Background
Gastric cancer (GC) is the fourth most common cancer and the second leading cause of cancer-related death worldwide [
1]. Effective treatment is lagging for this disease, and the underlying mechanism is still not generally understood. Metabolic plasticity is considered one of the defining hallmarks of cancer, suggesting that this could be a potential target for GC diagnosis and therapeutics [
2].
MACC1 is a transcriptional regulator of MET, and has emerged as a master oncogene that is upregulated in a variety of tumors, promoting proliferation, invasion, and chemotherapy resistance [
3]. Our previous study demonstrated that MACC1 expression is upregulated and promotes glycolysis under metabolic stress, a state characterized by nutrient deprivation and often occurring in mature tumor tissues due to abnormal vascularization [
4,
5]. However, the specific regulation of MACC1 under metabolic stress remains poorly understood.
LncRNAs have emerged as a concept in biology, playing integral roles in the control of a wide variety of cellular functions through transcriptional and post-transcriptional regulation of gene activity [
6]. MACC1-AS1 is the cognate antisense lncRNA of MACC1, and to date, its function and molecular mechanism have not been investigated. Generally, antisense lncRNA exerts regulatory effects on the expression of its sense counterpart. This relationship is also supported through the analysis of MACC1-AS1 and MACC1 expression using the TCGA database and tumor specimens from patients. Based on this, we suspected that MACC1-AS1 might regulate MACC1 expression, and the potential mechanism was further explored in this study.
Herein, we provide new evidence that MACC1-AS1 lncRNA is highly expressed in GC, which was found to be correlated with advanced clinical stage and poor prognosis. MACC1-AS1 potentiated metabolic plasticity through enhanced glycolysis and antioxidant activity under metabolic stress, which was found to be mediated through post-translational regulation of MACC1 mRNA stability.
Methods
Reagents
AMPK inhibitor dorsomorphin was purchased from Selleck (Houston, TX, USA); 2-Deoxy-D-glucose (2-DG), Cycloheximide (CHX), Dactinomycin D from MedChem Express (Princeton, MA, USA); 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide(MTT) from Biosharp (China); 2-NBDG (2-(N-(7-Nitrobenz-2-oxa-1,3-diazol-4-yl)Amino)-2 Deoxyglucose) and 4′,6-diamidino-2-phenylindole(DAPI) from Invitrogen; H2O2, N-acetyl-L-cysteine (NAC) from Sigma-Aldrich; Other reagents were mentioned for below procedure.
Clinical samples
A total of 123 formalin-fixed and paraffin-embedded (FFPE) GC tissue samples from patients who had underwent an operation in Nanfang Hospital (Guangzhou, China) between January 2006 and December 2010 were included in the cohort. All patients were staged based on the criteria of the 7th Edition of the AJCC Cancer Staging Manual: Stomach (2010) [
7]. No recruited patients received any preoperative treatment. For stage I–III patients, prognosis was assessed based on disease-free survival (DFS), whereas for stage IV patients were analyzed based on overall survival (OS). All patients provided written informed consent before the study, which was approved by the Nanfang Hospital Ethics Review Committee (Guangzhou, China).
Immunohistochemical (IHC) and in situ hybridization (ISH) staining
Briefly, IHC was applied to detect the expression of MACC1 (Abcam, San Francisco, CA), whereas ISH was utilized to detect the presence of MACC1-AS1 using a double digoxigenin (DIG)-labeled mercury locked nucleic acid probe (5’DigN/TCAATGCAGATCTAATACTCCT/3’Dig_N) (miRCURY LNA™,Exiqon, Vedbaek, Denmark). Staining score was assessed by two independent reviewers as: 0 = no staining, 1 = weak staining, 2 = moderate staining, 3 = strong staining. Tumor cells in five fields were selected randomly and scored based on the percentages of positive stained cells (1 = 0-25%, 2 = 26-50%, 3 = 51-75%, 4 = 76-100%). The final scores were computed by multiplying the intensity score and the percentage score of positive cells. According to scores, samples were divided into three groups, the negative expression (score 0-2), low expression (score 3-7) and high expression group (score 8-12). In the survival analysis, negative and low scores group were defined as MACC1-AS1 negative group (−), while high expression group were defined as MACC1-AS1 positive group (+).
Cell culture and treatment
Five human GC cell lines (AGS, BGC803, BGC823, MKN45, SGC7901) and immortalized human gastric epithelial cell line (GES-1) were obtained from Foleibao Biotechnology Development Co. (Shanghai, China). All cell lines were cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum (HyClone, Logan, UT, USA) at 37 °C under 5% CO2. Glucose deprivation was used to induce metabolic stress using glucose-free 1640 medium (GIBCO, USA). MKN45 and BGC803 cells stably transfected with pLenti-EF1a-Luc-F2A-puro-CMV-MACC1-AS1 constructs or pLKD-U6-shRNA-EF1a-LUC-F2A-Puro-MACC1 (Obio Technology, shanghai, China) were grown in 1 μg/ml puromycin (Invitrogen).
Vector construction
The cDNA of MACC1-AS1 was PCR-amplified and subcloned into pcDNA 3.1 (+) vector (Invitrogen). The pcDNA3.1-antisense-MACC1-AS1 was constructed by subcloning the antisense MACC1-AS1 sequence into the pcDNA3.1 (−) vector (Invitrogen). The putative MACC1 promoter was PCR-amplified and cloned into the pGL3-control vector (Promega) to construct the pGL3-MACC1-3′UTR vector. All products were validated by DNA sequencing.
Small interfering RNA transfection
Cells were plated and cultured in growth media until cell density reached to 30-40% prior to siRNA transfection using Lipofectamine 2000 (Invitrogen). siRNA sequences are as following: AMPK (5’-CTATGCTGCACCAGAAGTA-3′), Lin 28 (5′- GAAGAAATCCACAGCCCTA-3′), and synthetic sequence-scrambled siRNA was used as a negative control.
Biotin pull-down assays and mass spectrometry
MACC1-AS1 and its antisense control sequence were in vitro transcribed from pGEM-T plasmid and biotin-labeled with Biotin RNA Labeling Mix (Roche) and T7 RNA polymerase (Roche). 1 mg of whole-cell lysates from MKN45 cells were incubated with purified biotinylated transcripts for 1 h at room temperature. Complexes were isolated with streptavidin agarose beads (Invitrogen). The RNA present in the pull-down material was detected by qRT-PCR analysis, and retrieved proteins were detected by western blot analysis or resolved in gradient gel electrophoresis followed by mass spectrometry (MS) identification (TripleTOF 5600 LCMS, AB SCIEX).
Next-generation RNA sequencing (RNA-seq)
RNA-seq of MACC1-AS1-overexpressing or vector-transfected MKN45 cells were carried out using the platform BGIseq-500 (Shenzhen, China). Gene set enrichment analysis (GSEA) was carried out using GSEA v3.0.
RNA isolation and qPCR analysis
Total RNA of cells were extracted by TRIZOL reagent (Invitrogen), the cDNA was synthesized by using the first strand cDNA synthesis kit (Takara Shuzo, Kyoto, Japan). Quantitative PCR was applied using the SYBR Green dye (Takara, Japan) on LightCycler 480 system (Roche, Penzberg, Germany). The primer sequences were listed in the Additional file
1: Table S7. The relative expression was normalized to snRNA U6 or β-actin by 2
-ΔΔCt method.
Western blotting
Cells were harvested and lysed on ice in RIPA buffer (50 mM Tris, 100 mM NaCl, pH 8.0, 10 mM EDTA, pH 7.0, 0.4% v/v Triton X-100, 10 mM nicotinamide) containing protease and phosphatase inhibitors (Beyotime, Shanghai, China). Lysates were centrifuged at 15,000 g at 4 °C for 15 min. Proteins were resolved using SDS–polyacrylamide gel electrophoresis and transferred onto PVDF membrane. Membranes were blocked in 5% milk, incubated with primary antibody at recommended concentration in TBST with 5% BSA overnight at 4 °C, then incubated with fluorescent-tagged secondary antibody at a concentration of 1:15,000 and read using a LI-COR Odyssey infrared imaging system. The primary antibody used were: anti-MACC1, anti-GLUT1, anti-HK2, anti-Ecadherin, anti-Ki67 (Abcam, Cambridge, MA); anti-pAMPK (Thr 172), anti-GAPDH (ImmunoWay, New York, DE, USA); anti-LDHA, anti-Lin28, anti-β-actin and anti-Histone H3 (Proteintech Group, Chicago, IL, USA); anti-caspase 3, anti-bax, anti-Ncadherin (Abclonal, Cambridge, MA, USA); anti-8-OHdG (Japan Institute for the Control of Aging, Shizuoka, Japan). β-actin or Histone H3 was used as a loading control.
Luciferase assays
Cells were transfected with the pGL3-MACC1-3′UTR vector and the renilla luciferase plasmid (pRL-TK). Then, the cells were harvested after 36 h for firefly/renilla luciferase assays using the Dual-Luciferase Reporter Assay System (Promega). Luciferase activities were normalized to the pRL-TK plasmid.
FISH and Immunofluorescence analysis
To detect MACC1-AS1 expression in cells, Cy3-labeled probes was synthesized for fluorescence in situ hybridization (FISH). Cells were rinsed in PBS and fixed in paraformaldehyde (Solarbio, China) for 15 min at room temperature. Next, cells were permeabilized in 0.5% Triton X-100 on ice for 10 min, and washed in PBS for 3 times. Hybridization was carried out using Fluorescent in Situ Hybridization Kit (RIBO Bio, Guangzhou, PR China) in a moist dark chamber at 37 °C for 12 h according to the protocol. After serial saline-sodium citrate (SSC) buffer washing, MACC1-AS1 was detected under confocal laser microscope (Olympus Optical, Tokyo, Japan). For co-localization study, cells were fixed again in paraformaldehyde and subjected to immunofluorescence, 4′, 6-diamidino-2-phenylindole (DAPI) was used to stain the nucleus.
Cell nuclear and cytoplasmic RNA and protein isolation
For cytoplasmic fraction, cells were washed with ice-cold PBS for three times, then the PARIS Kit (Life Technologies, Carlsbad, CA, USA) was applied. Briefly, cells were lysed by cell fractionation buffer and incubated on ice for 10 min before centrifuging at 4 °C, 500 g for 5 min. The supernatant was collected as cytoplasmic fraction while the remaining nuclear pellets were lysed again by cell disruption buffer as nuclear fraction. Both the cytoplasmic and nuclear fraction were proceeded to isolate the RNA and protein by following binding and washing procedure according to the protocol. The nuclear and cytoplasmic RNA and protein was further analyzed by western blotting or qPCR.
2-NBDG uptake assay
After 12 h culture in conditions of either normal or glucose-free medium, recipient cells were labelled with 100 μM 2-NBDG diluted in glucose-free medium, and incubated for 30 min at 37 °C. Fluorescence quantification was carried out using flow cytometry at emission 465/excitation 540 nm.
MACC1-AS1-overexpressing or vector-transfected MKN45 cells were seeded on 6-well plates at a density of 2 × 105 cells per well. 48 h later, the culture medium was changed to glucose-free medium and cultured for another 12 h. Cell supernatant was collected to measure lactate concentration (Nanjing Jiancheng Bioengineering Institute, Nanjing, China), while the cell pellets to be lysed and measured ATP (Beyotime, Haimen, China) and the activity of HK and LDH (Comin, Suzhou, China). ROS was measured by fluorescent 2′, 7′-dichlorofluorescin diacetate (DCF-DA) as described by manufacturer’s protocol of commercial kit (Nanjing Jiancheng, China).
NADP+/NADPH and GSH/GSSG assays
After 12 h of glucose deprivation, the recipient cells were collected and analyzed using NADP+/NADPH quantification kit (Biovision K347, CA, USA) and GSH/GSSG quantification kit (Biovision, USA) according to the manufacturer’s instructions.
Cell apoptosis, viability and cell cycle analysis
Cell apoptosis was measured with Annexin/PI kit (Key Gen Bio TECH, Nanjing, China) by flow cytometry analysis (BD Biosciences, San Jose, CA, USA). Cell cycle was measured with cell cycle detection kit (Key Gen, China). Cell viability was measured by MTT assays and EdU assays (RIBO, China) according to the protocol.
The MACC1-AS1-overexpressing or vector-transfected MKN45 cells were seeded in 6-well plates at 1× 103 per well and cultured for 2 days. The medium was changed to culture in low concentration of glucose medium for another 2 weeks. After that, the cells were washed twice with PBS, fixed with paraformaldehyde and stained with 0.5% crystal violet. After washing again, the plates were photographed under a microscope.
Cell migration/ invasion assay
For transwell migration assays, The MACC1-AS1-overexpressing or vector-transfected MKN45 cells were plated in the top chamber with the non-coated membrane (BD Biosciences, San Jose, CA, USA). For invasion assays, matrigel (BD biosciences) was polymerized in transwell inserts for 2 h at 37 °C. Transfected cells were plated in the top chamber in medium without serum. In both assays, the lower chamber filled with 10% fetal bovine serum. Cells were incubated for 24 h and the cells that did not migrate or invade through the pores were removed by a cotton swab. Cells on the lower surface of the membrane were stained with crystal violet and photographed.
Ten 4-week-old BALB/c nude mice were injected with MACC1-AS1-overexpressing or vector-transfected MKN45 cells in each group. Briefly, 5*105 cells were injected intravenously through the tail vein. Six weeks after injection, the mice were sacrificed and examined by H&E and IHC staining.
The expression of MACC1-AS1 was profiled based on normalized RNA-seq expression data sets of the stomach adenocarcinoma (STAD) cohort study from the Broad TCGA GDAC website (
http://gdac.broadinstitute.org/). Up until March 2017, STAD had collected data on 32 normal and 375 cancerous tissues. We defined significant and differential expression of a gene, between tumors and normal samples, based on two conditions: false discovery rate adjusted
P-value < 0.01 and fold change > 2 [
8]. The Bioconductor package edgeR was used to analyze
P values and fold-changes using an R statistical environment (version 3.3.2).
Statistical analysis
Statistical analysis was performed using SPSS version 20 software (SPSS, Chicago, IL, USA) or R software (version 3.3.2). Differences between experimental groups were assessed by Student’s t-test or one-way analysis of variance. Chi-squared test and Mann-Whitney test were applied where appropriate to analyse the association between MACC1-AS1 expression and clinicopathological parameters. Spearman correlation were used to analyse the R2 between MACC1-AS1 and MACC1 for the serial staining. The log-rank test was performed to compare the survival curves of individual groups. Cox regression was used for univariate and multivariate analysis. The reported results included hazard ratios (HR) and 95% confidence intervals (CI). All values were expressed as mean ± SD, and statistical significance was noted as P < 0.05.
Discussion
We previously reported that MACC1 is overexpressed in GC and associated with GC metastasis, chemoresistance, and poor prognosis [
15,
16]. Our present study identifies that MACC1-AS1 is highly elevated and exhibits a robust correlation with MACC1 expression in GC. In addition, high expression of MACC1-AS1 was associated with poor prognosis, suggesting it could be a biomarker to identify patients at a higher risk of GC progression. Mechanistically, MACC1-AS1 regulates MACC1 expression and promotes metabolic plasticity by enhancing glycolysis and antioxidant capacity, conferring a proliferation advantage under metabolic stress. Specifically, MACC1-AS1 was shown to promote MACC1 mRNA stability through physical binding to MACC1 mRNA, and that was shown to be concomitantly coordinated by AMPK activation and subsequent Lin28 translocation (Fig.
6l).
MACC1 functions as a key regulator of the HGF/c-MET axis, which is a promising target for cancer therapy [
17]. However, the results of clinical trials targeting c-MET have failed to achieve better prognosis for GC patients [
18]. The reason for this is unclear, and understanding a c-MET-associated regulator might help us to better explore more effective ways to target HGF/c-MET. In our study, the MACC1-AS1/MACC1 axis functioned as a modulator of metabolic plasticity by promoting glycolysis and antioxidant production. The HGF/c-MET has been reported to drive metabolic adaptation to prevent glucose deprivation-induced apoptosis [
19]. In addition, inhibition of this pathway perturbs redox homeostasis by downregulating NADPH production [
20]; thus we suggest that inhibition of MACC1-AS1/MACC1 might result in synergetic anti-tumor effects, when combined with the clinical application of HGF/c-MET inhibitors, by targeting metabolic plasticity.
Extensive research has indicated that metabolic plasticity is essential for cancer cells to survive hostile nutrient-deprived environments [
21‐
23]. Cancer cells might select for highly metabolically plastic cells to survive. A recent study indicated that hypoglycemic tumor microenvironment increases resistance to cytotoxic agents in GC, especially in tumor cells with an enhanced glycolysis phenotype [
24]; glucose deprivation was also shown to drive the acquisition of KRAS mutations in colorectal cancer, which mediates tumor plasticity in multiple ways [
25‐
27]. Glucose deprivation could induce a mesenchymal phenotype through metabolic plasticity to overcome apoptosis during metastatic colonization and the acquisition of chemotherapy resistance [
10,
28‐
30]. Therefore, targeting metabolic plasticity could affect tumor growth, metastasis, and treatment efficacy [
2,
31]. Our current study indicated that MACC1-AS1 is a stress-responsive gene, as it was shown to promote metabolic adaptation by upregulating GLUT1, HK2, and LDHA expression. Consistent with the fact that MACC1-AS1 predicts poor survival, the stress-responsive genes implicated in our study, including AMPK, Lin28, MACC1, and GLUT1, also predict poor outcomes in GC (Additional file
9: Figure S5A–D). These results indicate that such genes might promote plasticity by modulating tumor metabolism, highlighting the association between MACC1-AS1 and metabolic plasticity in GC and suggesting a new strategy for GC treatment.
The mRNA decay and stability response under metabolic stress is essential for mediating tumor plasticity. Compared to the lag time required to produce proteins de novo, modulating mRNA stability is a fast-acting strategy that cancer cells could utilize to for rapid adaptation and maximum cell survival [
32]. AMPK is important for this process through its ability to regulate mRNA translation, elongation, and stability [
33]. It was reported that AMPK promotes the mRNA stability of VEGF and COX2 by modulating RBPs [
34,
35]. Lin28 is also an evolutionarily-conserved RBP. Although it is particularly known to regulate the Lin28/let-7 oncogenic pathway, its role in mRNA stability cannot be ignored. Lin28 can physically bind bulk mRNAs, many of which have been implicated in the regulation of metabolic processes, such as IGF2, IMP, and OCT4. It is also a component of stress granules where mRNAs are sequestered in RNA-protein complexes to adapt to conditions of stress [
36,
37]. However, whether AMPK is involved in the function of Lin28 has not been reported. In our study, MACC1-AS1 promoted AMPK activation and subsequent Lin28 translocation, which indicated that AMPK requires Lin28 subcellular trafficking to exert its roles in the maintenance of RNA stability. In addition, our present study demonstrated that MACC1-AS1 can physically bind MACC1 mRNA, forming RNA-RNA complexes, to stabilize MACC1 mRNA [
38]. However, whether MACC1-AS1 associates with Lin28 and mediates its recruitment to MACC1 mRNA requires further investigation.