The H3122 cell line (NSCLC cell line with an EML4-ALK rearrangement) was maintained in RPMI-1640 medium with 10% FBS (Sigma-Aldrich, St. Louis, MO, USA). For the normoxic state, the cell line was maintained in a 5% CO₂-humidified atmosphere at 37°C; for the hypoxic state, the cell line was maintained in 5% CO₂-humidified 0.2% O₂ at 37°C. Crizotinib and alectinib (ALK inhibitors) were purchased from Selleck Chemicals (Houston, TX, USA).

The growth-inhibitory effects of crizotinib and alectinib were examined using a 3, 4, 5-dimethyl- 2H-tetrazolium bromide assay (MTT; Sigma-Aldrich), as described previously (15). The experiment was performed in triplicate.

The migration assays were performed using Boyden chamber methods and polycarbonate membranes with an 8-μm pore size (Chemotaxicell; Kurabo, Osaka, Japan), as previously described (16). The membranes were coated with fibronectin on the outer side and were dried for 2 h at room temperature. The cells (2×10⁴ cells/well) were then seeded onto the upper chambers with 200 ml of migrating medium (RPMI containing 0.5% FBS), and the upper chambers were placed into the lower chambers of 24-well culture dishes containing 600 ml of RPMI with 10% FBS. After incubation for 48 h under a normoxic or hypoxic state, the media in the upper chambers were aspirated and the non-migrated cells on the inner sides of the membranes were removed using a cotton swab. The cells that had migrated to the outer side of the membranes were fixed with 4% paraformaldehyde for 10 min, stained with 0.1% crystal violet stain solution for 15 min, and then counted using a light microscope. The number of migrated cells was averaged from 5 fields per 1 chamber, and 3 chambers were used for each experiment. The experiment was performed in triplicate.

One microgram of total RNA from cultured cell lines was converted to cDNA using the GeneAmp RNA-PCR kit (Applied Biosystems, Foster City, CA, USA). Real-time PCR was performed using SYBR Premix Ex Taq and Thermal Cycler Dice (Takara, Shiga, Japan), as described previously (16). The glyceraldehyde 3-phosphate dehydrogenase (GAPD, NM_002046) gene was used to normalize the expression levels in subsequent quantitative analyses. The experiment was performed in triplicate. To amplify the target genes encoding E-cadherin, vimentin, fibronectin, and slug (E-cadherin, VIM, FN1 and SLUG gene), the following primers were used: E-cadherin-F, TTAAACT CCTGGCCTCAAGCAATC; E-cadherin-R, TCCTATCTT GGGCAAAGCAACTG; VIM-F, TGAGTACCGGAGACA GGTGCAG; VIM-R, TAGCAGCTTCAACGGCAAAGTTC; FN1-F, GGAGCAAATGGCACCGAGATA; FN1-R, GAGCT GCACATGTCTTGGGAAC; SLUG-F, ATGCATATTCG GACCCACACATTA; SLUG-R, AGATTTGACCTGTCTGC AAATGCTC; GAPD-F, GCACCGTCAAGGCTGAGAAC; GAPD-R, ATGGTGGTGAAGACGCCAGT.

Antibodies specific for AKT, phospho-AKT, ERK1/2, phospho-ERK1/2, PARP, cleaved PARP, caspase-3, cleaved caspase-3, E-cadherin, vimentin, slug, and β-actin were obtained from Cell Signaling (Beverly, MA, USA). An antibody specific for fibronectin was obtained from Abcam (Cambridge, UK). An antibody specific for hypoxia-inducible factor 1α (HIF1α) was obtained from Novus (Littleton, CO, USA).

A western blot analysis was performed as described previously (15). Briefly, subconfluent cells were washed with cold phosphate-buffered saline (PBS) and harvested with Lysis A buffer containing 1% Triton X-100, 20 mM Tris-HCl (pH 7.0), 5 mM EDTA, 50 mM sodium chloride, 10 mM sodium pyrophosphate, 50 mM sodium fluoride, 1 mM sodium orthovanadate, and the protease inhibitor mix Complete™ (Roche Diagnostics, Basel, Switzerland). Wholecell lyses were separated using SDS-PAGE and were blotted onto a polyvinylidene fluoride membrane. After blocking with 3% bovine serum albumin in a TBS buffer (pH 8.0) with 0.1% Tween-20, the membrane was probed with the primary antibody. After rinsing twice with TBS buffer, the membrane was incubated with a horseradish peroxidase-conjugated secondary antibody and washed, followed by visualization using an ECL detection system and LAS-4000 (GE Healthcare, Buckinghamshire, UK). When the phosphorylation was evaluated in cells treated with an inhibitor, the cells were incubated under a normoxic or hypoxic state for 48 h and the inhibitor was added three hours before the sample collection.

Cells were transfected with siRNA for HIF1A or a non-specific target (scramble) as follows: GGAAUUAACUCAGUUUGAACUAACU (si-HIF1A) and AAACCUUCAGACGUUAGUUUAUAGA for a scramble of HIF1A (si-Scr). siRNA transfection was performed using RNAiMAX (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions, as previously described (17), and was allowed to proceed for 48–96 h before the growth-inhibitory test or the collection of the whole-cell extract. Knockdown was confirmed using western blot analyses.

Continuous variables were analyzed using the Student’s t-test, and the results were expressed as the average and standard deviations (SD). The statistical analyses were two-tailed and were performed using Microsoft Excel (Microsoft, Redmond, WA, USA). P-values of <0.05 were considered statistically significant.

To examine the sensitivity of the H3122 cell line to ALK inhibitors, we performed an MTT assay (Fig. 1A). Under a normoxic state, the 50% inhibitory concentration (IC₅₀) of crizotinib and alectinib was 0.096 and 0.033 μM, respectively (Fig. 1B). Under a hypoxic state, however, the IC₅₀ was 0.75 and 1.30 μM, respectively. These results indicate that hypoxia induces resistance to ALK inhibitors in the H3122 cell line.

Next, to examine the effect of hypoxia on the downstream signal of ALK, we evaluated the phosphorylation of AKT and ERK under normoxic and hypoxic states. The cells were incubated under normoxia or hypoxia for 48 h, and the inhibitor was then added at the indicated concentrations three hours before the sample collection. The phosphorylation of AKT and ERK was dose-dependently reduced by crizotinib under a normoxic state. Under hypoxia, however, crizotinib reduced the phosphorylation to a lesser extent (Fig. 2A).

Furthermore, to evaluate apoptosis, western blot analyses for apoptosis-related molecules were performed. The samples were analyzed 48 h after DMSO (control) or crizotinib (0.1 μM) stimulation under normoxia or hypoxia. Under a normoxic state, the expressions of both cleaved PARP and cleaved caspase-3 were elevated by crizotinib, whereas the expressions were not elevated under hypoxia. These results suggest that the downstream signals are inactivated by crizotinib to a lesser degree and that crizotinib-induced apoptosis is inhibited under a hypoxic state.

To investigate the mechanism of the resistance induced by hypoxia, the morphologic changes of cells subjected to hypoxia were observed. The morphology of the cells changed to a scattered pattern of spindle-shaped cells in a time-dependent manner (Fig. 3A). Then, to evaluate the migration ability, a migration assay was performed using the Boyden chamber method. Under hypoxia, the number of migrating cells increased significantly, compared with the situation under normoxia (8.67±3.5 vs. 1.33±1.53/Field, ^*P=0.026) (Fig. 3B).

The EMT is characterized by an increase in cell scattering and an elongation of the cell shape (18,19). To evaluate whether hypoxia mediates the EMT in the H3122 cell line, changes in the mRNA expression levels of EMT-related genes were evaluated using real-time RT-PCR. Hypoxia time-dependently upregulated SLUG mRNA expression, which is considered to be a master regulator of the EMT (Fig. 4A). The downregulation of E-cadherin is also known to be a pivotal cellular event in the EMT (19,20). E-cadherin mRNA expression was clearly downregulated under hypoxia (Fig. 4A). The expression of mesenchymal marker VIM and the FN1 mRNA was also upregulated (Fig. 4A). Consistent with the mRNA changes, hypoxia time-dependently upregulated the protein expression of slug, fibronectin, and vimentin and downregulated the expression of E-cadherin, along with HIF1α upregulation, which plays a central role in the hypoxic cellular responses (21) (Fig. 4B).

As a major transcription factor, HIF1α plays a central role in hypoxic cellular responses, and this transcription factor is reportedly related to the EMT (21,22). HIF1A was knocked down using specific siRNA to investigate whether it was a key factor in hypoxia-induced EMT and resistance to ALK inhibitors. HIF1A-knockdown abolished the hypoxia-induced morphologic changes (si-HIF1A), whereas the control-siRNA did not (si-Scr) (Fig. 5A). Similarly, HIF1A-knockdown abolished the hypoxia-induced downregulation of E-cadherin and the upregulation of slug, vimentin, and fibronectin (Fig. 5B). That is, HIF1A-knockdown cancelled the hypoxia-induced EMT, resulting in a mesenchymal-epithelial transition (MET). In addition, the resistance to crizotinib was abolished by HIF1A-knockdown (Fig. 5C). These findings indicate that HIF1α transcriptionally regulates the expressions of EMT-related molecules during hypoxia and is implicated in hypoxia-induced resistance to ALK inhibitors in the H3122 cell line.