Background
Gastric cancer (GC for short), also recognized as stomach adenocarcinoma, is the fourth frequent cancer and the third leading cause of cancer-related deaths worldwide according to the GLOBOCAN database [
1].Even with the development of the earlier diagnosis techniques and new therapeutic approaches, the 5-year overall survival remains poor, and less than 30% in most countries. GC is characterized as a highly heterogeneous solid tumor and featured with high autophagy, metabolic disturbance and hypoxia [
2,
3]. Moreover, highly proliferative GC cells have to change their metabolism type to adjust to the severe micro-environment, eventually satisfying their needs for survival and metastasis [
4].Therefore, it is an urgent need for us to understand the deeper molecular mechanism responsible for the GC progression and find new therapeutic targets or reliable diagnosis markers.
Autophagy plays a crucial role in sustaining the homeostasis of cells and various malignant diseases. For example, cancers may result from autophagy disfunction [
5]. Autophagy has been reported to promote tumor development in multiple cancers. In addition to the role of autophagy on promoting resistance to chemotherapy [
6], it is also known that autophagy may activate in hypoxic and energy-lacking conditions, which eases the stress on metabolism. By means of degrading intracellular particles including amino acids, nucleotides, fatty acids, sugars and aged organisms, autophagy can recycle the aforementioned particles into bioenergetics and biosynthesis pathways, finally enabling the survival of tumor cells [
7]. Even with the sufficient energy, many tumor types demonstrate elevated autonomous basal autophagy. It was considered that the tumor metabolism mostly relied on glycolysis, also called Warburg Effect [
8,
9], to meet its need for malignant behaviors. However, recent reports found that in certain tumor types, glycolysis itself could not entirely enable the survival of tumor cells under starvation, and glutaminolysis could fully rescue the survival of tumor cells instead. The pyruvate produced from the glycolysis has been finally transformed into lactate which does not directly replenish the tricarboxylic acid (TCA) cycle [
10,
11]. Glutaminolysis enables glutamine transformed into glutamate by glutaminase (GLS), and then further being transformed into alpha-ketoglutarate (α-KG) by glutamate dehydrogenase(GDH) to produce adequate ATP and provide nitrogen and carbon to replenish the TCA cycle. Additionally, the glutaminolysis can help produce biosynthetic precursors which is necessary for cell growth and produce glutathione(GSH) to keep the ROS at a low level suitable for tumor cell proliferation [
12‐
14]. Autophagy has been reported to mediate several types of metabolism to provide substrates for tumor survival including glycolysis and glutaminolysis [
15‐
17]. However, the specific metabolic pathway and the regulation mechanism of glutaminolysis mediated by autophagy have not been explored in depth. Besides, there were reports proving that tumor metastasis mediated by epithelial–mesenchymal transition needs autophagy to sustain the viability of cancer cells [
18‐
20]. Furthermore,PI3K/AKT/mTOR axis is publicly recognized to mediate the autophagy activation directly [
21,
22] and promote EMT progression [
23,
24] as well. However, the specific interaction between autophagy and EMT on GC metastasis remains to be explored in depth.
Autophagy-related genes (ATGs) are involved in multiple stages of autophagy which regulate the autophagosome formation [
25]. Moreover, microRNAs (miRNAs) have been reported to regulate different ATGs to influence the autophagy [
26‐
28].MicroRNAs (miRNAs) is a kind of noncoding RNA that functions mostly at the post-transcriptional regulation level and further regulates the targeting gene expression. They bind the 3′ untranslated region (3′-UTR) of their target mRNAs to work. Many studies have reported miRNAs play important roles in regulating carcinogenesis [
29‐
31]. Some of them can affect proliferation, migration and invasion of tumor cells which make miRNAs themselves closely associated with clinical features including tumor stage, lymphatic invasion and patient survival. More and more studies discovered the diagnosis and prognosis value of certain miRNAs. Among these miRNAs, miR-133a-3p is reported to be a tumor suppressor in numerous cancer types including prostate cancer [
32], gastric cancer [
33], and bladder cancer [
34] through multiple signaling pathways including PI3K/AKT/mTOR axis. However, the deeper regulation mechanism of miR-133a-3p in GC remains largely unknown. Recently, many people have reported miRNA could affect the activation of autophagy in stressful conditions such as starvation [
35,
36].Therefore, we want to verify whether miR-133a-3p could break metabolic balance and EMT progression in GC by means of affecting the activation of autophagy, and ultimately limit GC cell proliferation, migration and invasion. We successfully found that GC cells use autophagy to recycle glutamine for glutaminolysis, and eventually ensure the survival of themselves and affect the EMT progression to promote metastasis. To summarize, miR-133a-3p acts as a crucial blocker on autophagy-derived pools of glutamine to inhibit GC development.
Methods
Tissue specimens
The tumor tissues and adjacent normal stomach mucosa tissues were collected from the GC patients who received radical gastrectomy at Department of Gastrointestinal Surgery, the First Affiliated Hospital of Nanjing Medical University from 2013 to 2017. All patients did not receive radiotherapy and chemotherapy before surgery. All the specimens were taken under the guidance of HIPAA protocol [
37] and proved by the ethics committee. The TNM stage classification was followed TNM classification system of the International Union Against Cancer (7th edition) [
38]. The disease free survival (DFS) is defined as the duration between the recurrence of GC and the primary GC. The overall survival(OS) is defined as the duration between the death of GC patients and the primary GC. We calculated the grouping of ROC curve according to the cut-off value of miR-133a-3p relative expression and the middle expression of miR-133a-3p.
Cell culture and lentivirus transfection
The human GC cell lines BGC-823,SGC-7901,MGC-803,MKN-45,HGC-27,AGS and normal stomach mucosa epithelium cells GES-1 were all purchased from the Cell Center of Shanghai Institutes for Biological Sciences. AGS cells were cultured in F12K medium with the supplementation of 10% fetal bovine serum (Invitrogen), while the GC cells were all cultured in RPMI-1640 medium with the supplementation of 10% fetal bovine serum (Invitrogen). Moreover Antibiotics(1% streptomycin/penicillin, Gibco, USA) were added in the aforementioned culture system which were set at 37 °C in a moisture air under 5% CO2. We seeded 2 × 105 BGC-823 and MKN-45 cells in 6-well plate overnight and then washed these cells with PBS for three times and these cells were finally both received overexpression lentivirus(LV-miR-133a-3p) and interference lentivirus (LV-miR-133a-3p IN) (MOI, 100 for BGC-823; 20 for MKN-45) by the supplementation of Polybrene (8 mg/ml). Subsequently we used RT-PCR to verify the transfection efficiency. The sequence we used for miR-133a-3p overexpression (5’ UUUGGUCCCCUUCAACCAGCUG 3′) and the sequence we used for miR-133a-3p-IN (5’ CAGCUGGUUGAAGGGGACCAAA 3′) were cloned into LV12-U6/luci05/puro vector (GenePharma, Shanghai, China). The LV12 empty lentiviral construct(LV-NC) served as a negative control and was taken to transfect BGC-823 and MKN-45 GC cells in the same approach. We successfully kept the stable clones of miR-133a-3p in GC cell lines by adding puromycin up to 5 μg/ml gradually (Sigma, Aldrich) until 5 days. Genepharma Biotech (Shanghai, China) constructed shRNA vectors targeting GLS (GLS siRNA).
Reagents and antibodies
Hydroxychloroquine (HCQ), BPTES(H0915, SML0601) were purchased from Sigma-Aldrich (Shanghai, China). Regarding the primary antibodies used in the study, anti-GLS(ab93434), anti-ULK1(ab128859),anti-ATG5(ab108327), anti-GABARAPL1(ab86497), and anti-ATG13(ab214074) were purchased from Abcam (Cambridge, UK); Anti-GDH was purchased from Shybio (Shanghai, China); Anti-LC3A/B (12741), anti-Beclin-1 (3495), anti-SQSTM/p62(8025), anti-ATG7(2631),anti-mTOR (2983), anti-phospho-mTOR(5536,1230), anti-AKT(4685), anti-phospho-AKT(4060), anti-E-Cadherin (3195), anti-N-Cadherin(13116), anti-Slug(9585), anti-Snail(3879) and anti-TWSIT1(46072) were purchased from Cell Signaling Technology (Danvers, PA USA).
Clonogenic assay and clonogenic survival assay
For clonogenic assay,we seeded LV-miR-133a-3p, LV-miR-133a-3p-IN and LV-miR-NC GC cells in 6-well plates (500 cells per well) for 2 weeks. For clonogenic survival assay, we seeded aforementioned cells in 6-well plates and cultured them with normal growth medium PRMI1640 with 10% fetal bovine serum. Subsequently we replaced the original culture medium with FBS-free RPMI1640 when GC cells were at 80% confluence in the well for 2 days. Furthermore, all the aforementioned cells in plates were fixed in 2 ml of methanol for 30 mins and stained with crystal violet for 20 mins.
Quantitative real-time reverse transcription polymerase chain reaction(qRT-PCR)
Trizol Reagent(Invitrogen,15,596,018) were used to extract RNA from GC cell lines and tumor tissues. We used PrimeScript RT Master Mix Kit (TaKaRa, RR036A, Japan) to reverse-transcribe the mRNA and New Poly(A) Tailing Kit (ThermoFisher Scientific, China) to reverse-transcribed the miRNA both into cDNA. Subsequently we used Universal SYBR Green Master Mix (4,913,914,001, Roche, Shanghai, China) to perform polymerase chain reactions in order to amplify the cDNA to the same quantity level. We used 2-△△CT analysis method to calculate the Ct value during the exponential amplification phase. Meanwhile, we used β-actin and RNU6–1 as the internal references for mRNA and miRNA respectively.
Protein extraction and western blot assay
All proteins within cells and tissues were extracted by the iced lysis buffer(Cell Signaling Technology, Danvers, Massachusetts, U.S.) under the supplementation of a mixture of protease inhibitor(Calbiochem, Darmstadt, Germany). Subsequently, we used BCA assay Kit (Pierce, Rockford, IL) to determine the protein concentration of different samples and used the lysis buffer to adjust all the samples to the same concentration. Furthermore, equal quantity of protein extracts were added with 10% SDS–PAGE and were then boiled until denaturation.
For the western blot, we transferred the aforementioned protein extracts into nitrocellulose membranes. Subsequently we blocked the membranes with 0.05% Tween in TBST buffer, and then we used the specific primary antibody to incubate at 4 °C overnight. The membranes were washed in TBST for 10 mins for 3 times at room temperature on the next day. Furthermore, we used the second antibody targeting the primary antibody for 2 h at room temperature and also washed with the aforementioned protocol. Finally, we used chemiluminescence detection system to determine the bands.
Transmission electron microscopy (TEM)
We fixed the MKN-45 and BGC-823 GC cells with 2.5% glutaraldehyde at 4 °C overnight. Subsequently, we used 1% OsO4 to continue to fix the cell samples. In the following step, we used ethanol and propylene oxide to dehydrate the cell samples. We cut samples into slide sections and stained them with 0.3% lead citrate before we finally took a JEM-1010 electron microscope (JEOL, Tokyo, Japan, 2500x or 8800x magnification) to detect the autophagosomes,
ROS detection with DHE staining
The cells were washed with PBS buffer twice and then were stained with DHE (Sigma, D7008) (20 μmol/L) at 37 °C for 30 mins. Subsequently, we took fluorescence microscope to observe the red fluorescence to determine the ROS level.
Dual-luciferase reporter assay
We cloned into pMIR-REPORT plasmid (H306, Obio Technology, Shanghai, China) with 3’UTRs containing wild-type or mutant miR-133a-3p response elements from GABARAPL1 and ATG13. We used Lipofectamine 3000 to transfect the reporter plasmid into GC cells. For the last step, we used DualLuciferase Reporter System Kit (E1910, Promega, USA) to indicate the Firefly and Renilla luciferase activity.
5-Ethynyl-2′-deoxyuridine (EdU) assay
We used EdU assay kit (RiboBio, China) to detect the DNA synthesis and cell proliferation. We seeded 10,000 treated GC cells into a 96-well plate overnignt. On the next day, we added Edu solution (25 μM) into the 96-well plates for 24 h.Subseuently we used 4% formalin to fix the GC cells at room temperature for 2 h. In the following step, we used 0.5% TritonX-100 to permeabilize the GC cells for 10 mins. We then added 1× Apollo reaction solution(200 μl) to stain the EdU for 30 mins and Hoechest33342 (200 μl) to stain the nuclei. Finally, we used Nikon microscope (Nikon, Japan) to observe the DNA synthesis and cell proliferation reflected by red and blue signals.
The detection of oxygen consumption rate (OCR)
We used Seahorse Biosciences metabolism analyzer (XF24) to detect the oxygen consumption rate of GC cells. We seeded 2.5 × 104 GC cells per well in the 24-well plates and we submitted them to metabolism assay after 24 h. Real-time OCR was measured every 15 mins for 3 h. We normalized the relative oxygen consumption rate to the 0-min time point.
LC-MS detection for glutaminolysis products
We washed treated GC cells in the plates with PBS twice to wipe off the dead cells. In the next step, we add cell samples in 2 ml premixed 80% methanol solution under − 80 °C. Subsequently, we used 0.5 ml additional 80% methanol solution to re-extract the insoluble particles for 5 min on dry ice. Continuously, we combined supernatants which collected at each extraction to dry under N2 and reconstituted them in specific LC-MS testing water. After this task, we took reversed-phase ion-pairing chromatography coupled by negative mode electrospray ionization to a stand-alone Orbitrap mass spectrometer (Thermo Scientific) to detect the amino acids within the samples. It used solvent A (97:3 H2O/MeOH with 10 mM tributylamine, 15 mM acetic acid) and solvent B (100% MeOH) to detect from m/z 85–1000 at 1 Hz at 100,000 resolution with LC separation on a Synergy Hydro-RP column (100 mm × 2 mm, 2.5-μm particle size; Phenomenex). The concentration of several amino acids were finally determined with Maven software.
CCK-8 assay
We used the Cell Counting Kit-8 assay (Dojindo Laboratories, Kumamoto, Japan) to determine the proliferation rate of GC cells. Firstly, we seeded 1000 cells into 96-well plates to observe for 7 days. Each day we added 10 ml CCK-8 solution to a new well at 37 °C for 2 h. Following this,we took automatic microplate reader (BioTek, Winooski, VT, USA) to measure at the absorbance at 450 nm to compare numbers on each day.
Cell migration and invasion assay
The methods used for migration assay and invasion assay are almost the same. We put transwell assay inserts (Millipore, Billerica, MA, USA) in a well of 24-well plates. However,the difference is the membrane in the upper chamber of transwell assay in invasion assay is Matrigel-coated membrane(BD Biosciences) while the one in migration assay is only a normal membrane. In the experiment, firstly, we set 500ul serum-free RPMI1640 at bottom chamber. Following this, we seeded 10,000 cells in 200ul RPMI1640 with 10% FBS at upper chamber. After 24 h or 48 h, we used methanol to fix the cells within the membrane and stain them with crystal violet. Our final task is to observe these cells by microscope.
Human gastric organoid model
We followed the previously reported method [
39] to construct human gastric organoid model and investigate the regulation role of miR-133a-3p in a micro-environment similar to human condition. We transfected the organoids with LV-miR-133a-3p-IN,LV-miR-133a-3p and LV-NC, and used microscope to observe the growth of organoids on each day. We harvested the GC organoids from the Matrigel to be fixed in 70% ethanol and wrapped into paraffin sections to prepare for immunofluorescence staining.
Immunofluorescence analysis of GC cells
We seeded 2000 GC cells in specially designed confocal dish cultured with RPMI1640 with 10% fetal bovine serum for one night. We used PBS to wash the plates twice and then fixed the GC cell with 4% formaldehyde solution for 30 mins at room temperature. Following this, we used PBS to wash plates twice again before we used 0.2% Triton X-100 to make GC cells more permeable.1% BSA was used to block the non-specific binding sites before we used primary antibody from different species to bind the specific location within GC cells or human gastric organoid at 4 °C. At last, we used second antibody Cy™ 3-Goat Anti-Rabbit IgG (Jackson, 1:100) to bind the primary antibody for 2 h and DAPI to stain nucleus for 15mins. At last, we used confocal microscope(Nikon, Japan) to observe the fluorescence signal.
Immunohistochemical (IHC) analysis of tissue samples
We used 10% formalin to fix GC tissues and xenografted tissues and embedded them with paraffin. After the embedment, we cut GC tissues into slides and used primary body to specifically bind the respective targets overnight at 4 °C. On the next day, we used PBS to wash them twice and used HRP-Polymer-conjugated secondary antibody (Abcam, UK) to incubate tissues at room temperature for 1 h. Subsequently, we used 3,3-diaminobenzidine solution and hematoxylin to stain the samples. Lastly, we used microscope to observe the positive tumors and intensity of stained cells.
Hematoxylin and eosin staining of tissue
Firstly, we used a microscope slide to rehydrate the tissue samples fixed in alcohol. Subsequently we agitated the slides for 30s in deionized water to hydrate the tissues. Subsequently, we put the slides into a bottle filled with hematoxylin and agitate for 30s and washed them in deionized water for 30s. After the previous steps, we used 1% eosin Y solution to stain the slide and rehydrated the sample with 95% alcohol and 100% alcohol. We took xylene to extract the alcohol continuously. In the last step, we covered the slides and used microscope to observe.
Patient derived xenograft model(PDX model)
Firstly, we kept the tissues in iced RPMI1640 with 10% fetal bovine serum and cut them into 2*2*3 Mm3 and then used the fresh RPMI1640 to wash tissues twice. Before we finished residure procedures, we kept tissues in PRMI1640 with the supplementation with penicillin and streptomycin.NOD/SCID mice was chosen to be the first generation PDX mice that carried the patients’ tissues. We used 10% chloral hydrate(0.004 ml/g) to make mice into anesthetized condition. By sterile operation, we buried tumor tissues on the mice back subcutaneously with the supplementation of penicillin and streptomycin at the same time. The rest PDX generation mice were BLAB/c-nude mice. When each xenografted tumor tissue grew up to 1–2 cm3,we followed the same protocols to harvest the tissues and transplanted them into the next generation mice immediately for four times until the 4th BLAB/c-nude. We injected the recombinant lentivirus vectors or drugs into tumor tissues continuously from day 0 to day 20, and then harvested the tumor tissues to received further analysis at day 40.
Firefly luciferase gene was stably transduced into miR-133a-3p lentivirus vectors. We injected the vectors through tail vein of BALB/c nude mice. At the last step, we observed the bioluminescent signal after injecting 100 mg/kg of D-luciferin (Xenogen, Hopkinton, MA) into mice with an IVIS 100 Imaging System (Xenogen).
Discussion
Glutaminolysis has now been recognized as an important hallmark of tumor metabolism. Researchers have focused their studies on glutamine rather than glucose in the terms of tumor cell metabolism over the past few years. Biosynthesis, bioenergetics and ROS homeostasis are three major ways that glutaminolysis takes to promote tumor cell proliferation, migration and invasion. Glutaminolysis could replenish the TCA cycle by providing α-KG to produce both adequate ATP and substrates for biosynthesis. In addition, glutaminolysis also helps synthesis of nucleotides by donating nitrogen to purines and pyrimidines and synthesis of certain nonessential amino acids. In this case, glutamine acts as a carbon and nitrogen donor which is necessary for survival of tumor cells. Concerning the tumor promotor role of glutaminolysis, FDA approved many kinds of glutaminolysis inhibitors to go into clinical trials [
54‐
56]. Apart from the metabolism alterations in tumor cells, it is also discovered that tumor can be usually identified as elevated autophagy to buffer the nutrients limitations. Many researchers reported the higher autophagy level within GC tumor were correlated with the shorter survival of GC patient. According to our investigation, the basal autophagy level of gastric cancer cells is higher than normal stomach epithelial cells. In the past few years, connections between autophagy and metabolism have been discovered, but the deeper regulation mechanisms of them on GC development have not been understood comprehensively. According to our study, autophagy involved in the glutaminolysis provides necessary glutamine metabolism products which eventually rescue GC cell survival. Inhibition of autophagy almost phenocopied the effect of directly blocking glutaminolysis on suppressing tumor deveolopment. Additionally, it was reported that the positive correlation between autophagy and EMT mediated metastasis. Considering the promotion effect of autophagy on glutaminolysis and EMT, we tried to found a method targeting autophagy to block GC growth and metastasis.
MiRNA is a kind of noncoding RNA that can post-transcriptionally regulate its targeting gene expression. Many researchers have reported that miR-133a-3p could function in prostate cancer, gastric cancer and bladder cancer as a tumor suppressor. As several researchers reported that miR-133a-3p is crucial in a series pathological and biological processes by regulating different genes, such as Her-2, EGFR, Snai1, and Evi1.However, the investigation about the exact role of miR-133a-3p on regulating autophagy, glutaminolysis and EMT have not been clarified. As we observed, miR-133a-3p blocked the activation of autophagy and glutaminolysis in GC cells, which prompts our further exploration on the deeper interaction between autophagy and glutaminolysis under the mediation of miR-133a-3p.The advancement of our reports is the utilization of certain models to investigate the biological role of miR-133a-3p.We combined organoid model and PDX model to verify the tumor suppressor role of miR-133a-3p both in vitro and in vivo. Organoid model consists of several types of GC cells that grow with three dimensional structures and phenocopies the micro-environment of primary GC tissues. PDX model is also retained the vivo micro environment and can more easily forecast the therapeutic efficiency. Therefore, these two models reduce the possibility of GC cells from intentionally or unintentionally population selecting. We used organoid model and PDX model to verify the inhibition role of miR-133a-3p role on autophagy mediated glutaminolysis. Additionally, we determined the suppress role of miR-133a-3p on EMT in GC cells and detected the distant organ metastasis via scan on BALB/c nude mice which were injected luciferase-labeled lentivirus.
Over the past few years, researchers have reported many miRNAs could interact the glutaminolysis including miR-204,miR-137,miR-192 and miR-23a/b [
57‐
59]. According to our study, it is found that miR-133a-3p could block the activation of glutaminolysis by reducing the expression level of core enzymes including GLS and GDH, but not directly targeting them at the posttranscriptional level. Besides, we found that miR-133a-3p could mediate autophagy in GC cells by directly regulating ATGs, including GABARAPL1 and ATG13.It is reported that GABARAP subfamily proteins can, more efficiently than LC3–2, serve as scaffolding proteins by recruiting ULK1 and beclin-1 (complex) to the site of autophagosome nucleation and directly participant autophagy. ATG13 is an autophagy initiator which are contained in ATG13-ULK1-RB1CC1 complex. Its regulation on ULK1 activity plays an important role in mediating the kinase activity of mTORC1 and cell proliferation. We found that LV-miR-133a-3p GC cells could block the expression of these two targets at both mRNA and protein level further demonstrating that miR-133a-3p is an autophagy regulator.