Loss of RACK1 promotes glutamine addiction via activating AKT/mTOR/ASCT2 axis to facilitate tumor growth in gastric cancer

Abstract

Background

Metabolic reprogramming is closely related to the development of gastric cancer (GC), which remains as the fourth leading cause of cancer-related death worldwide. As a tumor suppressor for GC, whether receptor for activated C-kinase 1 (RACK1) play a modulatory role in metabolic reprogramming remains largely unclear.

Methods

GC cell lines and cell-derived xenograft mouse model were used to identify the biological function of RACK1. Flow cytometry and Seahorse assays were applied to examine cell cycle and oxygen consumption rate (OCR), respectively. Western blot, real-time PCR and autophagy double fluorescent assays were utilized to explore the signaling. Immunohistochemistry was performed to detect the expression of RACK1 and other indicators in tissue sections.

Results

Loss of RACK1 facilitated the viability, colony formation, cell cycle progression and OCR of GC cells in a glutamine-dependent manner. Further investigation revealed that RACK1 knockdown inhibited the lysosomal degradation of Alanine-serine-cysteine amino acid transporter 2 (ASCT2). Mechanistically, depletion of RACK1 remarkably decreased PTEN expression through up-regulating miR-146b-5p, leading to the activation of AKT/mTOR signaling pathway which dampened autophagy flux subsequently. Moreover, knockdown of ASCT2 could reverse the promotive effect of RACK1 depletion on GC tumor growth both in vitro and in vivo. Tissue microarray confirmed that RACK1 was negatively correlated with the expression of ASCT2 and p62, as well as the phosphorylation of mTOR.

Conclusion

Together, our results demonstrate that the suppressive function of RACK1 in GC is associated with ASCT2-mediated glutamine metabolism, and imply that targeting RACK1/ASCT2 axis provides potential strategies for GC treatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13402-023-00854-1.

Introduction

Gastric cancer (GC) is an aggressive disease which continues to a daunting impact on global health [1]. According to the latest global cancer statistics data for 2020, GC remains the fifth most common type of cancer and the fourth leading cause of cancer-related death worldwide [2]. Although the survival rates for patients with GC have significantly boosted over the past few decades, the prognosis was still unsatisfactory due to its usually diagnosed at a late stage and high recurrence rates [3]. A large number of research have indicated that early diagnosis combined with advanced surgical treatment could effectively improve GC patients’ prognosis and survival rate [4]. This highlights the urgent need to unearth more potential molecular-driven therapeutic strategies, and in fact, various small molecule targeting drugs have been translated into clinical treatment [5].

The receptor for activated C-kinase 1 (RACK1) is a member of the Trp-Asp repeating protein family that contains a seven-bladed propeller structure [6]. As a classical scaffold protein for multifarious kinases and receptors, RACK1 plays a key role in a wide range of biological responses including cell growth, migration, differentiation and immune response [7]. Dysregulation of RACK1 has been demonstrated in several kinds of cancers, and RACK1 was considered as a oncogenic factor with upregulation in multiple malignant tumors such as breast cancer [8], non-small cell lung cancer (NSCLC) [9], pulmonary adenocarcinoma [10] and hepatocellular carcinoma (HCC) [11, 12]. On the contrary, RACK1 was reported to be a suppressor with downregulation in GC [13], and we have also certified that loss of RACK1 stimulated the invasion and metastasis of GC cells [14]. As chronic infection with Helicobacter pylori was identified as the strongest risk factor for GC and the regulation of diverse bacteria on RACK1 has been reported [15, 16]. Thus, the special characters including expression and function of RACK1 in GC might be modulated by Helicobacter pylori and/or gastric acid environment. Whereas the mechanisms by which RACK1 modulates cancer progression may be multifaceted, and in particularly, the role of RACK1 in GC need to be further investigated.

Interest in the relationship between signal transduction and metabolism has been aroused recently. High-throughput uptake of glucose and glutamine provides carbon and nitrogen for the biosynthesis of non-essential amino acids in growing cells [17]. Hence, metabolic reprogramming is recognized to be closely associated with carcinogenesis. And one of the major features of metabolic reprogramming was excessive glucose and amino acid uptake [18]. Increasing evidence for an alternative concept were provided over the past several years, that activation of oncogenes or inactivation of tumor suppressors are continued to reprogram cellular metabolism [19]. Efforts have been undertaken to reveal the mechanism by which oncogenes modulate metabolic pathways that favor cancer cell growth [20]. Nevertheless, as a tumor suppressor, it has not been reported whether RACK1 play a role in modulating glucose or glutamine metabolism in GC cells. In the present study, we found that depletion of RACK1 could improve Alanine-serine-cysteine amino acid transporter 2 (ASCT2) stability by activating AKT/mTOR signaling and inhibiting downstream autophagy, thus driving glutamine addiction in GC cells. And we identified the RACK1/ASCT2 axis as a potential target for GC treatment.

Materials and methods

Patients and specimens

For tissue microarray detection, tumor specimens of 120 GC patients were obtained from the Department of General Surgery, Zhongshan Hospital (Fudan University, Shanghai, China). The diagnosis of gastric carcinoma was confirmed by pathological examination. All the patients' demographic characteristics including date of surgery, survival time and other relevant data were extracted from the hospital records. This study was approved by the Ethics Committee of Zhongshan Hospital, Fudan University (B2021–711). Written informed consent was acquired from all patients.

Cell culture and reagents

The human GC cell lines AGS, MGC80-3 and HGC-27 were provided by Cell Bank of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). MKN-45 and SGC-7901 were obtained from FuDan IBS Cell Center (Shanghai, China). The human gastric mucosal epithelial cell line GES-1 was purchased from iCell Bioscience Inc (Shanghai, China). These cells were routinely cultured in RPMI-1640 or Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum (FBS) (Gibco, Waltham, MA, USA) at 5% CO₂ and 37 ℃. Reagents including MG-132 (MedchemExpress, Monmouth Junction, NJ, USA), Cycloheximide (CHX, MedchemExpress), Chloroquine (CQ, Sigma-Aldrich, St Louis, MO, USA), 3-Methyladenine (3-MA, Selleck, Houston, TX, USA), Rapamycin (Sirolimus, Selleck), V9302 (Selleck) and GPNA hydrochloride (MedChemExpress) were purchased and treated as indicated.

Lentiviral production, plasmids and microRNAs

All lentiviral were purchased from Genomeditech (Shanghai, China). Lentiviral-based shRNA plasmids were used: RACK1-specific targeting shRNAs sequences were as follows: shRACK1#1, TACCCTGGGTGTGTGCAAATA; shRACK1#2, GCAGCAACCCTATCATCGTCT. ASCT2-specific targeting shRNA sequence was GAAGCACAGAGCCTGAGTTGA. And the negative control scrambled target sequence was TTCTCCGAACGTGTCACGT. For RACK1 and ASCT2 plasmids construction, human RACK1 and ASCT2 cDNA were in frame subcloned into pcDNA3.1-Myc/His vector (Invitrogen) and p3 × FLAG-CMV-14 vector (Sigma Aldrich), respectively. pDEST-CMV mCherry-GFP-LC3B WT plasmid was purchased from Robin Ketteler (Addgene, Cat #123230). The hsa-miR-146b-5p inhibitor and mimics were purchased from GenePharma (Shanghai, China). Sequences for hsa-miR-146b-5p inhibitor were: 5’- CAGCCUAUGGAAUUCAGUUCUCA-3’, and for control microRNA inhibitor were: 5’-CAGUACUUUUGUGUAGUACAA-3’. Sequences for hsa-miR-146b-5p mimics were 5’-UGAGAACUGAAUUCCAUAGGCUG-3’, and for corresponding control were: 5’-UUCUUCGAACGUGUCACGUTT-3’. Lipofectamine 2000 (LifeTechnologies, Carlsbad, CA, USA) was used for plasmids and microRNAs transfection according to the manufacturer's instructions.

CCK-8

GC cell viability was detected by CCK-8 assays. Briefly, AGS and MKN-45 cells with stable RACK1 knockdown and/or ASCT2 knockdown or transfected with RACK1 and/or ASCT2 plasmids were plated into 96-well plates at a seeding density of 3000 cells per well. For glucose or glutamine deprivation, the medium was without glucose or glutamine, respectively. Then the cells were cultured for 24 h, 48 h and 72 h. For GPNA or V-9302 treatment, after cell seeding for 24 h, GPNA or V-9302 (20 μM) was added into the medium for another 48 h. Finally, 10 μL CCK-8 solution (Meilunbio, Dalian, China) was added into each well and the cells were incubated for another 2 h at 37 °C, and the absorbance was recorded at 450 nm by a microplate reader (Thermo Fisher Scientific).

Colony formation assay

For examination of cell colony formation ability, approximately 1500 stable GC cells were seeded into 12-well plates and then treated as indicated. Fresh medium was changed every 3 days. After 9 days of incubation, colonies were fixed with 4% paraformaldehyde and stained with crystal violet. After washing with PBS for three times, the wells were photographed by a camera. Then the number of colonies and the intensity of violet were evaluated by ImageJ software in an automated way. Experiments were performed in three replicates.

Flow cytometry

AGS cells with stable RACK1 knockdown or control scrambled shRNA were cultured with normal medium or glucose or glutamine depletion medium for 48 h. Then the cells were collected and stained with Propidium (PI, Meilunbio), and the intensity of PI was determined by flow cytometry.

Measurement of glutamine consumption

Normal medium with 2 mM glutamine was used to incubate GC cells for 72 h. Glutamine uptake/comsumption was measured with Glutamine Assay Kit (Dojindo) following the manufacturer’s instructions. Glutamine uptake was calculated by deducting the measured glutamine concentration in the medium from the original glutamine concentration.

Measurement of oxygen consumption rate (OCR)

OCR were measured by Agilent Seahorse XFe96 Analyzer, as instructed by the Mito Stress Test Kit from Agilent Technologies (CA, USA). Briefly, 20,000 cells were plated in Cell Culture Microplates the day before measurement. Next, the addition of oligomycin, FCCP and rotenone was performed to determine the OCR value.

Western blot

Total protein of cells was extracted by using SDS lysis buffer with protease and phosphatase inhibitors on ice, and aliquots of protein were separated by 10% or 15% SDS-PAGE and resolved proteins were transferred to polyvinylidene fluoride (PVDF) membranes (EMD Millipore, Billerica. MA, USA). Then the membranes were incubated with primary antibodies, including RACK1 (1:1000; Santa Cruz Biotechnology, Cat #sc-17754); ASCT2 (1:1000; Cell Signaling Technology, Cat #8057); Flag-tag (1:2000; Sigma-Aldrich, St Louis. MO, USA, Cat #F1804); p62 (1:1000; Cell Signaling Technology. Cat #23214); LC3B (1:1000; Cell Signaling Technology, Cat #3868); p-mTOR (1:1000; Cell Signaling Technology, Cat #5536); mTOR (1:1000; Cell Signaling Technology, Cat #2983); p-AKT (1:1000; Cell Signaling Technology, Cat #4060); AKT (1:1000; Cell Signaling Technology, Cat #4691); p-PDK1 (1:1000; Cell Signaling Technology. Cat #3438); PDK1 (1:1000; Cell Signaling Technology. Cat #3062); PTEN (1:1000; Santa Cruz Biotechnology, Cat #sc-7974); p85 (1:1000; Cell Signaling Technology, Cat #4257); p110 (1:1000; Cell Signaling Technology, Cat #4249); p-p70S6K^T389 (1:1000; Cell Signaling Technology, Cat #97596); p70S6K (1:1000; Cell Signaling Technology, Cat #34475); p-RPS6^S235/236 (1:1000; Cell Signaling Technology, Cat #5316); RPS6 (1:1000; Cell Signaling Technology, Cat #64108); p-4E-BP1^T37/46(1:1000; Cell Signaling Technology, Cat #2855); 4E-BP1 (1:1000; Cell Signaling Technology, Cat #9644); p-eIF4E (1:1000; Cell Signaling Technology, Cat #9741); eIF4E (1:1000; Cell Signaling Technology, Cat #2067); followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibody (1:2000; Proteintech, Chicago, USA, Cat #SA00001–1 or SA00001–2). β-Actin (1:2000; Proteintech, Cat #HRP-60008) was used as endogenous control. Protein expression was visualized by enhanced chemiluminescence assay.

Real-time PCR analysis

Total RNA was prepared with Trizol reagent (TaKaRa, Dalian, China) according to the manufacturer’s protocol. Subsequent reverse transcription was performed using commercial kit (TaKaRa). Then cDNA was amplified and quantified on ABI StepOne Plus detection system using SYBR Premix Ex Taq (Takara, Tokyo, Japan). GAPDH was used as endogenous control. The sequences and Tm for each pair of primers were shown in Table S2. A Track miRNA qRT-PCR Starter Kit (RiboBio, Shanghai, China, Cat # C10712) were utilized to detect the expression of miR-146b-5p. The primers for miR-146b-5p and endogenous control U6 were also provided by RiboBio.

Autophagy double fluorescent assay

AGS and MKN-45 cells with stable expressing control shRNA or RACK1 shRNA were seeded and cultured in 12-well plates (2 × 10⁵ cells in each well) for 12 h to 18 h. Then pDEST-CMV mCherry-GFP-LC3B WT plasmid was transfected into these cells using lipofectamine 2000 and the transfection medium was switched to normal medium after 6 h. 72 h later, cells were washed for 2–3 times with PBS following by fixed with 4% paraformaldehyde. And the fluorescence was detected by inverted fluorescence microscope (Olympus Corporation, Germany, BX53F).

Immunohistochemical staining and scoring

Briefly, the tissue microarrays or slides were baked at 60 °C for 4–6 h, deparaffinized in xylene, and subsequently soaked in a graded ethanol series for rehydration. Then the endogenous peroxidase activity was blocked by 0. 3% hydrogen peroxide. 10 mM citrate buffer was utilized to antigen retrieval under 121 °C for 20 min in a microwave oven. Then the sections were incubated with the primary antibody including RACK1 (1:100), ASCT2 (1:200), Ki-67 (1:300), p62 (1:100), p-mTOR (1:100) at 4 °C overnight. After treated with Primary Antibody Amplifier Quanto and HRP Polymer Quanto (Absin, CAS#abs957), the sections were stained by DAB solution and counterstained with hematoxylin. Clinical tissue microarrays were assessed by 3 experienced and independent pathologists whose were blinded to the patient’s status. The staining intensity was scored from 0 to 3, and its heterogeneity was scored from 0 to 4 depending on the percentage of positively stained tumor cells. To obtain an immunohistochemical (IHC) score which takes into consideration of both IHC signal intensity and frequency of positive cells, we generated composite expression scores (CES) with a full range from 0 to 12. Then receiver operating characteristic (ROC) curve analysis was employed to evaluate and determine the optimal cut-off value of RACK1 expression. As shown in Fig. 7c, the area under ROC curve (AUC) of CES 4 was the largest (0.623). Thus, CES > 4 indicated high expression of RACK1, while CES ≤ 4 indicated RACK1 low expression. Subcutaneous tumor sections were counted for five fields under 200 × magnification.

Fig. 7.

Loss of RACK1 is inversely correlated to ASCT2 and predicts poor prognosis in GC patient. (a) The representative staining images of RACK1, ASCT2, p-mTOR and p62 in a tissue microarray from the same patients. Scale bar, 50 µm. (b) The correlation between RACK1 expression in tumor tissues and the staining of ASCT2, p-mTOR and p62 was determined by Spearman’s correlation analysis, respectively. (c) ROC curve analysis of RACK1 composite expression scores (CES). (d) Association of RACK1 expression in tumor tissues with overall survival was determined by Kaplan–Meier analysis in GC patients. (e) Hypothetical scheme for the mechanism of RACK1 downregulation on modulating glutamine addiction of GC

Animal studies

5–6-weeks-old male BALB/c nude mice were purchased from the Slack company (Shanghai, China). AGS cells with stable RACK1 knockdown and/or ASCT2 knockdown (5 × 10⁶ in 100 μL PBS) were subcutaneously injected into the right flanks of BALB/c nude mice, respectively (n = 6 for each group). Tumor volumes were measured using a caliper every 6 days and calculated by the formula: Volume = 0.5 × length × width². After 24 days, all mice were firstly anesthetized with isoflurane and then euthanized by cervical dislocation. Then the subcutaneous tumors were separated, weighed, and fixed in a formalin solution. All these tissues were then embedded with paraffin and sectioned for further examination.

Bioinformatics analysis of gene expression

The mRNA expression data are publicly available from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/) and Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/) databases. Gene set enrichment analysis (GSEA) was performed using GSEA v4.0 software obtained from public database (https://www.gsea-msigdb.org/gsea/index.jsp). Human Gene Set “GOBP_POSITIVE_ REGULATION_OF_AUTOPHAGY” and “HALLMARK_MTORC1_SIGNALING” were utilized to perform autophagy enrichment analysis and mTOR signaling enrichment analysis based on RACK1 expression, respectively. The parameters for statistically significant differences were employed as Nominal p-value < 0.05, and FDR q-value < 0.05.

Statistical analysis

Results are presented as the mean ± standard deviation (SD). Spearman’s correlation analysis and chi-square test were conducted using SPSS 20.0 software. Data from quantitative PCR (qPCR) experiments, colony formation assay, double immunofluorescence, IHC and tumor weight were analyzed by Student’s two-tailed t-test, data from CCK-8 and tumor growth curve were estimated by two-way analysis of variance in GraphPad Prism 9. 0. The statistically significant differences were considered as P < 0. 05.

Results

Knockdown of RACK1 facilitates the viability, colony formation, cell cycle progression and oxygen consumption rate of GC cells in a glutamine-dependent manner

As endogenous expression of RACK1 is quite high in stomach tissues, AGS and MKN-45 which own relatively low expression of RACK1 and high expression of glutamine-related enzymes (Glutaminase-1 and Glutaminase-2) were selected to investigate whether RACK1 is involved in metabolic reprogramming (Fig. S1). Firstly, glucose or glutamine deprivation was applied to explore the relevance between RACK1 knockdown-mediated acceleration of GC cell proliferation and glucose or glutamine metabolism. CCK-8 assays showed that RACK1 knockdown remarkably promoted GC cell viability under glucose and glutamine sufficient condition (Fig. 1a and Fig. S2a). And the facilitative influence by RACK1 knockdown was little affected in the absence of glucose (Fig. 1b and Fig. S2b), but was blocked by glutamine deletion (Fig. 1c and Fig. S2c). Further statistical analysis also showed that the fold change of cell viability indued by RACK1 knockdown was significantly decreased after glutamine deprivation (Fig. 1d and Fig. S2d). Similarly, clonogenic assays demonstrated that RACK1 knockdown significantly enhanced the colony formation abilities of GC cells, and was reversed by glutamine deficiency but not glucose (Fig. 1e and Fig. S2e-f).

Fig. 1.

Knockdown of RACK1 facilitates the viability, colony formation, cell cycle progression and oxygen consumption rate of GC cells in a glutamine-dependent manner. (a-c) The effect of RACK1 knockdown on the viabilities of AGS cells under glucose and glutamine sufficient condition (a), glucose deficiency condition (b), and glutamine deficiency condition (c) were determined by CCK-8 assays, respectively. (d) The fold change of cell viabilities by RACK1 knockdown under indicated conditions. (e) Colony formation assays were used to detect the effect of RACK1 knockdown on the colony formation abilities of AGS and MKN-45 cells under indicated condition. (f-i) Cell cycle of AGS were examined by flow cytometry with PI staining under indicated condition. (j) The effect of RACK1 knockdown on glutamine uptake. (k) Seahorse assays were utilized to measure the oxygen consumption rate of GC cells with indicated treatment. In (a-e) and (i-k), error bars represented as mean ± SD. *, P < 0. 05; **, P < 0. 01; ***, P < 0. 001; ns, no significance; Scr, scramble; G, glucose; Q, glutamine

Furthermore, we found that loss of RACK1 could notably accelerated cell cycle progression from G1 phase to S phase, and the promotive effect was slightly altered by glucose deprivation, but was completely reversed by glutamine deprivation (Fig. 1f-i). These data suggest a potential modulation of RACK1 in glutamine addiction in GC. Hence, we next detect the influence of RACK1 depletion on glutamine uptake. And as expected, RACK1 knockdown dramatically increased the glutamine uptake (Fig. 1j). Next we aimed to determine whether RACK1 play an important role in oxidative phosphorylation of GC cells. As shown in Fig. 1k and Fig. S3, RACK1 knockdown could enhance oxygen consumption rate of AGS and MKN-45 cells, but when glutamine was removed, the promotive effect was totally abrogated. Taken together, these results indicate a dependence of RACK1 knockdown-induced malignant phenotype of GC cells on glutamine.

RACK1 promotes lysosomal degradation of ASCT2 in GC cells

Given that RACK1-knockdown GC cells exhibited glutamine addiction properties, we evaluated the effects of RACK1 on Alanine-serine-cysteine amino acid transporter 2 (ASCT2; SLC1A5), a major glutamine transporter which motivates malignant phenotypes in several cancer cells [21]. Western blot and real-time PCR results revealed that depletion of RACK1 conspicuously augmented ASCT2 protein level, whereas the mRNA expression of ASCT2 was not affected (Fig. 2a-b). Consistently, overexpression of RACK1 only decreased ASCT2 protein abundance but not affect its mRNA expression (Fig. 2c-d). Thus, we speculated that the regulation of RACK1 on ASCT2 expression might be at a post-translational level. To verify this hypothesis, we treated GC cells with cycloheximide (CHX) to inhibit de novo protein synthesis. And results indicated that RACK1 overexpression dramatically reduced ASCT2 protein half-lives both in AGS and MKN-45 cells (Fig. 2e-f). To further evaluate how RACK1 regulated ASCT2 protein stability, GC cells were treated with the lysosome inhibitor Chloroquine (CQ) or the proteasome inhibitor MG132, respectively. We found that application of CQ restored ASCT2 protein content when RACK1 overexpression (Fig. 2g), and conversely, administration of MG132 could not rescue the suppressive effect of RAKC1 on ASCT2 (Fig. 2h). All these results suggest that RACK1 represses ASCT2 expression through inducing its lysosomal degradation.

Fig. 2.

RACK1 promotes lysosomal degradation of ASCT2 in GC cells. (a) The effect of RACK1 knockdown on ASCT2 protein level were examined by western blot and quantification of relative ASCT2 protein intensity was shown at the right panel. (b) The effect of RACK1 knockdown on ASCT2 mRNA level were examined by real-time PCR. (c-d) AGS and MKN-45 cells were transfected with empty vector or myc-tagged RACK1, followed by western blot (c) and real-time PCR analysis (d). Quantification of relative ASCT2 protein intensity was shown at the right panel of (c). (e-f) AGS (e) and MKN-45 (f) cells were transfected with empty vector or myc-tagged RACK1, and then applied to CHX chase analysis at indicated time points. Quantification of relative ASCT2 protein levels were shown at the bottom panel. (g-h) Immunoblots of lysates from transfected GC cells with or without 50 μM chloroquine (g) or 10 μM MG132 (h) treatment for 24 h. Quantification of relative ASCT2 protein intensity was shown at the right panel of (g-h). In (a-h), error bars represented as mean ± SD. *, P < 0. 05; **, P < 0. 01; ***, P < 0. 001; ns, no significance

To further verify whether RACK1-mediated suppressive effect on GC tumor growth was ASCT2 downregulation-dependent, we co-transfected RACK1 and ASCT2 (Fig. S4a). Subsequent CCK-8 assays exhibited that RACK1 overexpression significantly inhibited the viability of GC cells, while ASCT2 displayed opposite effect (Fig. S4b). What’s more, ASCT2 addition faithfully reversed RACK1-induced growth arrest (Fig. S4b). These data reveal that RACK1 inhibits GC tumor growth through reducing ASCT2 expression.

RACK1 suppresses ASCT2 by facilitating autophagy in GC cells

As autophagy-lysosomal pathway is one of the two major pathways responsible for protein lysosomal degradation. We firstly estimated the association between RACK1 and autophagy by KEGG enrichment analysis in three GEO datasets (GSE54129, GSE26942, GSE35809), and found that RACK1 high expression was positively correlated with autophagy signature (Fig. 3a). Further analysis demonstrated the downregulation of autophagy-related genes following RACK1 depletion in GC cells (Fig. S5). Next we detected the effect of RACK1 on LC3-II/I ratio and p62, two recognized indicators for autophagy. As shown in Fig. 3b, LC3-II/I ratio was obviously reduced while p62 expression was boosted when RACK1 was stably knocked down. In addition, mCherry-GFP-LC3B double standard system was used to test the formation of autolysosome. In this system, the cells showed yellow fluorescence due to the co-expression of mCherry and GFP when only autophagosomes were existed. After autophagosomes and lysosomes fuse to form autolysosomes, the acidic environment of lysosomes will lead to a quenching of the acid-sensitive GFP, while mCherry is not affected, and cells displayed red fluorescence. Based on the fluorescence, we observed that the autolysosomes in shRACK1 cells were conspicuously lessened compared with that in control cells (Fig. 3c-d). Furthermore, we found that autophagy inhibitor 3-MA robustly reversed RACK1-induced destruction of ASCT2 (Fig. 3e). These results indicate that RACK1 accelerates the lysosomal degradation of ASCT2 by promoting autophagy.

Fig. 3.

RACK1 suppresses ASCT2 by facilitating autophagy in GC cells. (a) Gene set enrichment analysis of autophagy signature for comparison RACK1 high expression group with RACK1 low expression group in GSE54129, GSE26942 and GES35809 datasets, respectively. (b) The lysates of AGS and MKN-45 cells with stable expressing control shRNA or RACK1 shRNA were applied to western blot. Quantification of relative p62 protein intensity and LC3-II/I ratio were shown at the middle and right panels, respectively. (c) After transfected with pDEST-CMV mCherry-GFP-LC3B WT plasmid, GC cells were undergone microscopic observation and white short arrows indicated autophagosomes fluorescent dots (mCherry⁺/GFP⁺). (d) Quantification of relative fluorescent dots of autolysosomes in cells (Autolysosomes, AL, mCherry⁺/GFP⁻, upper panel) and total autophagosomes and autolysosomes in cells (Total, mCherry⁺, bottom panel). (e) Immunoblots of lysates from AGS and MKN-45 cells transfected as indicated and with or without 100 μM 3-MA treatment for 24 h. Quantification of relative ASCT2 and p62 protein intensity and LC3-II/I ratio were shown at the right panel. In (b-e), error bars represented as mean ± SD. *, P < 0. 05; **, P < 0. 01; ***, P < 0. 001; ns, no significance; Scr, scramble

RACK1 knockdown inhibits autophagy through activating AKT/mTOR signaling pathway in GC cells

Next, we aimed to investigate how RACK1 modulated autophagy in GC. KEGG enrichment analysis based on above GEO datasets showed that RACK1 high expression was distinctly negative correlated with mTOR signaling (Fig. 4a). Further investigation confirmed that RACK1 depletion could induce the phosphorylation of mTOR downstream proteins including p70S6K, RPS6, 4E-BP1 and eIF4E (Fig. S6). As mTOR is a major regulator of autophagy, inhibiting mTOR is an important way to increase autophagy level [22]. Therefore, we treated GC cells with rapamycin, a classic depressor of mTOR. Results displayed that impeding of mTOR could eliminate the autophagy braking effect as well as ASCT2 upregulation caused by RACK1 knockdown (Fig. 4b). It is well known that PI3K/AKT axis is one of the main signaling pathways for regulating mTOR [22]. Then the key molecules of PI3K/AKT pathway were detected, we observed an increase of PDK1 and AKT phosphorylation when RACK1 was knocked down (Fig. 4c). However, the protein levels of p85 and p110, two major subunits of PI3K exhibited little affected by RACK1 depletion; on the contrary, the protein level of PTEN, a PIP3-phosphatase that performs the opposite function of PI3K by converting PIP3 to PI-4, 5-P2 through dephosphorylation [23], was remarkably declined along with RACK1 knockdown (Fig. 4c). Real-time PCR analysis also confirmed that RACK1 depletion repressed the mRNA level of PTEN but had little effect on p85 and p110 (Fig. 4d). Spearman’s rank correlation analysis also revealed a positive correlation between RACK1 and PTEN in 343 clinical stomach adenocarcinoma samples (TCGA-STAD) (R = 0.287, P < 0.001) (Fig. 4e). These data suggest that RACK1 knockdown restrains autophagy through PTEN downregulation-mediated activation of AKT/mTOR signaling pathway in GC.

Fig. 4.

RACK1 knockdown inhibits autophagy through activating AKT/mTOR signaling pathway in GC cells. (a) Gene set enrichment analysis of mTOR signaling for comparison RACK1 high expression group with RACK1 low expression group in GSE54129, GSE26942 and GES35809 datasets, respectively. (b) Stable AGS and MKN-45 cells expressing control shRNA or RACK1 shRNA were treated with 1 μM rapamycin for 24 h, and applied to western blot. Quantification of relative ASCT2 and p62 protein intensity, mTOR phosphorylation and LC3-II/I ratio in AGS and MKN-45 cells were shown at the middle and bottom panels, respectively. (c) The effect of RACK1 knockdown on protein levels of the indicators were analyzed by western blot. Quantification of relative PTEN protein intensity, AKT and PDK1 phosphorylation in AGS and MKN-45 cells were shown at the middle and bottom panels, respectively. (d) The effect of RACK1 knockdown on mRNA levels of the indicators were analyzed by real-time PCR. (e) The correlation between GNB2L1 and PTEN mRNA levels in 343 patients from TCGA-STAD database were performed by Spearman’s correlation test. In (b-d), error bars represented as mean ± SD. *, P < 0. 05; **, P < 0. 01; ***, P < 0. 001; ns, no significance

RACK1 promotes PTEN expression via suppressing miR-146b-5p

Next we explored the mechanism by which loss of RACK1 reduced the expression of PTEN. Considering that RACK1 depletion affected PTEN mRNA level, and our previous studies found that multiple microRNAs were upregulated following RACK1 knockdown by a microRNA array screen (Fig. 5a) [14]. Moreover, among the top 10 microRNAs increased by RACK1 depletion, miR-146b-5p were demonstrated to have putative target site within the 3’-UTR of PTEN [24]. So we speculated that miR-146b-5p might be the mediator between RACK1 and PTEN. To further ascertain this assumption, we firstly examined the influence of RACK1 on miR-146b-5p. Results showed that RACK1 depletion significantly increased the expression of miR-146b-5p while RACK1 overexpression exhibited the opposite effect (Fig. S7). Next we addressed whether loss of RACK1-mediated PTEN downregulation and ASCT2 upregulation was miR-146b-5p-dependent. As shown in Fig. 5b-d, miR-146b-5p inhibition with specific inhibitor abolished RACK1 knockdown-induced the decrease of PTEN and the increase of ASCT2. Furthermore, loss of RACK1 could not stimulate oxidative phosphorylation of GC cells when miR-146b-5p inhibitor was co-transfected (Fig. 5e-f). Consistently, miR-146b-5p activation with corresponding mimics blocked the promotive effect of RACK1 on PTEN expression, and the suppressive effects on ASCT2 expression and OCR of GC cells (Fig. 5g-j). All these data indicate that RACK1 could augment PTEN mRNA abundance and inhibit ASCT2 expression and OCR of GC cells via repressing miR-146b-5p.

Fig. 5.

RACK1 promotes PTEN expression via suppressing miR-146b-5p in GC cells. (a) microRNA array E-MTAB-2619 was utilized to screen microRNAs affected by RACK1 knockdown, and the top 10 microRNAs with the most upregulated foldchanges were shown. (b-d) RACK1 knockdown or control GC cells were transfected with or without miR-164b-5p inhibitor, and then were applied to real-time PCR (b-c) and western blot (d), respectively. (e–f) Seahorse assays were utilized to measure the effect of miR-164b-5p inhibitor on RACK1 knockdown-induced oxygen consumption rate of GC cells with indicated treatment. (g-i) RACK1 overexpression or control GC cells were transfected with or without miR-164b-5p mimics, and then were applied to real-time PCR (g-h) and western blot (i), respectively. (j) Seahorse assays were utilized to measure the effect of miR-164b-5p mimics on RACK1 overexpression-inhibited oxygen consumption rate of GC cells with indicated treatment.In (b-c) and (e–h) and (j), error bars represented as mean ± SD. **, P < 0. 01; ***, P < 0. 001; ns, no significance; Scr, scramble

ASCT2 inhibition strikingly blocks the promotive function of RACK1 knockdown on GC tumor growth in vitro and in vivo

Since stimulant tumor growth by RACK1 knockdown leads to a strict dependence on glutamine in AGS and MKN-45 cells (Fig. 1 and Fig. S2), we hypothesized that inhibition of glutamine transporter ASCT2 could eliminate the accelerative effect by RACK1 depletion. Here, ASCT2 shRNA was employed. As expected, RACK1 knockdown failed to promote the viability and colony formation of GC cells along with ASCT2 knockdown (Fig. 6a-d and Fig. S8). Consistently, RACK1 knockdown significantly facilitated tumor growth in subcutaneous tumor model, but was reversed by ASCT2 depletion (Fig. 6e-f). Further immumohistochemical staining of RACK1, ASCT2 and Ki-67 (a marker reflects cellular proliferation) was performed on these tumor tissues. As shown in Fig. 6g-j, loss of RACK1 obviously increase ASCT2 and Ki-67 expression whereas the stimulative effect was distinctly counteracted when ASCT2 was knocked down. Together, the promotive role of RACK1 downregulation in GC is ASCT2-dependent.

Fig. 6.

ASCT2 inhibition remarkably blocks the promotive function of RACK1 knockdown on GC tumor growth in vitro and in vivo. (a-b) CCK-8 assays were applied to detect the effect of RACK1 knockdown on cell viability along with or without ASCT2 knockdown. (c-d) Colony formation assays were applied to detect the effect of RACK1 knockdown on colony formation along with or without ASCT2 knockdown. (e–f) Nude mice were subcutaneously injected with stable AGS cells as indicated for 24 days. (e) Tumor volumes were measured every 6 days. After sacrifice, tumor xenografts were excised and weighed (f). (h-j) The expression of RACK1, ASCT2 and Ki-67 in subcutaneous tumor tissues was detected by immunohistochemistry. In (a-f) and (h-j), error bars represented as mean ± SD. **, P < 0. 01; ***, P < 0. 001; ns, no significance; Scr, scramble

Loss of RACK1 is inversely correlated to ASCT2 and predicts poor prognosis in GC patient

To further confirm the regulation of RACK1 on mTOR/autophagy/ASCT2 axis in GC, we employed clinical tissue microarray. Immunohistochemistry results indicated that the expression of RACK1 is negatively correlated to the staining of ASCT2 and p62 as well as the phosphorylation level of mTOR (Fig. 7a-b). Then we assessed the value of RACK1 in predicting the prognosis of GC. After evaluating the optimal cut-off value of RACK1 expression by ROC curve analysis (Fig. 7c), we performed chi-square test to assess the correlation between RACK1 expression and clinicopathologic features in 120 GC patients. And we found that patients with low RACK1 expression were prone to have poorer differentiation, higher T stage and higher TNM stage (Table S1). Further Kaplan–Meier analysis revealed that patients with RACK1 low expression have a strikingly poor prognosis than those with RACK1 high expression (P = 0.003, HR = 2.49, 95% IC = 46.402–57.351) (Fig. 7d).

Discussion

Cancer cells exhibit excessive proliferation and require a constant supply of fuel such as glucose and glutamine. Thus, cancer cells coordinate metabolic pathways to harmonize the high demand for these nutrients, which is called metabolic reprogramming [25]. RACK1 has been reported to play an important role in regulating various features of diverse cancer cells. Nevertheless, the function of RACK1 on modulating metabolism in cancer cells, especially in GC remains largely unknown. In the present study, we surprisingly found that the promotive effect of RACK1 knockdown on the viability, colony formation, cell cycle progression and OCR of GC cells is glutamine-dependent while glucose-independent (Fig. 1 and Fig. S2-3). Downregulation of RACK1 inhibited lysosomal degradation of ASCT2 by stimulating miR-146b-5p to reduce PTEN, which subsequently led to an activation of AKT/mTOR signaling and impairment of downstream autophagy in GC cells (Fig. 7e). These results reveal the involvement of RACK1 in metabolic reprogramming of tumor cells for the first time, and blocking ASCT2 and/or modulating RACK1 expression is a potential experimental therapy for advanced GC cases in human subjects.

Glutamine is a non-essential amino acid (NEAA) that is currently considered to be essential for cancer cells to maintain exceeded glutamine biosynthesis [26]. Hence, glutamine deprivation inhibits tumor growth and even induces cell death in several cancers [27]. ASCT2, a major transporter for glutamine, has been reported to be overexpressed in highly proliferating cancer cells to fulfill enhanced glutamine demands [28]. Although the regulation of ASCT2 expression has been studied, a comprehensive view of the pathway is far from clear. The few studies that have shown that SLC1A5 expression is regulated by several transcription factors such as MYC, retinoblastoma/E2F transcription factor pathway (RB/E2F), androgen receptor (AR), and ATF4 have been shown to regulate ASCT2 mRNA level [29]. On the other hand, one recent study confirmed that short-term treatment of epidermal growth factor (EGF) could enhance the protein stability of membranal ASCT2 in enterocytes [30]. Similarly, in the present study, we revealed another potential mechanism that loss of RACK1 promote ASCT2 protein expression by activating AKT/mTOR signaling to inhibit the autophagolysosomal degradation pathway in GC cells.

Eukaryotic cells have two main degradation systems, autophagy-lysosomes and proteasomes. Autophagy is a highly dynamic self-degrading process initiated by the formation of double-membraned autophagosomes, which paly an essential role in maintaining cell homeostasis by its phagocytosis of cytoplasmic substrates or cargo (such as proteins, lipids, and organelles), and favor to energy recycled and reused [31]. As an important component of autophagy machinery, RACK1 has been shown to be an autophagy trigger in physiology, and its interaction with ATG5 is a key event in recruiting itself into the autophagy pathway [32]. Another study reported that RACK1 participated in the formation of autophagosome biogenesis complex upon its phosphorylation by AMPK at Thr50, which enhances its direct binding to Vps15, Atg14L, and Beclin1, thereby promoting the assembly of the autophagy-initiation complex [33]. In addition to a series of autophagy related genes, mTOR signaling pathway is also a recognized modulator for autophagy, which could regulate different steps in the autophagy process including nucleation, autophagosome extension, autophagosome maturation and termination [34]. Our data provide more evidence that down-regulated RACK1 inhibited autophagy flux by promoting AKT/mTOR pathway, thus affecting ASCT2 stability.

Differences in the consumption and metabolism of essential nutrients between cancer and normal cells have been considered as a promising target for cancer treatment [35]. An increased need for specific amino acids is observed in several cancers, makes it more dependent on their exogenous supply or new synthesis. Strategies to utilize such "metabolic addiction" include removing amino acids in serum, blocking uptake by transporters and inhibiting biosynthetic or catalytic enzymes [36]. To date, targeting glutamine metabolism is one of the most advanced strategies for cancer therapy. Recently, available preclinical data suggest that further efforts to pharmacologically inhibit ASCT2 are warranted. V-9302 and GPNA are two specific small molecule inhibitors of ASCT2, which were applied in our research. In fact, the suppressive effects of these two inhibitors were relatively slight in RACK1 high MGC80-3 cells compared with RACK1 lower AGS and MKN-45 cells (Fig. S9). These data suggest that GC patients with RACK1 low expression might obtained more clinical benefit based on ASCT2 inhibition than those with RACK1 high expression. In addition, miR-146b-5p has been found to play an important role in regulating various tumor progression via different pathways [37, 38]. However, our present study reported the modulation of miR-146b-5p on glutamine metabolism for the first time, which provides another potential target for glutamine metabolism-related therapies.

As of now, GC remains the fourth leading cause of cancer-related death worldwide. Current treatment strategies include single agent chemotherapy, combination chemotherapy, radiotherapy, targeted agents (either alone or in combination with conventional chemotherapy) and best supportive care [39]. Among them, chemotherapy is the main treatment approach for advanced GC and adjuvant post-operative GC; however, multi-drug resistance (MDR) often leads to the failure of chemotherapy [40]. Compared with traditional chemotherapy, targeted therapy has the characteristics of “high efficiency and low toxicity”. Therefore, it will be the trend to provide precise individual treatment with molecularly targeted drugs based on gene molecular expression characteristics. Our data suggest that loss of RACK1 promotes glutamine addiction via activating AKT/mTOR/ASCT2 axis to facilitate tumor growth in GC. These results suggest an underlying therapeutic target for GC patients, and glutamine starvation or ASCT2 inhibition provide a potential specific therapeutic method for GC patients with RACK1 low expression.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

M. Chen: wrote the original draft and prepared Figs. 2, 3, 4, 5, 6, and 7. G. Wang: prepared Figs. 1, 2, 3, 4, 5, and 6. Z. Xu: statistical analysis and prepared supplementary information. J. Sun: collected clinical data and prepared Fig. 7. B. Liu and L. Chang: resources. J. Gu and Y. Ruan: conceptualization, supervision and funding acquisition. X. Gao: resources and supervision. S. Song: data curation, supervision, funding acquisition, writing-review and editing. All authors reviewed the manuscript.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval

In this work, animal care and experiments were performed in strict accordance with the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health, and all animal experiments were approved by the ethics committee of School of Basic Medical Sciences, Fudan University (no. 20210907–002).

Competing interest

The authors have declared no potential conflicts of interest.