Transmembrane protein GRINA modulates aerobic glycolysis and promotes tumor progression in gastric cancer

This retrospective analysis includes 569 patients with resected primary gastric cancer who were treated at Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University from 2005 to 2011. Tumours were assigned by tumour pathology, node, and metastasis stage as defined by the American Joint Committee on Cancer (7th edition). For each case, the diagnosis was confirmed by a review of the H&E stained slides, and a representative block from each specimen was chosen for immunohistochemical (IHC) analysis. Patient characteristics were obtained from the Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University electronic medical records. The study was approved by the Research Ethics Committee of Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University and was carried out in accordance with the ethical standards of the World Medical Association’s Declaration of Helsinki. Signed informed consent was obtained from all the patients included in this study. The inclusion criteria for our study were as follows: 1) an obvious pathologic diagnosis of gastric cancer; 2) primary gastric cancer cases without a history of other solid tumours; 3) accepted radical surgery treatment without residual tumour; 4) no exposure to chemotherapy, radiotherapy, or other anti-cancer therapies before surgery; and 5) availability of complete clinicopathologic and follow-up data.

In total, 569 gastric cancer cases, diagnosed pathologically and treated from 2005 to 2011, were retrospectively identified from the hospitalization archives of the Department of Gastrointestinal Surgery, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University. The paraffin-embedded tissue samples of these patients were used for tissue microarray construction and immunohistochemical staining. Briefly, after tissue sections were deparaffinised, rehydrated with graded ethanol, incubated with 0.3% hydrogen peroxide for 30 min, and blocked with 10% bovine serum albumin (BSA) (Sangon, Shanghai, China), slides were first incubated overnight with an antibody against GRINA (dilution 1:50, ABGENT, AP13558c) at 4 °C, and then labelled with an HRP-conjugated (rabbit) secondary antibody (ThermoFisher Scientific, USA) at room temperature for 1 h. Finally, positive staining was visualized using diaminobenzidine (DAB) substrate liquid (Gene Tech, Shanghai), followed by counterstaining with haematoxylin. All sections were observed and photographed with a microscope (Carl Zeiss, Germany). Scoring was performed based on the ratio of positively stained cells: 0–5% scored 0; 6–35% scored 1; 36–70% scored 2; more than 70% scored 3, and staining intensity: no staining scored 0, weakly stained scored 1, moderately stained scored 2 and strongly stained scored 3. The final score was determined using the percentage of the positive cell score × staining intensity score as follows: “–” for a score of 0–1, “+” for a score of 2–3, “+ +” for a score of 4–6, and “+ + +” for a score of > 6; low expression was defined as a total score < 4 and high expression with a total score ≥ 4. These scores were determined independently by two senior pathologists in a blinded manner.

Human gastric cancer cell lines SGC-7901, MGC-803, BGC-823, MKN45, HGC27, AGS, N87, and GES-1 were all maintained at Shanghai Cancer Institute, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University. Cells were cultured in the suggested medium according to American Type Culture Collection (ATCC, Manassas, VA) protocols, supplemented with 10% (v/v) foetal bovine serum (FBS) and 1% antibiotics at 37 °C in a humidified incubator under 5% CO₂ conditions.

Total RNA was extracted using Trizol reagent (Takara), and reverse transcribed with the PrimeScript RT-PCR kit (Takara) according to the manufacturer’s instructions. Real-time PCR analysis was performed using the SYBR Premix Ex Taq (Takara) on a 7500 Real-time PCR system (Applied Biosystems) at the recommended thermal cycling settings: one initial cycle at 95 °C for 2 min followed by 40 cycles of 10 s at 95 °C and 34 s at 60 °C. Relative mRNA expression was calculated using the 2^(-ΔCT) method and normalized to 18S mRNA levels. The primer sequences used in this study are listed in Additional file 1: Table S1.

Cell lysis was performed using a total protein extraction buffer (Beyotime, Shanghai, China) and protein concentration was measured using a BCA Protein Assay Kit (Pierce Biotechnology). Cell lysates were separated by 6–12% SDS-PAGE and transferred to a PVDF membrane. After blocking with 5% skimmed milk, the membrane was probed with one of the following primary antibodies: GRINA (1:1000, ABGENT, AP13558c), β-ACTIN (1:1000, Servicebio, GB13001–3), Bcl-2 (1:1000, Cell Signaling Technology, #4223), Bax (1:1000, Cell Signaling Technology, #5023), Cleaved-caspase3 (1:1000, Cell Signaling Technology, #9664), Cleaved-caspase7 (1:1000, Cell Signaling Technology, #8438), Cleaved-caspase9 (1:1000, Cell Signaling Technology, #9505), Cyclin D1 (1:10000, Abcam, ab134175), Cyclin E (1:1000, Abcam, ab3927), Akt (1:1000, Cell Signaling Technology, #4685), p-Akt (1:2000, Cell Signaling Technology, #4060), P70S6K (1:1000, Cell Signaling Technology, #9202), p-P70S6K (1:1000, Cell Signaling Technology, #9204), 4EBP1 (1:1000, Cell Signaling Technology, #9644), p-4EBP1 (1:1000, Cell Signaling Technology, #2855), AMPK (1:1000, Cell Signaling Technology, #2532), p-AMPK (1:1000, Cell Signaling Technology, #2535), and respective species-specific secondary antibodies (ThermoFisher Scientific). The bound secondary antibodies were detected using the Odyssey imaging system (LI-COR Biosciences, Lincoln, NE).

Cells were seeded in 6-well plates in complete RPMI 1640 medium. The monolayers were scratched with a 200 μl plastic pipette tip to create a uniform wound. The monolayers were then washed with 1 × PBS and incubated in culture medium without FBS. The wound margin distances between the two edges of the migrating cell sheets were photographed after scratching using phase-contrast microscopy. All experiments were performed in triplicate.

The invasive potential of gastric cancer cells was measured by a Transwell assay (Corning, NY, USA) according to the manufacturer’s instructions. For the migration assay, 2 × 10⁴ cells in 200 μl medium were seeded into the upper chamber of the Transwell inserts. RPMI 1640 medium containing 10% (v/v) FBS was added to the bottom chamber. Cells were incubated at 37 °C and allowed to migrate for 24 h. At the designated time points, the non-invading cells that remained on the upper surface were removed. The migrated cells were fixed with 4% paraformaldehyde and stained with 0.5% crystal violet. The number of cells on the lower surface was counted under a light microscope in six random fields. Each experiment was performed in triplicate and repeated twice.

For the apoptosis assay, 10 × 10⁵ cells per well were cultured under serum-deprivation conditions in 6-well plates. Adherent cells were detached with 0.25% trypsin without EDTA in 1 × PBS. Cells were harvested in complete RPMI 1640 medium and centrifuged at 1000 rpm for 5 min. Each of the cells were washed with 1 × PBS, stained with 50 μg/ml propidium iodide (PI) and Annexin V-FITC (BD Pharmingen, USA) following the manufacturer’s instructions. The percentage of Annexin V (+) and PI (−) cells was analysed by flow cytometry.

Cells were cultured in 6-well plates in complete RPMI 1640 medium. In a different set of experiments, cells were cultured for 48 h after transfection with the shRNA against GRINA or the negative control shRNA. Cells were then harvested by trypsin treatment and fixed with 70% ethanol at 4 °C. Fixed cells were incubated with RNase A for 1 h at 37 °C and stained with propidium iodide for 20 min on ice. The number of cells in each phase of the cell cycle was determined at an excitation of 488 nm and an emission of 585 nm using a flow cytometer (FACSCalibur; Becton Dickinson, San Jose, CA, USA) and analysed with Mod Fit LT software (Verity Software House, Topsham, USA).

The sequences of the GRINA short hairpin RNA (shRNA) are shown in Additional file 2: Table S2. The shRNA-containing plasmids and a negative control plasmid were purchased from GenePharma (Shanghai, China). The plasmid containing GRINA-HA and a negative control plasmid were obtained from Asia-Vector Biotechnology (Shanghai, China). Lentivirus packaging was performed in 293 T cells with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) and virus titres were determined according to standard protocols. Cells were infected with 1 × 10⁶ recombinant lentivirus-transducing units in the presence of 6 mg/ml polybrene (Sigma-Aldrich, H9268). When grown at 60–70% confluence, the indicated cells were infected with the supernatant containing viral particles. Stable GRINA knockdown or overexpression cells were then cultured in the presence of 2 mg/ml puromycin (Gibco, A1113802).

Cells were plated at 60–70% confluence in 6-well plates. The sequences of the small interfering RNA (siRNA) used are shown in Additional file 3: Table S3. Scrambled siRNA targeting no known gene sequence was used as the negative control. Lipofectamine® RNAiMAX reagent (ThermoFisher Scientific, #13778030) was used to perform siRNA transfection according to the manufacturer’s protocol.

Athymic nu/nu mice aged 6 to 8 weeks were kept on a 12-h day/night cycle with free access to food and water. For subcutaneous xenografts in the genetic inhibition study, 2 × 10⁶ sh-Ctrl or sh-GRINA AGS cells in 200 ml Hanks buffered saline were injected subcutaneously in the lower back. For subcutaneous xenografts in the genetic overexpression study, 2 × 106 GRINA-Ctrl or GRINA-OE N87 cells in 200 ml Hanks buffered saline were injected subcutaneously in the lower back. After 4 weeks, the mice were sacrificed, the tumour was isolated and tumour weight was measured. Tumour volumes were calculated as volume = length × width²/2.

Paraffin sections (5 mm) of mice tumors were deparaffinized, rehydrated with graded ethanol, and subjected to antigen retrieval. After blocking with 10% BSA, sections were incubated with primary antibodies for 1 h followed by incubation with secondary antibodies for 30 min at room temperature. For staining of Ki67, primary and secondary antibodies were rabbit anti-Ki67 (1:200, Abcam, ab15580), donkey anti-rabbit Alexa Fluor 594 (1:400, Jackson ImmunoResearch, #711–585-152). ApopTag Red In Situ Apoptosis Detection Kit (S7165, Millipore) was used for apoptosis detection. Nuclei were counterstained with DAPI (40,6-diamidino-2-phenylindole dihydrochloride; AppliChem, A4099). Digital images were acquired with fluorescence or confocal microscopes equipped with a digital camera (Nikon).

A Pierce Agarose ChIP Kit was purchased from Thermo Fisher Scientific. Cells were cross-linked with 1% formaldehyde for 10 min at room temperature and quenched by adding glycine (1.25 M). Fixed cells were harvested in SDS buffer with a protease inhibitor and then sonicated to generate DNA fragments of 200–1000 base pairs (bp) in length. The sheared chromatin-lysed extracts were incubated overnight at 4 °C with the anti-c-Myc antibody or control IgG with rotation. After immunoprecipitation (IP) of the cross-linked protein/DNA, the immunocomplexes were reversed to release the DNA. PCR was performed with the input DNA or the immunoprecipitates. The PCR products were separated by agarose gel electrophoresis. The primers used for ChIP analysis are listed in Additional file 4: Table S4.

In vitro cell metabolic alternations were monitored with the Seahorse XF96 Flux Analyser (Seahorse Bioscience), according to the manufacturer’s instructions. Cells were seeded in a XF96-well plate at a density of 2 × 10⁴ per well and allowed to attach overnight, followed by serum starvation for 24 h. For assessment of the real-time glycolytic rate (ECAR), which is an indicator of net proton loss during glycolysis, cells were incubated in unbuffered medium followed by a sequential injection of 10 mM glucose, 1 mM oligomycin (Sigma-Aldrich), and 80 mM 2-deoxyglucose (2-DG, Sigma-Aldrich, D8375). The mitochondrial respiration (OCR) was assessed using a sequential injection of 1 mM oligomycin, carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, Sigma-Aldrich, C2920), and 2 mM antimycin A and rotenone (Sigma-Aldrich). Oligomycin A is applied to inhibit proton flow through ATP synthase and induce maximal glycolytic metabolism. FCCP, an uncoupler of electron transport, leads to collapse of membrane potential and peak oxygen consumption. To achieve maximal OCR, the concentration of FCCP used in gastric cancer cells was 500 nM. Antimycin A and rotenone were used to inhibit mitochondrial respiration by blocking complex III (Ubiquinone: Cytochrome b-c complex). Both ECAR and OCR measurements were normalized to total protein content and reported as mPH/min for ECAR and pmol/min for OCR. Each datum was determined in triplicate.

The gene expression data for stomach adenocarcinoma (STAD) were downloaded from TCGA, which is maintained by Broad Institute’s TCGA workgroup. The RNA-seq level 4 gene expression data contained log2-transformed RNA-seq by expectation maximization (RSEM) values summarized at the gene level. The data from the Gene Expression Omnibus (GEO) database had the series accession GSE13911.

Data were presented as means ± SD. The SPSS software program (version 19.0, IBM Corporation) was used for statistical analysis. Graphical representations were prepared with GraphPad Prism 5 (San Diego, CA, USA) software. The correlation of GRINA expression with clinicopathological parameters in patients with gastric cancer was evaluated by the chi-square test or Fisher’s exact test. Survival curves were evaluated using the Kaplan-Meier method and differences between survival curves were tested by the log-rank test. All statistical tests were two-sided. The Student’s t-test or one-way ANOVA was used for comparison between groups. P < 0.05 was considered statistically significant.

Comparing gene expression in 32 pairs of cancerous and noncancerous tissues from patients with gastric cancer in The Cancer Genome Atlas (TCGA) database revealed 36 dysregulated genes on chromosome 8q24 (fold change > 1.5 or fold change < − 1.5, q value < 0.001) (Fig. 1a, Additional file 5: Figure S1A), 5 of which were downregulated and 31 upregulated in gastric cancer tissues (Additional file 6: Table S5, Additional file 7: Table S6). GRINA was one of the most significantly upregulated genes on chromosome 8q24.3 (Additional file 8: Figure S1B).

In samples from TCGA database and Gene Expression Omnibus (GEO) database, GRINA expression was found to be significantly increased in gastric cancer samples compared with that in normal samples (matched or not matched) (Figs 1b-1d). However, no such difference was found for the other five members of the TMBIM family (Additional file 9: Figure S1C-G). These results indicated that GRINA was significantly upregulated in gastric cancer.

To validate the results of bioinformatics analysis, 20 fresh gastric cancer samples and their matched normal tissues were collected. The mRNA and protein levels of GRINA was assessed (Figs 1e-1f). The results demonstrated elevated expression of GRINA in tumour tissues consistent with the bioinformatics results.

Analysis of 876 gastric cancer patients with 33 months of follow-up data in KMplotter database demonstrated that patients with high GRINA expression had worse overall survival (OS) than those with low GRINA expression (Additional file 10: Figure S2), suggesting that GRINA indicates poor prognosis in gastric cancer.

To further understand GRINA expression in gastric cancer, tissue microarrays (TMA) containing 569 gastric cancer samples were constructed and subjected to IHC staining (Fig. 2a). The correlation analysis between GRINA expression and clinicopathological parameters in gastric cancer demonstrated that GRINA expression was significantly correlated with histological differentiation (P = 0.04), T stage (P = 0.027), N stage (P = 0.03), distant metastasis (P = 0.032), blood vessel invasion (P = 0.04) and perineuronal invasion (P = 0.005), but not with age and gender (Table 1).

We then analysed the correlation between GRINA expression and OS by the Kaplan-Meier method in the aforementioned 569 gastric cancer patients. The Kaplan-Meier analysis of overall survival related to histological differentiation, TNM stage (stage I and II, stage III and IV), lymphatic metastasis, vascular invasion, and neural invasion also suggested that patients with higher GRINA expression had worse prognosis than those with lower GRINA expression (P < 0.05) (Figs 2B-2L).

The mRNA levels of GRINA in 7 gastric cancer cell lines (SGC-7901, MGC-803, BGC-823, MKN45, HGC27, AGS, N87) and in a normal immortalized gastric mucosal epithelial cell line (GES-1) were examined by Real-time PCR. We found higher expression of GRINA in the 7 gastric cancer cell lines than in GES-1 (Fig. 3a). The results demonstrated that GRINA expression was higher in AGS and BGC-823 cell lines while lower in HGC27 and N87 cell lines. Therefore, AGS, BGC-823, HGC27 and N87 cell lines were selected for further experiments in vitro.

Lentiviral transfection of two short hairpin RNAs (shRNA) in AGS and BGC-823 cells significantly downregulated the GRINA mRNA levels by over 85% as well as the protein levels (Additional file 11: Figure S3A-D). In addition, cell proliferation was decreased in the shRNA-1 and shRNA-2 groups compared with that in the control groups (Figs. 3b-3c). Then we overexpressed GRINA in HGC27 and N87 cell lines (Additional file 3: Fig. S3E-H). Cell proliferation was increased in the GRINA-OE group compared with that in the control group (Figs. 3d-3e). The results indicated that GRINA remarkably promoted gastric cancer cell proliferation.

Next, we aimed to determine the role of GRINA in tumour growth in vivo. AGS cells stably transfected with GRINA-shRNA or vector-control were respectively injected subcutaneouslly into nude mice. After 4 weeks, we found that tumour weights and sizes were significantly smaller in the shRNA group than in the control group (P < 0.01) (Figs. 3F-3H). Then N87 cells stably transfected with GRINA-OE or vector-control were also injected subcutaneouslly into nude mice. Instead, tumour weights and sizes were markedly larger in the GRINA-OE group than in the control group 4 weeks later (P < 0.01) (Figs. 3I-3K) Immunofluorescence staining for Ki67 reaveled that tumor tissues in the control group showed higher positivity rate than those in the shRNA group (Fig. 3L).

To determine the influence of GRINA on cell migratory and invasive capacity, we performed scratch wound assays. In both AGS and BGC-823 cells, cell migration distances were shorter in the shRNA groups than that in the control groups. The results demonstrated that cell migration and invasion were restrained upon GRINA knockdown (Figs. 4A-4B).

Next, the results of transwell experiments showed that the number of migrated cells was significantly reduced after GRINA was knocked down in AGS and BGC-823 cell lines (Figs. 4C-4D). On the contrary, the number of migrated cells was remarkably increased after GRINA was overexpressed in HGC27 and N87 cells (Figs 4E-4F). These data suggested that GRINA could promote cell migratory and invasive capacity.

To further understand the mechanism by which GRINA promoted gastric cancer cell proliferation, we added 1 mM thymidine to each cell group for 24 h and cultured the cells in RPMI 1640 medium without FBS for another 24 h, and then counted the proportion of cells in each cell cycle phase. The results demonstrated that the proportion of cells in G1/S phase was larger in the shRNA groups than in the control groups (Figs. 5A-5B), indicating that GRINA knockdown induced cell cycle arrest in the G1 phase and suppressed cell proliferation.

Oncogenes can be activated by factors such as transcription factor regulation, chromosome translocation, genome copy number amplification, and point mutation [17]. We inferred that GRINA upregulation in gastric cancer was mainly induced by transcription factors. Thus, we searched for putative transcription factors that possibly bound the promoter of GRINA in the JASPAR (http://​jaspar.​genereg.​net/) database and found a few transcription factors that probably regulate GRINA expression (Additional file 11: Table S7). The oncogene c-Myc aroused our interest [18]. Gene set enrichment analysis (GSEA) of TCGA data using the Hallmarks gene sets also showed that GRINA expression was related to Myc targets alteration (Fig. 5C). Additional analysis revealed a positive correlation between GRINA and Myc expression in TCGA database (Fig. 5D). We then designed 10 pairs of primers based on the 0–3000 bp nucleotide sequence in the GRINA promoter. The results of ChIP analysis indicated that c-Myc can bind the GRINA promoter both at the − 2400 bp to − 2100 bp and − 900 bp to − 600 bp stretches (Fig. 5E). To further confirm the ChIP results, we knocked down c-Myc in AGS and BGC-823 cells. Strikingly, GRINA expression was also decreased with decreased c-Myc (Figs. 5F-5G). These results suggested that c-Myc could transcriptionally regulate GRINA expression.

Next, we assessed CyclinD1 and CyclinE, two key members of cell cycle, and found that their expression were decreased when GRINA was knocked down (Fig. 5H). These data suggested that GRINA was transcriptionally mediated by c-Myc and promoted cell cycle transition.

The results of GSEA using TCGA database also showed striking alterations in metabolic processes including upregulation of glycolysis and reactive oxygen species pathways (Fig. 6A, Additional file 4: Fig. S4A), which indicated a correlation between GRINA expression and enhanced glycolytic metabolism in gastric cancer. Consistently, real-time qualitative PCR confirmed that several metabolic enzymes involved in glycolysis (Figs. 6B-6E), phosphate pentose pathway (PPP) (Additional file 4: Fig. S4B-C, G-H), hexosamine biosynthesis pathway (HBP) (Additional file 4: Fig. S4D, I), and glutamine (Gln) metabolism (Additional file 4: Fig. S4E-F, J-K) were significantly downregulated by GRINA knockdown and upregulated by GRINA overexpression. To further study the glucose metabolism modulated by GRINA, we measured the extracellular acid ratio (ECAR) and oxygen consumption ratio (OCR) upon GRINA silencing. The results showed that silencing of GRINA attenuated whereas overexpression of GRINA enhanced the glycolytic capability of gastric cancer cells but not oxidative phosphorylation, indicating that GRINA principally affects the glycolytic component of glucose metabolism (Figs 6F-6I, Additional file 4: Fig. S4L-O). To determine the signalling pathway that allows GRINA to affect glycolysis, we determined the activity of AMP-activated protein kinase (AMPK) and mammalian target of rapamycin (mTOR) signalling, which are known to be influenced by the intracellular energy status [19, 20]. GRINA knockdown did not influence phosphorylation of AMPK (Additional file 4: Fig. S4P); on the contrary, GRINA knockdown inhibited the phosphorylation of Akt and the downstream effectors of mTOR, P70S6K and 4EBP1, indicating that the mTOR pathway could be activated by GRINA (Fig. 6J). These findings indicated that GRINA facilitates gastric cancer cell survival by modulating aerobic glycolysis.

Studies have shown that GRINA could participate in biological processes such as apoptosis and had a close relationship with Bcl-2 family members [16]. Therefore, we examined apoptosis in the GRINA-shRNA and control groups by the Annexin V/PI double labelling method. The percentages of apoptotic cells were obviously augmented in shRNA groups compared with those in control groups after cells were cultured in medium without foetal bovine serum (FBS) for 48 h (Figs 7A-7B). TUNEL staining also reaveled that mice tumor tissues in the shRNA group showed higher apoptosis rate than those in the control group (Fig. 7C). Then we assessed Bcl-2 and Bax expression after GIRNA knockdown in AGS and BGC-823 cells. We found that Bcl-2 expression was significantly decreased whereas Bax expression was obviously increased in both cell lines (Fig. 7D). The levels of Cleaved-caspase3, Cleaved-caspase7, and Cleaved-caspase9 were significantly increased (Fig. 7D). These results suggested GRINA could regulate apoptosis via Bcl-2 family members.