. However, the high frequency of EGFR mutations in NSCLC confers different efficacy to reversible EGFR-TKIs 12 . Although in sensitizing EGFR mutations such as exon 19 deletion, reversible EGFR-TKIs result in longer progression-free survival (PFS) than conventional chemotherapy, the overall survival (OS) benefit was not obtained. Moreover, Treatment with the first-generation EGFR-TKIs show minimal responses in patients with acquired TKI resistance mutations such as T790M mutation 11 .
As an irreversible inhibitor of the ErbB family of tyrosine kinases, the second-generation TKIs afatinib is a valuable option for patients with advanced lung adenocarcinoma harboring activating EGFR mutations 37 . Although targeted treatment of these patients with afatinib results in improved anti-tumor responses, acquired resistance is also a major clinical issue in lung cancer therapy with this irreversible pan-ErbB inhibitor 14 . Autophagy is reported to be associated with resistance to afatinib, but its exact effect and mechanisms has not been clearly elucidated especially in lung adenocarcinoma with activating EGFR mutations 25 .
In the present study, we chose a L858R/T790M mutant H1975 cell and an exon 19 deletion mutant H1650 cell to investigate the autophagy induction and the effect on anti-tumor activity of afatinib. In addition, the in The activity of caspase 3 were detected by caspase 3 activity kit. All these data were presented as mean ± SD (*P < 0.05).
vivo tumor responses were evaluated in H1975 lung adenocarcinoma xenograft model. We first observed growth inhibition and apoptosis in afatinib-treated H1975 and H1650 cells. Many studies prove that the TKI-induced apoptosis is caspase 3 dependent 38,39 , and it has been reported that afatinib could induce H1975, H1650 and RB116 cell apoptosis by activation of caspase 3 40,41 . Our experiments proved that afatinib dramatically increased casepase 3 activity, and pan-caspase inhibitor z-VAD-fmk could rescue the cells and decrease caspase 3 activity as well as the level of cleaved-PARP. Therefore, we concluded that afatinib-induced apoptosis was partially caspase 3-dependent in H1975 and H1650 cells.
Autophagy has been considered as an adaptive catabolic process responding to various stress conditions including chemotherapy drugs 42 , and serves as one of important resistance mechanisms to TKI treatment in many cancer cells such as lung cancer, chronic myeloid leukaemia (CML), tongue cancer and pancreatic cancer [25][26][27][28] . In this study, we focused on whether afatinib induced autophagy and its role in afatinib's anti-tumor activity in H1975 and H1650 cells. Through TEM, confocal Immunofluorescence and western blot analysis, we confirmed that afatinib did induce autophagy in lung adenocarcinoma cells with activating EGFR mutations 23,43 .
Chemotherapeutics can initiate either cytoprotective autophagy or autophagic cell death in tumor cells depending on different cell types and microenvironment 38,44 . We found that suppression of autophagy using two kinds of autophagy inhibitors CQ and 3-MA could significantly enhance afatinib-induced cytotoxicity as well as apoptosis in H1975 and H1650 cells. These observations proved the cytoprotective role of autophagy in afatinib-treated H1975 and H1650 cells, suggesting that abolishing autophagy could enhance the anti-lung adenocarcinoma effect of afatinib in vitro. We also established a H1975 cell xenograft model to further confirm the protective effect of autophagy on afatinib therapy in vivo. The results displayed that combination treatment with afatinib and CQ enhanced afatinib's therapeutic efficacy, which further proved that blocking autophagy could overcome resistance against afatinib in lung adenocarcinoma with double mutations.
In the study of the underlying mechanisms of afatinib-induced autophagy, we observed up-regulations of intracellular ROS accumulation and Erk signaling, as well as a down-regulations of Akt/mTOR signaling in H1975 and H1650 cells after afatinib treatment. Intracellular ROS triggered by diverse stress conditions has been considered as critical mediator of autophagy and apoptosis 23,35 . Through fluorescence microscope, we found that afatinib first initiated intracellular ROS accumulation which then induced autophagy. After scavenging ROS by NAC, the levels of autophagy-related protein LC3-II and apoptosis-related protein cleaved-PARP decreased, and cell viability restored. Therefore, afatinib-triggered ROS not only exerted cytotoxic effect on H1975 and H1650 cells, but also initiated autophagic responses to protect tumor cells from apoptosis and growth inhibition, which can explain our in vitro and in vivo experiment results that combination treatment with autophagy inhibitors enhanced afatinib's therapeutic efficacy.
Besides, antioxidant NAC also reversed afatinib-induced changes of Akt/mTOR and Erk signaling pathway. Previous studies reported that the Akt/mTOR signaling pathway negatively regulates autophagy through phosphorylation of Akt, downstream effector mTOR and its substrates p70S6K and 4E-BP1 45 , and Erk can be activated through phosphorylation of threonine and tyrosine residues, then induces the formation of cytoplasmic macrovacuoles, a symbol of autophagy 46 . In the present study, western blot analyses demonstrated that both Akt/mTOR and Erk signaling pathways were involved in afatinib-induced autophagy in H1975 and H1650 cells, and what's more, afatinib-triggered ROS accumulation initiated autophagic responses through influencing Akt/ mTOR and Erk signaling pathways.
In conclusion, our research proves that afatinib can induce cytotoxicity as well as autophagy in H1975 and H1650 cells. More importantly, inhibition of autophagy enhances the cytotoxicity of afatinib in vitro, suggesting the cytoprotective role of afatinib-induced autophagy. Further study suggests that Akt/mTOR and Erk signaling pathways are involved in autophagy triggered by afatinib, and intracellular ROS acts as a critical mediator of autophagy and cytotoxicity. Additionally, we demonstrate that autophagy suppression can potentiate the therapeutic efficacy of afatinib in vivo. Our research highlights that combination of afatinib and autophagy inhibitor may be a promising approach to increase the sensitivity of lung adenocarcinoma cells harboring activating EGFR mutations to afatinib treatment. Cell viability assay. MTT assay was used to detect the cell viability. About 1 × 10 4 cells per well were planted in 96-well plates and then exposed to CQ or 3-MA (Sigma-Aldrich) for 2 hours, then added the indicated concentrations of afatinib (Shanghai Hanxiang Biotech) into different wells. After certain time, cells were incubated with MTT solution (0.5 mg/ml) (Beyotime Institute of Biotechnology) for 4 h at 37 °C. Then 150 μl DMSO was added to each well to dissolve formazan for measurement. The optical density was recorded at the wavelength of 570 nm by microplate reader.
Some wells were added with 10 μM afatinib for 24 h, or treated with 50 nM rapamycin (Sigma-Aldrich) for 6 h as positive controls, and the left wells were used as negative control. After that, cells were washed twice by culture medium and then added with Cyto-ID ® green dye and nuclear dye Hoechst 33342 (Enzo Life Sciences), or ROS detection kit Mito Sox TM red dye (Invitrogen) at 37 °C for 25 min. Cells were immediately detected with fluorescence microscope after incubation. These procedures were done in the dark room.
Western blot analysis. Cells were washed with cold phosphate-buffered saline (PBS) after added to different concentrations of afatinib for different time. Then, cells were lysed by RIPA buffer (Beyotime Institute of Biotechnology) on ice for 30 min. The lysates were centrifuged at 12,000 rpm at 4 °C for 10 min, and the supernatant was collected. Then the lystates protein was quantized by BCA protein quantitation kit (Beyotime Institute of Biotechnology). The equal amount of protein per lane was separated by SDS-PAGE, then the protein was transferred to PVDF membranes. After blocked by 3% bovine saline albumin, the membranes were washed with 1 × TBST (150 mM NaCl, 10 mM TRIS-Cl, pH 8.0), containing 0.05% (v/v) Tween-20 (Dalian Meilun Biotech) for three times. After incubated overnight with specialized primary antibodies (Cell Signaling), membranes were washed for three times with 1 × TBST and suspended in the secondary antibodies buffer (Beyotime Institute of Biotechnology) for 2 h at room temperature. At last detection was performed with enhanced chemiluminescence reagents (Beyotime Institute of Biotechnology).
Transmission electron microscopy analysis. H1650 and H1975 cells were incubated with or without afatinib (10 μM) for 24 h. The harvested cells were fixed in ice-cold glutaraldehyde, then sliced and analyzed by Detection of intracellular ROS. Intracellular ROS production was analyzed in H1650 and H1975 cells with ROS assay kit (Beyotime Institute of Biotechnology). ROS can oxidize dichlorofluorescein diacetate (DCFH-DA), a cell infiltrative and non-fluorescent dye, to fluorescent carboxy dichlorofluorescein. H1650 and H1975 cells were co-incubated with afatinib (10 μM) and 10 μM DCFH-DA at 37 °C for indicated time, then washed twice by PBS. The Tecan Infinite ® 200 PRO microplate reader was used to quantify DCF fluorescence intensity at excitation wavelength of 488 nm and emission wavelength of 525 nm.
Caspase 3 activity analysis. After treated with specific concentrations of afatinib in the presence or absence of autophagy inhibitors, H1650 and H1975 cells were lysed and the activity of apoptosis-related protein caspase 3 was measured by Caspase 3 Activity detection kit (Beyotime Institute of Biotechnology).
Tumor xenograft model. H1975 cells (1 × 10 6 /ml) were inoculated subcutaneously into the right thigh of male 4 to 6 weeks old of Nu/Nu mice (Charles River). When the average tumor volume reached 200 mm 3 , the mice were randomized into 5 groups to receive the following treatments: (a) vehicle (methycellulose/Tween 80, per os) (Dalian Meilun Biotech); (b) CQ (50 mg/kg, intraperitoneal injection); (c) afatinib (20 mg/kg/day, per os); (d) afatinib (20 mg/kg/day, per os) + CQ (50 mg/kg, intraperitoneal injection) (e) cisplatin (10 mg/kg, every other day, intraperitoneal injection) (Dalian MeilunBiotech). Tumors were measured every 3 days, and tumor volumes were determined from caliper measurements of tumor length (L) and width (W) according to the formula V = (L × W 2 )/2. All animal procedures were performed in accordance with institutional guidelines and with approval from Fudan University Committee on Animal Care. Statistical analysis. Statistical analysis was performed by GraphPad Prism 5 and the data were presented as mean ± S.D. The two-tailed Student's t test was used to compare multiple sets of data. P < 0.05 were considered to be statistically remarkable.