Glycolysis-related genes predict prognosis and indicate immune microenvironment features in gastric cancer

Glycolysis-related genes predict prognosis and indicate immune microenvironment features in gastric cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Glycolysis-related genes predict prognosis and indicate immune microenvironment features in gastric cancer Lu Xu, Jin Liu, Yuanqing An, Lei Zhou, Hui Sun, Zhen Xu, Deqiang Wang, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4130368/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 08 Aug, 2024 Read the published version in BMC Cancer → Version 1 posted 4 You are reading this latest preprint version Abstract Background Gastric cancer (GC) is one of the leading causes of cancer-related death. Glycolysis plays a pivotal role in tumor microenvironment (TME) reprogramming. This study assessed the roles of glycolysis-related genes (GRGs) in predicting prognosis and indicating the immune microenvironment features in gastric cancer patients. Methods Gene expression data and clinical data of GC patients were obtained from The Cancer Genome Atlas (TCGA)-stomach adenocarcinoma (STAD) cohort and validated using datasets acquired from the Gene Expression Omnibus (GEO). A total of 326 GRGs were identified from the Molecular Signatures Database (MSigDB). Subtypes of GC were delineated via consensus clustering based on GRG expression. A multigene risk score model was developed using multivariate Cox regression analysis. The CIBERSORT and ESTIMATE algorithms were used to evaluate the immune microenvironment. To probe the biological function of critical genes, wound healing assays, transwell invasion assays, and MTT assays were used. Results The patients were divided into two groups, namely, the metabolic subtype (cluster A) and immune subtype (cluster B), based on the expression patterns of the GRGs. Patients in cluster B had a worse prognosis. A risk score model based on the expression of six GRGs, including ME1, PLOD2, NUP50, CXCR4, SLC35A3, and SRD35A3, could predict patient prognosis. Knockdown of CXCR4 significantly attenuated the glycolytic capacity, as well as the migration, invasion, and proliferation of GC cells. Interestingly, although both the immune subtype (cluster B) and high-risk groups had unfavorable prognosis, these two cohorts had favorable immune microenvironment and increased expression of immune checkpoint genes. We found that high expression of CXCR4 and low expression of ME1 were positively correlated with the infiltration of CD8 + T cells and the response to treatment with an anti-PD-1 immune checkpoint inhibitor. Conclusions In the present study, we identified that the expression patterns of GRGs could be used to predict the prognosis of GC patients and may be helpful in guiding clinical treatment decisions. glycolysis gastric cancer tumor microenvironment immune cell infiltration prognostic model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1 Introduction Gastric cancer (GC) is the most prevalent malignant neoplasm of the digestive system. Approximately one million people are diagnosed with GC annually worldwide [ 1 ]. GC patients have an unfavorable prognosis, and less than 30% of them may survive for five years after diagnosis [ 2 ]. The poor clinical prognosis of these patients is primarily due to the advanced disease stage and unfavorable tumor microenvironment (TME) [ 3 ]. Tumor cells primarily use glycolysis as a source of energy. A high glycolytic capacity in tumor cells is related to poor prognosis and drug resistance in various cancers, including GC [ 4 ]. Several studies have shown interactions between tumor glycolysis and the TME [ 5 – 7 ]. Tumor cells can produce lactic acid via glycolysis, and an increase in lactic acid makes the TME acidic, thus facilitating tumor cell proliferation, invasion, and migration [ 8 ]. Moreover, several studies have reported that metabolic reprogramming contributes to the development of an immunosuppressive tumor microenvironment [ 9 ]. Many studies have focused on the systematic investigation of glycolysis and tumors, including liver cancer [ 10 ], glioblastoma [ 11 ], and endometrial cancer [ 12 ]. In the present study, we evaluated the roles of glycolysis-related genes (GRGs) in predicting prognosis and indicating immune microenvironment features in gastric cancer patients. 2 Materials and methods 2.1 Acquisition of the mRNA expression dataset The transcriptome (fragments per kilobase million, FPKM) and clinical information were obtained from The Cancer Genome Atlas (TCGA) database ( https://portal.gdc.cancer.gov/ ) [ 13 ]. The expression data of 407 samples included 32 normal tissues and 375 tumor tissues. Concurrently, the GSE84437 dataset [ 14 ] and the GSE13763 dataset [ 15 ], which include microarray data (Illumina HumanHT-12 V4.0 expression bead chip platform) and clinical data, were acquired from the Gene Expression Omnibus (GEO) ( https://www.ncbi.nlm.nih.gov/geo/ ) and used as the validation cohort. The probe IDs were converted to gene symbols using the relevant annotation files (GPL6947 and GPL570), and the average expression values of numerous probes for the same gene were calculated. The data were standardized and processed with R (version 3.5.1) software and R Bioconductor components. We used data from the immunotherapy cohort PRJEB25780 ( https://www.ebi.ac.uk/ena/browser/home ) [ 16 ] to evaluate the relationship between gene expression and the response to immunotherapy. Gene expression data were obtained using the limma and ggpubr packages in R software. ROC curves were generated to predict the response of GC patients to anti-PD-1 therapy through the pROC package in R software. 2.2 Analysis of differentially expressed glycolysis-related genes (GRGs) According to the keyword “glycolysis”, five glycolysis-related sets (GO_GLYCOLYTIC_PROCESS, KEGG_GLYCOLYSIS_GLUCONEOGENESIS, BIOCARTA_GLYCOLYSIS_PATHWAY, HALLMARK_GLYCOLYSIS, and REACTOME_GLYCOLYSIS) were obtained from the Molecular Signatures Database (MSigDB) ( https://www.gseamsigdb.org/gsea/msigdb/ind-ex.jsp ) [ 17 ]. The genes in the above five sets were defined as glycolysis-related genes (GRGs). A total of 326 GRGs were identified from the TCGA-STAD cohort data. The differentially expressed genes (DEGs), defined as those with a false discovery rate (FDR) 1, were obtained by the "edgeR" R package [ 18 ]. Furthermore, gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were carried out to assess the potential functions and related pathways of the differentially expressed GRGs via the “ClusterProfiler” R package [ 19 ]. 2.3 Consensus clustering analysis of GRGs The "ConsensuClusterPlus" R package [ 20 ] was used to assess the expression profiles of 326 GRGs for consensus clustering. One thousand permutations were conducted to ensure consistency in the classification. A consensus heatmap and the cumulative distribution function (CDF) were used to determine the ideal k value. To analyze the biological functions between the subgroups, gene set variation analysis (GSVA) was performed by using the R package “GSVA”. The enrichment results were then visualized by a heatmap using the R package “heatmap”. By using the R package “limma”, the differences were statistically significant at adjusted p < 0.05 and a false discovery rate (FDR) < 0.25. The Kyoto Encyclopedia of Genes and Genomes (KEGG) gene set collection (c2.cp.kegg.v7.1.symbols.gmt) was used as the input file for both gene set enrichment analysis (GSEA) and gene set variation analysis (GSVA) analyses. 2.4 Construction of a glycolysis-related risk model Univariate Cox regression analysis was performed to identify the survival-related GRGs. The “glmnet” R package [ 21 ] was used to minimize the dimension by applying the least absolute shrinkage and selection operator (LASSO) for the survival-related GRGs. Then, a glycolysis-related risk model was generated by using the following formula: \(Risk score={\sum }_{i=1}^{n}Expi \beta i\) , where Exp denotes the expression levels of the genes and β represents the regression coefficient [ 22 ]. According to the median risk score, Kaplan-Meier curves were generated to evaluate the prognosis of patients in different groups. Survival ROC curves were generated using the “Survival ROC” R package [ 23 ]. 2.5 Development of a nomogram based on GRGs and clinical features Univariate Cox regression analysis was performed to assess the relationships between GRG expression and clinical features (age, sex, clinical stage, grade, tumor-node-metastasis (TNM) stage, and G grade). Multivariate Cox regression analysis was employed to identify independent prognostic variables. Taking GRG expression and clinical features into account, a nomogram scoring system was developed. Univariate and multivariate Cox regression analyses were conducted to identify factors associated with prognosis. Each item was given a score, and the scores were added together. Calibration curves were applied to assess the accuracy of the nomogram, and the concordance index (C-index) was used to quantify the discrimination capacity of the nomogram. 2.6 Cell culture and treatment The human gastric cancer cell line BGC-823 was obtained from the Shanghai Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured in RPMI-1640 medium (Gibco, Grand Island, New York, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, Grand Island, New York, USA) and 1% penicillin‒streptomycin solution (Beyotime, Nantong, China). The cells were maintained at 37°C in a humidified environment containing 5% CO 2 . The cells were passaged every 2–3 days. Lentiviruses encoding shRNAs for CXCR4 (VSVG-Lentai-hU6-shRNA-CXCR4-BSD-hEF1a-3xFlag) and negative control (NC) lentivirus (VSVG-Lentai-hU6-shRNA-NC-BSD-hEF1a-3xFlag) were purchased from Shanghai Taitool Bioscience Co. (Shanghai, China). The sequences of the shRNAs targeting CXCR4 are summarized in Table 1 . Cells were infected with the lentivirus and selected by using blasticidin (BSD, Sigma) 72 h after infection. Table 1 ShRNA sequences targeting CXCR4. Name Sense Antisense CXCR4-homo-1 GAAGCATGACGGACAAGTA TACTTGTCCGTCATGCTTC CXCR4-homo-2 GGAAGCTGTTGGCTGAAAA TTTTCAGCCAACAGCTTCC 2.7 Quantitative real-time PCR TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) was used to extract total RNA. After spectrophotometric quantification, 1 µg of total RNA in a final volume of 20 µl was reverse transcribed with a PrimeScript RT Reagent Kit (TaKaRa, Shiga, Japan) according to the manufacturer's protocol. Aliquots of complementary DNA (cDNA) corresponding to equal amounts of RNA were used for quantification of messenger RNA (mRNA) by RT-PCR using the Light Cycler 96 Real-time Quantitative PCR Detection System (Roche, Indianapolis, IN, USA). The reaction system contained the corresponding cDNA, forward and reverse primers, and SYBR Green PCR master mix (Roche). β-Actin was used as an internal standard. The primer sequences are summarized in Table 2 . Table 2 Primers used for real-time PCR. Primer name Forward Sequence Reverse Sequence CXCR4 CTCCTCTTTGTCATCACGCTTCC GGATGAGGACACTGCTGTAGAG β-actin TCATGAAGTGTGACGTGGACAT CTCAGGAGGAGCAATGATCTTG 2.8 Glucose Uptake Assa y Glucose uptake was defined as the uptake of 2-deoxyglucose (2-DG) using a glucose uptake assay kit (fluorometric, ab136956, Abcam) according to the manufacturer’s instructions. Cells were seeded at a density of 2.5×10 4 cells/well in 96-well plates overnight. The cells were serum-starved for an additional 24 h, and the cells in fresh complete media were incubated for 48 h. The cells were then incubated in 0.5% bovine serum albumin (BSA)/phosphate-buffered saline (PBS) with or without 2-DG for 1 h. The relative fluorescence units (RFU) were measured at Ex/Em = 535/587 nm using a Fluoroskan microplate fluorometer (Thermo Fisher Scientific, San Jose, CA, USA). The uptake of 2-DG was calculated by the 2-DG-6-phosphate (2-DG6P) standard curve and the RFU of the samples. 2.9 Lactate measurement assay Lactate levels were measured using an L-Lactate Assay Kit (Abcam, ab169557). Cell extracts were specifically oxidized to form an intermediate that reacts with a colorless probe to generate fluorescence that was measured at 530 nm (excitation)/590 nm (emission) using a Fluoroskan microplate fluorometer (Thermo Fisher Scientific, San Jose, CA, USA). The intensity was directly proportional to the amount of lactate measured in nmol/ml. 2.10 Extracellular acidification rate (ECAR) assay The glycolysis assay (extracellular acidification rate, ECAR) was measured with a test kit (Abcam, ab197244). Specifically, the cells were incubated in a 96-well plate at a density of 5×10 4 cells/well, and 150 µL of respiration buffer was added before the test. The addition of 10 µL of carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) was used as a positive or negative control. Then, 10 µL of reconstituted glycolysis assay reagent was added, and the glycolysis assay signal was measured at 1.5 min intervals for ≥ 120 min using excitation and emission wavelengths of Ex/Em 380/615 nm, respectively, by using a Fluoroskan microplate fluorometer (Thermo Fisher Scientific, San Jose, CA, USA). 2.11 Wound healing assay The cells were placed in a six-well plate at a density of 2×10 5 /well. The cells were expected to grow to 80%~90% confluence. Cell monolayers were scratched with a 200 µL pipette tip to produce lesions of a consistent length and then cultured in basal medium. After the cells were washed with phosphate-buffered saline (PBS) to remove cellular fragments, each wound was imaged at 0, 12, and 24h by inversion microscopy. Cell migration was quantified by measuring the relative wound areas using ImageJ (National Institutes of Health, Bethesda, MD, USA). 2.12 Transwell invasion assay Matrigel (1:8 dilution of 50 mg/L) was coated on the upper surface of the bottom membrane of the transwell chamber (Corning, NY, USA) and allowed to dry at 4°C. The cells were resuspended in serum-free DMEM and seeded into the upper chamber. The bottom wells were filled with complete DMEM. The cells in the upper chamber were removed after 24 h of incubation. The cells that migrated through the Matrigel matrix membrane were fixed in 4% paraformaldehyde and stained with 0.1% crystal violet. Finally, representative images were obtained under a microscope, and the number of cells in the picture was calculated through ImageJ. 2.13 MTT Cell growth was evaluated using the MTT [3-(4,5-dimethyl-2-yl)-2,5-diphenyl tetrazolium bromide] assay. The cells were seeded in 96-well plates at a density of 3,000 cells/well and treated with GEM for 24 hours. Methyl thiazolyl diphenyl tetrazolium bromide (MTT; Sigma‒Aldrich, St. Louis, MO, USA) (5 mg/mL) was added to the medium at 0, 1, 2, 3, 4, and 5 days. After incubation at 37°C for 4 hours, the medium was removed, and 200 ml of dimethyl sulfoxide (DMSO, Sigma) was added to each well for dissolution. The absorbance of DMSO was measured at 490 nm using a microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). 2.14 Subcutaneous xenograft nude mouse model Four-week-old female BALB/c thymic nude mice (Hangzhou Ziyuan Experimental Animal Technology Co., Ltd., Zhejiang, China) were nurtured in the animal laboratory in a specific pathogen-free environment. Cells were injected into the left axilla of the nude mice. At the end of the experiment, the mice were anesthetized, and the tumors were excised. The length (L) and width (W) of each subcutaneous tumor were measured using calipers. The tumor volume (TV) was calculated as TC = (L × W 2 )/2. 2.15 Statistical analysis R version 3.5.1 or GraphPad Prism 8.0 was used for all the statistical analyses. There were a minimum of three runs of each experiment. The results are shown as the average ± standard deviation (SD). 3 Results 3.1 Characteristics of GRGs in GC The expression of 326 glycolysis-related genes (GRGs) in normal tissues and in the TCGA-STAD cohort was visualized via a heatmap (Fig. 1 A) and a volcano map (Fig. 1 B). Somatic copy number alterations were detected in a total of 73 differentially expressed genes (DEGs). A decrease in CNV (copy number variation) was detected in STMN2, NUP205, VCAN, HS2S11, DCN, and LDHAL6B, while CNV increased in VEGFA, ALDOC, EFNA3, COL5A1, NIP188, and NUP155 (Fig. 1 C). The sites of CNV abnormalities in the DEGs are presented in the circle diagram (Fig. 1 D). GO and KEGG pathway enrichment analyses were used to predict the potential biological effects of these DEGs. GO analysis revealed that the DEGs were clustered into metabolic GO categories such as nucleic acid transport, RNA transport, mRNA transportation, RNA localization, and nucleobase-containing molecule transport (Fig. 1 E-F). KEGG analysis revealed that genes associated with glycolysis, fructose and mannose metabolism, nucleocytoplasmic transport, pyruvate metabolism, and the HIF-1 signaling pathway were most prevalent (Fig. 1 G-H). 3.2 Subtype identification based on the expression patterns of GRGs in GC To evaluate whether the prognosis of patients with GC was associated with the expression of GRGs, these 326 glycolysis-related genes (GRGs) were subjected to consensus cluster analysis. The consensus matrix heatmap presented a distinct boundary and low interference when k = 2, indicating that the samples could be stably clustered according to the expression patterns of the GRGs (Fig. 2 A-C). Two different groups, namely, cluster A (n = 291) and cluster B (n = 43), were obtained. The differences in survival between these two clusters were then assessed. By using gene set variation analysis (GSVA) (Fig. 2 D) and gene set enrichment analysis (GSEA) (Fig. 2 E and F) of these two groups, we found that the crucial pathways in cluster A were mainly enriched in metabolic-related pathways, while the corresponding pathways enriched in cluster B were immune-related pathways. Therefore, cluster A was named the metabolic subtype, and cluster B was termed the immune subtype. Furthermore, we found that the prognosis of patients with the metabolic subtype (cluster A) was better than that of patients with the immune subtype (cluster B) ( P = 0.016, Fig. 2 G). 3.3 Construction of a glycolysis-related risk model based on the GRG signature Since the expression of GRGs was related to survival, we developed a signature model according to the expression of GRGs in the TCGA-STAD cohort to predict the prognosis of GC patients. First, univariate Cox analysis was used to identify OS-related GRGs, and ten genes were identified (Fig. 3 A). Then, LASSO regression was performed to avoid overfitting (Fig. 3 B-C), and six genes, namely, ME1, PLOD2, NUP50, CXCR4, SLC35A3 and SRD5A3, were screened out. Finally, a predictive model was constructed according to the expression of these six genes, and the corresponding regression coefficients were calculated. Risk Score was determined by using the following formula: Risk Score = (-0.057 × expression of ME1) + (0.062 × expression of PLOD2) + (-0.088 × expression of NUP50) + (0.007 × expression of CXCR4) + (-0.050 × expression of SLC35A3) + (0.026 × expression of SRD5A3). The risk score of each patient was calculated, and the TCGA-STAD cohort was divided into high- and low-risk groups based on the median risk score. The expression of the six genes in the high- and low-risk groups was visualized using a heatmap (Fig. 3 D). The risk scores of each patient are depicted in Fig. 3 E. Patients in the high-risk group had shorter OS (Fig. 3 F). Principal component analysis (PCA) revealed that the high- and low-risk groups were well separated from each other (Fig. 3 G). According to the Kaplan‒Meier survival curves, patients in the low-risk group had a better prognosis ( P < 0.05, Fig. 3 H). Furthermore, a receiver operating characteristic (ROC) curve was drawn to verify the performance of the model (Fig. 3 I). The area under the curve (AUC) values were 0.712 (1-year), 0.689 (3-year), and 0.681 (5-year), indicating the good sensitivity and specificity of this risk model. Furthermore, this risk model was validated by using the GSE84437 dataset. Patients in the validation cohort were divided into low- and high-risk groups based on the median risk score obtained from the TCGA-STAD cohort. The expression of the six genes in the validation set was depicted using a heatmap (Fig. 3 J). The risk scores of each patient and survival time distribution are presented in Fig. 3 K and L. Consistent with the results of the TCGA-STAD cohort, patients in the low-risk group of the validation set also had a better prognosis (Fig. 3 M). The AUC values in the validation set were 0.742 (1-year), 0.738 (3-year), and 0.740 (5-year) (Fig. 3 N). According to the above analysis, the glycolysis-related risk model based on the 6-GRG signature is a promising model for predicting the prognosis of GC patients. 3.4 Establishment of a nomogram according to the glycolysis-related risk model and other risk factors By using univariate Cox analysis, the 6-GRG signature risk score was found to be an independent prognostic factor for GC patients in the TCGA-STAD cohort (HR = 1.263, P = 0.001; Fig. 4 A). Multivariate Cox analysis revealed that the risk score remained an independent predictor after considering clinical characteristics (HR = 1.262, P = 0.001; Fig. 4 B). Multivariate ROC analysis was used to compare the predictive accuracy of the risk score with that of various clinical factors. Overall, the risk score was the most accurate prognostic indicator (AUC = 0.658, Fig. 4 C). Consistent with poorer prognosis in cluster B (immune subtype), subjects in cluster B were located more frequently in the high-risk group (Fig. 4 D). Moreover, correlations were found between the risk score and other factors, including age, grade, and M stage (Fig. 4 D). Finally, a nomogram (Fig. 4 E) incorporating several risk factors was developed to estimate 1-, 3-, and 5-year OS. The calibration curves showed excellent agreement between the probabilities of 1-, 3-, and 5-year OS predicted by the nomogram and the actual OS rates of patients in the TCGA-STAD cohort (Fig. 4 F). 3.5 CXCR4 promoted the metastasis and growth of GC cells To analyze the characteristics of the six genes included in the risk model, we investigated the mutation profiles of the six GRGs by using the "maftools" R package [ 24 ]. As shown in Fig. 5 A, mutations were not frequent in these GRGs. The mutation frequencies of PLOD2, SRD5A3, SLC35A3, NUP50, ME1 and CXCR4 were 3%, 1%, 1%, 1%, 0% and 0%, respectively. A heatmap was generated to show how the risk score and the expression of the six GRGs are related to each other (Fig. 5 B). ME1, PLOD2, and CXCR4 were highly expressed in the high-risk group, while NUP50 and SLC35A3 were highly expressed in the low-risk group (Fig. 5 C). K‒M analysis was then subsequently performed (Fig. 5 D-I). High expression of CXCR4 ( P = 0.002) and PLOD2 ( P = 0.043) indicated a relatively poor prognosis (Fig. 5 E and G), and high expression of NUP50 ( P = 0.018) and SLC35A3 ( P = 0.043) indicated a better prognosis (Fig. 5 F and H). However, the expression of SRD5A3 and ME1 was not related to prognosis (Fig. 5 D and I). Since CXCR4 was the most significantly related gene to prognosis among these six GRGs, we further focused on the biological functions of CXCR4 in GC. CXCR4 expression was greater in tumor tissues than in normal tissues according to data in the TCGA-STAD cohort (Fig. 5 J). Knockdown of CXCR4 in BGC-823 cells (Fig. 5 K) was associated with reduced glucose consumption (Fig. 5 L), lactate production (Fig. 5 M) and ECAR (Fig. 5 N-O). MTT assays showed that the downregulation of CXCR4 significantly inhibited the proliferation of BGC-823 cells (Fig. 5 P). Subcutaneous xenograft nude mouse models further confirmed that the growth of tumors was weakened by knocking down CXCR4 in vivo (Fig. 5 Q and R). The wound healing assay showed that cell migration was repressed after CXCR4 was knocked down (Fig. 5 S). Transwell invasion assays showed that the downregulation of CXCR4 weakened the invasion ability of BGC-823 cells (Fig. 5 T). 3.6 The expression of ME1 and CXCR4 predicted immune cell infiltration and response to immunotherapy Several studies have reported that metabolic reprogramming of cancer cells leads to the development of an immunosuppressive tumor microenvironment [ 9 ]. Since GRGs participate in the progression of GC, we further investigated whether the GRGs were related to the modeling of the immune microenvironment. The ImmuneScore was calculated and compared between metabolic subtype (cluster A) and immune subtype (cluster B) or between low-risk and high-risk groups by using the "ESTIMATE" R package [ 23 ]. Surprisingly, although patients in cluster B and the high-risk group had worse prognosis in TCGA-STAD cohort, both had significantly greater ImmuneScore (Fig. 6 A and B), suggesting that patients in the immune subtype (cluster B) or high-risk group might have a favorable immune microenvironment. To confirm this hypothesis, differences in the expression levels of immunological checkpoint molecules were compared between clusters A and B. Cluster B samples mainly expressed immunological checkpoint molecules, including CD274 (also known as PD-L1), at higher levels than did cluster A samples (Fig. 6 C). Similarly, the expression of most immune checkpoint molecules, including PD-L1, was greater in the high-risk group than in the low-risk group (Fig. 6 D). The 22 immune cell subsets were then analyzed by using the CIBERSORT algorithm [ 25 ]. M2 macrophages, activated mast cells, and eosinophils were found at higher levels in cluster A samples than in cluster B samples. In comparison, the proportions of memory B cells, CD8 + T cells, regulatory T cells (Tregs), and naive B cells were greater in cluster B samples (Fig. 6 E). Activated CD4 + memory T cells, follicular helper T cells, and M1 macrophages were more enriched in the low-risk group, while there was more infiltration of regulatory T cells (Tregs), monocytes, M2 macrophages, and resting dendritic cells in the high-risk group (Fig. 6 F). Cluster B and the high-risk group preserved a favorable immune microenvironment and increased expression of immune checkpoints, including PD-L1, suggesting that GRGs might indicate the response to immunotherapy. To confirm this hypothesis, we used RNA-Seq data from the immunotherapy cohort PRJEB25780 from the European Nucleotide Archive (ENA) ( https://www.ebi.ac.uk/ena/browser/home ). In this cohort, researchers examined 45 patients with metastatic or recurrent gastric cancer who received treatment with anti-programmed cell death protein 1 (PD-1) [ 16 ]. The treatment responses were classified as follows: complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). Patients were divided into two groups according to their treatment response: those with a confirmed response (CR/PR) and those without a clinical response (PD/SD). The expression levels of PLOD2, SRD5A3, SLC35A3, NUP50, ME1 and CXCR4 were then compared between these two groups (Fig. 6 G). The expression of ME1 was significantly greater in the group without a response, while the expression of CXCR4 was greater in the group with a confirmed response (Fig. 6 G). The ROC curve was then used to evaluate the prognostic performance of the expression of the GRGs (Fig. 6 H-O). The areas under the ROC curves for ME1 and CXCR4 were 0.727 and 0.687, respectively (Fig. 6 J and M). The expression of PLOD2 (Fig. 6 L), SRD5A3 (Fig. 6 N), SLC35A3 (Fig. 6 I) and NUP50 (Fig. 6 K), as well as the 6-GRG signature risk score (Fig. 6 H), indicated no difference. Moreover, we developed a score based on the expression of CXCR4 and ME1 (score = 0.189 + 2.516×CXCR4-3.351×ME1) (Fig. 6 O). The area under the ROC curve for this score was 0.814, suggesting that the score worked more effectively than CXCR4 or ME1 alone (Fig. 6 O). Based on the optimal threshold of expression levels of CXCR4 and ME1 or the ME1/CXCR4 score for the maximum ROC curve values, the patients were then dichotomized into high- and low-risk subgroups (Fig. 6 P-R). The expression of CXCR4 (Fig. 6 P) and ME1 (Fig. 6 Q), as well as the ME1/CXCR4 score (Fig. 6 R), significantly correlated with the response to immunotherapy. To investigate the mechanisms involved in the predictive value of the expression of CXCR4 and ME1 upon immunotherapy, we evaluated the immune microenvironment and immune cell infiltration according to the expression of CXCR4 and ME1. The expression of CXCR4 was positively correlated with the ImmuneScore, while the expression of ME1 was negatively correlated with the ImmuneScore (Fig. 6 S). Consistently, higher expression of CXCR4 predicted more infiltration of CD8 + T cells, while expression of ME1 presented a negative correlation with infiltration of CD8 + T cells (Fig. 6 T). Therefore, the high expression of CXCR4 and low expression of ME1 suggested a favorable immune microenvironment. 4 Discussion Metabolic reprogramming is one of the hallmarks of cancer [ 26 ]. To sustain continuous replication and high proliferation rates, metabolic reprogramming of cancer cells must switch their metabolic reprogramming to a ‘glycolysis-dominant’ metabolic profile. This shift promotes cell survival by meeting energy, synthesis, and redox demands, and it also renders the tumor microenvironment more conducive to cancer progression. [ 27 , 28 ]. The classic example of metabolic switching in cancer cells was first reported by Otto Warburg in the 1920s. This phenomenon, which was subsequently named the Warburg effect, was characterized by a shift in the metabolic machinery of cancer cells toward glycolysis and lactic acid fermentation from oxidative metabolism even under normoxic conditions [ 29 ]. Under physiological conditions in aerobic environments, cells metabolize glucose through glycolysis, the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS). Normal cells preferentially perform oxidation through OXPHOS because it produces more adenosine triphosphate (ATP). However, the limited oxygen supply in the hypoxic tumor microenvironment causes cancer cells to prefer aerobic glycolysis [ 30 ]. Therefore, cancer cells exhibit increased glycolytic dependency and exhibit increased glucose uptake and fermentation of glucose to lactate to meet the increased anabolic requirements for a malignant phenotype [ 31 ]. The Warburg effect has been investigated for approximately a century, and multiple studies have elucidated the mechanisms governing the increased glycolytic dependency of cancer cells. In addition to energy supply, this aberrant energy utilization mode in cancers critically influences the tumor microenvironment (TME). It does so by creating a hypoxic, acidic, nutrient-poor, and immune-modulatory metabolite-rich environment (e.g., lactate, kynurenine), which fosters a favorable tumor niche and impairs effective cancer treatment [ 30 , 32 ]. Given the usefulness of these metabolic intermediates of glycolysis, cancer cells tend to maintain the Warburg effect through mutation. Several oncogenic proteins and pathways, including hypoxia-inducible factor (HIF-1) and the PI3K/Akt/mTOR pathway, have been implicated in regulating cancer cell-specific metabolic reprogramming [ 9 ]. Given the critical role of the Warburg effect in the progression of cancer, increasing research has been dedicated to revealing the mechanisms involved and identifying novel targets for clinical application. In the present study, we elucidated global alterations in glycolysis-related genes (GRGs) at the transcriptional and genetic levels in GC. Based on 326 GRGs, we identified two distinct molecular subtypes. Patients in the immune subtype (cluster B) had more advanced clinicopathological characteristics and shorter OS than did those in the metabolic subtype (cluster A). Furthermore, both univariate and multivariate Cox regression analyses revealed the prognostic value of a risk core based on the expression of six GRGs, namely, ME1, PLOD2, NUP50, CXCR4, SLC35A3, and SRD35A3. Survival analysis revealed that the high-risk group had a worse prognosis. These results suggested that the GRG signature could be helpful in guiding clinical treatment decisions. The deviation of intermediate molecules of glycolysis and the TCA cycle for the synthesis of nucleotides, lipids and nonessential amino acids is pivotal for generating the products necessary for cell proliferation and survival [ 30 ]. Additionally, increased glycolysis also triggers chemo- and radio-resistance, suggesting that therapeutics targeting elevated glycolysis in cancer cells could be a promising treatment modality to increase the sensitivity of cancer cells to treatment via other conventional strategies [ 9 ]. Although several therapeutics targeting enhanced glycolysis in cancer cells have been developed and several of these strategies are in different stages of preclinical and clinical investigations, there is still a lack of effective solutions specifically targeting cancer cell glycolysis [ 9 ]. In our present study, we showed that CXCR4 might be a promising target for strategies against elevated glycolysis in GC. CXCR4 is one of the most prevalent chemokine receptors and is also known as CD184. CXCR4 has been linked to the occurrence and progression of a variety of cancers, including breast cancer, melanoma, prostate cancer, and gastric cancer [ 33 – 36 ]. CXCR4 expression is also associated with lymph node metastasis in GC [ 37 ]. CXCR4 is a G protein-coupled receptor that binds to the CXCR4 ligand CXC ligand 12 (CXCL12, stromal cell-derived factor-1, SDF-1). CXCL12 is generated by stromal cells and released into the TME, where it combines with CXCR4 on the tumor surface. In GC, cancer-associated fibroblasts influence the CXCL12/CXCR4 axis [ 38 ]. The CXCL12/CXCR4 axis activates the MAPK cascade, which promotes chemotaxis and cell proliferation [ 39 ], and stimulates the PLC/PKC and PI3K/Akt signaling pathways, enhancing cell migration and survival [ 40 ]. In the present study, we demonstrated that knockdown of CXCR4 suppressed the migration and invasion of GC cells. Moreover, the downregulation of CXCR4 inhibited the growth of GC cells both in vitro and in vivo , suggesting that CXCR4 could be a potential target for further drug development. Lactate, the end product of aerobic glycolysis, decreases the pH of the TME, facilitating the invasion and metastasis of cancer cells [ 31 , 41 ]. Lactate is also metabolized in human tumors in vivo , fueling the TCA cycle and contributing to energy production [ 42 ]. Previous studies have shown that CXCR4 participates in the Warburg effect in acute myeloid leukemia (AML) cells via the CXCL12/CXCR4/mTOR pathway [ 43 ]. The mTOR inhibitor rapamycin significantly suppressed the upregulation of glucose transporters in the matrix, enhanced glucose influx, and decreased lactate generation [ 43 ]. Our present study revealed that knockdown of CXCR4 in GC cells reduced glycolytic capability, glucose intake, and lactate generation. Therefore, the anticancer effect of CXCR4 knockdown could be mediated through repression of the Warburg effect. Interestingly, although both the immune subtype (cluster B) and the high-risk group identified in the present study by using the TCGA database were associated with an unfavorable prognosis, these two cohorts presented a favorable immune microenvironment and increased expression of immune checkpoints, suggesting that the application value of GRGs still needs further elucidation. The reason for this inconsistency may be that the data from the TCGA-STAD cohort were collected mainly from before the era of immunotherapy. Without immune checkpoint inhibitor treatment, increased expression of immune checkpoints suggests immune escape, resulting in unfavorable outcomes [ 44 ]. Immune cells in the TME share a similar switch to a glycolytic metabolic profile, inducing competition between cancer cells and tumor-infiltrating cells over nutrients [ 9 ]. The metabolic dysregulation of the tumor microenvironment can halt the infiltration of immune cells and inhibit antitumor immunity through the production of immune-suppressive byproducts. Furthermore, tumor-derived lactate plays direct and indirect immune-suppressive roles, such as accelerating the accumulation of Tregs and MDSCs, as well as the polarization of M2 tumor-associated macrophages [ 45 , 46 ]. Increased tumor glycolysis leads to resistance to adoptive T cell therapy in melanoma, suggesting the favorable benefits of dual targeting cancer immunity and metabolism [ 47 ]. In addition, metabolic reprogramming in cancer cells can also alter the expression of several cell surface markers, which interferes with immune surveillance [ 30 ]. Glycolytic activity can upregulate immune checkpoint expression, suggesting that glycolysis might promote immunotherapy response [ 48 ] and could be a useful indicator for immunotherapy efficacy. Immune checkpoint inhibitors (ICIs) have revolutionized therapeutic models for several types of tumors. The best-studied such inhibitors are antibodies that block the cytotoxic T-lymphocyte antigen-4 (CTLA-4) and programmed cell death protein-1 (PD-1) proteins. Treatment with anti-CTLA-4 and anti-PD-1 agents led to significantly improved survival in patients with various tumor types [ 30 ]. The phase 2 KEYNOTE-059 trial suggested the safety and effectiveness of pembrolizumab as a third-line treatment for locally advanced or metastatic gastric or gastroesophageal junction (G/GEJ) adenocarcinoma [ 49 ]. In this single-arm, multicohort trial, the objective response rate (ORR) was 11.6%, and complete responses (CRs) were noted in 2.3% of patients [ 49 ]. The phase 3 CheckMate-649 trial demonstrated the acceptable safety profile and efficacy of nivolumab as a first-line treatment for patients with untreated, unresectable, human epidermal growth factor receptor 2 (HER2)-negative G/GEJ cancer or esophageal adenocarcinoma [ 50 ]. The CheckMate-649 trial demonstrated significant improvements in overall survival (OS) and progression-free survival (PFS) for patients with a combined positive score (CPS) ≥ 5 [ 50 ]. Based on this progress, immune checkpoint inhibitors have become standard therapeutics for the treatment of GC. The success of immunotherapy relies on its potential to “rewire” the immune cycle through the generation of long-lasting immune responses. However, not all patients respond to this type of intervention [ 30 ]. Therefore, it is important to identify target biomarkers for predicting response to ICIs. For now, PD-L1 immunohistochemistry (IHC), microsatellite instability (MSI)/mismatch repair (MMR), Epstein-Barr virus (EBV), and tumor mutational burden (TMB) have been considered promising predictive biomarkers for the response to ICIs in patients with GC [ 51 ]. However, no single biomarker can be used as a reproducible proxy for predicting the effects of immunotherapy [ 51 ]. Hence, developing an integrated predictive model that considers the complex components affecting host-TME interactions by reflecting the heterogeneity of GC is necessary. In our present study, we showed that GRGs exhibit potential for predicting response to ICIs. The expression of CXCR4 and ME1 was correlated with the infiltration of CD8 + T cells and the response to treatment with an anti-PD-1 ICI. Moreover, the combination of CXCR4 and ME1 expression had a more effective prognostic prediction effect. Interestingly, although increased CXCR4 expression indicates a poor prognosis when ICIs are not used, CXCR4 could be a promising biomarker for the response to ICIs. Therefore, the prognostic value of specific subjects should consider different historical backgrounds. The roles of GRGs in the age of immunotherapy still require further investigation, and how to best utilize glycolytic targets to boost antitumor immunity and improve immunotherapies still needs further elucidation. This work revealed that GRGs play a nonnegligible role in determining TME diversity and complexity. Evaluating the GRG modification pattern of individual tumors will contribute to enhancing our understanding of TME infiltration and guide more effective immunotherapy strategies. Declarations • Funding: This work was supported by the National Natural Science Foundation of China (grant number: 81873587). • Conflict of interest: The authors declare no conflicts of interest in this study or manuscript. • Data availability: Data of the transcriptome and clinical information were obtained from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/) [13]. The expression data of 407 samples included 32 normal tissues and 375 tumor tissues. Concurrently, the GSE84437 dataset [14] and the GSE13763 dataset [15], which include microarray data and clinical data, were acquired from the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) and used as validation cohort. Data of immunotherapy cohort PRJEB25780 were downloaded from The European Nucleotide Archive (ENA) (https://www.ebi.ac.uk/ena/browser/home) [16]. All the datasets presented in this study can be found in online repositories. Author Contribution Lu Xu, Jin Liu and Yuanqing An: Data curation (equal); Investigation (equal); Methodology (equal); Writing – original draft (equal). Lei Zhou and Hui Sun: Data curation (equal); Formal analysis (equal). Zhen Xu, Deqiang Wang and Zhanwen Liang: Data curation (equal); Investigation (equal). Caihua Xu and Bingyi Wang: Data curation (equal); Investigation (equal); Resources (equal); Supervision (equal); Writing – review & editing (equal). Wei Li: Conceptualization (equal); Data curation (equal); Investigation (equal); Project administration (equal); Writing – review & editing (equal). References Thrift AP, El-Serag HB. Burden of Gastric Cancer. Clin Gastroenterol Hepatol. 2020;18(3):534–42. 10.1016/j.cgh.2019.07.045 . Zhou J, Ma X, Bi F, Liu M. Clinical significance of circulating tumor cells in gastric cancer patients. Oncotarget. 2017;8(15):25713–20. 10.18632/oncotarget.14879 . Correa P. Gastric cancer: overview. Gastroenterol Clin North Am. 2013;42(2):211–7. 10.1016/j.gtc.2013.01.002 . Bonomo P, Merlotti A, Olmetto E, et al. What is the prognostic impact of FDG PET in locally advanced head and neck squamous cell carcinoma treated with concomitant chemo-radiotherapy? A systematic review and meta-analysis. Eur J Nucl Med Mol Imaging. 2018;45(12):2122–38. 10.1007/s00259-018-4065-5 . Cascone T, McKenzie JA, Mbofung RM, et al. Increased Tumor Glycolysis Characterizes Immune Resistance to Adoptive T Cell Therapy. Cell Metab. 2018;27(5):977–e9874. 10.1016/j.cmet.2018.02.024 . Chang CH, Qiu J, O'Sullivan D, et al. Metabolic Competition in the Tumor Microenvironment Is a Driver of Cancer Progression. Cell. 2015;162(6):1229–41. 10.1016/j.cell.2015.08.016 . Justus CR, Sanderlin EJ, Yang LV. Molecular Connections between Cancer Cell Metabolism and the Tumor Microenvironment. Int J Mol Sci. 2015;16(5):11055–86. 10.3390/ijms160511055 . Published 2015 May 15. Chen S, Cao G, Wu W, et al. Mining novel cell glycolysis related gene markers that can predict the survival of colon adenocarcinoma patients. Biosci Rep. 2020;40(8):BSR20201427. 10.1042/BSR20201427 . Chelakkot C, Chelakkot VS, Shin Y, Song K. Modulating Glycolysis to Improve Cancer Therapy. Int J Mol Sci . 2023;24(3):2606. Published 2023 Jan 30. 10.3390/ijms24032606 . Jiang L, Zhao L, Bi J, et al. Glycolysis gene expression profilings screen for prognostic risk signature of hepatocellular carcinoma. Aging. 2019;11(23):10861–82. 10.18632/aging.102489 . Chen C, Shi Y, Li Y, et al. A glycolysis-based ten-gene signature correlates with the clinical outcome, molecular subtype and IDH1 mutation in glioblastoma. J Genet Genomics. 2017;44(11):519–30. 10.1016/j.jgg.2017.05.007 . Wang ZH, Zhang YZ, Wang YS, Ma XX. Identification of novel cell glycolysis related gene signature predicting survival in patients with endometrial cancer. Cancer Cell Int. 2019;19:296. 10.1186/s12935-019-1001-0 . Hoadley KA, Yau C, Hinoue T, et al. Cell-of-Origin Patterns Dominate the Molecular Classification of 10,000 Tumors from 33 Types of Cancer. Cell. 2018;173(2):291–e3046. 10.1016/j.cell.2018.03.022 . Yoon SJ, Park J, Shin Y et al. Deconvolution of diffuse gastric cancer and the suppression of CD34 on the BALB/c nude mice model. BMC Cancer . 2020;20(1):314. Published 2020 Apr 15. 10.1186/s12885-020-06814-4 . Kulbe H, Chakravarty P, Leinster DA, et al. A dynamic inflammatory cytokine network in the human ovarian cancer microenvironment. Cancer Res. 2012;72(1):66–75. 10.1158/0008-5472.CAN-11-2178 . Kim ST, Cristescu R, Bass AJ, et al. Comprehensive molecular characterization of clinical responses to PD-1 inhibition in metastatic gastric cancer. Nat Med. 2018;24(9):1449–58. 10.1038/s41591-018-0101-z . Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545–50. 10.1073/pnas.0506580102 . Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26(1):139–40. 10.1093/bioinformatics/btp616 . Wu T, Hu E, Xu S et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. Innovation (Camb) . 2021;2(3):100141. Published 2021 Jul 1. 10.1016/j.xinn.2021.100141 . Wilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010;26(12):1572–3. 10.1093/bioinformatics/btq170 . Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw. 2010;33(1):1–22. Cao M, Cai J, Yuan Y, et al. A four-gene signature-derived risk score for glioblastoma: prospects for prognostic and response predictive analyses. Cancer Biol Med. 2019;16(3):595–605. 10.20892/j.issn.2095-3941.2018.0277 . Heagerty PJ, Lumley T, Pepe MS. Time-dependent ROC curves for censored survival data and a diagnostic marker. Biometrics. 2000;56(2):337–44. 10.1111/j.0006-341x.2000 . 00337.x. Hope C, Foulcer S, Jagodinsky J, et al. Immunoregulatory roles of versican proteolysis in the myeloma microenvironment. Blood. 2016;128(5):680–5. 10.1182/blood-2016-03-705780 . Newman AM, Liu CL, Green MR, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12(5):453–7. 10.1038/nmeth.3337 . Hanahan D. Hallmarks of Cancer: New Dimensions. Cancer Discov. 2022;12(1):31–46. 10.1158/2159-8290.CD-21-1059 . DeBerardinis RJ, Chandel NS. Fundamentals of cancer metabolism. Sci Adv . 2016;2(5): e1600200. Published 2016 May 27. 10.1126/sciadv.1600200 . Pavlova NN, Thompson CB. The Emerging Hallmarks of Cancer Metabolism. Cell Metab. 2016;23(1):27–47. 10.1016/j.cmet.2015.12.006 . WARBURG O. On the origin of cancer cells. Science. 1956;123(3191):309–14. 10.1126/science.123.3191.309 . Ramapriyan R, Caetano MS, Barsoumian HB, et al. Altered cancer metabolism in mechanisms of immunotherapy resistance. Pharmacol Ther. 2019;195:162–71. 10.1016/j.pharmthera.2018.11.004 . Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029–33. 10.1126/science.1160809 . DePeaux K, Delgoffe GM. Metabolic barriers to cancer immunotherapy. Nat Rev Immunol. 2021;21(12):785–97. 10.1038/s41577-021-00541-y . Yang F, Takagaki Y, Yoshitomi Y, et al. Inhibition of Dipeptidyl Peptidase-4 Accelerates Epithelial-Mesenchymal Transition and Breast Cancer Metastasis via the CXCL12/CXCR4/mTOR Axis. Cancer Res. 2019;79(4):735–46. 10.1158/0008-5472.CAN-18-0620 . Mannavola F, Tucci M, Felici C, Passarelli A, D'Oronzo S, Silvestris F. Tumor-derived exosomes promote the in vitro osteotropism of melanoma cells by activating the SDF-1/CXCR4/CXCR7 axis. J Transl Med . 2019;17(1):230. Published 2019 Jul 19. 10.1186/s12967-019-1982-4 . Miwa S, Mizokami A, Keller ET, Taichman R, Zhang J, Namiki M. The bisphosphonate YM529 inhibits osteolytic and osteoblastic changes and CXCR-4-induced invasion in prostate cancer. Cancer Res. 2005;65(19):8818–25. 10.1158/0008-5472.CAN-05-0540 . Yu C, Zhang Y. Characterization of the prognostic values of CXCR family in gastric cancer. Cytokine. 2019;123:154785. 10.1016/j.cyto.2019.154785 . Lee HJ, Kim SW, Kim HY, et al. Chemokine receptor CXCR4 expression, function, and clinical implications in gastric cancer. Int J Oncol. 2009;34(2):473–80. Qin Y, Wang F, Ni H, et al. Cancer-associated fibroblasts in gastric cancer affect malignant progression via the CXCL12-CXCR4 axis. J Cancer. 2021;12(10):3011–23. 10.7150/jca.49707 . Published 2021 Mar 19. Kukreja P, Abdel-Mageed AB, Mondal D, Liu K, Agrawal KC. Up-regulation of CXCR4 expression in PC-3 cells by stromal-derived factor-1alpha (CXCL12) increases endothelial adhesion and transendothelial migration: role of MEK/ERK signaling pathway-dependent NF-kappaB activation. Cancer Res. 2005;65(21):9891–8. 10.1158/0008-5472.CAN-05-1293 . Hinton CV, Avraham S, Avraham HK. Role of the CXCR4/CXCL12 signaling axis in breast cancer metastasis to the brain. Clin Exp Metastasis. 2010;27(2):97–105. 10.1007/s10585-008-9210-2 . Gatenby RA, Gillies RJ. Why do cancers have high aerobic glycolysis? Nat Rev Cancer. 2004;4(11):891–9. 10.1038/nrc1478 . Martínez-Reyes I, Chandel NS. Waste Not, Want Not: Lactate Oxidation Fuels the TCA Cycle. Cell Metab. 2017;26(6):803–4. 10.1016/j.cmet.2017.11.005 . Braun M, Qorraj M, Büttner M, et al. CXCL12 promotes glycolytic reprogramming in acute myeloid leukemia cells via the CXCR4/mTOR axis. Leukemia. 2016;30(8):1788–92. 10.1038/leu.2016.58 . Zhou AL, Wang X, Yu W, et al. Expression level of PD-L1 is involved in ALDH1A1-mediated poor prognosis in patients with head and neck squamous cell carcinoma. Pathol Res Pract. 2020;216(9):153093. 10.1016/j.prp.2020.153093 . Colegio OR, Chu NQ, Szabo AL, et al. Functional polarization of tumour-associated macrophages by tumour-derived lactic acid. Nature. 2014;513(7519):559–63. 10.1038/nature13490 . Husain Z, Huang Y, Seth P, Sukhatme VP. Tumor-derived lactate modifies antitumor immune response: effect on myeloid-derived suppressor cells and NK cells. J Immunol. 2013;191(3):1486–95. 10.4049/jimmunol.1202702 . Cascone T, McKenzie JA, Mbofung RM, et al. Increased Tumor Glycolysis Characterizes Immune Resistance to Adoptive T Cell Therapy. Cell Metab. 2018;27(5):977–e9874. 10.1016/j.cmet.2018.02.024 . Jiang Z, Liu Z, Li M, Chen C, Wang X. Increased glycolysis correlates with elevated immune activity in tumor immune microenvironment. EBioMedicine. 2019;42:431–42. 10.1016/j.ebiom.2019.03.068 . Fuchs CS, Doi T, Jang RW et al. Safety and Efficacy of Pembrolizumab Monotherapy in Patients with Previously Treated Advanced Gastric and Gastroesophageal Junction Cancer: Phase 2 Clinical KEYNOTE-059 Trial [published correction appears in JAMA Oncol. 2019;5(4):579]. JAMA Oncol . 2018;4(5):e180013. 10.1001/jamaoncol.2018.0013 . Janjigian YY, Shitara K, Moehler M, et al. First-line nivolumab plus chemotherapy versus chemotherapy alone for advanced gastric, gastro-oesophageal junction, and oesophageal adenocarcinoma (CheckMate 649): a randomised, open-label, phase 3 trial. Lancet. 2021;398(10294):27–40. 10.1016/S0140-6736(21)00797-2 . Kim M, Jeong JY, Seo AN. Biomarkers for Predicting Response to Personalized Immunotherapy in Gastric Cancer. Diagnostics (Basel). 2023;13(17):2782. 10.3390/diagnostics13172782 . Published 2023 Aug 28. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 08 Aug, 2024 Read the published version in BMC Cancer → Version 1 posted Editorial decision: Revision requested 28 Mar, 2024 Submission checks completed at journal 27 Mar, 2024 Editor assigned by journal 27 Mar, 2024 First submitted to journal 19 Mar, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4130368","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":284937606,"identity":"f89d4c5a-c4c9-41fe-8b96-92b593cd422d","order_by":0,"name":"Lu Xu","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Lu","middleName":"","lastName":"Xu","suffix":""},{"id":284937607,"identity":"2a6564e8-c664-4dd1-98c4-018dfe17c65b","order_by":1,"name":"Jin Liu","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Jin","middleName":"","lastName":"Liu","suffix":""},{"id":284937608,"identity":"0e78bc2d-08e4-4864-9b08-fef6aed96f83","order_by":2,"name":"Yuanqing An","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Yuanqing","middleName":"","lastName":"An","suffix":""},{"id":284937609,"identity":"911ecfd0-250e-4ba5-8f9e-c0086f25002f","order_by":3,"name":"Lei Zhou","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Lei","middleName":"","lastName":"Zhou","suffix":""},{"id":284937610,"identity":"c52ca687-3f90-4174-9cc5-deb480acec08","order_by":4,"name":"Hui Sun","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Hui","middleName":"","lastName":"Sun","suffix":""},{"id":284937611,"identity":"f6e01c85-51b6-4796-8187-525c4fe5d313","order_by":5,"name":"Zhen Xu","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Zhen","middleName":"","lastName":"Xu","suffix":""},{"id":284937612,"identity":"906daf7f-4095-4885-8f51-86f7774510fe","order_by":6,"name":"Deqiang Wang","email":"","orcid":"","institution":"Affiliated Hospital of Jiangsu University","correspondingAuthor":false,"prefix":"","firstName":"Deqiang","middleName":"","lastName":"Wang","suffix":""},{"id":284937613,"identity":"6702d02b-4e60-4c97-897f-c6020fb8f86b","order_by":7,"name":"Zhanwen Liang","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Zhanwen","middleName":"","lastName":"Liang","suffix":""},{"id":284937614,"identity":"1a789399-c516-495f-aaf4-6d4c83691471","order_by":8,"name":"Caihua Xu","email":"","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Caihua","middleName":"","lastName":"Xu","suffix":""},{"id":284937615,"identity":"264a2a30-3766-46d3-9dc5-97c4b0191121","order_by":9,"name":"Bingyi Wang","email":"","orcid":"","institution":"Changshu No.1 People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Bingyi","middleName":"","lastName":"Wang","suffix":""},{"id":284937616,"identity":"e3094b43-ad82-49f6-8016-96c4b3422160","order_by":10,"name":"Wei Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzUlEQVRIiWNgGAWjYBACPmYgkWAgIQekDICYmbAWNrCWAhtjErSAyQ9piQ3Ea2HnMXzwwOBw+vwZyRs/MFRYJzawnz1AwGE8xgYJBodzN9xIK5ZgOJOe2MCTl0BAC+82CbAWiRwDCca2w4kNEjwGhLRs/wHUki4/I8f4B+M/4rRsAwZyWgLDjRwzCcYGorTwfwY6zMZww5lnZRYJx9KN23hy8Gvh5z+W+PHHHwl5+fbkzTc+1FjL9rOfwa8FFSQwwGJqFIyCUTAKRgFFAABOsz67nZ3wPgAAAABJRU5ErkJggg==","orcid":"","institution":"the First Affiliated Hospital of Soochow University","correspondingAuthor":true,"prefix":"","firstName":"Wei","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2024-03-19 12:42:50","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4130368/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4130368/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s12885-024-12747-z","type":"published","date":"2024-08-08T15:57:51+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":53876629,"identity":"c069d68e-7ae6-4b82-9d5b-394bb93540df","added_by":"auto","created_at":"2024-04-01 16:41:59","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3452229,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCharacteristics of GRGs in GC. (A\u003c/strong\u003e) Heatmap showing DEGsbetween GC tissues and normal tissues. \u003cstrong\u003e(B\u003c/strong\u003e) The volcano map shows DEGs between GC and normal tissues. Red dots indicate significantly upregulated genes, light blue dots represent downregulated genes, and black dots indicate genes with no significant difference between groups. \u003cstrong\u003e(C)\u003c/strong\u003eFrequencies of CNV gain, CNV loss, and non-CNV among the GRGs.\u003cstrong\u003e (D) \u003c/strong\u003eLocations of CNV alterations in GRGs on chromosomes. \u003cstrong\u003e(E-F)\u003c/strong\u003e GO enrichment analysis of DEGs. \u003cstrong\u003e(G-H) \u003c/strong\u003eKEGG enrichment analysis of DEGs.\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-4130368/v1/a114d67f306e2827f7456b6c.png"},{"id":53876633,"identity":"23c11515-f93a-4c3d-af9e-c169285169d9","added_by":"auto","created_at":"2024-04-01 16:42:00","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":8443831,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSubtypeidentification based on GRGs. (A)\u003c/strong\u003e The item-consensus plot shows the chosen optimal cluster number (k = 2) (cluster A, n = 291; cluster B n = 43).\u003cstrong\u003e (B) \u003c/strong\u003eConsensus values, ranging from 0 to 1.\u003cstrong\u003e(C) \u003c/strong\u003eThe corresponding relative changes in area under the CDF curves when the cluster number changed from k to k+1. The range of k changed from 2 to 9, and the optimal k was 2. \u003cstrong\u003e(D)\u003c/strong\u003e Survival curves of patients in cluster A and cluster B (cluster A, n = 291; cluster B, n = 43).\u003cstrong\u003e (E)\u003c/strong\u003e Validation of subtypeidentification using the GSE84437 dataset. The selected optimum cluster number was illustrated by the item-consensus plot. (k = 2) (cluster A, n = 253; cluster B, n = 180).\u003cstrong\u003e (B) \u003c/strong\u003eConsensus values of the GSE84437 dataset. ranging from 0 to 1\u003cstrong\u003e. (C) \u003c/strong\u003eThe corresponding relative changes in the areas under the cumulative distribution function (CDF) curves when the cluster number changed from k to k+1. The range of k changed from 2 to 9, and the optimal k was 2. (D) Gene set variation analysis (GSVA) revealed the differentially enriched crucial pathways between clusters A and B. (E) Gene set enrichment analysis (GSEA) indicated the pathways enriched in cluster A. (F) GSEA indicated the pathways enriched in cluster B. \u003cstrong\u003e(G)\u003c/strong\u003eSurvival curves of patients in cluster A and cluster B (cluster A, n = 252; cluster B, n =180).\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-4130368/v1/9350359739d2644ca7390f94.png"},{"id":53877682,"identity":"897a07fa-0232-481b-bf39-583bb9bd364d","added_by":"auto","created_at":"2024-04-01 16:49:59","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3904113,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction of the GRG signature as a risk model. (A)\u003c/strong\u003eUnivariate Cox regression analysis of prognostic GRGs in the TCGA-STAD cohort. \u003cstrong\u003e(B) \u003c/strong\u003eLASSO coefficient values of 10 prognosis-related GRGs. Each colored line represents the change trend of each independent variable coefficient.\u003cstrong\u003e (C) \u003c/strong\u003eAn elastic net regularization course with a partial likelihood deviance plot. The vertical dashed line with the minimum partial likelihood deviance value is optimally logarithmic (lambda). Lambda is the parameter that controls the regulation degree of LASSO regression complexity. The ordinate is the value of the coefficient, the lower abscissa is the log (lambda), and the upper abscissa is the number of nonzero coefficients in the model. \u003cstrong\u003e(D) \u003c/strong\u003eDistribution of the risk score of patients in the TCGA-STAD cohort. \u003cstrong\u003e(D) \u003c/strong\u003eHeatmap of GRG expression profiles in the prognostic signature of the TCGA-STAD cohort.\u003cstrong\u003e (E)\u003c/strong\u003eSurvival time and status of patients in the TCGA-STAD cohort. The black dotted line represents the optimal cutoff value for dividing patients into low-risk and high-risk groups. \u003cstrong\u003e(F) \u003c/strong\u003ePCA (principal component analysis) based on the prognostic signature of patients in the TCGA-STAD cohort. The high- and low-risk patients are represented by red and light green dots, respectively. \u003cstrong\u003e(H)\u003c/strong\u003eKaplan-Meier analysis of OS between the high- and low-risk patients in the TCGA-STAD cohort. \u003cstrong\u003e(I) \u003c/strong\u003eROC curves predicting the sensitivity and specificity of the risk score for predicting the 1-, 3-, and 5-year survival of patients in the TCGA-STAD cohort. \u003cstrong\u003e(J) \u003c/strong\u003eHeatmap of the GRG expression profiles associated with the prognostic signature in the GSE84437 dataset. \u003cstrong\u003e(K) \u003c/strong\u003eDistribution of the risk score of patients in the GSE84437 dataset. \u003cstrong\u003e(L)\u003c/strong\u003eSurvival time and status of patients in the GSE84437 dataset. \u003cstrong\u003e(M)\u003c/strong\u003e Kaplan-Meier analysis of OS between the two groups of patients in the GSE84437 dataset. \u003cstrong\u003e(N) \u003c/strong\u003eROC curves predicting the sensitivity and specificity of the risk score for predicting the 1-, 3-, and 5-year survival of patients in the GSE84437 dataset.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-4130368/v1/b843af3e2a95607c70581827.png"},{"id":53877683,"identity":"11e56466-3f9b-442b-bfa8-c39c42d3a3d6","added_by":"auto","created_at":"2024-04-01 16:49:59","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1067170,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA nomogram was established according to the 6-GRG signature risk score and other risk factors. (A) \u003c/strong\u003eUnivariate regression analysis of risk factors, including the 6-GRG signature risk score. \u003cstrong\u003e(B) \u003c/strong\u003eMultivariate regression analysis of risk factors.\u003cstrong\u003e (C) \u003c/strong\u003eMulti-index ROC curve of the 6-GRG signature risk score and other indicators. \u003cstrong\u003e(D)\u003c/strong\u003e The distributions of clinicopathological features were compared between the low- and high-risk groups. \u003cstrong\u003e(E)\u003c/strong\u003e Development of a nomogram to predict 1-, 3-, and 5-year OS. \u003cstrong\u003e(F)\u003c/strong\u003eCalibration plots of the nomograms in terms of the agreement between the nomogram-predicted and observed 1-, 3-, and 5-year survival outcomes in the TCGA-STAD cohort. The 45° dashed line represents the ideal performance. The lines represent the actual performances of the models, and the different colored lines represent the 1-, 3-, and 5-year results. *\u003cem\u003eP\u003c/em\u003e\u0026lt; 0.05, **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01, ***\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-4130368/v1/5d94865f65e453e95f3ba05e.png"},{"id":53876632,"identity":"be9edb26-13f2-4df1-9ddd-fa7174242750","added_by":"auto","created_at":"2024-04-01 16:41:59","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":5131778,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCharacteristics of the six genes recruited in the risk model. (A)\u003c/strong\u003e The landscape of the mutation profiles of patients in the TCGA-STAD cohort. The waterfall plot shows mutation information for the GRGs. The corresponding colors indicate different mutation types. The upper barplot shows the TMB (tumor mutation burden) of each subject. The numbers on the right represent individual mutation frequencies. \u003cstrong\u003e(B)\u003c/strong\u003e Correlation matrix of the six prognostic GRGs and the 6-GRG signature risk score. The purple background represents a positive correlation, and the blue background represents a negative correlation. \u003cstrong\u003e(C)\u003c/strong\u003e Differential expression of GRGs in the low- and high-risk groups of the TCGA-STAD cohort. \u003cstrong\u003e(D-I) \u003c/strong\u003eKaplan‒Meier survival curves for the six prognostic GRGs in the TCGA-STAD cohort. \u003cstrong\u003e(J) \u003c/strong\u003eThe expression of CXCR4 in GC and adjacent tissues in the TCGA-STAD cohort. \u003cstrong\u003e(K) \u003c/strong\u003eVerification of CXCR4 knockdown in BGC-823 cells. \u003cstrong\u003e(L)\u003c/strong\u003e Glucose consumption of BGC-823 cells after CXCR4 knockdown. \u003cstrong\u003e(M)\u003c/strong\u003e Lactate production in BGC-823 cells after CXCR4 knockdown. \u003cstrong\u003e(N-O)\u003c/strong\u003e ECAR of BGC-823 cells after knockdown of CXCR4. \u003cstrong\u003e(P) \u003c/strong\u003eProliferation of BGC-823 cells after CXCR4 knockdown. \u003cstrong\u003e(Q) \u003c/strong\u003eGrowth curve of transplanted tumors in nude mice. \u003cstrong\u003e(R) \u003c/strong\u003eImages and weights of nude mice bearing transplanted tumors after dissection. \u003cstrong\u003e(S) \u003c/strong\u003eMigration ability of BGC-823 cells after knockdown of CXCR4. \u003cstrong\u003e(T) \u003c/strong\u003eInvasive ability of BGC-823 cells after knockdown of CXCR4. ns, not significant, *\u003cem\u003ep\u003c/em\u003e\u0026lt; 0.05, **\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01, ***\u003cem\u003ep\u003c/em\u003e \u0026lt; 0.001.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-4130368/v1/e26a341c427683f82fa5e6d0.png"},{"id":53876634,"identity":"ce4e687d-ee31-4098-a63c-2c81f1e63495","added_by":"auto","created_at":"2024-04-01 16:42:00","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":2416653,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEvaluation of the immune microenvironment and response to immunotherapy in the GRG-related groups. (A)\u003c/strong\u003e ImmuneScore of subjectes in cluster A and B. \u003cstrong\u003e(B) \u003c/strong\u003eImmuneScores of subjects in the low- and high-risk groups.\u003cstrong\u003e (C) \u003c/strong\u003eImmune checkpoint expression in cluster A and B. \u003cstrong\u003e(D) \u003c/strong\u003eImmune checkpoint expression in the low- and high-risk groups. \u003cstrong\u003e(E) \u003c/strong\u003eInfiltration of immune cells in cluster A and B. \u003cstrong\u003e(F)\u003c/strong\u003eInfiltration of immune cells in the low- and high-risk groups. \u003cstrong\u003e(G)\u003c/strong\u003e Expression of SLC35A3, ME1, NUP50, PLOD2, CXCR4 and SRD5A3 in groups that did or did not respond to treatment with anti-PD-1. \u003cstrong\u003e(H-O)\u003c/strong\u003e The ROC curve of the 6-GRG signature risk score; expression of SLC35A3, ME1, NUP50, PLOD2, CXCR4 and SRD5A3; and the ME1/CXCR4 score. \u003cstrong\u003e(P)\u003c/strong\u003eRelationship between the expression of ME1 and the response to treatment with anti-PD-1. \u003cstrong\u003e(Q)\u003c/strong\u003eRelationship between the expression of CXCR4 and the response to treatment with anti-PD-1. \u003cstrong\u003e(R)\u003c/strong\u003eRelationship between the ME1/CXCR4 score and response to treatment with anti-PD-1. \u003cstrong\u003e(S) \u003c/strong\u003eRelationship between the ImmuneScore and the expression of SLC35A3, ME1, NUP50, PLOD2, CXCR4 and SRD5A3. \u003cstrong\u003e(T) \u003c/strong\u003eRelationship between the infiltration of immune cells and the expression of SLC35A3, ME1, NUP50, PLOD2, CXCR4 and SRD5A3. *\u003cem\u003eP\u003c/em\u003e\u0026lt; 0.05, **\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01, ***\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001.\u003c/p\u003e","description":"","filename":"Figure6.png","url":"https://assets-eu.researchsquare.com/files/rs-4130368/v1/be07bc4a68d5d6d519782d98.png"},{"id":62298457,"identity":"b225c65c-790a-4545-9849-91611c57d433","added_by":"auto","created_at":"2024-08-12 16:13:34","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":30412808,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4130368/v1/6461254f-a022-4515-afb4-1475695a1ce8.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Glycolysis-related genes predict prognosis and indicate immune microenvironment features in gastric cancer","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eGastric cancer (GC) is the most prevalent malignant neoplasm of the digestive system. Approximately one million people are diagnosed with GC annually worldwide [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. GC patients have an unfavorable prognosis, and less than 30% of them may survive for five years after diagnosis [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. The poor clinical prognosis of these patients is primarily due to the advanced disease stage and unfavorable tumor microenvironment (TME) [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eTumor cells primarily use glycolysis as a source of energy. A high glycolytic capacity in tumor cells is related to poor prognosis and drug resistance in various cancers, including GC [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Several studies have shown interactions between tumor glycolysis and the TME [\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Tumor cells can produce lactic acid via glycolysis, and an increase in lactic acid makes the TME acidic, thus facilitating tumor cell proliferation, invasion, and migration [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Moreover, several studies have reported that metabolic reprogramming contributes to the development of an immunosuppressive tumor microenvironment [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Many studies have focused on the systematic investigation of glycolysis and tumors, including liver cancer [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], glioblastoma [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], and endometrial cancer [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. In the present study, we evaluated the roles of glycolysis-related genes (GRGs) in predicting prognosis and indicating immune microenvironment features in gastric cancer patients.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e"},{"header":"2 Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Acquisition of the mRNA expression dataset\u003c/h2\u003e \u003cp\u003eThe transcriptome (fragments per kilobase million, FPKM) and clinical information were obtained from The Cancer Genome Atlas (TCGA) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://portal.gdc.cancer.gov/\u003c/span\u003e\u003cspan address=\"https://portal.gdc.cancer.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. The expression data of 407 samples included 32 normal tissues and 375 tumor tissues. Concurrently, the GSE84437 dataset [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e] and the GSE13763 dataset [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], which include microarray data (Illumina HumanHT-12 V4.0 expression bead chip platform) and clinical data, were acquired from the Gene Expression Omnibus (GEO) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and used as the validation cohort. The probe IDs were converted to gene symbols using the relevant annotation files (GPL6947 and GPL570), and the average expression values of numerous probes for the same gene were calculated. The data were standardized and processed with R (version 3.5.1) software and R Bioconductor components. We used data from the immunotherapy cohort PRJEB25780 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ebi.ac.uk/ena/browser/home\u003c/span\u003e\u003cspan address=\"https://www.ebi.ac.uk/ena/browser/home\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e] to evaluate the relationship between gene expression and the response to immunotherapy. Gene expression data were obtained using the limma and ggpubr packages in R software. ROC curves were generated to predict the response of GC patients to anti-PD-1 therapy through the pROC package in R software.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Analysis of differentially expressed glycolysis-related genes (GRGs)\u003c/h2\u003e \u003cp\u003eAccording to the keyword \u0026ldquo;glycolysis\u0026rdquo;, five glycolysis-related sets (GO_GLYCOLYTIC_PROCESS, KEGG_GLYCOLYSIS_GLUCONEOGENESIS, BIOCARTA_GLYCOLYSIS_PATHWAY, HALLMARK_GLYCOLYSIS, and REACTOME_GLYCOLYSIS) were obtained from the Molecular Signatures Database (MSigDB) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gseamsigdb.org/gsea/msigdb/ind-ex.jsp\u003c/span\u003e\u003cspan address=\"https://www.gseamsigdb.org/gsea/msigdb/ind-ex.jsp\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. The genes in the above five sets were defined as glycolysis-related genes (GRGs). A total of 326 GRGs were identified from the TCGA-STAD cohort data. The differentially expressed genes (DEGs), defined as those with a false discovery rate (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and |logFC (fold-change) | \u0026gt;1, were obtained by the \"edgeR\" R package [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Furthermore, gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were carried out to assess the potential functions and related pathways of the differentially expressed GRGs via the \u0026ldquo;ClusterProfiler\u0026rdquo; R package [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Consensus clustering analysis of GRGs\u003c/h2\u003e \u003cp\u003eThe \"ConsensuClusterPlus\" R package [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] was used to assess the expression profiles of 326 GRGs for consensus clustering. One thousand permutations were conducted to ensure consistency in the classification. A consensus heatmap and the cumulative distribution function (CDF) were used to determine the ideal k value.\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eTo analyze the biological functions between the subgroups, gene set variation analysis (GSVA) was performed by using the R package \u0026ldquo;GSVA\u0026rdquo;. The enrichment results were then visualized by a heatmap using the R package \u0026ldquo;heatmap\u0026rdquo;. By using the R package \u0026ldquo;limma\u0026rdquo;, the differences were statistically significant at adjusted p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and a false discovery rate (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.25. The Kyoto Encyclopedia of Genes and Genomes (KEGG) gene set collection (c2.cp.kegg.v7.1.symbols.gmt) was used as the input file for both gene set enrichment analysis (GSEA) and gene set variation analysis (GSVA) analyses.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Construction of a glycolysis-related risk model\u003c/h2\u003e \u003cp\u003eUnivariate Cox regression analysis was performed to identify the survival-related GRGs. The \u0026ldquo;glmnet\u0026rdquo; R package [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] was used to minimize the dimension by applying the least absolute shrinkage and selection operator (LASSO) for the survival-related GRGs. Then, a glycolysis-related risk model was generated by using the following formula: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(Risk score={\\sum }_{i=1}^{n}Expi \\beta i\\)\u003c/span\u003e\u003c/span\u003e, where Exp denotes the expression levels of the genes and β represents the regression coefficient [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. According to the median risk score, Kaplan-Meier curves were generated to evaluate the prognosis of patients in different groups. Survival ROC curves were generated using the \u0026ldquo;Survival ROC\u0026rdquo; R package [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Development of a nomogram based on GRGs and clinical features\u003c/h2\u003e \u003cp\u003eUnivariate Cox regression analysis was performed to assess the relationships between GRG expression and clinical features (age, sex, clinical stage, grade, tumor-node-metastasis (TNM) stage, and G grade). Multivariate Cox regression analysis was employed to identify independent prognostic variables.\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eTaking GRG expression and clinical features into account, a nomogram scoring system was developed. Univariate and multivariate Cox regression analyses were conducted to identify factors associated with prognosis. Each item was given a score, and the scores were added together. Calibration curves were applied to assess the accuracy of the nomogram, and the concordance index (C-index) was used to quantify the discrimination capacity of the nomogram.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Cell culture and treatment\u003c/h2\u003e \u003cp\u003eThe human gastric cancer cell line BGC-823 was obtained from the Shanghai Cell Bank of the Chinese Academy of Sciences (Shanghai, China) and cultured in RPMI-1640 medium (Gibco, Grand Island, New York, USA) supplemented with 10% fetal bovine serum (FBS, Gibco, Grand Island, New York, USA) and 1% penicillin‒streptomycin solution (Beyotime, Nantong, China). The cells were maintained at 37\u0026deg;C in a humidified environment containing 5% CO\u003csub\u003e2\u003c/sub\u003e. The cells were passaged every 2\u0026ndash;3 days. Lentiviruses encoding shRNAs for CXCR4 (VSVG-Lentai-hU6-shRNA-CXCR4-BSD-hEF1a-3xFlag) and negative control (NC) lentivirus (VSVG-Lentai-hU6-shRNA-NC-BSD-hEF1a-3xFlag) were purchased from Shanghai Taitool Bioscience Co. (Shanghai, China). The sequences of the shRNAs targeting CXCR4 are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Cells were infected with the lentivirus and selected by using blasticidin (BSD, Sigma) 72 h after infection.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eShRNA sequences targeting CXCR4.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eName\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSense\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAntisense\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCXCR4-homo-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGAAGCATGACGGACAAGTA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTACTTGTCCGTCATGCTTC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCXCR4-homo-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGGAAGCTGTTGGCTGAAAA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTTTTCAGCCAACAGCTTCC\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Quantitative real-time PCR\u003c/h2\u003e \u003cp\u003eTRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) was used to extract total RNA. After spectrophotometric quantification, 1 \u0026micro;g of total RNA in a final volume of 20 \u0026micro;l was reverse transcribed with a PrimeScript RT Reagent Kit (TaKaRa, Shiga, Japan) according to the manufacturer's protocol. Aliquots of complementary DNA (cDNA) corresponding to equal amounts of RNA were used for quantification of messenger RNA (mRNA) by RT-PCR using the Light Cycler 96 Real-time Quantitative PCR Detection System (Roche, Indianapolis, IN, USA). The reaction system contained the corresponding cDNA, forward and reverse primers, and SYBR Green PCR master mix (Roche). β-Actin was used as an internal standard. The primer sequences are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePrimers used for real-time PCR.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePrimer name\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eForward Sequence\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eReverse Sequence\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCXCR4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCTCCTCTTTGTCATCACGCTTCC\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eGGATGAGGACACTGCTGTAGAG\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eβ-actin\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTCATGAAGTGTGACGTGGACAT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCTCAGGAGGAGCAATGATCTTG\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e2.8 Glucose Uptake Assa\u003c/b\u003ey\u003c/h2\u003e \u003cp\u003eGlucose uptake was defined as the uptake of 2-deoxyglucose (2-DG) using a glucose uptake assay kit (fluorometric, ab136956, Abcam) according to the manufacturer\u0026rsquo;s instructions. Cells were seeded at a density of 2.5\u0026times;10\u003csup\u003e4\u003c/sup\u003e cells/well in 96-well plates overnight. The cells were serum-starved for an additional 24 h, and the cells in fresh complete media were incubated for 48 h. The cells were then incubated in 0.5% bovine serum albumin (BSA)/phosphate-buffered saline (PBS) with or without 2-DG for 1 h. The relative fluorescence units (RFU) were measured at Ex/Em\u0026thinsp;=\u0026thinsp;535/587 nm using a Fluoroskan microplate fluorometer (Thermo Fisher Scientific, San Jose, CA, USA). The uptake of 2-DG was calculated by the 2-DG-6-phosphate (2-DG6P) standard curve and the RFU of the samples.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Lactate measurement assay\u003c/h2\u003e \u003cp\u003eLactate levels were measured using an L-Lactate Assay Kit (Abcam, ab169557). Cell extracts were specifically oxidized to form an intermediate that reacts with a colorless probe to generate fluorescence that was measured at 530 nm (excitation)/590 nm (emission) using a Fluoroskan microplate fluorometer (Thermo Fisher Scientific, San Jose, CA, USA). The intensity was directly proportional to the amount of lactate measured in nmol/ml.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Extracellular acidification rate (ECAR) assay\u003c/h2\u003e \u003cp\u003eThe glycolysis assay (extracellular acidification rate, ECAR) was measured with a test kit (Abcam, ab197244). Specifically, the cells were incubated in a 96-well plate at a density of 5\u0026times;10\u003csup\u003e4\u003c/sup\u003e cells/well, and 150 \u0026micro;L of respiration buffer was added before the test. The addition of 10 \u0026micro;L of carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP) was used as a positive or negative control. Then, 10 \u0026micro;L of reconstituted glycolysis assay reagent was added, and the glycolysis assay signal was measured at 1.5 min intervals for \u0026ge;\u0026thinsp;120 min using excitation and emission wavelengths of Ex/Em 380/615 nm, respectively, by using a Fluoroskan microplate fluorometer (Thermo Fisher Scientific, San Jose, CA, USA).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Wound healing assay\u003c/h2\u003e \u003cp\u003eThe cells were placed in a six-well plate at a density of 2\u0026times;10\u003csup\u003e5\u003c/sup\u003e/well. The cells were expected to grow to 80%~90% confluence. Cell monolayers were scratched with a 200 \u0026micro;L pipette tip to produce lesions of a consistent length and then cultured in basal medium. After the cells were washed with phosphate-buffered saline (PBS) to remove cellular fragments, each wound was imaged at 0, 12, and 24h by inversion microscopy. Cell migration was quantified by measuring the relative wound areas using ImageJ (National Institutes of Health, Bethesda, MD, USA).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e2.12 Transwell invasion assay\u003c/h2\u003e \u003cp\u003eMatrigel (1:8 dilution of 50 mg/L) was coated on the upper surface of the bottom membrane of the transwell chamber (Corning, NY, USA) and allowed to dry at 4\u0026deg;C. The cells were resuspended in serum-free DMEM and seeded into the upper chamber. The bottom wells were filled with complete DMEM. The cells in the upper chamber were removed after 24 h of incubation. The cells that migrated through the Matrigel matrix membrane were fixed in 4% paraformaldehyde and stained with 0.1% crystal violet. Finally, representative images were obtained under a microscope, and the number of cells in the picture was calculated through ImageJ.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.13 MTT\u003c/h2\u003e \u003cp\u003eCell growth was evaluated using the MTT [3-(4,5-dimethyl-2-yl)-2,5-diphenyl tetrazolium bromide] assay. The cells were seeded in 96-well plates at a density of 3,000 cells/well and treated with GEM for 24 hours. Methyl thiazolyl diphenyl tetrazolium bromide (MTT; Sigma‒Aldrich, St. Louis, MO, USA) (5 mg/mL) was added to the medium at 0, 1, 2, 3, 4, and 5 days. After incubation at 37\u0026deg;C for 4 hours, the medium was removed, and 200 ml of dimethyl sulfoxide (DMSO, Sigma) was added to each well for dissolution. The absorbance of DMSO was measured at 490 nm using a microplate reader (Thermo Fisher Scientific, Waltham, MA, USA).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.14 Subcutaneous xenograft nude mouse model\u003c/h2\u003e \u003cp\u003eFour-week-old female BALB/c thymic nude mice (Hangzhou Ziyuan Experimental Animal Technology Co., Ltd., Zhejiang, China) were nurtured in the animal laboratory in a specific pathogen-free environment. Cells were injected into the left axilla of the nude mice. At the end of the experiment, the mice were anesthetized, and the tumors were excised. The length (L) and width (W) of each subcutaneous tumor were measured using calipers. The tumor volume (TV) was calculated as TC = (L \u0026times; W\u003csup\u003e2\u003c/sup\u003e)/2.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e2.15 Statistical analysis\u003c/h2\u003e \u003cp\u003eR version 3.5.1 or GraphPad Prism 8.0 was used for all the statistical analyses. There were a minimum of three runs of each experiment. The results are shown as the average\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD).\u003c/p\u003e \u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Characteristics of GRGs in GC\u003c/h2\u003e \u003cp\u003eThe expression of 326 glycolysis-related genes (GRGs) in normal tissues and in the TCGA-STAD cohort was visualized via a heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA) and a volcano map (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). Somatic copy number alterations were detected in a total of 73 differentially expressed genes (DEGs). A decrease in CNV (copy number variation) was detected in STMN2, NUP205, VCAN, HS2S11, DCN, and LDHAL6B, while CNV increased in VEGFA, ALDOC, EFNA3, COL5A1, NIP188, and NUP155 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC). The sites of CNV abnormalities in the DEGs are presented in the circle diagram (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). GO and KEGG pathway enrichment analyses were used to predict the potential biological effects of these DEGs. GO analysis revealed that the DEGs were clustered into metabolic GO categories such as nucleic acid transport, RNA transport, mRNA transportation, RNA localization, and nucleobase-containing molecule transport (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE-F). KEGG analysis revealed that genes associated with glycolysis, fructose and mannose metabolism, nucleocytoplasmic transport, pyruvate metabolism, and the HIF-1 signaling pathway were most prevalent (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eG-H).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Subtype identification based on the expression patterns of GRGs in GC\u003c/h2\u003e \u003cp\u003eTo evaluate whether the prognosis of patients with GC was associated with the expression of GRGs, these 326 glycolysis-related genes (GRGs) were subjected to consensus cluster analysis. The consensus matrix heatmap presented a distinct boundary and low interference when k\u0026thinsp;=\u0026thinsp;2, indicating that the samples could be stably clustered according to the expression patterns of the GRGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA-C). Two different groups, namely, cluster A (n\u0026thinsp;=\u0026thinsp;291) and cluster B (n\u0026thinsp;=\u0026thinsp;43), were obtained. The differences in survival between these two clusters were then assessed. By using gene set variation analysis (GSVA) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD) and gene set enrichment analysis (GSEA) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE and F) of these two groups, we found that the crucial pathways in cluster A were mainly enriched in metabolic-related pathways, while the corresponding pathways enriched in cluster B were immune-related pathways. Therefore, cluster A was named the metabolic subtype, and cluster B was termed the immune subtype. Furthermore, we found that the prognosis of patients with the metabolic subtype (cluster A) was better than that of patients with the immune subtype (cluster B) (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.016, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eG).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Construction of a glycolysis-related risk model based on the GRG signature\u003c/h2\u003e \u003cp\u003eSince the expression of GRGs was related to survival, we developed a signature model according to the expression of GRGs in the TCGA-STAD cohort to predict the prognosis of GC patients. First, univariate Cox analysis was used to identify OS-related GRGs, and ten genes were identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Then, LASSO regression was performed to avoid overfitting (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB-C), and six genes, namely, ME1, PLOD2, NUP50, CXCR4, SLC35A3 and SRD5A3, were screened out. Finally, a predictive model was constructed according to the expression of these six genes, and the corresponding regression coefficients were calculated. Risk Score was determined by using the following formula: Risk Score = (-0.057 \u0026times; expression of ME1) + (0.062 \u0026times; expression of PLOD2) + (-0.088 \u0026times; expression of NUP50) + (0.007 \u0026times; expression of CXCR4) + (-0.050 \u0026times; expression of SLC35A3) + (0.026 \u0026times; expression of SRD5A3).\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eThe risk score of each patient was calculated, and the TCGA-STAD cohort was divided into high- and low-risk groups based on the median risk score. The expression of the six genes in the high- and low-risk groups was visualized using a heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). The risk scores of each patient are depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE. Patients in the high-risk group had shorter OS (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF). Principal component analysis (PCA) revealed that the high- and low-risk groups were well separated from each other (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG). According to the Kaplan‒Meier survival curves, patients in the low-risk group had a better prognosis (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eH). Furthermore, a receiver operating characteristic (ROC) curve was drawn to verify the performance of the model (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI). The area under the curve (AUC) values were 0.712 (1-year), 0.689 (3-year), and 0.681 (5-year), indicating the good sensitivity and specificity of this risk model.\u003c/p\u003e\u003cp\u003eFurthermore, this risk model was validated by using the GSE84437 dataset. Patients in the validation cohort were divided into low- and high-risk groups based on the median risk score obtained from the TCGA-STAD cohort. The expression of the six genes in the validation set was depicted using a heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eJ). The risk scores of each patient and survival time distribution are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eK and L. Consistent with the results of the TCGA-STAD cohort, patients in the low-risk group of the validation set also had a better prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eM). The AUC values in the validation set were 0.742 (1-year), 0.738 (3-year), and 0.740 (5-year) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eN).\u003c/p\u003e\u003cp\u003eAccording to the above analysis, the glycolysis-related risk model based on the 6-GRG signature is a promising model for predicting the prognosis of GC patients.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Establishment of a nomogram according to the glycolysis-related risk model and other risk factors\u003c/h2\u003e \u003cp\u003eBy using univariate Cox analysis, the 6-GRG signature risk score was found to be an independent prognostic factor for GC patients in the TCGA-STAD cohort (HR\u0026thinsp;=\u0026thinsp;1.263, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Multivariate Cox analysis revealed that the risk score remained an independent predictor after considering clinical characteristics (HR\u0026thinsp;=\u0026thinsp;1.262, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). Multivariate ROC analysis was used to compare the predictive accuracy of the risk score with that of various clinical factors. Overall, the risk score was the most accurate prognostic indicator (AUC\u0026thinsp;=\u0026thinsp;0.658, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Consistent with poorer prognosis in cluster B (immune subtype), subjects in cluster B were located more frequently in the high-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). Moreover, correlations were found between the risk score and other factors, including age, grade, and M stage (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). Finally, a nomogram (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE) incorporating several risk factors was developed to estimate 1-, 3-, and 5-year OS. The calibration curves showed excellent agreement between the probabilities of 1-, 3-, and 5-year OS predicted by the nomogram and the actual OS rates of patients in the TCGA-STAD cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e3.5 CXCR4 promoted the metastasis and growth of GC cells\u003c/h2\u003e \u003cp\u003eTo analyze the characteristics of the six genes included in the risk model, we investigated the mutation profiles of the six GRGs by using the \"maftools\" R package [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, mutations were not frequent in these GRGs. The mutation frequencies of PLOD2, SRD5A3, SLC35A3, NUP50, ME1 and CXCR4 were 3%, 1%, 1%, 1%, 0% and 0%, respectively. A heatmap was generated to show how the risk score and the expression of the six GRGs are related to each other (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). ME1, PLOD2, and CXCR4 were highly expressed in the high-risk group, while NUP50 and SLC35A3 were highly expressed in the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). K‒M analysis was then subsequently performed (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD-I). High expression of CXCR4 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.002) and PLOD2 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.043) indicated a relatively poor prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE and G), and high expression of NUP50 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.018) and SLC35A3 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.043) indicated a better prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF and H). However, the expression of SRD5A3 and ME1 was not related to prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD and I).\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eSince CXCR4 was the most significantly related gene to prognosis among these six GRGs, we further focused on the biological functions of CXCR4 in GC. CXCR4 expression was greater in tumor tissues than in normal tissues according to data in the TCGA-STAD cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eJ). Knockdown of CXCR4 in BGC-823 cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eK) was associated with reduced glucose consumption (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eL), lactate production (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eM) and ECAR (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eN-O). MTT assays showed that the downregulation of CXCR4 significantly inhibited the proliferation of BGC-823 cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eP). Subcutaneous xenograft nude mouse models further confirmed that the growth of tumors was weakened by knocking down CXCR4 \u003cem\u003ein vivo\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eQ and R). The wound healing assay showed that cell migration was repressed after CXCR4 was knocked down (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eS). Transwell invasion assays showed that the downregulation of CXCR4 weakened the invasion ability of BGC-823 cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eT).\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e3.6 The expression of ME1 and CXCR4 predicted immune cell infiltration and response to immunotherapy\u003c/h2\u003e \u003cp\u003eSeveral studies have reported that metabolic reprogramming of cancer cells leads to the development of an immunosuppressive tumor microenvironment [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Since GRGs participate in the progression of GC, we further investigated whether the GRGs were related to the modeling of the immune microenvironment. The ImmuneScore was calculated and compared between metabolic subtype (cluster A) and immune subtype (cluster B) or between low-risk and high-risk groups by using the \"ESTIMATE\" R package [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Surprisingly, although patients in cluster B and the high-risk group had worse prognosis in TCGA-STAD cohort, both had significantly greater ImmuneScore (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA and B), suggesting that patients in the immune subtype (cluster B) or high-risk group might have a favorable immune microenvironment.\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eTo confirm this hypothesis, differences in the expression levels of immunological checkpoint molecules were compared between clusters A and B. Cluster B samples mainly expressed immunological checkpoint molecules, including CD274 (also known as PD-L1), at higher levels than did cluster A samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC). Similarly, the expression of most immune checkpoint molecules, including PD-L1, was greater in the high-risk group than in the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). The 22 immune cell subsets were then analyzed by using the CIBERSORT algorithm [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. M2 macrophages, activated mast cells, and eosinophils were found at higher levels in cluster A samples than in cluster B samples. In comparison, the proportions of memory B cells, CD8\u003csup\u003e+\u003c/sup\u003e T cells, regulatory T cells (Tregs), and naive B cells were greater in cluster B samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE). Activated CD4\u003csup\u003e+\u003c/sup\u003e memory T cells, follicular helper T cells, and M1 macrophages were more enriched in the low-risk group, while there was more infiltration of regulatory T cells (Tregs), monocytes, M2 macrophages, and resting dendritic cells in the high-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eF).\u003c/p\u003e\u003cp\u003eCluster B and the high-risk group preserved a favorable immune microenvironment and increased expression of immune checkpoints, including PD-L1, suggesting that GRGs might indicate the response to immunotherapy. To confirm this hypothesis, we used RNA-Seq data from the immunotherapy cohort PRJEB25780 from the European Nucleotide Archive (ENA) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ebi.ac.uk/ena/browser/home\u003c/span\u003e\u003cspan address=\"https://www.ebi.ac.uk/ena/browser/home\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). In this cohort, researchers examined 45 patients with metastatic or recurrent gastric cancer who received treatment with anti-programmed cell death protein 1 (PD-1) [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. The treatment responses were classified as follows: complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). Patients were divided into two groups according to their treatment response: those with a confirmed response (CR/PR) and those without a clinical response (PD/SD). The expression levels of PLOD2, SRD5A3, SLC35A3, NUP50, ME1 and CXCR4 were then compared between these two groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eG). The expression of ME1 was significantly greater in the group without a response, while the expression of CXCR4 was greater in the group with a confirmed response (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eG). The ROC curve was then used to evaluate the prognostic performance of the expression of the GRGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eH-O). The areas under the ROC curves for ME1 and CXCR4 were 0.727 and 0.687, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eJ and M). The expression of PLOD2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eL), SRD5A3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eN), SLC35A3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eI) and NUP50 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eK), as well as the 6-GRG signature risk score (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eH), indicated no difference. Moreover, we developed a score based on the expression of CXCR4 and ME1 (score\u0026thinsp;=\u0026thinsp;0.189\u0026thinsp;+\u0026thinsp;2.516\u0026times;CXCR4-3.351\u0026times;ME1) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eO). The area under the ROC curve for this score was 0.814, suggesting that the score worked more effectively than CXCR4 or ME1 alone (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eO). Based on the optimal threshold of expression levels of CXCR4 and ME1 or the ME1/CXCR4 score for the maximum ROC curve values, the patients were then dichotomized into high- and low-risk subgroups (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eP-R). The expression of CXCR4 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eP) and ME1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eQ), as well as the ME1/CXCR4 score (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eR), significantly correlated with the response to immunotherapy.\u003c/p\u003e\u003cp\u003eTo investigate the mechanisms involved in the predictive value of the expression of CXCR4 and ME1 upon immunotherapy, we evaluated the immune microenvironment and immune cell infiltration according to the expression of CXCR4 and ME1. The expression of CXCR4 was positively correlated with the ImmuneScore, while the expression of ME1 was negatively correlated with the ImmuneScore (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eS). Consistently, higher expression of CXCR4 predicted more infiltration of CD8\u003csup\u003e+\u003c/sup\u003e T cells, while expression of ME1 presented a negative correlation with infiltration of CD8\u003csup\u003e+\u003c/sup\u003e T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eT). Therefore, the high expression of CXCR4 and low expression of ME1 suggested a favorable immune microenvironment.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eMetabolic reprogramming is one of the hallmarks of cancer [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. To sustain continuous replication and high proliferation rates, metabolic reprogramming of cancer cells must switch their metabolic reprogramming to a \u0026lsquo;glycolysis-dominant\u0026rsquo; metabolic profile. This shift promotes cell survival by meeting energy, synthesis, and redox demands, and it also renders the tumor microenvironment more conducive to cancer progression. [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eThe classic example of metabolic switching in cancer cells was first reported by Otto Warburg in the 1920s. This phenomenon, which was subsequently named the Warburg effect, was characterized by a shift in the metabolic machinery of cancer cells toward glycolysis and lactic acid fermentation from oxidative metabolism even under normoxic conditions [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Under physiological conditions in aerobic environments, cells metabolize glucose through glycolysis, the tricarboxylic acid (TCA) cycle and oxidative phosphorylation (OXPHOS). Normal cells preferentially perform oxidation through OXPHOS because it produces more adenosine triphosphate (ATP). However, the limited oxygen supply in the hypoxic tumor microenvironment causes cancer cells to prefer aerobic glycolysis [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Therefore, cancer cells exhibit increased glycolytic dependency and exhibit increased glucose uptake and fermentation of glucose to lactate to meet the increased anabolic requirements for a malignant phenotype [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eThe Warburg effect has been investigated for approximately a century, and multiple studies have elucidated the mechanisms governing the increased glycolytic dependency of cancer cells. In addition to energy supply, this aberrant energy utilization mode in cancers critically influences the tumor microenvironment (TME). It does so by creating a hypoxic, acidic, nutrient-poor, and immune-modulatory metabolite-rich environment (e.g., lactate, kynurenine), which fosters a favorable tumor niche and impairs effective cancer treatment [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]. Given the usefulness of these metabolic intermediates of glycolysis, cancer cells tend to maintain the Warburg effect through mutation. Several oncogenic proteins and pathways, including hypoxia-inducible factor (HIF-1) and the PI3K/Akt/mTOR pathway, have been implicated in regulating cancer cell-specific metabolic reprogramming [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eGiven the critical role of the Warburg effect in the progression of cancer, increasing research has been dedicated to revealing the mechanisms involved and identifying novel targets for clinical application. In the present study, we elucidated global alterations in glycolysis-related genes (GRGs) at the transcriptional and genetic levels in GC. Based on 326 GRGs, we identified two distinct molecular subtypes. Patients in the immune subtype (cluster B) had more advanced clinicopathological characteristics and shorter OS than did those in the metabolic subtype (cluster A). Furthermore, both univariate and multivariate Cox regression analyses revealed the prognostic value of a risk core based on the expression of six GRGs, namely, ME1, PLOD2, NUP50, CXCR4, SLC35A3, and SRD35A3. Survival analysis revealed that the high-risk group had a worse prognosis. These results suggested that the GRG signature could be helpful in guiding clinical treatment decisions.\u003c/p\u003e\u003cp\u003eThe deviation of intermediate molecules of glycolysis and the TCA cycle for the synthesis of nucleotides, lipids and nonessential amino acids is pivotal for generating the products necessary for cell proliferation and survival [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Additionally, increased glycolysis also triggers chemo- and radio-resistance, suggesting that therapeutics targeting elevated glycolysis in cancer cells could be a promising treatment modality to increase the sensitivity of cancer cells to treatment via other conventional strategies [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Although several therapeutics targeting enhanced glycolysis in cancer cells have been developed and several of these strategies are in different stages of preclinical and clinical investigations, there is still a lack of effective solutions specifically targeting cancer cell glycolysis [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eIn our present study, we showed that CXCR4 might be a promising target for strategies against elevated glycolysis in GC. CXCR4 is one of the most prevalent chemokine receptors and is also known as CD184. CXCR4 has been linked to the occurrence and progression of a variety of cancers, including breast cancer, melanoma, prostate cancer, and gastric cancer [\u003cspan additionalcitationids=\"CR34 CR35\" citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]. CXCR4 expression is also associated with lymph node metastasis in GC [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. CXCR4 is a G protein-coupled receptor that binds to the CXCR4 ligand CXC ligand 12 (CXCL12, stromal cell-derived factor-1, SDF-1). CXCL12 is generated by stromal cells and released into the TME, where it combines with CXCR4 on the tumor surface. In GC, cancer-associated fibroblasts influence the CXCL12/CXCR4 axis [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. The CXCL12/CXCR4 axis activates the MAPK cascade, which promotes chemotaxis and cell proliferation [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e], and stimulates the PLC/PKC and PI3K/Akt signaling pathways, enhancing cell migration and survival [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]. In the present study, we demonstrated that knockdown of CXCR4 suppressed the migration and invasion of GC cells. Moreover, the downregulation of CXCR4 inhibited the growth of GC cells both \u003cem\u003ein vitro\u003c/em\u003e and \u003cem\u003ein vivo\u003c/em\u003e, suggesting that CXCR4 could be a potential target for further drug development. Lactate, the end product of aerobic glycolysis, decreases the pH of the TME, facilitating the invasion and metastasis of cancer cells [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. Lactate is also metabolized in human tumors \u003cem\u003ein vivo\u003c/em\u003e, fueling the TCA cycle and contributing to energy production [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e]. Previous studies have shown that CXCR4 participates in the Warburg effect in acute myeloid leukemia (AML) cells via the CXCL12/CXCR4/mTOR pathway [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. The mTOR inhibitor rapamycin significantly suppressed the upregulation of glucose transporters in the matrix, enhanced glucose influx, and decreased lactate generation [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]. Our present study revealed that knockdown of CXCR4 in GC cells reduced glycolytic capability, glucose intake, and lactate generation. Therefore, the anticancer effect of CXCR4 knockdown could be mediated through repression of the Warburg effect.\u003c/p\u003e\u003cp\u003eInterestingly, although both the immune subtype (cluster B) and the high-risk group identified in the present study by using the TCGA database were associated with an unfavorable prognosis, these two cohorts presented a favorable immune microenvironment and increased expression of immune checkpoints, suggesting that the application value of GRGs still needs further elucidation. The reason for this inconsistency may be that the data from the TCGA-STAD cohort were collected mainly from before the era of immunotherapy. Without immune checkpoint inhibitor treatment, increased expression of immune checkpoints suggests immune escape, resulting in unfavorable outcomes [\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eImmune cells in the TME share a similar switch to a glycolytic metabolic profile, inducing competition between cancer cells and tumor-infiltrating cells over nutrients [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. The metabolic dysregulation of the tumor microenvironment can halt the infiltration of immune cells and inhibit antitumor immunity through the production of immune-suppressive byproducts. Furthermore, tumor-derived lactate plays direct and indirect immune-suppressive roles, such as accelerating the accumulation of Tregs and MDSCs, as well as the polarization of M2 tumor-associated macrophages [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]. Increased tumor glycolysis leads to resistance to adoptive T cell therapy in melanoma, suggesting the favorable benefits of dual targeting cancer immunity and metabolism [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]. In addition, metabolic reprogramming in cancer cells can also alter the expression of several cell surface markers, which interferes with immune surveillance [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Glycolytic activity can upregulate immune checkpoint expression, suggesting that glycolysis might promote immunotherapy response [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e] and could be a useful indicator for immunotherapy efficacy.\u003c/p\u003e\u003cp\u003eImmune checkpoint inhibitors (ICIs) have revolutionized therapeutic models for several types of tumors. The best-studied such inhibitors are antibodies that block the cytotoxic T-lymphocyte antigen-4 (CTLA-4) and programmed cell death protein-1 (PD-1) proteins. Treatment with anti-CTLA-4 and anti-PD-1 agents led to significantly improved survival in patients with various tumor types [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. The phase 2 KEYNOTE-059 trial suggested the safety and effectiveness of pembrolizumab as a third-line treatment for locally advanced or metastatic gastric or gastroesophageal junction (G/GEJ) adenocarcinoma [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. In this single-arm, multicohort trial, the objective response rate (ORR) was 11.6%, and complete responses (CRs) were noted in 2.3% of patients [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e]. The phase 3 CheckMate-649 trial demonstrated the acceptable safety profile and efficacy of nivolumab as a first-line treatment for patients with untreated, unresectable, human epidermal growth factor receptor 2 (HER2)-negative G/GEJ cancer or esophageal adenocarcinoma [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. The CheckMate-649 trial demonstrated significant improvements in overall survival (OS) and progression-free survival (PFS) for patients with a combined positive score (CPS)\u0026thinsp;\u0026ge;\u0026thinsp;5 [\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]. Based on this progress, immune checkpoint inhibitors have become standard therapeutics for the treatment of GC.\u003c/p\u003e\u003cp\u003eThe success of immunotherapy relies on its potential to \u0026ldquo;rewire\u0026rdquo; the immune cycle through the generation of long-lasting immune responses. However, not all patients respond to this type of intervention [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Therefore, it is important to identify target biomarkers for predicting response to ICIs. For now, PD-L1 immunohistochemistry (IHC), microsatellite instability (MSI)/mismatch repair (MMR), Epstein-Barr virus (EBV), and tumor mutational burden (TMB) have been considered promising predictive biomarkers for the response to ICIs in patients with GC [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. However, no single biomarker can be used as a reproducible proxy for predicting the effects of immunotherapy [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. Hence, developing an integrated predictive model that considers the complex components affecting host-TME interactions by reflecting the heterogeneity of GC is necessary.\u003c/p\u003e\u003cp\u003eIn our present study, we showed that GRGs exhibit potential for predicting response to ICIs. The expression of CXCR4 and ME1 was correlated with the infiltration of CD8\u0026thinsp;+\u0026thinsp;T cells and the response to treatment with an anti-PD-1 ICI. Moreover, the combination of CXCR4 and ME1 expression had a more effective prognostic prediction effect. Interestingly, although increased CXCR4 expression indicates a poor prognosis when ICIs are not used, CXCR4 could be a promising biomarker for the response to ICIs. Therefore, the prognostic value of specific subjects should consider different historical backgrounds. The roles of GRGs in the age of immunotherapy still require further investigation, and how to best utilize glycolytic targets to boost antitumor immunity and improve immunotherapies still needs further elucidation.\u003c/p\u003e\u003cp\u003eThis work revealed that GRGs play a nonnegligible role in determining TME diversity and complexity. Evaluating the GRG modification pattern of individual tumors will contribute to enhancing our understanding of TME infiltration and guide more effective immunotherapy strategies.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u0026bull;\u0026nbsp;Funding: This work was supported by the National Natural Science Foundation of\u003c/p\u003e\n\u003cp\u003eChina (grant number: 81873587).\u003c/p\u003e\n\u003cp\u003e\u0026bull; Conflict of interest: The authors declare no conflicts of interest in this study or manuscript.\u003c/p\u003e\n\u003cp\u003e\u0026bull; Data availability: Data of the transcriptome and clinical information were obtained from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/) [13]. The expression data of 407 samples included 32 normal tissues and 375 tumor tissues. Concurrently, the GSE84437 dataset [14] and the GSE13763 dataset [15], which include microarray data and clinical data, were acquired from the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) and used as validation cohort. Data of immunotherapy cohort PRJEB25780 were downloaded from The European Nucleotide Archive (ENA) (https://www.ebi.ac.uk/ena/browser/home) [16]. All the datasets presented in this study can be found in online repositories.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eLu Xu, Jin Liu and Yuanqing An: Data curation (equal); Investigation (equal); Methodology (equal); Writing \u0026ndash; original draft (equal). Lei Zhou and Hui Sun: Data curation (equal); Formal analysis (equal). Zhen Xu, Deqiang Wang and Zhanwen Liang: Data curation (equal); Investigation (equal). Caihua Xu and Bingyi Wang: Data curation (equal); Investigation (equal); Resources (equal); Supervision (equal); Writing \u0026ndash; review \u0026amp; editing (equal). Wei Li: Conceptualization (equal); Data curation (equal); Investigation (equal); Project administration (equal); Writing \u0026ndash; review \u0026amp; editing (equal).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eThrift AP, El-Serag HB. Burden of Gastric Cancer. Clin Gastroenterol Hepatol. 2020;18(3):534\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cgh.2019.07.045\u003c/span\u003e\u003cspan address=\"10.1016/j.cgh.2019.07.045\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou J, Ma X, Bi F, Liu M. Clinical significance of circulating tumor cells in gastric cancer patients. Oncotarget. 2017;8(15):25713\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.18632/oncotarget.14879\u003c/span\u003e\u003cspan address=\"10.18632/oncotarget.14879\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCorrea P. Gastric cancer: overview. Gastroenterol Clin North Am. 2013;42(2):211\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.gtc.2013.01.002\u003c/span\u003e\u003cspan address=\"10.1016/j.gtc.2013.01.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBonomo P, Merlotti A, Olmetto E, et al. What is the prognostic impact of FDG PET in locally advanced head and neck squamous cell carcinoma treated with concomitant chemo-radiotherapy? A systematic review and meta-analysis. Eur J Nucl Med Mol Imaging. 2018;45(12):2122\u0026ndash;38. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00259-018-4065-5\u003c/span\u003e\u003cspan address=\"10.1007/s00259-018-4065-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCascone T, McKenzie JA, Mbofung RM, et al. Increased Tumor Glycolysis Characterizes Immune Resistance to Adoptive T Cell Therapy. Cell Metab. 2018;27(5):977\u0026ndash;e9874. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cmet.2018.02.024\u003c/span\u003e\u003cspan address=\"10.1016/j.cmet.2018.02.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChang CH, Qiu J, O'Sullivan D, et al. Metabolic Competition in the Tumor Microenvironment Is a Driver of Cancer Progression. Cell. 2015;162(6):1229\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2015.08.016\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2015.08.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJustus CR, Sanderlin EJ, Yang LV. Molecular Connections between Cancer Cell Metabolism and the Tumor Microenvironment. Int J Mol Sci. 2015;16(5):11055\u0026ndash;86. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijms160511055\u003c/span\u003e\u003cspan address=\"10.3390/ijms160511055\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2015 May 15.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen S, Cao G, Wu W, et al. Mining novel cell glycolysis related gene markers that can predict the survival of colon adenocarcinoma patients. Biosci Rep. 2020;40(8):BSR20201427. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1042/BSR20201427\u003c/span\u003e\u003cspan address=\"10.1042/BSR20201427\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChelakkot C, Chelakkot VS, Shin Y, Song K. Modulating Glycolysis to Improve Cancer Therapy. \u003cem\u003eInt J Mol Sci\u003c/em\u003e. 2023;24(3):2606. Published 2023 Jan 30. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijms24032606\u003c/span\u003e\u003cspan address=\"10.3390/ijms24032606\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang L, Zhao L, Bi J, et al. Glycolysis gene expression profilings screen for prognostic risk signature of hepatocellular carcinoma. Aging. 2019;11(23):10861\u0026ndash;82. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.18632/aging.102489\u003c/span\u003e\u003cspan address=\"10.18632/aging.102489\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen C, Shi Y, Li Y, et al. A glycolysis-based ten-gene signature correlates with the clinical outcome, molecular subtype and IDH1 mutation in glioblastoma. J Genet Genomics. 2017;44(11):519\u0026ndash;30. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jgg.2017.05.007\u003c/span\u003e\u003cspan address=\"10.1016/j.jgg.2017.05.007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang ZH, Zhang YZ, Wang YS, Ma XX. Identification of novel cell glycolysis related gene signature predicting survival in patients with endometrial cancer. Cancer Cell Int. 2019;19:296. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12935-019-1001-0\u003c/span\u003e\u003cspan address=\"10.1186/s12935-019-1001-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHoadley KA, Yau C, Hinoue T, et al. Cell-of-Origin Patterns Dominate the Molecular Classification of 10,000 Tumors from 33 Types of Cancer. Cell. 2018;173(2):291\u0026ndash;e3046. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2018.03.022\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2018.03.022\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoon SJ, Park J, Shin Y et al. Deconvolution of diffuse gastric cancer and the suppression of CD34 on the BALB/c nude mice model. \u003cem\u003eBMC Cancer\u003c/em\u003e. 2020;20(1):314. Published 2020 Apr 15. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12885-020-06814-4\u003c/span\u003e\u003cspan address=\"10.1186/s12885-020-06814-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKulbe H, Chakravarty P, Leinster DA, et al. A dynamic inflammatory cytokine network in the human ovarian cancer microenvironment. Cancer Res. 2012;72(1):66\u0026ndash;75. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/0008-5472.CAN-11-2178\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-11-2178\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim ST, Cristescu R, Bass AJ, et al. Comprehensive molecular characterization of clinical responses to PD-1 inhibition in metastatic gastric cancer. Nat Med. 2018;24(9):1449\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41591-018-0101-z\u003c/span\u003e\u003cspan address=\"10.1038/s41591-018-0101-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSubramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1073/pnas.0506580102\u003c/span\u003e\u003cspan address=\"10.1073/pnas.0506580102\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRobinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26(1):139\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/bioinformatics/btp616\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btp616\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu T, Hu E, Xu S et al. clusterProfiler 4.0: A universal enrichment tool for interpreting omics data. \u003cem\u003eInnovation (Camb)\u003c/em\u003e. 2021;2(3):100141. Published 2021 Jul 1. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.xinn.2021.100141\u003c/span\u003e\u003cspan address=\"10.1016/j.xinn.2021.100141\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWilkerson MD, Hayes DN. ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking. Bioinformatics. 2010;26(12):1572\u0026ndash;3. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/bioinformatics/btq170\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btq170\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFriedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw. 2010;33(1):1\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCao M, Cai J, Yuan Y, et al. A four-gene signature-derived risk score for glioblastoma: prospects for prognostic and response predictive analyses. Cancer Biol Med. 2019;16(3):595\u0026ndash;605. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.20892/j.issn.2095-3941.2018.0277\u003c/span\u003e\u003cspan address=\"10.20892/j.issn.2095-3941.2018.0277\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeagerty PJ, Lumley T, Pepe MS. Time-dependent ROC curves for censored survival data and a diagnostic marker. Biometrics. 2000;56(2):337\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.0006-341x.2000\u003c/span\u003e\u003cspan address=\"10.1111/j.0006-341x.2000\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. 00337.x.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHope C, Foulcer S, Jagodinsky J, et al. Immunoregulatory roles of versican proteolysis in the myeloma microenvironment. Blood. 2016;128(5):680\u0026ndash;5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1182/blood-2016-03-705780\u003c/span\u003e\u003cspan address=\"10.1182/blood-2016-03-705780\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNewman AM, Liu CL, Green MR, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12(5):453\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nmeth.3337\u003c/span\u003e\u003cspan address=\"10.1038/nmeth.3337\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanahan D. Hallmarks of Cancer: New Dimensions. Cancer Discov. 2022;12(1):31\u0026ndash;46. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/2159-8290.CD-21-1059\u003c/span\u003e\u003cspan address=\"10.1158/2159-8290.CD-21-1059\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeBerardinis RJ, Chandel NS. Fundamentals of cancer metabolism. \u003cem\u003eSci Adv\u003c/em\u003e. 2016;2(5): e1600200. Published 2016 May 27. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/sciadv.1600200\u003c/span\u003e\u003cspan address=\"10.1126/sciadv.1600200\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePavlova NN, Thompson CB. The Emerging Hallmarks of Cancer Metabolism. Cell Metab. 2016;23(1):27\u0026ndash;47. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cmet.2015.12.006\u003c/span\u003e\u003cspan address=\"10.1016/j.cmet.2015.12.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWARBURG O. On the origin of cancer cells. Science. 1956;123(3191):309\u0026ndash;14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.123.3191.309\u003c/span\u003e\u003cspan address=\"10.1126/science.123.3191.309\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamapriyan R, Caetano MS, Barsoumian HB, et al. Altered cancer metabolism in mechanisms of immunotherapy resistance. Pharmacol Ther. 2019;195:162\u0026ndash;71. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.pharmthera.2018.11.004\u003c/span\u003e\u003cspan address=\"10.1016/j.pharmthera.2018.11.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.1160809\u003c/span\u003e\u003cspan address=\"10.1126/science.1160809\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDePeaux K, Delgoffe GM. Metabolic barriers to cancer immunotherapy. Nat Rev Immunol. 2021;21(12):785\u0026ndash;97. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41577-021-00541-y\u003c/span\u003e\u003cspan address=\"10.1038/s41577-021-00541-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang F, Takagaki Y, Yoshitomi Y, et al. Inhibition of Dipeptidyl Peptidase-4 Accelerates Epithelial-Mesenchymal Transition and Breast Cancer Metastasis via the CXCL12/CXCR4/mTOR Axis. Cancer Res. 2019;79(4):735\u0026ndash;46. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/0008-5472.CAN-18-0620\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-18-0620\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMannavola F, Tucci M, Felici C, Passarelli A, D'Oronzo S, Silvestris F. Tumor-derived exosomes promote the in vitro osteotropism of melanoma cells by activating the SDF-1/CXCR4/CXCR7 axis. \u003cem\u003eJ Transl Med\u003c/em\u003e. 2019;17(1):230. Published 2019 Jul 19. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12967-019-1982-4\u003c/span\u003e\u003cspan address=\"10.1186/s12967-019-1982-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMiwa S, Mizokami A, Keller ET, Taichman R, Zhang J, Namiki M. The bisphosphonate YM529 inhibits osteolytic and osteoblastic changes and CXCR-4-induced invasion in prostate cancer. Cancer Res. 2005;65(19):8818\u0026ndash;25. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/0008-5472.CAN-05-0540\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-05-0540\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu C, Zhang Y. Characterization of the prognostic values of CXCR family in gastric cancer. Cytokine. 2019;123:154785. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cyto.2019.154785\u003c/span\u003e\u003cspan address=\"10.1016/j.cyto.2019.154785\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee HJ, Kim SW, Kim HY, et al. Chemokine receptor CXCR4 expression, function, and clinical implications in gastric cancer. Int J Oncol. 2009;34(2):473\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQin Y, Wang F, Ni H, et al. Cancer-associated fibroblasts in gastric cancer affect malignant progression via the CXCL12-CXCR4 axis. J Cancer. 2021;12(10):3011\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7150/jca.49707\u003c/span\u003e\u003cspan address=\"10.7150/jca.49707\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2021 Mar 19.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKukreja P, Abdel-Mageed AB, Mondal D, Liu K, Agrawal KC. Up-regulation of CXCR4 expression in PC-3 cells by stromal-derived factor-1alpha (CXCL12) increases endothelial adhesion and transendothelial migration: role of MEK/ERK signaling pathway-dependent NF-kappaB activation. Cancer Res. 2005;65(21):9891\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/0008-5472.CAN-05-1293\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-05-1293\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHinton CV, Avraham S, Avraham HK. Role of the CXCR4/CXCL12 signaling axis in breast cancer metastasis to the brain. Clin Exp Metastasis. 2010;27(2):97\u0026ndash;105. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10585-008-9210-2\u003c/span\u003e\u003cspan address=\"10.1007/s10585-008-9210-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGatenby RA, Gillies RJ. Why do cancers have high aerobic glycolysis? Nat Rev Cancer. 2004;4(11):891\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nrc1478\u003c/span\u003e\u003cspan address=\"10.1038/nrc1478\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMart\u0026iacute;nez-Reyes I, Chandel NS. Waste Not, Want Not: Lactate Oxidation Fuels the TCA Cycle. Cell Metab. 2017;26(6):803\u0026ndash;4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cmet.2017.11.005\u003c/span\u003e\u003cspan address=\"10.1016/j.cmet.2017.11.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBraun M, Qorraj M, B\u0026uuml;ttner M, et al. CXCL12 promotes glycolytic reprogramming in acute myeloid leukemia cells via the CXCR4/mTOR axis. Leukemia. 2016;30(8):1788\u0026ndash;92. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/leu.2016.58\u003c/span\u003e\u003cspan address=\"10.1038/leu.2016.58\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou AL, Wang X, Yu W, et al. Expression level of PD-L1 is involved in ALDH1A1-mediated poor prognosis in patients with head and neck squamous cell carcinoma. Pathol Res Pract. 2020;216(9):153093. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.prp.2020.153093\u003c/span\u003e\u003cspan address=\"10.1016/j.prp.2020.153093\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColegio OR, Chu NQ, Szabo AL, et al. Functional polarization of tumour-associated macrophages by tumour-derived lactic acid. Nature. 2014;513(7519):559\u0026ndash;63. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nature13490\u003c/span\u003e\u003cspan address=\"10.1038/nature13490\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHusain Z, Huang Y, Seth P, Sukhatme VP. Tumor-derived lactate modifies antitumor immune response: effect on myeloid-derived suppressor cells and NK cells. J Immunol. 2013;191(3):1486\u0026ndash;95. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.4049/jimmunol.1202702\u003c/span\u003e\u003cspan address=\"10.4049/jimmunol.1202702\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCascone T, McKenzie JA, Mbofung RM, et al. Increased Tumor Glycolysis Characterizes Immune Resistance to Adoptive T Cell Therapy. Cell Metab. 2018;27(5):977\u0026ndash;e9874. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cmet.2018.02.024\u003c/span\u003e\u003cspan address=\"10.1016/j.cmet.2018.02.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang Z, Liu Z, Li M, Chen C, Wang X. Increased glycolysis correlates with elevated immune activity in tumor immune microenvironment. EBioMedicine. 2019;42:431\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ebiom.2019.03.068\u003c/span\u003e\u003cspan address=\"10.1016/j.ebiom.2019.03.068\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFuchs CS, Doi T, Jang RW et al. Safety and Efficacy of Pembrolizumab Monotherapy in Patients with Previously Treated Advanced Gastric and Gastroesophageal Junction Cancer: Phase 2 Clinical KEYNOTE-059 Trial [published correction appears in JAMA Oncol. 2019;5(4):579]. \u003cem\u003eJAMA Oncol\u003c/em\u003e. 2018;4(5):e180013. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jamaoncol.2018.0013\u003c/span\u003e\u003cspan address=\"10.1001/jamaoncol.2018.0013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJanjigian YY, Shitara K, Moehler M, et al. First-line nivolumab plus chemotherapy versus chemotherapy alone for advanced gastric, gastro-oesophageal junction, and oesophageal adenocarcinoma (CheckMate 649): a randomised, open-label, phase 3 trial. Lancet. 2021;398(10294):27\u0026ndash;40. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S0140-6736(21)00797-2\u003c/span\u003e\u003cspan address=\"10.1016/S0140-6736(21)00797-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim M, Jeong JY, Seo AN. Biomarkers for Predicting Response to Personalized Immunotherapy in Gastric Cancer. Diagnostics (Basel). 2023;13(17):2782. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/diagnostics13172782\u003c/span\u003e\u003cspan address=\"10.3390/diagnostics13172782\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Published 2023 Aug 28.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcan","sideBox":"Learn more about [BMC Cancer](http://bmccancer.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcan/default.aspx","title":"BMC Cancer","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"glycolysis, gastric cancer, tumor microenvironment, immune cell infiltration, prognostic model","lastPublishedDoi":"10.21203/rs.3.rs-4130368/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4130368/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eGastric cancer (GC) is one of the leading causes of cancer-related death. Glycolysis plays a pivotal role in tumor microenvironment (TME) reprogramming. This study assessed the roles of glycolysis-related genes (GRGs) in predicting prognosis and indicating the immune microenvironment features in gastric cancer patients.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eGene expression data and clinical data of GC patients were obtained from The Cancer Genome Atlas (TCGA)-stomach adenocarcinoma (STAD) cohort and validated using datasets acquired from the Gene Expression Omnibus (GEO). A total of 326 GRGs were identified from the Molecular Signatures Database (MSigDB). Subtypes of GC were delineated via consensus clustering based on GRG expression. A multigene risk score model was developed using multivariate Cox regression analysis. The CIBERSORT and ESTIMATE algorithms were used to evaluate the immune microenvironment. To probe the biological function of critical genes, wound healing assays, transwell invasion assays, and MTT assays were used.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe patients were divided into two groups, namely, the metabolic subtype (cluster A) and immune subtype (cluster B), based on the expression patterns of the GRGs. Patients in cluster B had a worse prognosis. A risk score model based on the expression of six GRGs, including ME1, PLOD2, NUP50, CXCR4, SLC35A3, and SRD35A3, could predict patient prognosis. Knockdown of CXCR4 significantly attenuated the glycolytic capacity, as well as the migration, invasion, and proliferation of GC cells. Interestingly, although both the immune subtype (cluster B) and high-risk groups had unfavorable prognosis, these two cohorts had favorable immune microenvironment and increased expression of immune checkpoint genes. We found that high expression of CXCR4 and low expression of ME1 were positively correlated with the infiltration of CD8\u0026thinsp;+\u0026thinsp;T cells and the response to treatment with an anti-PD-1 immune checkpoint inhibitor.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eIn the present study, we identified that the expression patterns of GRGs could be used to predict the prognosis of GC patients and may be helpful in guiding clinical treatment decisions.\u003c/p\u003e","manuscriptTitle":"Glycolysis-related genes predict prognosis and indicate immune microenvironment features in gastric cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-04-01 16:41:54","doi":"10.21203/rs.3.rs-4130368/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-03-28T10:57:39+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-03-27T08:26:27+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-03-27T08:26:27+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Cancer","date":"2024-03-19T12:41:34+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-cancer","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcan","sideBox":"Learn more about [BMC Cancer](http://bmccancer.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcan/default.aspx","title":"BMC Cancer","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1b8572da-c514-4c66-835a-4632740c9cdf","owner":[],"postedDate":"April 1st, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-08-12T16:04:42+00:00","versionOfRecord":{"articleIdentity":"rs-4130368","link":"https://doi.org/10.1186/s12885-024-12747-z","journal":{"identity":"bmc-cancer","isVorOnly":false,"title":"BMC Cancer"},"publishedOn":"2024-08-08 15:57:51","publishedOnDateReadable":"August 8th, 2024"},"versionCreatedAt":"2024-04-01 16:41:54","video":"","vorDoi":"10.1186/s12885-024-12747-z","vorDoiUrl":"https://doi.org/10.1186/s12885-024-12747-z","workflowStages":[]},"version":"v1","identity":"rs-4130368","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4130368","identity":"rs-4130368","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}