Hypoxia-induced autophagy mediates cisplatin resistance in lung cancer cells

Hypoxia which commonly exists in solid tumors, leads to cancer cells chemoresistance via provoking adaptive responses including autophagy. Therefore, we sought to evaluate the role of autophagy and hypoxia as well as the underlying mechanism in the cisplatin resistance of lung cancer cells. Our study demonstrated that hypoxia significantly protected A549 and SPC-A1 cells from cisplatin-induced cell death in a Hif-1α- and Hif-2α- dependent manner. Moreover, compared with normoxia, cisplatin-induced apoptosis under hypoxia was markedly reduced. However, when autophagy was inhibited by 3-MA or siRNA targeted ATG5, this reduction was effectively attenuated, which means autophagy mediates cisplatin resisitance under hypoxia. In parallel, we showed that hypoxia robustly augmented cisplatin-induced autophagy activation, accompanying by suppressing cisplatin-induced BNIP3 death pathways, which was due to the more efficient autophagic process under hypoxia. Consequently, we proposed that autophagy was a protective mechanism after cisplatin incubation under both normoxia and hypoxia. However, under normoxia, autophagy activation ‘was unable to counteract the stress induced by cisplatin, therefore resulting in cell death, whereas under hypoxia, autophagy induction was augmented that solved the cisplatin-induced stress, allowing the cells to survival. In conclusion, augmented induction of autophagy by hypoxia decreased lung cancer cells susceptibility to cisplatin-induced apoptosis.


Hypoxia blocked cisplatin-induced subcellular redistribution of pro-apoptotic proteins.
Cisplatin-induced cell death in lung cancer cells potentially involves multiple signaling pathways. Among these pathways, mitochondrial apoptosis events (Bax translocation and Cyto C release) play a critical role. Here, we investigated the effect of hypoxia on these two cisplatin-induced mitochondrial apoptosis events. Cytosolic and mitochondrial proteins were separated and analyzed by western blot. The effect of separation was measured by detection of the expression of COX IV ( Fig. 2A). The total protein levels of Cyto C and Bax showed no significant differences in normoxic group and hypoxic group with or without cisplatin treatment (Fig. 2B). However, Bax was detected primarily in the cytoplasm in the control group, but translocated from cytoplasm to mitochondria after 12 h hypoxia or cisplatin incubation, of interest was that the cisplatin-induced Bax translocation was significantly decreased in the presence of hypoxia in both A549 and SPC-A1 cells (Fig. 2C,D). On the other hand, hypoxia or cisplatin significantly induced Cyto C released from mitochondria into the cytoplasm, but the cisplatin-induced Cyto C release was suppressed by simultaneous treatment of hypoxia (Fig. 2C,D). Identical results were obtained by Bax and Cyto C immunocytochemistry with Confocal Laser Scanning Microscopy (CLSM) (Fig. 3A,B and supplementary Fig. 3). These results showed that hypoxia blocked cisplatin-induced subcellular redistribution of pro-apoptotic proteins, suggesting that hypoxia induces cisplatin resistance mainly by reducing lung cancer cells apoptotic potential.

Inhibition of autophagy restores lung cancer cells sensitivity to cisplatin under hypoxia. Data
above proved that hypoxia protected A549 and SPC-A1 cells from cisplatin, and this may be due to the decreased apoptotic potential. But how does it work? It has recently been reported that under adverse conditions such as hypoxia, cancer cells make themselves be adaptive by activating autophagy, thus we hypothesized that autophagy may attribute to hypoxia-induced cisplatin resistance in lung cancer cells. In order to test this hypothesis, we inhibited autophagy by 3-MA or siRNA targeted the substantial autophagic gene ATG5, then the occurrence of apoptosis was determined by Annexin-V-FITC/PI assay, the expression of activated caspase-3 and retinoblastoma (Rb) protein was analyzed by western blot.
First, the percentage of apoptotic cells was analyzed by flow cytometry, as shown in Fig. 4A,B, cisplatin significantly induced apoptosis in A549 and SPC-A1 cells in both normoxic and hypoxic conditions. Compared with that in normoxia, cisplatin-induced apoptosis in A549 and SPC-A1 cells was significantly decreased in hypoxic conditions ( Fig. 4A-D). Interestingly, the addition of 3-MA (10 μ M) dramatically restored the rate of apoptosis under hypoxia in both A549 and SPC-A1 cells (Fig. 4A-D), comparable to that under normoxia with cisplatin treatment. To further confirm this restoration, siRNA targeted ATG5 was transfected into A549 and SPC-A1 cells. Remarkably, knockdown of ATG5 almost completely abrogated the effect of hypoxia on cisplatin-induced cell apoptosis ( Supplementary Fig. 4). The knockdown efficiency of ATG5 siRNA in both A549 and SPC-A1 cells was showed in supplementary Fig. 4F.
Next, we found that cisplatin significantly induced the expression of activated caspase-3 and Rb which both widely observed during apoptosis 29,30 in A549 and SPC-A1 cells under both normoxic and hypoxic conditions. Compared with that in normoxia, cisplatin-induced expression of activated caspase-3 and Rb was significantly decreased in hypoxic conditions ( Fig. 4E-G). Interestingly, the addition of 3-MA dramatically restored the expression of activated caspase-3 and Rb under hypoxia in both A549 and SPC-A1 cells (Fig. 4E-G). The similar results were obtained by ATG5 siRNA treatment (Supplementary The knockdown efficiency of Hif-1α or Hif-2α siRNA in the SPC-A1 cell. (E) Cells incubated with hypoxia at difference time point were harvested for protein extraction and then subjected to western blot using Hif-1α and Hif-2α antibody. The results were shown as means ± SD of four independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001. Blot images were cropped for comparison. Hypoxia augmented cisplatin-induced autophagy in lung cancer cells. We then assessed the effect of hypoxia on autophagy in lung cancer cells upon treatment of cisplatin. First, following 12h incubation under normoxia or hypoxia with or without cisplatin treatment, A549 and SPC-A1 cells were stained by MDC, a specific autophagolysosome marker. As shown in Fig. 5A,C, cisplatin significantly induced MDC localization in vacuoles in both A549 and SPC-A1 cells. Compared with that in normoxia, cells under hypoxia exhibited significantly high percentage of cells containing MDC-labeled vacuoles (Fig. 5B,D).
Sencond, we transfected A549 and SPC-A1 cells with GFP-LC3 plasmids, and analyzed the occurrence of autophagy by monitoring the GFP dots. After transfection, cells were incubated under normoxia or hypoxia in the presence or absence of cisplatin. Then cells were observed by CLSM, and the cells with diffused or punctuate GFP were counted. As shown in Fig. 6A,B, cisplatin increased the GFP-LC3 puncta (as shown by white arrows) under conditions of both normoxia and hypoxia in A549 cells (Fig. 6A,B). However, there were significantly more cells containing GFP-LC3 puncta induced by cisplatin under condition of hypoxia as compared to that in normoxia (Fig. 6A,B). Interestingly, 3-MA significantly reduced the hypoxia or/and cisplatin-induced GFP-LC3 puncta accumulation (Fig. 6A,B). The similar results were showed in SPC-A1 cells (Fig. 6C,D).
As another independent assay of autophagy, cells were processed for transmission electron microscopy after incubated under normaxia or hypoxia with or without cisplatin treatment. Representative electron micrographs shown in Fig. 7 demonstrated that cisplatin induced more autophagic vacuoles, and the number of cisplatin-induced autophagic vacuoles per cell was significantly increased by hypoxia in both A549 (Fig. 7A,B) and SPC-A1 cells (Fig. 7C,D). To further confirm the involvement of autophagy by additionally independent assay, we analyzed the expression of Beclin-1, p-Beclin-1, p62 and LC3-II, which are hallmarks of autophagy by western blot 31 . Compared to that in normoxia, hypoxia markedly increased the levels of Beclin-1, p-Beclin-1 and LC3-II induced by cisplatin in both A549 and SPC-A1 cells ( Hypoxia suppressed cisplatin-induced BNIP3 death pathways. As for BNIP3 (BCL2/adenovirus E1B 19kDa-interacting protein 3) and its close family member BNIP3L have been identified as apoptotic mediators under hypoxia 32,33 , and BNIP3 has been identified as a hypoxia inducible regulator of autophagy 26 . We further studied whether BNIP3 and/or BNIP3L were involved in the hypoxia-induced cisplatin resistance. A549 and SPC-A1 cells were incubated under normoxia or hypoxia with or without cisplatin. It had to be noted that, in these conditions, hypoxia did lead to Hif-1α and Hif-2α activation but cisplatin had no influence on this process ( Fig. 9A-C). On the other hand, BNIP3 and BNIP3L abundance was markedly increased by cisplatin or hypoxia, but this abundance was decreased when cells treated with cisplatin under hypoxia (Fig. 9A,D,E). However, the addition of 3-MA completely restored the expression of BNIP3 and BNIP3L (Fig. 9A,D,E).

Discussion
Hypoxia is a largely studied factor promoting cancer cells chemoresistance. Moreover, autophagy has been highlighted as a cytoprotective mechanism against chemotherapy in these past years 23 . Here, our results demonstrated that robustly augmented-induction of autophagy by hypoxia mediated cisplatin resistance. Targeting autophagy was sufficient to restore lung cancer cells susceptibility to cisplatin.
It was generally acknowledged that, in hypoxic condition, cancer cells undergo a series of genetic and metabolic changes that allow them to be more resistant to chemotherapy 8,[34][35][36] . Consistent with these studies, we showed that lung cancer cells incubated with hypoxia showed a lower percentage of cell death compared with that in normoxia when treated with cisplatin. One of the key players for hypoxia-mediated effects is hypoxia-inducible factor (Hif) 27,37 . Several studies show that Hif-1α is involved in hypoxia-induced chemoresistance [38][39][40][41] . Our study further revealed that both Hif-1α and Hif-2α were critical for the hypoxia-induced cisplatin resistance.
There are several explanations for hypoxia-induced cisplatin resistance, of which reduced cellular susceptibility to apoptosis may be the most prevailing one 42 . Defective apoptosis underpins cisplatin resistance and is a hallmark of lung cancer 43 . Mitochondria are at the crossroads of apoptotic pathways induced by anticancer agents at several levels. In response to apoptotic stimuli, the outer mitochondrial membrane is permeabilized, causing tanslocation of Bax protein from cytosol to mitochondria and subsequent cytochrome c released from mitochondria into the cytosol where it helps to activate the caspases 31 . However, hypoxia inhibits cell apoptosis by blocking Bax translocation 44 . Consistant with these reports, we demonstrated that cisplatin promoted Bax translocation and Cyto C release, whereas this effect was suppressed by hypoxia. Our result further supported the view that hypoxia decreased lung cancer cells sensitization to cisplatin by affecting their apoptotic potential. At the same time, it has been reported that tumor resistance to many anticancer agents including cisplatin, tamoxifen can be enhanced through upregulation of autophagy in different tumor cell lines 45,46 . Moreover, inhibition of autophagy augments cytotoxicity in combination with several anticancer drugs in preclinical models by triggering apoptosis 47,48 . Therefore, the role of autophagy in the hypoxia-induced cisplatin resistance deserves attention. By our analysis, cisplatin induced a significant increase of apoptosis in both A549 and SPC-A1 cells, but the apoptotic rate and the level of active caspase 3 decreased a lot under hypoxic condition. Interestingly, when lung cancer cells were treated by 3-MA or ATG5 siRNA under hypoxia, they became sensitive to cisplatin, similar to the cells under normoxia. Despite present controversies on the exact role of autophagy in the process of tumor generation and progression, either by cell protection or contrarily by inducing cell death 49,50 , a majority of studies have been indicated that autophagy is a protective mechanism associated with increased resistance to chemotherapy 2,51 . Our results also indicated the protective role of autophagy in the process which hypoxia potentially prevents lung cancer cells from cisplatin-induced apoptosis. Meanwhile, our result demonstrated that cisplatin significantly increased the expression of Rb, while this increase was counteracted by simultaneous hypoxia treatment. However, the addition of 3-MA significantly restored the effect of cisplatin under hypoxia. Many studies have indicated the correlation between Rb and cell cycle control 52,53 . Rb has been reported to actively arrest cell cycle progression in G0 or G1 phase when it is unphosphorylated 54 . Therefore, our results indicating that cell survival may be the result of cell cycle arrest under hypoxia and/or cisplatin, while the addition of 3-MA may restore cell cycle progression.
Some studies have reported that hypoxia can activate autophagy [55][56][57] , and the hypoxia-induced autophagy may protect cells from apoptosis induced by chemotherapeutic agents 58 . Here, we showed that regardless of O 2 concentration, cisplatin treatment dramatically increased the accumulation of GFP-LC3 puncta, the formation of autophagic vacuoles, the protein level of Beclin1, p-Beclin1 and LC3-II, as well as the degradation of p62, which are typical features of autophagy. However, note that cisplatin-induced autophagy was markedly augmented by hypoxia. These evidences suggested that there may be a certain threshold value of autophagy activation in lung cancer cells stimulated by cisplatin as shown in Fig. 10. Under normoxia, autophagy activation was unable to excess the threshold, whereas under hypoxia, autophagy was augmented and exceed the threshold. The different regulation of autophagy process by normoxia and hypoxia may be the underlying mechanism for the cisplatin resistance.
As for the molecular mechanism, transcriptional factor Hif-1α and Hif-2α are considered to play foundamental roles under hypoxia 59,60 . Hif-1, by regulating the expression of its target downstream BNIP3 and BNIP3L 32 , regulates the activation of apoptosis, autophagy and necrosis under hypoxia 61 . Our study showed that either cisplatin or hypoxia markedly increased BNIP3 and BNIP3L expression. However, their simultaneous treatment reduced the elevation of BNIP3 and BNIP3L induced by cisplatin, and the addition of 3-MA completely restored the elevation. The expression of BNIP3 is reported to be tightly regulated, as its overexpression induces cell death 61 .
Consequently, we hypothesized that autophagy was a protective mechanism after cisplatin incubation under both normoxia and hypoxia, and there may be a certain threshold value of autophagy activation. Under normoxia, autophagy activation was unable to excess the threshold to counteract the stress induced by cisplatin, resulting in lower p62 degradation, more BNIP3 and BNIP3L abundance, leading to apoptosis activation and cell death. However, under hypoxia, autophagy induction was augmented that solved the stress, resulting in more p62 degradation, lower BNIP3 and BNIP3L abundance and lower apoptosis activation, allowing the cells to survival (Fig. 10). Therefore, this would require a better In conclusion, the present study demonstrated that cisplatin resistance in lung cancer cells under hypoxia can be explained by the augmented induction of autophagy, which suppressed BNIP3 death  The relative expressions of Hif-1α , Hif-2α , BNIP3 and BNIP3L were quantified by detecting densitometry. Data were shown as the means ± SD. *p < 0.05, **p < 0.01, ***p < 0.001. All studies were representative of three independent experiments. Blot images were cropped for comparison. pathway. Further studies on the underlying molecular mechanisms of hypoxia-induced autophagy in chemoresistance will provide novel chemotherapeutic agents for lung cancer therapy.

Materials and Methods
Cell culture. The human lung cancer cell lines A549 and SPC-A1 (China Centre for Type Culture Collection, China) were maintained in DMEM (GBICO, Invitrogen,12100-038) containing 10% FBS (GBICO, Invitrogen, 10099-141) at 37 °C in humidified atmosphere of 5% CO 2 . The modular incubation chamber (MIC-101) from Billups-Rothenberg Inc. was used for hypoxic cell culture conditions, the cells were pre-incubated with 1% O 2 for 1 h before treatment of cisplatin, the detailed protocol and PO 2 levels of the medium were described in our previous study 62 . Cell viability assay. Cell viability was assessed with the MTT assay. A549 and SPC-A1 cells were seeded in 96-well plates at a density of 1 × 10 4 cells/well and cultured in 1% or 21% O 2 in medium containing cisplatin (Sigma, P4394) as indicated for 24 h. Next, 10 μ l of MTT (5 mg/ml in PBS, Amersco, 0793) was added to each well and incubated for 4 h at 37 °C. Then, the formazan crystals were solubilized with 200 μ l DMSO (Sigma, D2650). The absorbance (A) at 570 nm was measured using an automatic multiwell spectrophotometer. The experiment was repeated four times for each group. PI staining. A549 and SPC-A1 cells were seeded in 12-well plates containing coverslips at a density of 1 × 10 4 cells/well and cultured in 1% or 21% O 2 in medium containing cisplatin (Sigma, P4394) as indicated for 24 h. 1 ug/ml PI (BD Pharmingen ™ , 556547) was added to each well and incubated for 10 min at 37 °C, then cells were rinsed three times and observed by confocal sanning microscopy.
Mitochondrial and cytosolic protein fractionation. In order to determine the release of cytochrome C from mitochondria to cytosol and the translocation of Bax from cytosol to mitochondira, the isolation of mitochondria and cytosol was performed using the Cell Mitochondria Isolation Kit (abcam, ab110170). Briefly, 5 × 10 7 cells were harvested and incubated in 100 μ l ice-cold mitochondrial lysis buffer on ice for 10 min. Cell suspension was placed into a glass homogenizer and homogenized for 50 strokes using a tight pestle on ice. The homogenate was then centrifuging at 600 g for 10 min at 4 °C to remove the nuclei and unbroken cells. The supernatant was then collected and centrifuged at 12000 g for 30 min at 4 °C to isolate the cytosol (supernatant) and mitochondria (deposition) fractions. Samples of cytosol and mitochondria were dissolved in lysis buffer and proteins were subjected to western blotting, respectively.
Immunofluorescence labeling and confocal microscopy. A549 and SPC-A1 cells were seeded in 12-well plates containing coverslips 24 h before indicated treatments. Cells were incubated for 12 h with or without cisplatin under normoxic or hypoxic conditions. After incubation, cells were incubated with 100 nM MitoTacker (Invitrogen, M7513) under growth conditions for 30 min. After staining was completed, the cells were fixed by 10% formaldehyde. Then cells were permeabilized in PBS containing 0.5% Figure 10. Schematic representation of the lower sensibility of lung cancer cells against cisplatininduced cell death under hypoxia. Cisplatin induced autophagy activation which was a protective mechanism against cisplatin-induced cell death under both normoxia and hypoxia, and there may be a certain threshold value of autophagy activation. Under normoxia, autophagy activation was unable to excess the threshold to counteract the stress induced by cisplatin, resulting in lower p62 degradation, more BNIP3 and BNIP3L abundance, leading to apoptosis activation and cell death. However, under hypoxia, autophagy induction was augmented that solve the stress, resulting in more p62 degradation, lower BNIP3 and BNIP3L abundance and lower apoptosis activation, allowing the cells to survival.

MDC staining.
A monolayer of cells were cultured for 48 h in 2-well glass-covered chamber slides and then treated with or without cisplatin under normoxic and hypoxic conditions for 12 h. Slides were washed with culture medium without serum. Then, cells were exposed to 50mM MDC (Sigma, 30432), an autofluorescent dye, for 10 min at 37 °C and visualized using confocal laser scanning microscope (Carl Zeiss 710). At least 3 areas per well were analyzed. Two wells were analyzed per treatment and per time. The experiment was repeated for four times. The number of MDC-labeled cells was counted.

Transient transfection and identification of autophagy. A549 and SPC-A1 cells were transfected
with GFP-LC3 plasmid using Lipofectamine 2000 (Invitrogen, 11668-019). After 24 h incubation, A549 and SPC-A1 cells were treated with 10 μ M and 30 μ M cisplatin for 12 h, respectively. Then, the GFP-LC3 punctate-structures were observed using a confocal laser scanning microscope. The experiment was repeated for four times, more than 100 cells were calculated.
Transmission electron microscopy (TEM). The cells were pre-fixed in a solution of 2.5% glutaraldehyde in 0.1M PBS (pH 7.4) for 2 h at room temperature, and post-fixed in 1% osmium tetroxide for 2 h. The samples were dehydrated in increasing concentrations of ethanol (50%, 70%, and 100%) and acetone, and then embedded in Araldite. Fifty to sixty nanometer sections were cut on a LKB-I ultramicrotome and transferred to copper grids, post-stained with uranyl acetate and lead citrate, and examined with a Philips CM-120 transmission electron microscopy.
Statistical methods. All of the experiments were repeated at least three times. The Data were presented as means ± SD. Statistical analysis was performed using SPSS for Windows, version 11.5. Statistical significance was determined using one-way ANOVA with a post hoc Bonferroni's test. Significance was set to p values < 0.05.