The human GC cell line, AGS, was purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). The AGS cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 U/ml penicillin, and 100 lg/ml streptomycin at 37°C in a humidified atmosphere containing 5% CO₂. To suppress and activate the c-Jun N-terminal kinase (JNK) pathway, SP600125 (SP, 10 μM) and Anisomycin (Ani, 10 μM) (both from Selleck Chemicals, Houston, TX, USA) were used, respectively to treat the cells for ~2 h at room temperature. To activate mitophagy, carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 3 μM; cat. no. C2920; Sigma-Aldrich, St. Louis, MO, USA) was used for ~1 h at room temperature.

The cells were washed with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 30 min. The cells were incubated with the primary antibody at 4°C overnight. The cells were then washed with PBS 3 times and stained with fluorescent secondary antibody (Alexa-Fluor 488 donkey anti-rabbit secondary antibody (1:1,000, cat. no. A-21206; Invitrogen, Carlsbad, CA, USA) at 37°C for 30 min in the dark. DAPI (cat. no. 28718-90-3; Sigma-Aldrich) was used for nuclear staining, as previously described (25). After fluorescence quenching, the images were acquired with the same exposure settings using a fluorescence microscope with standard excitation filters (Olympus Corp., Tokyo, Japan). The primary antibodies used in the present study were as follows: translocase of outer mitochondrial membrane 20 (Tom20; #ab78547), lysosomal-associated membrane protein 1 (LAMP1; #ab24170), cytochrome c (Cyt-c; (#ab133504) (all from Abcam, Camridge, MA, USA), Parkin (#2132) and p-JNK (#2656) (both from Cell Signaling Technology, Danvers, MA, USA). PI staining was conducted using live cells as previously described (26).

The cells were trypsinized, washed with PBS and then lysed. The lysates were incubated at 4°C for 20 min and centrifuged at 12,000 × g for 15 min. Equal amounts of lysate (20 or 30 μg) were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). The membranes were then blocked in 5% non-fat skim milk/TBST [20 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.1% Tween-20] at room temperature for 2 h and detected with primary antibodies at room temperature for 2 h, as previously described (27). The membranes were then blotted for 1 h at room temperature with an appropriate horseradish peroxidase-linked horseradish peroxidase-conjugated secondary antibodies (Beyotime Institute of Biotechnology, Shanghai, China), followed by enhanced chemiluminescence western blot detection reagents (Amersham Pharmacia Biotech, Piscataway, NJ, USA). The primary antibodies used were as follows: pro-caspase-3 (1:1,000, #9662), cleaved caspase-3 (1:1,000, #9664) (both from Cell Signaling Technology), survivin (1:1,000, #ab469), cellular inhibitor of apoptosis protein-1 (c-IAP1; 1:1,000, #ab25939), Bad (1:2,000, #ab90435) (all from Abcam), Bax (1:2,000, #5023; Cell Signaling Technology), caspase-9 (1:1,000, #ab32539; Abcam), LC3II (1:1,000, #3868; Cell Signaling Technology), p62 (1:1,000, #ab56416; Abcam), Beclin1 (1:1,000, #3495), autophagy-related 5 (ATG5; 1:1,000, #12994), Parkin (1:1,000, #2132) and JNK (1:1,000, #2656) (all from Cell Signaling Technology), complex III subunit core (CIII-core2, 1:1,000; cat. no. 459220; Invitrogen), complex II (CII-30, 1:1,000, #ab110410), complex IV subunit II (CIV-II, 1:1,000, #ab110268), complex I subunit NDUFB8 (CI-20, 1:1,000, #ab110242) (all from Abcam). The blots were visualized using enhanced chemiluminescence reagents (BeyoECL Plus; Beyotime Institute of Biotechnology). The mean densities of the bands were represented as the optical density (OD) in units per square millimeter and normalized to those of β-actin (Quantity One, version 4.6.2; Bio-Rad Laboratories, Inc., Hercules, CA, USA).

The apoptosis of the cells was detected using a TUNEL assay kit according to the manufacturer's instructions (Roche Applied Science, Indianapolis, IN, USA). In brief, the fixed cells were permeabilized with proteinase K for 15 min at room temperature, as previously described (28). The cells were then treated with 3% H₂O₂ to block endogenous peroxidase and then incubated with equilibration buffer and terminal deoxynucleotidyl transferase (TdT) enzyme. Finally, the cells were incubated with anti-digoxigenin-peroxidase conjugate. The cells were examined under a light microscope (magnification, ×100; BX51; Olympus Corp., Tokyo, Japan).

Intracellular ROS levels were measured by flow cytometry using the peroxide-sensitive fluorescent probe 2′7′-dichlorofluorescein diacetate (DCF-DA). Briefly, the cells (3.5×10⁶ cells/well) were plated in 6-well culture plates for 6 h. The cells were then incubated with DCF-DA (25 μM) in PBS at 37°C for 30 min, washed twice with PBS, and detached by treatment with trypsin-EDTA. The detached cells were collected and resuspended in PBS, and the fluorescence intensity of the cells was measured using a flow cytometer (BD FACSVerse; BD Biosciences, San Jose, CA, USA) (29).

The MitoProbe™ JC-1 assay kit (Thermo Fisher Scientific Inc., Waltham, MA, USA) was used to detect changes in MMP. The assay was performed according to the manufacturer's instructions, and the results of the assay were obtained using the BD FACSAria II flow cytometer (BD Biosciences). JC-1 forms J-aggregates emitting red fluorescence at 590 nm in healthy mitochondria and J-monomers emitting green fluorescence at 490 nm in depolarized mitochondria. An increased ratio of J-monomers indicates mitochondrial damage. Carbonyl cyanide m-chlorophenyl hydrazone (CCCP, 50 μM), a disruptor of MMP, was used as a positive control (30).

Cell viability was detected by determining caspase-3 and -9 activity, as well as by MMT assay. Caspase-9 and -3 activities were determined using caspase-assay kits (Beyotime Institute of Biotechnology), which detect the production of the chromophore p-nitroanilide after its cleavage from the peptide substrate DEVD-p-nitroanilide and LEHD-p-nitroanilide (31). MMT assay was conducted as previously described (26). Briefly, the cells were seeded in a 96-well plate at a density of 1×10⁶ cells in triplicate for 7 days at 37°C with 5% CO₂. Subsequently, 20 μl 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (5 mg/ml; pH 7.4; Sigma-Aldrich) were added to the cells for 4 h. The supernatants were then discarded and 100 μl dimethyl sulfoxide (Sigma-Aldrich) was added to each well for 10 min. The OD of the samples was measured at an absorbance of 490 nm using a spectrophotometer (Epoch 2; BioTek Instruments, Inc., Winooski, VT, USA). The assay was repeated 3 times.

Malondialdehyde (MDA) is an end product of lipid peroxidation that results from oxidative damage and reflects the level of cellular damage during oxidative injury. Glutathione (GSH), glutathione peroxidase (GPx) and superoxide dismutase (SOD) are important antioxidants that scavenge free radicals. MDA activity was assessed using a Lipid Peroxidation (MDA) assay kit (Sigma-Aldrich) according to the manufacturer's instructions. The results are expressed as nmol/g tissue. GPx activity (a marker of the systemic antioxidant status) was measured using a Glutathione Peroxidase assay kit (Cayman Chemical Co., Ann Arbor, MI, USA), as per the manufacturer's instructions. The SOD and GSH concentrations were measured using commercial kits (Beyotime Institute of Biotechnology) following the manufacturer's instructions and as previously described (32).

Extracellular lactate levels were measured using the cell culture medium with a lactate assay kit (#K607-100; BioVision, Milpitas, CA, USA). Intracellular glucose levels were measured using cell lysates with a glucose assay kit (#K606-100; BioVision). Adenosine triphosphate (ATP) levels were measured using an ATP assay kit (Celltiter-Glo Luminescent Cell Viability assay; Promega, Madison, WI, USA). The uptake of glucose, the production of lactate and the levels of ATP were all measured according to the manufacturer's instructions and as previously described (33). Mitochondrial respiration was initiated by the addition of glutamate/malate to a final concentration of 5 and 2.5 mmol/l, respectively. Signal transducer and activator of transcription 3 (Stat3) respiration was initiated by the addition of ADP (150 nmol/l); stat4 was measured as the rate of oxygen consumption following ADP phosphorylation.

To induce the overexpression of NR4A1, the pDC316-mCMV-NR4A1 plasmid was purchased from Vigene Bioscience (Rockville, MD, USA) and was transfected with the framework plasmid (1:1) into 293T cells (purchased from ATCC) using Lipofectamine 2000 (Thermo Fisher Scientific, Inc.). Transfection was carried out for 48 h and the viral supernatant was then collected and identified by PCR. Following amplification, the supernatant was acquired again and filtered through a 0.45-μm filter to obtain the adenovirus-NR4A1 (Ad-NR4A1). Subsequently, Ad-NR4A1 was used to transfect the cells and the cells stably expressing NR4A1 were examined by western blot analysis.

The data are expressed as the means ± SD of at least 3 independent experiments. Statistical analysis was performed using one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparison test. A P-value <0.05 was considered to indicate a statistically significant difference.

First, we used various concentrations of TNFα to induce a cell model of GC. Compared to the control group, TNFα reduced the viability of the human GC cell line, AGS, in a dose-dependent manner (Fig. 1A). As there was no difference observed between the 10 and the 20 ng/ml treatment groups, 10 ng/ml was used as the treatment concentration in the following experiments. To explore the role of NR4A1 in TNFα-mediated damage in GC, we examined changes in the expression of NR4A1. We found that NR4A1 expression progressively decreased following treatment with TNFα (Fig. 1B and C). Subsequently, to obtain information as to the role of NR4A1 in cancer cell damage, we overexpressed NR4A1 by transfection with adenovirus (Ad-NR4A1). We demonstrated that transfection with Ad-NR4A1 increased the expression of NR4A1 which had been decreased by TNFα (Fig. 1B and C). Subsequently, we performed TUNEL assay to observe the effects of NR4A1 overexpression on TNFα-induced cell death. Unexpectedly, the overexpression of NR4A1 further increased cellular death when compared to TNFα treatment alone (Fig. 1D and E). However, the number of TUNEL-positive cells in the control group transfected with control adenovirus (Ad-ctrl) did not differ significantly from that of the cells treated with TNFα alone (Fig. 1D and E).

To exclude the role of accidental factors, we also examined the activities of caspase-3 and -9 (Fig. 1F and G). The results revealed that caspase-3 and -9 activities were -in accordance with the above-mentioned finding that the re-introduction (overexpression) of NR4A1 enhanced the sensitivity of GC cells to TNFα-mediated damage. Apart from the changes in protein activity, we also detected alterations in protein expression. Compared to the control group, TNFα elevated the expression of caspase-3 and -9 (Fig. 1H–L). However, the overexpression of NR4A1 further increased the levels of pro-apoptotic proteins, but reduced the expression levels of anti-apoptotic factors (survivin and c-IAP1) (Fig. 1H–L). Taken tother, these data indicated that regaining NR4A1 expression further increased GC cell death induced by TNFα.

NR4A1 facilitates TNFα-evoked mitochondrial injury. It has been reported that NR4A1 induces mitochondrial damage (34). In this study, to determine whether mitochondrial damage is responsible for the promoting effects of NR4A1 on TNFα-induced cell death, we detected mitochondrial function. MMP is the source for the mitochondria to generate ATP, which fuels a cell's biological function. However, we observed that NR4A1 further reduced MMP, as revealed by less red fluorescence and more green fluorescence (Fig. 2A and B). Furthermore, the collapse of MMP may be a result of the mitochondrial permeability transition pore (mPTP) opening (35). As shown in Fig. 2C, compared to the control group, TNFα induced more mPTP opening. However, regaining NR4A1 expression further contributed to mPTP opening.

The consequence of mPTP opening and the reduction of MMP is evidence of mitochondrial metabolism disruption and the initiation of apoptotic signaling (36). In this study, we first focused on mitochondrial apoptosis. Mitochondrial apoptosis is characterized by excessive ROS production and subsequent cellular oxidative injury, which contributes to the leakage of pro-apoptotic factors from the mitochondria (29). In this study, using flow cytometric analysis, we found that TNFα treatment enhanced the levels of ROS (Fig. 2D and E). Furthermore, the overexpression of NR4A1 amplified ROS overproduction. In response to excessive ROS production, the levels of antioxidant factors, such as GSH, SOD and GPX, were reduced, suggesting an imbalance of the redox status (Fig. 2F–I). The enhancement of NR4A1 further consumed antioxidant factors, but generated greater levels of MDA, an end product of lipid peroxidation. Finally, we observed greater amounts of Cyt-c released into the cytoplasm in response to TNFα treatment, which was similar to the results in the NR4A1 overexpression group (Fig. 2J). Collectively, these data suggested that NR4A1 enhanced TNFα-induced mitochondrial damage.

In addition to cellular apoptosis, mitochondrial energy production is another determinant for cancer growth and cellular survival (37). Considering that NR4A1 promoted mitochondrial apoptosis, we therefore wished to determine whether NR4A1 hinders mitochondrial energy production in the presence of TNFα. First, compared to the control group, TNFα abated the content of ATP, which was further reduced once NR4A1 was expression was recovered (Fig. 3A), suggesting that NR4A1 has the ability to influence mitochondrial energy generation. Since ATP is produced via the mitochondrial respiratory complex, we therefore examined changes in the mitochondrial respiratory complex. Compared to the control group, TNFα suppressed the contents of the mitochondrial respiratory complex (Fig. 3B–F). However, NR4A1 further suppressed the expression of these factors of the mitochondrial respiratory complex (Fig. 3B–F). Correspondingly, mitochondrial Stat3 (Fig. 3G) and Stat4 (Fig. 3H) respiratory functions also declined when NR4A1 was overexpressed. These data indicated that NR4A1 impaired mitochondrial respiratory function. The reduced mitochondrial energy metabolism may be associated with a decrease in glucose intake and lactate production. Thus, we measured the content of glucose in medium. As shown in Fig. 3I and J, TNFα indeed reduced glucose consumption and lactate generation, suggesting the termination of glycometabolism in GC cells. By contrast, the overexpression of NR4A1 further limited glucose intake and lactate production. Taken together, these data indicated that NR4A1 enhanced the suppressive effects of TNFα on mitochondrial energy metabolism.

To determine the mechanisms through which NR4A1 enhanced TNFα-induced mitochondrial damage, we focused on the mitochondria repairing system itself, mitophagy. Damaged mitochondria activate mitophagy, which facilitates the removal and clearing of malfunctioning mitochondria via the lysosome, leading to the preservation of mitochondrial quantity and quality (38). Thus, mitophagy counteracts TNFα-induced mitochondrial damage (39). Based on this, we first investigated mitophagy activity. In response to TNFα treatment, the levels of mito-LC3II, Beclin1, ATG5 and p62 were increased (Fig. 4A–E), suggestive of mitophagy activation. However, the overexpression of NR4A1 reversed this tendency, illustrating the inhibitory effect of NR4A1 on mitophagy. To provide further direct evidence of mitophagy, we used immunofluorescence to label the mitochondria and lysosomes at the same time. As shown in Fig. 4F and G, in the control group, few mitochondria were merged with the lysosomes. Following treatment with TNFα, most of the mitochondria were contained by lysosomes, indicative of mitophagy activation. However, the re-introduction of NR4A1 suppressed the lysosome-mitochondria interaction, displaying the inhibitory role of NR4A1 on mitophagy. Subsequently, to determine whether mitophagy inhibition contributes to the excess mitochondrial damage following NR4A1 overexpression, we used FCCP to activate mitophagy. After the activation of mitophagy, caspase-9 activity increased despite the overexpression of NR4A1 (Fig. 4H). These data indicated that mitophagy, the protective system of the mitochondria, was activated by TNFα and contributed to mitochondrial protection and resistance to apoptosis. The recovery of NR4A1 expression inhibited mitophagy and enhanced the sensitivity of GC cells to TNFα-induced cell death.

Finally, we wished to determine the underlying signal that modulates mitophagy activity under TNFα treatment conditions. Parkin is the primary receptor for mitophagy, and it mediates the damaged mitochondrial removal in several types of cells (40). Therefore, we first examined changes in Parkin expression. We found that Parkin expression was upregulated by TNFα treatment, but was decreased upon the overexpression of NR4A1 (Fig. 5A–C). This information established the regulatory effects of TNFα on Parkin. Furthermore, to explain the mechanisms through which TNFα upregulates Parkin expression, we focused on JNK. JNK is considered the upstream trigger of Parkin expression (41). To determine whether TNFα regulates Parkin-required mitophagy via JNK, an inhibitor (SP600125, SP) and activator (Ani) of JNK were used. As shown in Fig. 5A and B, compared to the control group, TNFα treatment increased JNK activity as evidenced by greater levels of phosphorylated JNK. However, NR4A1 overexpression suppressed TNFα-induced JNK activation. Next, upon the inhibition of JNK via SP600125 under TNFα treatment, JNK activity was blocked, and Parkin expression was also inhibited (Fig. 5A–C). By contrast, the activation of JNK in the cells overexpressing NR4A1 increased Parkin expression (Fig. 5A–C). In addition, the results of co-immunofluorescence of p-JNK and Parkin were also in agreement with the above-mentioned results (Fig. 5F). These data indicated that the TNFα-induced activation of the JNK pathway was responsible for Parkin upregulation.

To provide further direct evidence of the role of JNK in Parkin-mediated mitophagy, we examined markers of mitophagy. After the blockade of JNK under TNFα treatment, mitophagy was reduced, as evidenced by less mito-LC3II and p62 expression (Fig. 5A, D and E), which was similar to the results observed in the NR4A1 group. By contrast, the activation of JNK re-evoked the expression of mito-LC3II and p62 despite the overexpression of NR4A1 (Fig. 5A, D and E).

Finally, to determine whether JNK is also involved in the resistance of GC cells to TNFα-induced damage, we used PI to detect cellular damage in cells treated with or without JNK activator. As shown in Fig. 5G, the inhibition of JNK increased the number of TNFα-induced PI-positive cells. However, following NR4A1 overexpression, the activation of JNK reduced the number of PI-positive cells. Taken together, these data suggested that JNK/Parkin-dependent mitophagy was responsible for the therapeutic resistance observed with TNFα and that NR4A1 enhanced the fatal effects of TNFα via the inhibition of JNK/Parkin-dependent mitophagy.