Identification of a lncRNA based ceRNA network signature to establish a prognostic model and explore potential therapeutic targets in gastric cancer

Identification of a lncRNA based ceRNA network signature to establish a prognostic model and explore potential therapeutic targets 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 Article Identification of a lncRNA based ceRNA network signature to establish a prognostic model and explore potential therapeutic targets in gastric cancer Yuanqing An, Xiaomeng Liu, Jin Liu, Deqiang Wang, Wenying Yan, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4989662/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 01 Jul, 2025 Read the published version in Scientific Reports → Version 1 posted 8 You are reading this latest preprint version Abstract Background Numerous studies have demonstrated that long non-coding RNA (lncRNA) play critical roles in regulating physiological processes and contributing to pathological diseases. This study aimed to develop lncRNA-based signatures to predict the prognostic risk of gastric cancer (GC) patients and provide therapeutic guidance. Methods Gene expression profiles and clinical information were obtained from The Cancer Genome Atlas (TCGA) database. Differentially expressed RNAs, including lncRNA, miRNA, and mRNA, in cancerous and adjacent non-cancerous tissues were analyzed using Weighted correlation network analysis (WGCNA) and construction of a lncRNA-miRNA-mRNA competing endogenous RNA (ceRNA) network. Then, a lncRNA-based risk model was constructed by Cox regression and Lasso regression analyses. Results A ceRNA network comprising 235 lncRNAs, 60 miRNAs, and 52 mRNAs was identified. Based on the expression of five lncRNAs (including AC010333.1, LINC01579, AP000695.2, LINC00922 and AL121772.1) screened from the ceRNA network, a lncRNA-based risk model was developed, which effectively predict the prognosis of GC patients. The expression of AP000695.2 was significantly associated with poor prognosis and higher T stage. The knockdown of AP000695.2 inhibited the growth of GC cells both in vitro and in vivo . Transfection with miR-144-3p and miR-7-5p mimics attenuate the up-regulation of targets genes, including CDH11, COL5A2, COL12A1, and VCAN, which was induced by AP000695.2, suggesting a ceRNA mechanism. Additionally, elevated VCAN expression was correlated with poorer survival and a reduced response to anti-PD-1 immune checkpoint inhibitor treatment of GC. Conclusion This study established a lncRNA-based risk model for predicting the prognosis of GC patients and identified a ceRNA mechanism involving AP000695.2-miR-144-3p-VCAN, presenting novel biomarkers and therapeutic targets for GC treatment. Biological sciences/Cancer Health sciences/Biomarkers gastric cancer lncRNA WGCNA prognostic model tumor microenvironment immunotherapy Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Gastric cancer (GC) ranks among the most prevalent malignant tumors globally and is the fourth leading cause of cancer-related mortality [ 1 ] . The primary treatment for early GC is endoscopic resection, while non-early operable gastric cancer is treated with surgery combination with chemoradiotherapy. Nevertheless, the 5-year survival rate for patients with stage III GC undergoing surgery is only 18–50% [ 2 ] . For those patients with advanced gastric cancer (AGC), chemotherapy and targeted therapy are the main treatment modalities. Unfortunately, some patients do not respond to chemotherapy or quickly develop chemoresistance, resulting in the 5-year survival rate of less than 30% [ 3 ] . Therefore, there is a critical need for novel, validated biomarkers to improve prognosis assessing and guide treatment strategies. Non-coding RNAs (ncRNAs) are a class of RNA molecules that do not encode protein but perform essential regulatory functions [ 4 ] . Studies have shown that ncRNAs are involved in the progression of various cancers and can serve as potential biomarkers for cancer diagnosis and prognosis assessment [ 5 ] . MicroRNAs (miRNAs) are a class of ncRNAs approximately 20 nucleotides in length [ 6 ] . The main function of miRNAs is to bind to sequences with partial complementarity on target RNA transcripts, known as microRNA recognition elements (MRE), to regulate gene expression [ 7 ] . Long non-coding RNAs (lncRNAs) are functional RNA molecules exceeding 200 nucleotides in length that can regulate the expression of protein-coding genes by competitively binding to the MREs of miRNAs [ 8 ] . Numerous studies have demonstrated that various lncRNAs and miRNAs are involved in the development of GC and impact the prognosis of patients [ 9 ] . According to the hypothesis of ceRNA regulatory network, lncRNAs and mRNAs can influence each other's levels by competing for limited miRNAs, which plays a crucial role in the regulation of both lncRNAs and mRNAs [ 10 ] . Therefore, constructing a prognosis model based on the ceRNA network is of significant importance for elucidating the molecular mechanisms and predicting the prognosis of GC. Exosomes, a subpopulation of the tumor microenvironment (TME), contain abundant ncRNAs [ 11 ] . Components within the TME significantly influence tumor cell invasion, metastasis and response to immunotherapy [ 12 ] . However, a prognosis model for GC patients that integrates ceRNA with immune status in the TME to exploring potential immunotherapy targets has rarely been reported. In this study, we constructed a prognostic model for GC based on the ceRNA network and further explored the relationship between the lncRNA-miRNA-mRNA axes with TME as well as its implications for immunotherapy. Materials and methods Data source and processing Stomach adenocarcinoma (STAD) RNA expression data and detailed clinicopathologic information were downloaded from TCGA ( https://portal.gdc.cancer.gov/ ). The miRNA expression data totaled 490 samples (445 GC tissues, 45 normal tissues). The lncRNA and mRNA expression data consisted of 407 samples (375 GC tissues,32 normal tissues), of which 340 GC samples had complete clinical data. Then, the data were collated and Gene ID transformed by using the dplyr and tidyr R packages of the R software. Normalized counts were calculated for all patients using the DESeq2 package. To evaluate the predictive potential of gene expression for immunotherapy, we also downloaded data from the immunotherapy cohort PRJEB25780 from the European Nucleotide Archive (ENA) ( https://www.ebi.ac.uk/ena/browser/home ) [ 13 ] . The ROC curves of gene expression were derived to predict the response to anti-PD-1 therapy in GC patients using the pROC package in R software (R Version 3.6.1). Samples of cancerous and normal gastric tissues were collected from 23 GC patients who underwent surgery for GC at the First Affiliated Hospital of Soochow University in 2019 for subsequent gene expression difference analysis. The clinical data of the patients are listed in Supplementary Table 1. The study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (approval no. 2022362) and was conducted in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent was waived by our Institutional Review Board because of the retrospective nature of our study. Analysis of differentially expressed RNAs Fold change (FC) in the expression of miRNAs, lncRNAs, and mRNAs between GC tissues and normal tissues were analyzed using the edgeR R package. Differentially expressed mRNAs, miRNAs, and lncRNAs were screened by | log2FC | > 1 and false discovery rate (FDR) 1 was defined as up-regulation and log2FC < -1 as down-regulation. Weighted correlation network analysis (WGCNA) WGCNA analyzes the correlation between the expression patterns of genes and then categorizes genes with similar expression patterns into the same module eigengenes (ME). WGCNA was constructed for miRNA, lncRNA, and mRNA. The least value for which the scale-free topology fit R^2 index > 0.85 was chosen as the soft threshold. The Pearson method was used to calculate the correlation between all genes and construct the network, and the similarity of gene was assessed by Topological Overlap Matrix (TOM), with 1-TOM as the measured value. The number of genes in each ME was set to ≥ 50, and genes with high correlation were clustered into the same ME after forming a co-expression network. Correlation coefficients and P -values between MEs and clinical features of GC patients were calculated and the results were visualized by heatmap. The R packages utilized in these operations included the WGCNA, hclust and cuttreeDynamic packages. Protein interaction (PPI) networks Protein interaction networks were formed using the Search Tool for the Retrieval of Interaction Gene/Proteins (STRING) database ( https://string-db.org/ ). Degree Centrality, Betweenness Centrality and Closeness Centrality of the network were calculated using Cytoscape software to assess the extent to which genes occupy central positions throughout the network. Competing endogenous RNA (ceRNA) network The miRNA-mRNA correspondence was obtained from starbase, targetscan and miRTarbase databases, and the miRNA-lncRNA correspondence was retrieved from lncbase and starbase databases. The ceRNA network associated with prognosis was constructed. Construction and validation of the risk model The samples in TCGA were randomly divided into two groups as training set (170 cases) and validation set (170 cases). The prognostic model was constructed using the training set, and the feasibility of the prognostic model was verified using the validation set. univariate Cox regression analysis was performed on lncRNAs in the ceRNA network, then Lasso regression was used to minimize prediction errors, the prognostic model was constructed in the multivariate Cox regression analysis and screened by the AIC method, finally. The risk score formula for each patient in the prognostic model: Risk score = ∑ (Expi * Coefi); Coefi and Expi denote the risk coefficient and RNA expression, respectively. The patients were divided into high- and low-risk groups based on the median risk score, the differences between the two risk groups in terms of gene expression, survival status, and risk scores were visualized. Kaplan–Meier survival curves were constructed to visualize differences of survival rate between the high- and low-risk groups. Five-year ROC curves were generated to evaluate the accuracy of the prognostic model. Univariate and multivariate Cox regression analyses were used to evaluate the independence of the risk score, and the results were shown in forest plots. Furthermore, a nomogram was constructed to predict the survival time of patients. All analyses were performed using R for which the following packages were employed: glmnet, survival, survminer, regplot, timeROC and rms packages. Gene Set Enrichment Analysis (GSEA) The GSEA enrichment analysis method allows the identification of biological processes and signaling pathways of genes. Signaling pathway enrichment analysis was performed on individual genes using GSEA 4.1.0 software. Functional enrichment results were filtered by P < 0.05 and FDR < 0.1. Correlation of gene expression levels with immunological checkpoints and degree of immune cell infiltration We used the CIBERSORT and XCELL methods to evaluate the abundance of immune cell infiltration among the samples in the TCGA–STAD dataset. The expression levels of immunological checkpoints between the high- and low-expression groups of key genes are shown in boxplots. Cell culture and gene knock-down techniques The human GC cell line MKN-45 was purchased from the American Typical Culture Collection (ATCC). MKN-45 cells were cultured in DMEM medium (HyClone, USA) with 10% FBS (Gibco, USA) and 1% penicillin/streptomycin (HyClone, USA) at 37°C in a humidified atmosphere with 5% CO2. Cells were passaged every 2–3 days to maintain growth. Lentivirus-encoding shRNAs for AP000695.2 (VSVG-Lentai-hU6-shRNA-AP000695.2-puro-hEF1a-3xFlag) and negative control (NC) lentivirus (VSVG-Lentai-hU6-shRNA-NC-puro-hEF1a-3xFlag) were compounded by Shanghai Taitool Bioscience Co (Shanghai, China). MKN-45 cells were harvested and used for further analysis after infecting lentivirus with 72 hours. RNA isolation and real-time quantitative reverse transcriptase PCR (qRT-PCR) Total RNA was extracted using Trizol Reagent (Thermo Fisher Scientific, USA) according to the manufacturer’s protocol. Reverse transcription was performed using the PrimeScript RT Reagent Kit (TaKaRa, Japan), yielding 1 µg of total RNA in a final volume of 20 µl according to the manufacturer's protocol. A Light Cycler 96 (Roche) real-time PCR system was used to detect complementary DNA (cDNA) for the same amount of RNA. The reaction system (13 µl) contained the reverse and forward primers, the corresponding cDNAs, and the SYBR Green PCR master mix (Roche, USA). All data were analyzed using GAPDH gene expression as an internal standard. Primers for specific genes are listed in Supplementary Table 2. Methyl thiazolyl tetrazolium (MTT) assay Cell proliferation was detected by the MTT [3-(4,5-dimethyl-2-yl)-2,5-diphenyl tetrazolium bromide] assay. Well-grown cells were harvested for digestion and centrifuged. Cells were plated in 96-well plates at 5000 cells per well and incubated for 24 hours. MTT (Beyotime, China) was added to each well at the same time every day during the experiment, to give a final concentration of 5 mg/ml, and the 96-well plate was incubated in an incubator. After 4 hours the medium was aspirated and 400 µl of dimethyl sulfoxide (DMSO, Sigma, USA) was added to each well to dissolve it completely. The absorbance of each well was measured at 490 nm using a microplate reader (Thermo Fisher Scientific, USA). Subcutaneous xenograft nude mouse model Five-week-old female BALA/c nude mice were purchased from Hangzhou Ziyuan Experimental Animal Technology Co, Ltd. (China) and housed under specific pathogen-free conditions. MKN-45 cells transfected with AP000695.2-shRNA lentivirus were injected into subcutaneous tissues of nude mice in a total volume of 5 × 10 6 cells, respectively. The long and short diameters of the subcutaneous tumors were measured every 3 days using vernier calipers. After the experiments were conducted for 3 weeks, the mice were anaesthetized and the tumors were resected. The length (L) and width (W) of each subcutaneous tumor were measured by a caliper. The calculation method for tumor volume (TV): TV = (L * W 2 )/2. Cell transfection Wild type and mutant type luciferase reporter gene vectors for AP000695.2 and VCAN, VCAN overexpression vectors, and miRNA-mimics were constructed by General Bio Co. (Anhui, China). Transfection of them were performed using a Lipofectamine Kit (Thermo Fisher Scientific, USA). The well-grown cells were seeded in 6-well plates and 24 hours later, 3 mL of fresh complete medium was added after the cells were attached to the wall. Then, a new EP tube was taken and prepared as follows: 3.75 µL lipo3000 + 125 µL basal medium + 50 nM miRNA mimic per well, and then added to the wells separately. The corresponding experiments could be performed after transfecting for 24 hours. Dual-luciferase reporter assay Dual-luciferase reporter assay was performed using Dual Luciferase Reporter Gene Assay Kit (Promega, America) and Promega Chemiluminescent Analyzer. After transfecting for 24 hours, the cells were lysed, centrifuged (12,000 g, 4℃, 10 min). The supernatant was retained, 100 µL firefly luciferase detection reagent was added to every 20 µL sample, and then detected using a Chemiluminescent Analyzer. Subsequently, renilla luciferase detection reagent was added to the sample and detected. Patient samples and immunohistochemistry (IHC) Formalin-fixed and paraffin-embedded tumor tissue collected from 67 clinical trials was retrospectively analyzed from the first Affiliated Hospital of Soochow University, from 2020 to 2023. All the patients’ survival intervals were available and dated to the end of December 2023. The study protocol was performed under the guidelines outlined in the Declaration of Helsinki and was approved by the Ethics Committee of the first Affiliated Hospital of Soochow University. Expressions of target genes were evaluated by using immunohistochemistry. The formalin-fixed and paraffin-embedded tissue blocks were cut to 5-µm sections, Paraffin-embedded tumor tissue sections were dewaxed and performed to Antigen retrieval using antigen repair buffer under high pressure and high-temperature conditions. Then, the sections were blocked with 5% BSA and incubated with primary antibodies against VCAN and CD8 at 4°C overnight. Next, the sections were incubated with secondary antibody for 20 min at room temperature and then visualized with diaminobenzidine (DAB) Substrate Kit. Statistical analysis All data were analyzed and plotted using R software (Version 3.6.1), GraphPad Prism software (Version 8.3.1), and SPSS (Version 19.0). t-tests were used to analyze the differences between the independent samples of the two groups. The P -value of < 0.05 was considered statistically significant. Results Identification of prognosis-associated DEmiRNAs, DElncRNsA and DEmRNsA, and construction of ceRNA networks The total number of differentially expressed lncRNAs (DElncRNAs) in the TCGA-STAD cohort was 5,578, of which 3,615 were upregulated and 1,963 were downregulated ( Fig. 1 A ) . To explore the DElncRNAs highly correlated with prognosis, we performed WCGNA to identify the gene modules significantly associated with survival status. The optimal soft-threshold power β was set as 3 to ensure the construction of a scale-free network (scale-free R 2 > 0.85) (Fig. 1 B, C). The minimum number of lncRNAs per module was set as 50, resulting in the clustering of lncRNAs with similar expression patterns into 11 modules (Fig. 1 D, E). Among these, the blue module showed the strongest correlation with survival status (R = 0.13, P = 0.02) (Fig. 1 F). This module contained 4,377 lncRNAs, which intersected with the 5,578 DElncRNAs to yield 1,259 prognosis-associated lncRNAs ( Fig. 1 G ) . In the TCGA-STAD cohort, there were 238 differentially expression miRNAs (DEmiRNAs), which 148 were upregulated and 90 were downregulated ( Fig. 1 H ) . We again used WCGNA to identify gene modules highly correlated with survival status, setting the optimal soft-threshold power β at 2 to ensure a scale-free network (scale-free R 2 > 0.85) (Fig. 1 I, J). The minimum number of miRNAs in each module was set at 50, clustering miRNAs with similar expression patterns into 5 modules ( Fig. 1 K, L ) . The turquoise module showed the strongest correlation with survival status (R = -0.12, P = 0.001) (Fig. 1 M). This module contained 234 miRNAs, which intersected with the 238 DEmiRNAs to yield 124 prognosis-associated miRNAs (Fig. 1 N). For differentially expressed mRNAs (DEmRNAs) in the TCGA-STAD cohort, the total was 4480, with 2125 upregulated and 2355 downregulated ( Fig. 1 O ) . Subsequently, we used WGCNA to identify gene modules significantly correlated with survival status, setting the optimal soft-threshold power β at 3 to ensure a scale-free network (scale-free R 2 > 0.85) (Fig. 1 P, Q). The minimum number of mRNAs in each module was set at 50, clustering mRNAs with similar expression patterns into 20 modules ( Fig. 1 R, S ) . The green module showed the strongest correlation with survival status (R = 0.16, P = 0.002) (Fig. 1 T). This module contained 763 mRNAs, which intersected with the 4,480 DEmRNAs to yield 161 prognosis-associated mRNAs (Fig. 1 U). The Protein-Protein Interaction (PPI) Network of the 161 prognosis-associated mRNAs was generated using the Search Tool for the Retrieval of Interacting Genes (STRING). Using Cytoscape software, we constructed a network and calculated network parameters, identifying 62 closely interacting mRNAs. The 62 mRNAs from the PPI network, along with 1259 prognosis-associated lncRNAs and 124 prognosis-associated miRNAs, were used to construct the ceRNA network. Finally, a ceRNA network containing 235 lncRNAs, 60 miRNAs, and 52 mRNAs was obtained ( Fig. 1 V ) . Construction and verification of a ceRNA-related prognostic model in GC The samples from the TCGA-STAD cohort were randomly divided into two groups: the training set (n = 170) and the validation set (n = 170). In the training set, 235 lncRNAs in the ceRNA network were analyzed by univariate Cox regression, and the obtained overall survival (OS)-associated lncRNAs were refined using Lasso regression to avoid overfitting ( Fig. 2 A, B ) . This process identified 12 lncRNAs associated with the prognosis of with GC patients. Subsequently, multivariate Cox regression analysis was performed on these 12 lncRNAs based on Akaike information criterion (AIC) values. Finally, 5 lncRNAs were selected out to construct the prognostic model, including AC010333.1, LINC01579, AL121772.1, AP000695.2 and LINC00922, and the corresponding regression coefficients were calculated. The Risk Score was determined by using the following formula: Risk Score = (1.70 × AC010333.1) + (0.48 × LINC01579) + (-0.54 × AL121772.1) + (0.37 × AP000695.2) + (1.28 × LINC00922). The expression of these 5 lncRNAs between two risk groups were visualized by a heatmap ( Fig. 2 C ) . Risk scores were calculated for each patient, and the training cohort was divided into high- and low-risk groups based on the median risk score. The risk score distribution and survival time distribution plots showed that patients in the high-risk group had shorter survival time ( Fig. 2 D, E ) . The Kaplan-Meier (K-M) survival curves demonstrated that patients in the low-risk group had a significant survival advantage over those in the high-risk group ( P = 7.842e-03) ( Fig. 2 F ) . To validate the model's performance, a receiver operator characteristic (ROC) curve predicting 5-year survival was plotted, yielding an area under the curve (AUC) value was 0.608 ( Fig. 2 G ) . To further evaluate the predictive efficacy of the risk model, risk scores for each patient in the validation set were calculated. Based on the median risk score, the validation cohort was also divided into high- and low-risk groups. The differential expression of 5 lncRNAs between the two groups were displayed in a heatmap ( Fig. 2 H ) . The risk score distribution and survival time distribution plots showed that patients in the high-risk group had shorter survival time ( Fig. 2 I, J ) . The K-M survival curves showed that patients in the low-risk group had a significant survival advantage over those in the high-risk group ( P = 3.198e-02) ( Fig. 2 K ) . An ROC curve predicting 5-year survival was plotted for the validation set, with an AUC value of 0.632 ( Fig. 2 L ) . Evaluation of the clinical values of the prognostic risk model The prognostic value of clinicopathologic features and risk scores was explored through univariate and multivariate Cox regression analyses on the TCGA-STAD cohort. Univariate Cox analysis revealed that age ( HR = 1.021, P = 0.03), stage ( HR = 1.500, P < 0.001), T stage ( HR = 1.302, P = 0.03), N stage ( HR = 1.267, P = 0.007) and risk score ( HR = 1.229, P < 0.001) were significantly associated with OS ( Fig. 2 M ) . In multivariate Cox analysis, age ( HR = 1.033, P < 0.002) and risk score ( HR = 1.228, P < 0.001) demonstrated independent prognostic value ( Fig. 2 N ) . To compare the predictive accuracy of the risk score with various clinical factors, multivariable ROC analysis was performed ( Fig. 2 O ) . Overall, the risk score emerged as the most accurate prognostic indicator (AUC = 0.617). Utilizing both clinical parameters and risk score, a nomogram model was established to predict OS ( Fig. 2 P ) . The 1-, 2-, and 3-year predicted OS values were calculated based on the total points from each parameter, with higher point totals indicating worse survival outcomes. These curves showed close alignment with the ideal diagonal line, indicating the nomogram's precision in predicting patient outcomes ( Fig. 2 Q ) . Characterization of lncRNAs in the prognostic risk model By analyzing K-M survival curves of the TCGA-STAD cohort, it was observed that GC patients with high expression of AL121772.1 ( P = 0.042) suggested a relatively better prognosis ( Fig. 3 A ) . Conversely, patients with high expression of AP000695.2 ( P = 0.039) demonstrated poor prognosis ( Fig. 3 B ) . However, the expression levels of AC010333.1, LINC00922, and LINC01579 did not affect patient prognosis ( Fig. 3 C-E ) . The expression levels of the five lncRNAs in the TCGA-STAD cohort were investigated, revealing that AL121772.1 ( P = 5.9e-09), AP000695.2 ( P = 1.3e-09), AC010333.1 ( P = 5.1e-06), and LINC00922 ( P = 3.9e-12) were expressed at higher levels in tumor tissues compared to normal tissues ( Fig. 3 F-I ) . In contract, LINC01579 was expressed at lower levels in tumor tissues than in normal tissues ( P = 0.03) ( Fig. 3 J ) . To validate these finding further, tumor tissues and adjacent tissues from 23 clinical GC patients were collected. The expression of the five lncRNAs in these samples was examined using qRT-PCR, and the differential expression profiles were consistent with those observed in the TCGA-STAD cohort ( Fig. 3 K-O ) . Following this, we focused on the expressing the five lncRNAs during tumor progression. Notably, AP000695.2 was signigicantly higher at T3 stage compared to other T stages ( P < 0.05) ( Fig. 3 P ) . There were no significant differences in the expression of the five lncRNAs across N and M stage ( Fig. 3 Q, R ) . Functional experiments were conducted using the MKN-45 cell line, which was infected with lentivirus carrying AP000695.2-shRNA. The efficiency of the lentiviral vectors was confirmed by qPCR ( Fig. 3 S ) . The MTT assay demonstrated that, knockdown of AP000695.2 significantly inhibited the proliferation of MKN-45 cells ( Fig. 3 T ) . This was further validated in vivo using a subcutaneous xenograft nude mouse model, where the growth of tumors was weakened by knocking down AP000695.2 ( Fig. 3 U, V ) . Based on analysis of the five lncRNAs in the prognostic model, AP000695.2 emerged as the most significant gene related to the prognosis and is a promising potential target for GC treatment. The lncRNA-miRNA-mRNA regulatory axes of AP000695.2 The Sankey diagram demonstrated the lncRNA-miRNA-mRNA axes associated with AP000695.2 in the ceRNA network, and obtained a total of 13 potential ceRNA-mRNAs ( Fig. 4 A, Supplemental Table 3 ) . K-M survival curves of the TCGA-STAD cohort revealed that patients with low expression of CDH11 ( P = 0.046) ( Fig. 4 B ) , COL12A1 ( P = 0.025) ( Fig. 4 D ) , COL5A2 ( P = 0.009) ( Fig. 4 G ) , and VCAN ( P = 0.004) ( Fig. 4 N ) had a significant survival advantage. Conversely, the expression levels of COL11A1 ( Fig. 4 C ) , COL1A2 ( Fig. 4 E ) , COL5A1 ( Fig. 4 F ) , LOXL2 ( Fig. 4 H ) , PRRX1 ( Fig. 4 I ) , STC2 ( Fig. 4 J ) , THBS2 ( Fig. 4 K ) , TIMP3 ( Fig. 4 L ) , and TNFSF11 ( Fig. 4 M ) did not appear to significantly affect the prognosis of GC patients ( P > 0.05). To validate the lncRNA-miRNA-mRNA axes illustrated in Fig. 4 A, MKN-45 cells were transfected with miR-144-3p-mimic and miR-7-5p-mimic. In AP000695.2 knockdown cells, the down-regulation capacity of miR-144-3p-mimic on CDH11 ( Fig. 4 O ) , COL5A2 ( Fig. 4 P ) , COL12A1 ( Fig. 4 Q ) and VCAN ( Fig. 4 R ) expression was diminished ( P < 0.01). Similarly, the knockdown of AP000695.2 also reduced the down-regulation effect of miR-7-5p-mimic on COL5A2 ( Fig. 4 S ) , COL12A1 ( Fig. 4 T ) and VCAN ( Fig. 4 U ) expression ( P < 0.01) but showed no similar function on expression of CDH11 (Supplementary Fig. 1). To elucidate the molecular mechanisms related to lncRNA-miRNA-mRNA axes based on CDH11, COL12A1, COL5A2 and VCAN in GC patients, GSEA analysis was conducted on high- and low-expression groups of these genes. GSEA enrichment analysis indicated that the CDH11 low-expression group was enriched in base excision repair, DNA replication, and mismatch repair pathways, whereas the high-expression group was primarily associated with the calcium signaling pathway, extracellular matrix (ECM) receptor interaction, and TGF-β signaling pathway ( Fig. 4 V ) . The high-expression groups of COL5A2 ( Fig. 4 W ) , COL12A1 ( Fig. 4 X ) and VCAN ( Fig. 4 Y ) were mainly enriched in ECM receptor interaction, JAK-STAT signaling pathway, MAPK signaling pathway, TGF-β signaling pathway, and Toll-like receptor signaling pathway. The aforementioned signaling pathways have been well elucidated in their role in regulating the tumor immune microenvironment [ 14 , 15 ] , making it reasonable to hypothesize that the AP000695.2-based lncRNA-miRNA-mRNA axis may be involved in tumor immune evasion. Estimation of immune response and immune cell infiltration using the key genes of lncRNA-miRNA-mRNA axes related to AP000695.2 To explore potential indicators for immunotherapy within the AP000695.2-related lncRNA-miRNA-mRNA axes, we utilized RNA-Seq data from the immunotherapy cohort PRJEB25780 available at 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 anti-programmed cell death protein 1 (PD-1) therapy, grouping them by clinical outcomes into complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD) [ 13 ] . We compared the expression of 13 mRNAs related to the AP000695.2 axis between the Pembrolizumab treatment-sensitive (CR/PR) group and the Pembrolizumab treatment-resistant (SD/PD) group. The expression levels of COL12A1 ( P < 0.05), COL1A2 ( P < 0.05), COL5A1 ( P < 0.05) and VCAN ( P < 0.001) in the CR/PR group were lower ( Fig. 5 A ) . The ROC curves for COL12A1 (AUC = 0.172), COL1A2 (AUC = 0.720), COL5A1 (AUC = 0.745) and VCAN (AUC = 0.816) demonstrated their predictive effect on immunotherapy response ( Fig. 5 B-E ) . Next, we calculated the proportions of infiltrating immune cells in the TCGA-STAD using the CIBERSORT method. We compared the proportions of each immune cell between the differential expression groups of COL12A1 ( Fig. 5 F ) , COL1A2 ( Fig. 5 H ) , COL5A1 ( Fig. 5 J ) and VCAN ( Fig. 5 L ) . Some immune cells, including naive B cells, memory B cells and Plasma cells, were more enriched in the low-expression groups, whereas M2 macrophages were more enriched in the high-expression groups. Notably, CD8 + T cells were more enriched only in the low-expression groups of COL12A1 and COL5A1. Proportions of infiltrating fibroblasts in the TCGA-STAD cohort were calculated using the XCELL method, revealing low fibroblasts infiltration in the low-expression groups of COL12A1 ( Fig. 5 G ) , COL1A2 ( Fig. 5 I ) , COL5A1 ( Fig. 5 K ) and VCAN ( Fig. 5 M ) . Additionally, the low-expression groups of COL12A1 ( Fig. 5 N ) , COL1A2 ( Fig. 5 O ) , COL5A1 ( Fig. 5 P ) and VCAN ( Fig. 5 Q ) preserved lower expression of immune checkpoints. Validation of AP000695.2/miR-144-3p/VCAN axis As it showed in Fig. 4 U, both AP000695.2 and miR-144-3p affect the expression of VCAN. To further investigate the regulatory network of AP000695.2, miR-144-3p and VCAN, we validated the AP000695.2/miR-144-3p/VCAN axis by dual-luciferase reporter experiments. The online database starbase ( http://starbase.sysu.edu.cn ) predicted the binding sites of miR-144-3p to AP000695.2, The mutant plasmid vectors were constructed and specific binding sites are shown ( Fig. 6 A ). The dual-luciferase reporter assay displayed that MKN-45 ( Fig. 6 B ) . knockdown of AP000695.2 significantly inhibited the proliferation of MKN-45 cells, transfection of miR-144-3p mimic in MKN-45 cells that had knocked down AP000695.2 resulted in further diminution of the proliferative capacity of MKN-45 cells cells ( Fig. 6 C ) . Next, by targetscan database ( https://www.targetscan.org/vert_80/ ), we predicted the binding site of miR-144-3p to the 3'UTR of VCAN ( Fig. 6 D ) . The luciferase report assay was applied to verify miR-144-3p directly targeted VCAN in MKN-45 cells. The results showed the luciferase activity was obviously declined when VCAN-wt and miR-144-3p mimic were transfected into MKN-45 cells, while VCAN-mut had no effect ( Fig. 6 E ) . Overexpression of VCAN promoted cell proliferation, which was impaired by transfection of miR-144-3p mimic ( Fig. 6 F ) . Based on the above results, The regulation of AP000695.2, miR-144-3p and VCAN interacted with each other to affect cell proliferation. In addition, the difference in VCAN expression levels between Pembrolizumab treatment-sensitive group and Pembrolizumab treatment-resistant group in the immunotherapy queue PRJEB25780 is most significant, also the expression of VCAN exhibiting the highest predictive power for immunotherapy response. CD8 + T cells are crucial anti-tumor immune cells within the tumor microenvironment and are the primary immune cells in tumor targeting therapy. PD-1 ligand (PD-L1) can be targeted to alleviate the exhaustion of CD8 + T cells [ 16 ] . Therefore, we used immunohistochemistry to detect the CD8 + T cell infiltration in tumor tissues of clinical gastric cancer patients with different levels of VCAN expression. The clinical characteristics of 67 gastric cancer patients and the IHC score of VCAN in tumor tissue were demonstrated (Supplementary Table 4). In tissues with higher VCAN expression, the infiltration of CD8 + T cell is comparatively lower ( Fig. 6 G, H ) K-M survival curves of the clinical gastric cancer patients showed Progression-Free-Survival (PFS) stratified on high and low expression of VCAN ( Fig. 6 I ) , which was consistent with the conclusion in the K-M survival curve of VCAN in the TCGA-STAD cohort in Fig. 4 N. Discussion Currently, numerous studies have focused on ceRNA networks in cancer. LncRNAs and miRNAs have been shown to play regulatory role in various biological processes and tumor-related signaling pathways. For example, researchers have found that the lncRNA-ATB can be activated by TGF-β, competitively binding to the miR-200 family, thereby inducing epithelial-mesenchymal transition (EMT) [ 17 ] . The lncRNA-MALAT1 stimulates PI3K/Akt axis to accelerate proliferation and invasion of tumor cells and to stimulate chemoresistance [ 18 ] . In gastric cancer, the lncRNA-THAP7-AS1 inhibits the transcription of miR-22-3p and miR-320a, playing a carcinogenic role [ 19 ] , and lncRNA-NKX2-1-AS1 activates the VEGFR-2 signal pathway by regulating miR-145-5p, promoting GC progression and angiogenesis [ 20 ] . In these processes, both lncRNAs and miRNAs act as ceRNAs. Therefore, studies on the ceRNA network in GC are expected to provide new insights the mechanism of GC occurrence and development and simultaneously guide the identification of new therapeutic targets for GC. Weighted correlation network analysis (WGCNA) can be used to find clusters (modules) of highly correlated genes and associate these gene modules with biological traits. Hence, we constructed a prognostic model for GC based on the ceRNA network using WGCNA. Firstly, we screened differentially expressed RNAs and identified prognosis-related RNA modules by WGCNA. The selected RNAs were used to construct the ceRNA network. Then, we developed a ceRNA-related prognostic model composed of 5 lncRNAs (AC010333.1, LINC01579, AP000695.2, LINC00922, AL121772.1) using Cox regression analysis and Lasso regression analysis. The accuracy of the prognostic model was predicted and validated by a validation set. In addition, univariate and multivariate Cox regression analysis suggested that the risk score could be serve as an independent prognostic factor for GC patients. Regarding the roles of the 5 lncRNAs (AC010333.1, LINC01579, AP000695.2, LINC00922, AL121772.1) in tumors within the ceRNA-related prognostic model, LINC00922 has been extensively studied. Ji et al. found that LINC00922 accelerated GC progression through the miR-204-5p/HMGA2 axis [ 21 ] . Additionally, LINC00922 promotes the proliferation, migration and invasion of lung cancer through the miRNA-204/CXCR4 axis [ 22 ] . In ovarian cancer, LINC00922 can competitively bind miR-361-3p to promote CLDN1 expression and activate the Wnt/β-catenin signaling pathway, accelerating ovarian cancer progression [ 23 ] . Moreover, in hepatocellular carcinoma, LINC00922 promotes cell proliferation, migration, invasion, and EMT by regulating miR-424-5p/ARK5 axis [ 24 ] . In addition, some researchers applied an APEX-seq method to identify plasma membrane-associated RNAs and found that RNA-AL121772.1 has a considerable affinity for sphingomyelin (SM) [ 25 ] . Previous studies have confirmed that LINC01579 competitively binds to miR-139-5p and regulates cell proliferation and apoptosis in glioblastoma by modulating EIF4G2 expression [ 26 ] . Subsequently, we plotted Kaplan-Meier (K-M) survival curves for the five lncRNAs in the ceRNA-related prognostic model in GC. The results indicated that the expression level of AP000695.2 was significantly negatively correlated with the prognosis of GC patients. Moreover, the expression level of AP000695.2 was significantly higher in GC tissues than in normal tissues and correlated with the T stage. Therefore, we speculated that AP000695.2 may promote the development of GC and serve as a potential target for GC treatment. Functional experiments with GC cells showed that the reducing AP000695.2 expression suppressed the proliferative ability of GC cells, which was also verified in vivo using the subcutaneous xenograft nude mouse model. Besides, we identified a ceRNA network comprising AP000695.2, two miRNAs (miR-144-3p, miR-7-5p), and thirteen mRNAs (CDH11, COL12A1, STC2, THBS2, COL1A2, COL5A1, COL5A2, LOXL2, PRRX1, TIMP3, COL11A1, TNFSF11, VCAN). The K-M survival curves showed that low expression levels of CDH11, COL5A2, COL12A1, and VCAN were associated with improved prognosis. CDH11 has been implicated in the development and progression of various cancers, including breast cancer, pancreatic cancer and gastric cancer [ 27 – 29 ] . Both COL5A2 and COL12A1, members of the collagen family, have been associated with the progression of gastric cancer [ 30 ] . VCAN exhibits pro-carcinogenic properties in several malignancies, such as breast cancer, ovarian cancer and bladder cancer [ 31 – 33 ] . In gastric cancer, VCAN is reportedly regulated by miRNA or cirRNA-miRNA axis to facilitate gastric cancer progression [ 34 ] , and high VCAN expression is associated with resistant to immunotherapy [ 35 ] . Through miR-mimic and real-time PCR, we revealed that AP000695.2 competitively binds miR-144-3p with CDH11, COL12A1, VCAN and miR-7-5p with COL5A2, COL12A1. The GSEA enrichment suggested that CDH11, COL5A2, COL12A1, VCAN may regulate downstream signaling pathways, including Hedgehog signaling pathway, JAK-STAT signaling pathway, MAPK signaling pathway, TGF-β signaling pathway, Wnt signaling pathway. Based on the above studies, we hypothesized that AP000695.2 enhances the expression of CDH11, COL5A2, COL12A1 and VCAN through competitive binding miR-144-3p and miR-7-5p. This interaction likely promotes the proliferation of GC cells through the tumor-associated signaling pathway, resulting in poorer patient survival. Cancer immunotherapy exhibits promising antitumor efficacy across various malignancies, particularly through immune checkpoint blockade with antibodies, demonstrating highly durable response rates in phase I studies involving patients with advanced melanoma, non-small-cell lung cancer, and other solid tumors [ 36 , 37 ] . The TME is crucial in cancer progression and therapeutic outcome [ 38 ] . According to CIBERSORT and XCELL algorithms, we predicted the abundance of infiltrating immune cells in patients with GC from the TCGA datasets. Low expression groups of COL12A1, COL1A2, COL5A1 and VCAN presented a favorable immune microenvironment and lower expressions of immune checkpoints. Increased expressions of immune checkpoints can directly inhibit specific immune cells activation and contribute to immune escape, resulting in unfavorable outcomes [ 39 ] . This observation elucidates why GC patients in the low expression groups of COL12A1 and VCAN exhibited better prognosis. Although there was no significant difference in PDCD1 expression between high- and low-expression groups of COL12A1, COL1A2, COL5A1 and VCAN, the low expression groups of COL12A1, COL1A2, COL5A1 and VCAN in the immunotherapy cohort PRJEB25780 from ENA indicated a promising response to anti-PD-1 immunotherapy, especially the VCAN. This may be attributed to the limited significance of PD-1 as a solitary evaluation indicator for the therapeutic response of immune checkpoint inhibitors [ 40 ] . Moreover, Due to the significant impact of VCAN expression on the outcomes of immunotherapy and its improved predictive value for immune responses, we demonstrated that elevated VCAN expression leads to reduced infiltration of CD8 + T cells in gastric cancer tissue, as well as a less favorable prognosis in clinical gastric cancer patients, as evidenced by IHC analysis. In summary, we employed WGCNA bioinformatic analysis and ceRNA network construct to develop a prognostic model. Further analysis identified AP000695.2 as a promoter of GC cell proliferation and elucidate its regulatory ceRNA network. We propose that AP000695.2 forms the AP000695.2/miR-144-3p/VCAN axis by competitively binding to miR-144-3p with VCAN, and forms the AP000695.2/miR-144-3p, miR-7-5p/COL12A1 axis by competitively binding to miR-144-3p and miR-7-5p with COL12A1, thereby influencing the expression of VCAN and COL12A1. Changes in VCAN and COL12A1 expression affects tumor-related signaling pathways such as MAPK signaling pathway, TGF-β signaling pathway, Wnt signaling pathway, etc., which in turn impact the prognosis of GC patients. In addition, analysis of immune response and TME revealed that the expression of VCAN affects the immune microenvironment and exhibits a high predictive ability for immunotherapy response. Therefore, we further validated the binding between RNAs in the AP000695.2/miR-144-3p/VCAN axis. Subsequently, we demonstrated in clinical gastric cancer patients that VCAN expression inhibits the infiltration of CD8 + T cells within the TME and influences patient prognosis. However, the specific regulatory mechanisms of the AP000695.2/miR-144-3p/VCAN axis as well as the pathways they regulate, require further verified. Whether these mechanisms can be exploited to enhance anti-tumor immunity and improve immunotherapy responses remains to be elucidated. Declarations Author Contribution L.X., W.L. and G.H. designed the research. Y.A., X.L. and J.L. performed the majority of experiments and analyzed data. D.W., W.Y. and W.Y. contributed to clinical samples and reagents. All authors reviewed the manuscript. Acknowledgement We are grateful for participation and cooperation from gastric cancer patients. This work was supported in part by Suzhou Medical Engineering Collaborative Innovation Research Project (SLJ2022001). Data Availability Transcriptome data and clinicopathologic information were downloaded from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). Data of immunotherapy cohort PRJEB25780 were obtained from the European Nucleotide Archive (ENA) (https://www.ebi.ac.uk/ena/browser/home). Protein interaction networks were constructed from the Search Tool for the Retrieval of Interaction Gene/Proteins (STRING) database (https://string-db.org/). Further details of all data generated or analyzed for this study can be obtained from the corresponding author upon reasonable request. Ethical Compliance Statement: All animal experiments were performed in accordance with the institutional guidelines of the Laboratory Animal Center of Soochow University and were approved by the Animal Ethics Committee of Soochow University. ARRIVE Guidelines Statement: This study strictly adhered to the ARRIVE guidelines 2.0 (https://arriveguidelines.org). Experimental procedures included a priori sample size calculation, randomized allocation of animals into treatment groups, and blinded assessment of outcomes to ensure scientific rigor and reproducibility. References Machlowska J, Baj J, Sitarz M, et al. Gastric Cancer: Epidemiology, Risk Factors, Classification, Genomic Characteristics and Treatment Strategies[J]. International Journal of Molecular Sciences, 2020, 21(11): 4012. Sexton R E, Al Hallak M N, Diab M, et al. Gastric cancer: a comprehensive review of current and future treatment strategies[J]. Cancer Metastasis Reviews, 2020, 39(4): 1179-1203. Zhou J, Ma X, Bi F, et al. Clinical significance of circulating tumor cells in gastric cancer patients[J]. Oncotarget, 2017, 8(15): 25713-25720. Djebali S, Davis C A, Merkel A, et al. Landscape of transcription in human cells[J]. Nature, 2012, 489(7414): 101-108. Xie S S, Jin J, Xu X, et al. Emerging roles of non-coding RNAs in gastric cancer: Pathogenesis and clinical implications[J]. World Journal of Gastroenterology, 2016, 22(3): 1213-1223. Taft R J, Glazov E A, Lassmann T, et al. Small RNAs derived from snoRNAs[J]. RNA (New York, N.Y.), 2009, 15(7): 1233-1240. Bartel D P. MicroRNAs: target recognition and regulatory functions[J]. Cell, 2009, 136(2): 215-233. McCabe E M, Rasmussen T P. lncRNA involvement in cancer stem cell function and epithelial-mesenchymal transitions[J]. Seminars in Cancer Biology, 2021, 75: 38-48. Zhu Y, Zhou B, Hu X, et al. LncRNA LINC00942 promotes chemoresistance in gastric cancer by suppressing MSI2 degradation to enhance c-Myc mRNA stability[J]. Clinical and Translational Medicine, 2022, 12(1): e703. Salmena L, Poliseno L, Tay Y, et al. A ceRNA hypothesis: the Rosetta Stone of a hidden RNA language?[J]. Cell, 2011, 146(3): 353-358. Théry C. Exosomes: secreted vesicles and intercellular communications[J]. F1000 Biology Reports, 2011, 3: 15. Binnewies M, Roberts E W, Kersten K, et al. Understanding the tumor immune microenvironment (TIME) for effective therapy[J]. Nature Medicine, 2018, 24(5): 541-550. Kim S T, Cristescu R, Bass A J, et al. Comprehensive molecular characterization of clinical responses to PD-1 inhibition in metastatic gastric cancer[J]. Nature Medicine, 2018, 24(9): 1449-1458. Fang Z, Meng Q, Xu J, et al. Signaling pathways in cancer-associated fibroblasts: recent advances and future perspectives[J]. Cancer Communications (London, England), 2023, 43(1): 3-41. Fitzgerald K A, Kagan J C. Toll-like Receptors and the Control of Immunity[J]. Cell, 2020, 180(6): 1044-1066. Farhood B, Najafi M, Mortezaee K. CD8+ cytotoxic T lymphocytes in cancer immunotherapy: A review[J]. Journal of Cellular Physiology, 2019, 234(6): 8509-8521. Yuan J hang, Yang F, Wang F, et al. A long noncoding RNA activated by TGF-β promotes the invasion-metastasis cascade in hepatocellular carcinoma[J]. Cancer Cell, 2014, 25(5): 666-681. Mao T L, Fan M H, Dlamini N, et al. LncRNA MALAT1 Facilitates Ovarian Cancer Progression through Promoting Chemoresistance and Invasiveness in the Tumor Microenvironment[J]. International Journal of Molecular Sciences, 2021, 22(19): 10201. Liu H T, Zou Y X, Zhu W J, et al. lncRNA THAP7-AS1, transcriptionally activated by SP1 and post-transcriptionally stabilized by METTL3-mediated m6A modification, exerts oncogenic properties by improving CUL4B entry into the nucleus[J]. Cell Death and Differentiation, 2022, 29(3): 627-641. Teng F, Zhang J X, Chen Y, et al. LncRNA NKX2-1-AS1 promotes tumor progression and angiogenesis via upregulation of SERPINE1 expression and activation of the VEGFR-2 signaling pathway in gastric cancer[J]. Molecular Oncology, 2021, 15(4): 1234-1255. Ji Z, Qiu Y, Cai Q, et al. LINC00922 acts as a novel oncogene in gastric cancer[J]. World Journal of Surgical Oncology, 2022, 20(1): 121. Liang T, Wang B, Li J, et al. LINC00922 Accelerates the Proliferation, Migration and Invasion of Lung Cancer Via the miRNA-204/CXCR4 Axis[J]. Medical Science Monitor: International Medical Journal of Experimental and Clinical Research, 2019, 25: 5075-5086. Wang L, Ren C, Xu Y, et al. The LINC00922 aggravates ovarian cancer progression via sponging miR-361-3p[J]. Journal of Ovarian Research, 2021, 14(1): 77. Ye Z, He Q, Wang Q, et al. LINC00922 promotes the proliferation, migration, invasion and EMT process of liver cancer cells by regulating miR-424-5p/ARK5[J]. Molecular and Cellular Biochemistry, 2021, 476(10): 3757-3769. Wu E, Guo X, Teng X, et al. Discovery of Plasma Membrane-Associated RNAs through APEX-seq[J]. Cell Biochemistry and Biophysics, 2021, 79(4): 905-917. Chai Y, Xie M. LINC01579 promotes cell proliferation by acting as a ceRNA of miR-139-5p to upregulate EIF4G2 expression in glioblastoma[J]. Journal of Cellular Physiology, 2019, 234(12): 23658-23666. Peran I, Dakshanamurthy S, McCoy M D, et al. Cadherin 11 Promotes Immunosuppression and Extracellular Matrix Deposition to Support Growth of Pancreatic Tumors and Resistance to Gemcitabine in Mice[J]. Gastroenterology, 2021, 160(4): 1359-1372.e13. Yang Z, Yan C, Yu Z, et al. Downregulation of CDH11 Promotes Metastasis and Resistance to Paclitaxel in Gastric Cancer Cells[J]. Journal of Cancer, 2021, 12(1): 65-75. Chen J H, Huang W C, Bamodu O A, et al. Monospecific antibody targeting of CDH11 inhibits epithelial-to-mesenchymal transition and represses cancer stem cell-like phenotype by up-regulating miR-335 in metastatic breast cancer, in vitro and in vivo[J]. BMC cancer, 2019, 19(1): 634. Chivu-Economescu M, Necula L G, Matei L, et al. Collagen Family and Other Matrix Remodeling Proteins Identified by Bioinformatics Analysis as Hub Genes Involved in Gastric Cancer Progression and Prognosis[J]. International Journal of Molecular Sciences, 2022, 23(6): 3214. Li R, Hou S, Zou M, et al. miR-543 impairs cell proliferation, migration, and invasion in breast cancer by suppressing VCAN[J]. Biochemical and Biophysical Research Communications, 2021, 570: 191-198. Ping Q, Wang C, Cheng X, et al. TGF-β1 dominates stromal fibroblast-mediated EMT via the FAP/VCAN axis in bladder cancer cells[J]. Journal of Translational Medicine, 2023, 21(1): 475. Salem M, O’Brien J A, Bernaudo S, et al. miR-590-3p Promotes Ovarian Cancer Growth and Metastasis via a Novel FOXA2-Versican Pathway[J]. Cancer Research, 2018, 78(15): 4175-4190. Yang L, Zhou Y N, Zeng M M, et al. Circular RNA Circ-0002570 Accelerates Cancer Progression by Regulating VCAN via MiR-587 in Gastric Cancer[J]. Frontiers in Oncology, 2021, 11: 733745. Song J, Wei R, Huo S, et al. Versican enrichment predicts poor prognosis and response to adjuvant therapy and immunotherapy in gastric cancer[J]. Frontiers in Immunology, 2022, 13: 960570. Topalian S L, Sznol M, McDermott D F, et al. Survival, durable tumor remission, and long-term safety in patients with advanced melanoma receiving nivolumab[J]. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 2014, 32(10): 1020-1030. Topalian S L, Hodi F S, Brahmer J R, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer[J]. The New England Journal of Medicine, 2012, 366(26): 2443-2454. Zhang Y, Zhang Z. The history and advances in cancer immunotherapy: understanding the characteristics of tumor-infiltrating immune cells and their therapeutic implications[J]. Cellular & Molecular Immunology, 2020, 17(8): 807-821. Zhou A L, 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[J]. Pathology, Research and Practice, 2020, 216(9): 153093. Lu S, Stein J E, Rimm D L, et al. Comparison of Biomarker Modalities for Predicting Response to PD-1/PD-L1 Checkpoint Blockade: A Systematic Review and Meta-analysis[J]. JAMA oncology, 2019, 5(8): 1195-1204. Additional Declarations No competing interests reported. Supplementary Files Supplementarytableandfigure.docx Cite Share Download PDF Status: Published Journal Publication published 01 Jul, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 30 Apr, 2025 Reviews received at journal 28 Apr, 2025 Reviewers agreed at journal 09 Apr, 2025 Reviews received at journal 06 Apr, 2025 Reviewers agreed at journal 01 Apr, 2025 Reviewers invited by journal 01 Apr, 2025 Submission checks completed at journal 01 Apr, 2025 First submitted to journal 22 Mar, 2025 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-4989662","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":437003686,"identity":"9a35a2f3-6188-4dcf-a61c-3352db810c24","order_by":0,"name":"Yuanqing An","email":"","orcid":"","institution":"First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Yuanqing","middleName":"","lastName":"An","suffix":""},{"id":437003687,"identity":"d3f8423f-8c4c-46c4-899f-732732fa746f","order_by":1,"name":"Xiaomeng Liu","email":"","orcid":"","institution":"First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Xiaomeng","middleName":"","lastName":"Liu","suffix":""},{"id":437003688,"identity":"9a700421-37ac-4e7a-b5ae-1d7e24c1b3ed","order_by":2,"name":"Jin Liu","email":"","orcid":"","institution":"First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Jin","middleName":"","lastName":"Liu","suffix":""},{"id":437003689,"identity":"8e7095a7-3203-48e6-be98-14802e0b6fc4","order_by":3,"name":"Deqiang Wang","email":"","orcid":"","institution":"Affiliated Hospital of Jiangsu University","correspondingAuthor":false,"prefix":"","firstName":"Deqiang","middleName":"","lastName":"Wang","suffix":""},{"id":437003692,"identity":"80a031ce-3843-4a5f-b9f9-8f290a33a252","order_by":4,"name":"Wenying Yan","email":"","orcid":"","institution":"Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Wenying","middleName":"","lastName":"Yan","suffix":""},{"id":437003693,"identity":"3b7942f8-ca68-420f-8836-3ba5cde2d22e","order_by":5,"name":"Guang Hu","email":"","orcid":"","institution":"Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Guang","middleName":"","lastName":"Hu","suffix":""},{"id":437003694,"identity":"4b5ca0e5-606c-4459-8ed0-91d35f54b85f","order_by":6,"name":"Lu Xu","email":"","orcid":"","institution":"First Affiliated Hospital of Soochow University","correspondingAuthor":false,"prefix":"","firstName":"Lu","middleName":"","lastName":"Xu","suffix":""},{"id":437003695,"identity":"f21d62b1-9e90-4fb1-92c3-50516f78c9a3","order_by":7,"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":"First Affiliated Hospital of Soochow University","correspondingAuthor":true,"prefix":"","firstName":"Wei","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2024-08-28 08:58:47","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4989662/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4989662/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-05105-x","type":"published","date":"2025-07-01T15:58:02+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":79811525,"identity":"6e386e21-89f7-4eea-a1a0-5ab814e248ed","added_by":"auto","created_at":"2025-04-03 06:45:50","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":51816651,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction of ceRNA network. (A)\u003c/strong\u003e The volcano plot of the 5578 DElncRNAs. \u003cstrong\u003e(B, C) \u003c/strong\u003eScale-free index analysis and mean connectivity analysis, for selecting the best soft threshold of WGCNA for lncRNAs. \u003cstrong\u003e(D, E) \u003c/strong\u003eClustering dendrograms of lncRNAs: Different colors below indicate different co-expression modules.\u003cstrong\u003e (F) \u003c/strong\u003eModule-trait associations: Each row represents a module eigengene and each column represents a clinical trait. Each cell includes the corresponding correlation and \u003cem\u003eP\u003c/em\u003e -value. \u003cstrong\u003e(G)\u003c/strong\u003e The intersected lncRNAs from MEblue and DElncRNAs. \u003cstrong\u003e(H)\u003c/strong\u003e The volcano plot of the 238 DEmiRNAs. \u003cstrong\u003e(I, J) \u003c/strong\u003eScale-free index analysis and mean connectivity analysis, for selecting the best soft power of WGCNA for miRNAs. \u003cstrong\u003e(K, L\u003c/strong\u003e) Clustering dendrograms of miRNAs: Different colors below indicate different co-expression modules. \u003cstrong\u003e(M) \u003c/strong\u003eModule-trait relationship. \u003cstrong\u003e(N)\u003c/strong\u003e The intersected miRNAs from MEturquoise and DEmiRNAs.\u003cstrong\u003e(O)\u003c/strong\u003e The volcano plot of the 4480 DEmRNAs. \u003cstrong\u003e(P, Q) \u003c/strong\u003eScale-free index analysis and mean connectivity analysis, for selecting the best soft power of WGCNA for miRNAs.\u003cstrong\u003e (R, S)\u003c/strong\u003e Clustering dendrograms of miRNAs. \u003cstrong\u003e(T)\u003c/strong\u003eModule-trait relationship.\u003cstrong\u003e (U)\u003c/strong\u003e The intersected mRNAs from MEgreen and DEmRNAs. \u003cstrong\u003e(V)\u003c/strong\u003e lncRNA-miRNA-mRNA regulatory network.\u003c/p\u003e","description":"","filename":"Figure1NF.png","url":"https://assets-eu.researchsquare.com/files/rs-4989662/v1/43c01dc7261bb8d613a47809.png"},{"id":79811529,"identity":"868dc6e3-2425-4c12-b869-ead9cbdc5bba","added_by":"auto","created_at":"2025-04-03 06:45:50","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":24336028,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction and validation of the lncRNA-related prognostic model based ceRNA network. (A) \u003c/strong\u003eCoefficient curve. Different colors represent different genes. No zero values were selected as a penalty coefficient.\u003cstrong\u003e (B) \u003c/strong\u003ePartial likelihood deviance of overall survival for the LASSO coefficient profiles. Twelvefeatures with non-zero coefficients were selected by optimal lambda.\u003cstrong\u003e \u0026nbsp;(C) \u003c/strong\u003eHeatmap expression of the prognostic lncRNAs in the different risk groups in the training cohort.\u003cstrong\u003e (D, E) \u003c/strong\u003eThe risk score distributions and the survival times in the training cohort. Patients were divided according to their median risk score. \u003cstrong\u003e(F)\u003c/strong\u003e Kaplan-Meier analysis of OS between the high- and low-risk groups in the training cohort.\u003cstrong\u003e (G) \u003c/strong\u003eThe time-dependent ROC curve for OS prediction at 5 years in the training cohort.\u003cstrong\u003e (H) \u003c/strong\u003eHeatmap expression of the prognostic lncRNAs in the different risk groups in the test cohort.\u003cstrong\u003e (I, J) \u003c/strong\u003eThe risk score distributions and the survival times in the test cohort. The patients were divided according to their median risk score. \u003cstrong\u003e(K)\u003c/strong\u003e Kaplan-Meier analysis of OS between the high- and low-risk groups in the test cohort.\u003cstrong\u003e (L) \u003c/strong\u003eThe time-dependent ROC curve for predicting OS at 5 years in the test cohort. \u003cstrong\u003e(M, N) \u003c/strong\u003eForrest plot of the univariate \u003cstrong\u003e(M)\u003c/strong\u003e and multivariate \u003cstrong\u003e(N) \u003c/strong\u003eassociation of the risk-score model and clinicopathological characteristics with overall survival.\u003cstrong\u003e (O) \u003c/strong\u003eMulti-index ROC curve of the 5-lncRNA signature risk score and other indicators. \u003cstrong\u003e(P) \u003c/strong\u003eThe nomogram is based on risk score, age, and clinicopathological characteristics to predict the 1-, 3-, and 5-year survival in STAD patients.\u003cstrong\u003e (Q) \u003c/strong\u003eCalibration curves for the 1, 3, and 5 years of training cohort.\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-4989662/v1/95eef965d93e38e69b6dede0.png"},{"id":79811552,"identity":"78c0c8ce-95df-43fa-bddb-48db9e456a48","added_by":"auto","created_at":"2025-04-03 06:45:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":20638115,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCharacteristics of the five lncRNAs included in the risk model. (A-E) \u003c/strong\u003eKaplan-Meier survival curves for the 5 prognostic lncRNAs, including AL121772.1 \u003cstrong\u003e(A), \u003c/strong\u003eAP000695.2 \u003cstrong\u003e(B), \u003c/strong\u003eAC010333.1 \u003cstrong\u003e(C), \u003c/strong\u003eLINC00922 \u003cstrong\u003e(D) \u003c/strong\u003eand LINC01579 \u003cstrong\u003e(E) \u003c/strong\u003ein the TCGA-STAD cohort. \u003cstrong\u003e(F-J) \u003c/strong\u003eDifferential expression of AL121772.1 \u003cstrong\u003e(F), \u003c/strong\u003eAP000695.2 \u003cstrong\u003e(G), \u003c/strong\u003eAC010333.1 \u003cstrong\u003e(H), \u003c/strong\u003eLINC00922 \u003cstrong\u003e(I) \u003c/strong\u003eand LINC01579 \u003cstrong\u003e(J)\u003c/strong\u003e in GC and adjacent tissues in TCGA-STAD cohort. \u003cstrong\u003e(K-O) \u003c/strong\u003eThe expression of AL121772.1 \u003cstrong\u003e(K), \u003c/strong\u003eAP000695.2 \u003cstrong\u003e(L), \u003c/strong\u003eAC010333.1 \u003cstrong\u003e(M), \u003c/strong\u003eLINC00922 \u003cstrong\u003e(N) \u003c/strong\u003eand LINC01579 \u003cstrong\u003e(O)\u003c/strong\u003e in GC and adjacent tissues in clinical GC patients. \u003cstrong\u003e(P-R) \u003c/strong\u003eThe expression of the 5 prognostic lncRNAs in differential T \u003cstrong\u003e(P)\u003c/strong\u003e, N \u003cstrong\u003e(Q) \u003c/strong\u003eand\u003cstrong\u003e \u003c/strong\u003eM\u003cstrong\u003e(R) \u003c/strong\u003estaging. \u003cstrong\u003e(S) \u003c/strong\u003eVerification of AP000695.2 knockdown in MKN-45-shAP000695.2 cells. \u003cstrong\u003e(T) \u003c/strong\u003eProliferation of MKN-45 cells after knockdown of AP000695.2. \u003cstrong\u003e(U) \u003c/strong\u003eGrowth curve of transplanted tumor in nude mice. \u003cstrong\u003e(V)\u003c/strong\u003e Visual image and weight map of the transplanted tumor in mice after dissection. \u003cem\u003eP\u003c/em\u003evalues were shown as: ns, not significant; 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":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-4989662/v1/72ed493fa120c3583185ee99.png"},{"id":79811536,"identity":"239b56b4-211b-4424-bf10-1791e250adc4","added_by":"auto","created_at":"2025-04-03 06:45:50","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":37299112,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFunction and downstream genes of AP000695.2. (A)\u003c/strong\u003eSankey diagram showing the lncRNA-miRNA-mRNA axis of AP000695.2. \u003cstrong\u003e(B-N)\u003c/strong\u003eKaplan-Meier survival curves for the 13 mRNAs co-expressed with AP000695.2, including CDH11 \u003cstrong\u003e(B)\u003c/strong\u003e, COL11A11 \u003cstrong\u003e(C)\u003c/strong\u003e, COL12A1 \u003cstrong\u003e(D)\u003c/strong\u003e, COL1A2 \u003cstrong\u003e(E)\u003c/strong\u003e, COL5A1 \u003cstrong\u003e(F)\u003c/strong\u003e, COL5A2 \u003cstrong\u003e(G)\u003c/strong\u003e, LOXL2 \u003cstrong\u003e(H)\u003c/strong\u003e, PRRX1\u003cstrong\u003e (I)\u003c/strong\u003e, STC2 \u003cstrong\u003e(J)\u003c/strong\u003e, THBS2 \u003cstrong\u003e(K)\u003c/strong\u003e, TIMP3 \u003cstrong\u003e(L)\u003c/strong\u003e, TNFSF11 \u003cstrong\u003e(M)\u003c/strong\u003e and VCAN \u003cstrong\u003e(N)\u003c/strong\u003e in TCGA-STAD cohort. \u003cstrong\u003e(O-Q) \u003c/strong\u003eThe relative expression level of CDH11 \u003cstrong\u003e(O),\u003c/strong\u003e COL12A1 \u003cstrong\u003e(P), \u003c/strong\u003eand VCAN \u003cstrong\u003e(Q)\u003c/strong\u003e in MKN-45-shAP000695.2 cells after transfection with miR-144-3p mimic. \u003cstrong\u003e(R, S)\u003c/strong\u003e The relative expression level of COL5A2 \u003cstrong\u003e(R)\u003c/strong\u003e and COL12A1 \u003cstrong\u003e(S)\u003c/strong\u003e in MKN-45-shAP000695.2 cells after transfection with miR-7-5p mimic.\u003cstrong\u003e (T-W) \u003c/strong\u003eEnrichment of GSEA in high and low-expression of CDH11 \u003cstrong\u003e(T), \u003c/strong\u003eCOL5A2 \u003cstrong\u003e(U)\u003c/strong\u003e, COL12A1 \u003cstrong\u003e(V) \u003c/strong\u003eand VCAN \u003cstrong\u003e(W).\u003c/strong\u003e *\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-4989662/v1/068c259387d7f025c45acf1f.png"},{"id":79812215,"identity":"ddd8131e-a6df-4422-b412-c5510f147856","added_by":"auto","created_at":"2025-04-03 06:53:50","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":24732484,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEffect of key genes of 13 mRNAs co-expressed with AP000695.2 on immune cell infiltration and immune treatment.\u003c/strong\u003e \u003cstrong\u003e(A) \u003c/strong\u003eDifferential expression of 13 genes co-expressed with AP000695.2 of GC patients in the Pembrolizumab treatment-sensitive group and treatment-resistant group. \u003cstrong\u003e(B-E) \u003c/strong\u003eThe\u003cstrong\u003e \u003c/strong\u003eROC curves for the impact of expression of COL12A1 \u003cstrong\u003e(B), \u003c/strong\u003eCOL1A2 \u003cstrong\u003e(C), \u003c/strong\u003eCOL5A1 \u003cstrong\u003e(D)\u003c/strong\u003e and VCAN \u003cstrong\u003e(E) \u003c/strong\u003eon ORR\u003cstrong\u003e. (F)\u003c/strong\u003e The differences in the proportions of 22 immune cells between high- and low-expression groups of COL12A1. \u003cstrong\u003e(G) \u003c/strong\u003eInfiltration of fibroblasts between high- and low-expression groups of COL12A1.\u003cstrong\u003e (H) \u003c/strong\u003eThe differences in the proportions of 22 immune cells between high- and low-expression groups of COL1A2.\u003cstrong\u003e (I) \u003c/strong\u003eInfiltration of fibroblasts between high- and low-expression groups of COL1A2.\u003cstrong\u003e (J) \u003c/strong\u003eThe differences in the proportions of 22 immune cells between high- and low-expression groups of COL5A1.\u003cstrong\u003e (K) \u003c/strong\u003eInfiltration of fibroblasts between high- and low-expression groups of COL5A1.\u003cstrong\u003e (L) \u003c/strong\u003eThe differences in the proportions of 22 immune cells between high- and low-expression groups of VCAN.\u003cstrong\u003e(M) \u003c/strong\u003eInfiltration of fibroblasts between high- and low-expression groups of VCAN.\u003cstrong\u003e (N-Q)\u003c/strong\u003e Expression of immune checkpoint in the low- and high-expression groups of COL12A1 \u003cstrong\u003e(N), \u003c/strong\u003eCOL1A2 \u003cstrong\u003e(O), \u003c/strong\u003eCOL5A1 \u003cstrong\u003e(P)\u003c/strong\u003eand VCAN \u003cstrong\u003e(Q)\u003c/strong\u003e. *\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":"Figure5NF.png","url":"https://assets-eu.researchsquare.com/files/rs-4989662/v1/e7665968f825d2c872294c99.png"},{"id":79812219,"identity":"d07d9efd-0aba-49ef-951a-4eb4ba880707","added_by":"auto","created_at":"2025-04-03 06:53:51","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":19062474,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eVerification of AP000695.2/miR-144-3p/VCAN axis. (A) \u003c/strong\u003eThe predicted binding sites of miR-144-3p to AP000695.2 sequence using starbase. \u003cstrong\u003e(B)\u003c/strong\u003e Dual-luciferase reporter assay of luciferase activity in MKN-45 cells co-transfected with miR-144-3p and AP000695.2-wt/AP000695.2-mut.\u003cstrong\u003e (C)\u003c/strong\u003e Proliferation of MKN-45 cells with knockdown of AP000695.2 or with AP000695.2 knockdown and transfection of miR-144-3p. \u003cstrong\u003e(D)\u003c/strong\u003e Binding sites of miR-144-3p to the 3'UTR of VCAN predicted in the TargetScan website. \u003cstrong\u003e(E)\u003c/strong\u003e Relative luciferase activity in MKN-45 cells co-transfected with miR-144-3p and either VCAN-wt or VCAN-mut. \u003cstrong\u003e(F)\u003c/strong\u003e Proliferation curve of MKN-45 cells with transfection of VCAN overexpression vector or transfection of VCAN overexpression vector and miR-144-3p mimic. \u003cstrong\u003e(G)\u003c/strong\u003e Representative images of CD8\u003csup\u003e+ \u003c/sup\u003eT cell infiltration in gastric cancer tissues with different level of VCAN expression. \u003cstrong\u003e(H) \u003c/strong\u003eThe level of CD8+ T cells infiltration in gastric cancer tissues with high and low VCAN expression. \u003cstrong\u003e(I) \u003c/strong\u003eKaplan-Meier survival curve of 65 clinical gastric cancer patients in high and low-expression groups of VCAN. *\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":"Figure6NF.png","url":"https://assets-eu.researchsquare.com/files/rs-4989662/v1/63f9cbbf1cc74db9bce16ace.png"},{"id":79811524,"identity":"5c292ff1-7abc-40c1-b667-b3787a12b574","added_by":"auto","created_at":"2025-04-03 06:45:50","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":47653,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarytableandfigure.docx","url":"https://assets-eu.researchsquare.com/files/rs-4989662/v1/2a17d80fb67840aae37e9125.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Identification of a lncRNA based ceRNA network signature to establish a prognostic model and explore potential therapeutic targets in gastric cancer","fulltext":[{"header":"Introduction","content":"\u003cp\u003eGastric cancer (GC) ranks among the most prevalent malignant tumors globally and is the fourth leading cause of cancer-related mortality\u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. The primary treatment for early GC is endoscopic resection, while non-early operable gastric cancer is treated with surgery combination with chemoradiotherapy. Nevertheless, the 5-year survival rate for patients with stage III GC undergoing surgery is only 18\u0026ndash;50%\u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. For those patients with advanced gastric cancer (AGC), chemotherapy and targeted therapy are the main treatment modalities. Unfortunately, some patients do not respond to chemotherapy or quickly develop chemoresistance, resulting in the 5-year survival rate of less than 30%\u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e. Therefore, there is a critical need for novel, validated biomarkers to improve prognosis assessing and guide treatment strategies.\u003c/p\u003e \u003cp\u003eNon-coding RNAs (ncRNAs) are a class of RNA molecules that do not encode protein but perform essential regulatory functions\u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. Studies have shown that ncRNAs are involved in the progression of various cancers and can serve as potential biomarkers for cancer diagnosis and prognosis assessment\u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e. MicroRNAs (miRNAs) are a class of ncRNAs approximately 20 nucleotides in length\u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/sup\u003e. The main function of miRNAs is to bind to sequences with partial complementarity on target RNA transcripts, known as microRNA recognition elements (MRE), to regulate gene expression\u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]\u003c/sup\u003e. Long non-coding RNAs (lncRNAs) are functional RNA molecules exceeding 200 nucleotides in length that can regulate the expression of protein-coding genes by competitively binding to the MREs of miRNAs\u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e. Numerous studies have demonstrated that various lncRNAs and miRNAs are involved in the development of GC and impact the prognosis of patients\u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e. According to the hypothesis of ceRNA regulatory network, lncRNAs and mRNAs can influence each other's levels by competing for limited miRNAs, which plays a crucial role in the regulation of both lncRNAs and mRNAs\u003csup\u003e[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e. Therefore, constructing a prognosis model based on the ceRNA network is of significant importance for elucidating the molecular mechanisms and predicting the prognosis of GC.\u003c/p\u003e \u003cp\u003eExosomes, a subpopulation of the tumor microenvironment (TME), contain abundant ncRNAs\u003csup\u003e[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e. Components within the TME significantly influence tumor cell invasion, metastasis and response to immunotherapy\u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]\u003c/sup\u003e. However, a prognosis model for GC patients that integrates ceRNA with immune status in the TME to exploring potential immunotherapy targets has rarely been reported. In this study, we constructed a prognostic model for GC based on the ceRNA network and further explored the relationship between the lncRNA-miRNA-mRNA axes with TME as well as its implications for immunotherapy.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eData source and processing\u003c/h2\u003e \u003cp\u003eStomach adenocarcinoma (STAD) RNA expression data and detailed clinicopathologic information were downloaded from TCGA (\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 type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e).\u003c/span\u003e The miRNA expression data totaled 490 samples (445 GC tissues, 45 normal tissues). The lncRNA and mRNA expression data consisted of 407 samples (375 GC tissues,32 normal tissues), of which 340 GC samples had complete clinical data. Then, the data were collated and Gene ID transformed by using the dplyr and tidyr R packages of the R software. Normalized counts were calculated for all patients using the DESeq2 package. To evaluate the predictive potential of gene expression for immunotherapy, we also downloaded 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)\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. The ROC curves of gene expression were derived to predict the response to anti-PD-1 therapy in GC patients using the pROC package in R software (R Version 3.6.1).\u003c/p\u003e \u003cp\u003eSamples of cancerous and normal gastric tissues were collected from 23 GC patients who underwent surgery for GC at the First Affiliated Hospital of Soochow University in 2019 for subsequent gene expression difference analysis. The clinical data of the patients are listed in Supplementary Table\u0026nbsp;1. The study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University (approval no. 2022362) and was conducted in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent was waived by our Institutional Review Board because of the retrospective nature of our study.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eAnalysis of differentially expressed RNAs\u003c/h3\u003e\n\u003cp\u003eFold change (FC) in the expression of miRNAs, lncRNAs, and mRNAs between GC tissues and normal tissues were analyzed using the edgeR R package. Differentially expressed mRNAs, miRNAs, and lncRNAs were screened by | log2FC | \u0026gt; 1 and false discovery rate (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.05, where log2FC\u0026thinsp;\u0026gt;\u0026thinsp;1 was defined as up-regulation and log2FC \u0026lt; -1 as down-regulation.\u003c/p\u003e\n\u003ch3\u003eWeighted correlation network analysis (WGCNA)\u003c/h3\u003e\n\u003cp\u003eWGCNA analyzes the correlation between the expression patterns of genes and then categorizes genes with similar expression patterns into the same module eigengenes (ME). WGCNA was constructed for miRNA, lncRNA, and mRNA. The least value for which the scale-free topology fit R^2 index\u0026thinsp;\u0026gt;\u0026thinsp;0.85 was chosen as the soft threshold. The Pearson method was used to calculate the correlation between all genes and construct the network, and the similarity of gene was assessed by Topological Overlap Matrix (TOM), with 1-TOM as the measured value. The number of genes in each ME was set to \u0026ge;\u0026thinsp;50, and genes with high correlation were clustered into the same ME after forming a co-expression network. Correlation coefficients and \u003cem\u003eP\u003c/em\u003e-values between MEs and clinical features of GC patients were calculated and the results were visualized by heatmap. The R packages utilized in these operations included the WGCNA, hclust and cuttreeDynamic packages.\u003c/p\u003e\n\u003ch3\u003eProtein interaction (PPI) networks\u003c/h3\u003e\n\u003cp\u003eProtein interaction networks were formed using the Search Tool for the Retrieval of Interaction Gene/Proteins (STRING) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://string-db.org/\u003c/span\u003e\u003cspan address=\"https://string-db.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Degree Centrality, Betweenness Centrality and Closeness Centrality of the network were calculated using Cytoscape software to assess the extent to which genes occupy central positions throughout the network.\u003c/p\u003e\n\u003ch3\u003eCompeting endogenous RNA (ceRNA) network\u003c/h3\u003e\n\u003cp\u003eThe miRNA-mRNA correspondence was obtained from starbase, targetscan and miRTarbase databases, and the miRNA-lncRNA correspondence was retrieved from lncbase and starbase databases. The ceRNA network associated with prognosis was constructed.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eConstruction and validation of the risk model\u003c/h2\u003e \u003cp\u003eThe samples in TCGA were randomly divided into two groups as training set (170 cases) and validation set (170 cases). The prognostic model was constructed using the training set, and the feasibility of the prognostic model was verified using the validation set. univariate Cox regression analysis was performed on lncRNAs in the ceRNA network, then Lasso regression was used to minimize prediction errors, the prognostic model was constructed in the multivariate Cox regression analysis and screened by the AIC method, finally. The risk score formula for each patient in the prognostic model: Risk score = \u0026sum; (Expi * Coefi); Coefi and Expi denote the risk coefficient and RNA expression, respectively. The patients were divided into high- and low-risk groups based on the median risk score, the differences between the two risk groups in terms of gene expression, survival status, and risk scores were visualized. Kaplan\u0026ndash;Meier survival curves were constructed to visualize differences of survival rate between the high- and low-risk groups. Five-year ROC curves were generated to evaluate the accuracy of the prognostic model. Univariate and multivariate Cox regression analyses were used to evaluate the independence of the risk score, and the results were shown in forest plots. Furthermore, a nomogram was constructed to predict the survival time of patients. All analyses were performed using R for which the following packages were employed: glmnet, survival, survminer, regplot, timeROC and rms packages.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eGene Set Enrichment Analysis (GSEA)\u003c/h3\u003e\n\u003cp\u003eThe GSEA enrichment analysis method allows the identification of biological processes and signaling pathways of genes. Signaling pathway enrichment analysis was performed on individual genes using GSEA 4.1.0 software. Functional enrichment results were filtered by P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.1.\u003c/p\u003e\n\u003ch3\u003eCorrelation of gene expression levels with immunological checkpoints and degree of immune cell infiltration\u003c/h3\u003e\n\u003cp\u003eWe used the CIBERSORT and XCELL methods to evaluate the abundance of immune cell infiltration among the samples in the TCGA\u0026ndash;STAD dataset. The expression levels of immunological checkpoints between the high- and low-expression groups of key genes are shown in boxplots.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eCell culture and gene knock-down techniques\u003c/h2\u003e \u003cp\u003eThe human GC cell line MKN-45 was purchased from the American Typical Culture Collection (ATCC). MKN-45 cells were cultured in DMEM medium (HyClone, USA) with 10% FBS (Gibco, USA) and 1% penicillin/streptomycin (HyClone, USA) at 37\u0026deg;C in a humidified atmosphere with 5% CO2. Cells were passaged every 2\u0026ndash;3 days to maintain growth. Lentivirus-encoding shRNAs for AP000695.2 (VSVG-Lentai-hU6-shRNA-AP000695.2-puro-hEF1a-3xFlag) and negative control (NC) lentivirus (VSVG-Lentai-hU6-shRNA-NC-puro-hEF1a-3xFlag) were compounded by Shanghai Taitool Bioscience Co (Shanghai, China). MKN-45 cells were harvested and used for further analysis after infecting lentivirus with 72 hours.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eRNA isolation and real-time quantitative reverse transcriptase PCR (qRT-PCR)\u003c/h2\u003e \u003cp\u003eTotal RNA was extracted using Trizol Reagent (Thermo Fisher Scientific, USA) according to the manufacturer\u0026rsquo;s protocol. Reverse transcription was performed using the PrimeScript RT Reagent Kit (TaKaRa, Japan), yielding 1 \u0026micro;g of total RNA in a final volume of 20 \u0026micro;l according to the manufacturer's protocol. A Light Cycler 96 (Roche) real-time PCR system was used to detect complementary DNA (cDNA) for the same amount of RNA. The reaction system (13 \u0026micro;l) contained the reverse and forward primers, the corresponding cDNAs, and the SYBR Green PCR master mix (Roche, USA). All data were analyzed using GAPDH gene expression as an internal standard. Primers for specific genes are listed in Supplementary Table\u0026nbsp;2.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMethyl thiazolyl tetrazolium (MTT) assay\u003c/h2\u003e \u003cp\u003eCell proliferation was detected by the MTT [3-(4,5-dimethyl-2-yl)-2,5-diphenyl tetrazolium bromide] assay. Well-grown cells were harvested for digestion and centrifuged. Cells were plated in 96-well plates at 5000 cells per well and incubated for 24 hours. MTT (Beyotime, China) was added to each well at the same time every day during the experiment, to give a final concentration of 5 mg/ml, and the 96-well plate was incubated in an incubator. After 4 hours the medium was aspirated and 400 \u0026micro;l of dimethyl sulfoxide (DMSO, Sigma, USA) was added to each well to dissolve it completely. The absorbance of each well was measured at 490 nm using a microplate reader (Thermo Fisher Scientific, USA).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eSubcutaneous xenograft nude mouse model\u003c/h2\u003e \u003cp\u003eFive-week-old female BALA/c nude mice were purchased from Hangzhou Ziyuan Experimental Animal Technology Co, Ltd. (China) and housed under specific pathogen-free conditions. MKN-45 cells transfected with AP000695.2-shRNA lentivirus were injected into subcutaneous tissues of nude mice in a total volume of 5 \u0026times; 10\u003csup\u003e6\u003c/sup\u003e cells, respectively. The long and short diameters of the subcutaneous tumors were measured every 3 days using vernier calipers. After the experiments were conducted for 3 weeks, the mice were anaesthetized and the tumors were resected. The length (L) and width (W) of each subcutaneous tumor were measured by a caliper. The calculation method for tumor volume (TV): TV = (L * W\u003csup\u003e2\u003c/sup\u003e)/2.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eCell transfection\u003c/h2\u003e \u003cp\u003eWild type and mutant type luciferase reporter gene vectors for AP000695.2 and VCAN, VCAN overexpression vectors, and miRNA-mimics were constructed by General Bio Co. (Anhui, China). Transfection of them were performed using a Lipofectamine Kit (Thermo Fisher Scientific, USA). The well-grown cells were seeded in 6-well plates and 24 hours later, 3 mL of fresh complete medium was added after the cells were attached to the wall. Then, a new EP tube was taken and prepared as follows: 3.75 \u0026micro;L lipo3000\u0026thinsp;+\u0026thinsp;125 \u0026micro;L basal medium\u0026thinsp;+\u0026thinsp;50 nM miRNA mimic per well, and then added to the wells separately. The corresponding experiments could be performed after transfecting for 24 hours.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eDual-luciferase reporter assay\u003c/h2\u003e \u003cp\u003eDual-luciferase reporter assay was performed using Dual Luciferase Reporter Gene Assay Kit (Promega, America) and Promega Chemiluminescent Analyzer. After transfecting for 24 hours, the cells were lysed, centrifuged (12,000 g, 4℃, 10 min). The supernatant was retained, 100 \u0026micro;L firefly luciferase detection reagent was added to every 20 \u0026micro;L sample, and then detected using a Chemiluminescent Analyzer. Subsequently, renilla luciferase detection reagent was added to the sample and detected.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003ePatient samples and immunohistochemistry (IHC)\u003c/h2\u003e \u003cp\u003eFormalin-fixed and paraffin-embedded tumor tissue collected from 67 clinical trials was retrospectively analyzed from the first Affiliated Hospital of Soochow University, from 2020 to 2023. All the patients\u0026rsquo; survival intervals were available and dated to the end of December 2023. The study protocol was performed under the guidelines outlined in the Declaration of Helsinki and was approved by the Ethics Committee of the first Affiliated Hospital of Soochow University. Expressions of target genes were evaluated by using immunohistochemistry. The formalin-fixed and paraffin-embedded tissue blocks were cut to 5-\u0026micro;m sections, Paraffin-embedded tumor tissue sections were dewaxed and performed to Antigen retrieval using antigen repair buffer under high pressure and high-temperature conditions. Then, the sections were blocked with 5% BSA and incubated with primary antibodies against VCAN and CD8 at 4\u0026deg;C overnight. Next, the sections were incubated with secondary antibody for 20 min at room temperature and then visualized with diaminobenzidine (DAB) Substrate Kit.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eAll data were analyzed and plotted using R software (Version 3.6.1), GraphPad Prism software (Version 8.3.1), and SPSS (Version 19.0). t-tests were used to analyze the differences between the independent samples of the two groups. The \u003cem\u003eP\u003c/em\u003e-value of \u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eIdentification of prognosis-associated DEmiRNAs, DElncRNsA and DEmRNsA, and construction of ceRNA networks\u003c/h2\u003e \u003cp\u003eThe total number of differentially expressed lncRNAs (DElncRNAs) in the TCGA-STAD cohort was 5,578, of which 3,615 were upregulated and 1,963 were downregulated \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e. To explore the DElncRNAs highly correlated with prognosis, we performed WCGNA to identify the gene modules significantly associated with survival status. The optimal soft-threshold power \u003cem\u003eβ\u003c/em\u003e was set as 3 to ensure the construction of a scale-free network (scale-free \u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.85) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB, C). The minimum number of lncRNAs per module was set as 50, resulting in the clustering of lncRNAs with similar expression patterns into 11 modules (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD, E). Among these, the blue module showed the strongest correlation with survival status (R\u0026thinsp;=\u0026thinsp;0.13, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.02) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF). This module contained 4,377 lncRNAs, which intersected with the 5,578 DElncRNAs to yield 1,259 prognosis-associated lncRNAs \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eG\u003cb\u003e)\u003c/b\u003e. In the TCGA-STAD cohort, there were 238 differentially expression miRNAs (DEmiRNAs), which 148 were upregulated and 90 were downregulated \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e. We again used WCGNA to identify gene modules highly correlated with survival status, setting the optimal soft-threshold power \u003cem\u003eβ\u003c/em\u003e at 2 to ensure a scale-free network (scale-free \u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.85) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eI, J). The minimum number of miRNAs in each module was set at 50, clustering miRNAs with similar expression patterns into 5 modules \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eK, L\u003cb\u003e)\u003c/b\u003e. The turquoise module showed the strongest correlation with survival status (R = -0.12, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eM). This module contained 234 miRNAs, which intersected with the 238 DEmiRNAs to yield 124 prognosis-associated miRNAs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eN). For differentially expressed mRNAs (DEmRNAs) in the TCGA-STAD cohort, the total was 4480, with 2125 upregulated and 2355 downregulated \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eO\u003cb\u003e)\u003c/b\u003e. Subsequently, we used WGCNA to identify gene modules significantly correlated with survival status, setting the optimal soft-threshold power \u003cem\u003eβ\u003c/em\u003e at 3 to ensure a scale-free network (scale-free \u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.85) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eP, Q). The minimum number of mRNAs in each module was set at 50, clustering mRNAs with similar expression patterns into 20 modules \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eR, S\u003cb\u003e)\u003c/b\u003e. The green module showed the strongest correlation with survival status (R\u0026thinsp;=\u0026thinsp;0.16, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.002) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eT). This module contained 763 mRNAs, which intersected with the 4,480 DEmRNAs to yield 161 prognosis-associated mRNAs (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eU).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe Protein-Protein Interaction (PPI) Network of the 161 prognosis-associated mRNAs was generated using the Search Tool for the Retrieval of Interacting Genes (STRING). Using Cytoscape software, we constructed a network and calculated network parameters, identifying 62 closely interacting mRNAs. The 62 mRNAs from the PPI network, along with 1259 prognosis-associated lncRNAs and 124 prognosis-associated miRNAs, were used to construct the ceRNA network. Finally, a ceRNA network containing 235 lncRNAs, 60 miRNAs, and 52 mRNAs was obtained \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eV\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eConstruction and verification of a ceRNA-related prognostic model in GC\u003c/h2\u003e \u003cp\u003eThe samples from the TCGA-STAD cohort were randomly divided into two groups: the training set (n\u0026thinsp;=\u0026thinsp;170) and the validation set (n\u0026thinsp;=\u0026thinsp;170). In the training set, 235 lncRNAs in the ceRNA network were analyzed by univariate Cox regression, and the obtained overall survival (OS)-associated lncRNAs were refined using Lasso regression to avoid overfitting \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, B\u003cb\u003e)\u003c/b\u003e. This process identified 12 lncRNAs associated with the prognosis of with GC patients. Subsequently, multivariate Cox regression analysis was performed on these 12 lncRNAs based on Akaike information criterion (AIC) values. Finally, 5 lncRNAs were selected out to construct the prognostic model, including AC010333.1, LINC01579, AL121772.1, AP000695.2 and LINC00922, and the corresponding regression coefficients were calculated. The Risk Score was determined by using the following formula: Risk Score = (1.70 \u0026times; AC010333.1) + (0.48 \u0026times; LINC01579) + (-0.54 \u0026times; AL121772.1) + (0.37 \u0026times; AP000695.2) + (1.28 \u0026times; LINC00922). The expression of these 5 lncRNAs between two risk groups were visualized by a heatmap \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e. Risk scores were calculated for each patient, and the training cohort was divided into high- and low-risk groups based on the median risk score. The risk score distribution and survival time distribution plots showed that patients in the high-risk group had shorter survival time \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD, E\u003cb\u003e)\u003c/b\u003e. The Kaplan-Meier (K-M) survival curves demonstrated that patients in the low-risk group had a significant survival advantage over those in the high-risk group (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;7.842e-03) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF\u003cb\u003e)\u003c/b\u003e. To validate the model's performance, a receiver operator characteristic (ROC) curve predicting 5-year survival was plotted, yielding an area under the curve (AUC) value was 0.608 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eG\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo further evaluate the predictive efficacy of the risk model, risk scores for each patient in the validation set were calculated. Based on the median risk score, the validation cohort was also divided into high- and low-risk groups. The differential expression of 5 lncRNAs between the two groups were displayed in a heatmap \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e. The risk score distribution and survival time distribution plots showed that patients in the high-risk group had shorter survival time \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eI, J\u003cb\u003e)\u003c/b\u003e. The K-M survival curves showed that patients in the low-risk group had a significant survival advantage over those in the high-risk group (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;3.198e-02) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eK\u003cb\u003e)\u003c/b\u003e. An ROC curve predicting 5-year survival was plotted for the validation set, with an AUC value of 0.632 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eL\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eEvaluation of the clinical values of the prognostic risk model\u003c/h2\u003e \u003cp\u003eThe prognostic value of clinicopathologic features and risk scores was explored through univariate and multivariate Cox regression analyses on the TCGA-STAD cohort. Univariate Cox analysis revealed that age (\u003cem\u003eHR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.021, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03), stage (\u003cem\u003eHR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.500, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), T stage (\u003cem\u003eHR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.302, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03), N stage (\u003cem\u003eHR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.267, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.007) and risk score (\u003cem\u003eHR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.229, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) were significantly associated with OS \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eM\u003cb\u003e)\u003c/b\u003e. In multivariate Cox analysis, age (\u003cem\u003eHR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.033, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.002) and risk score (\u003cem\u003eHR\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.228, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) demonstrated independent prognostic value \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eN\u003cb\u003e)\u003c/b\u003e. To compare the predictive accuracy of the risk score with various clinical factors, multivariable ROC analysis was performed \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eO\u003cb\u003e)\u003c/b\u003e. Overall, the risk score emerged as the most accurate prognostic indicator (AUC\u0026thinsp;=\u0026thinsp;0.617). Utilizing both clinical parameters and risk score, a nomogram model was established to predict OS \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eP\u003cb\u003e)\u003c/b\u003e. The 1-, 2-, and 3-year predicted OS values were calculated based on the total points from each parameter, with higher point totals indicating worse survival outcomes. These curves showed close alignment with the ideal diagonal line, indicating the nomogram's precision in predicting patient outcomes \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eQ\u003cb\u003e)\u003c/b\u003e.\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eCharacterization of lncRNAs in the prognostic risk model\u003c/h2\u003e \u003cp\u003eBy analyzing K-M survival curves of the TCGA-STAD cohort, it was observed that GC patients with high expression of AL121772.1 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.042) suggested a relatively better prognosis \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e. Conversely, patients with high expression of AP000695.2 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.039) demonstrated poor prognosis \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e. However, the expression levels of AC010333.1, LINC00922, and LINC01579 did not affect patient prognosis \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC-E\u003cb\u003e)\u003c/b\u003e. The expression levels of the five lncRNAs in the TCGA-STAD cohort were investigated, revealing that AL121772.1 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5.9e-09), AP000695.2 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.3e-09), AC010333.1 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5.1e-06), and LINC00922 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;3.9e-12) were expressed at higher levels in tumor tissues compared to normal tissues \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF-I\u003cb\u003e)\u003c/b\u003e. In contract, LINC01579 was expressed at lower levels in tumor tissues than in normal tissues (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.03) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eJ\u003cb\u003e)\u003c/b\u003e. To validate these finding further, tumor tissues and adjacent tissues from 23 clinical GC patients were collected. The expression of the five lncRNAs in these samples was examined using qRT-PCR, and the differential expression profiles were consistent with those observed in the TCGA-STAD cohort \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eK-O\u003cb\u003e)\u003c/b\u003e. Following this, we focused on the expressing the five lncRNAs during tumor progression. Notably, AP000695.2 was signigicantly higher at T3 stage compared to other T stages (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eP\u003cb\u003e)\u003c/b\u003e. There were no significant differences in the expression of the five lncRNAs across N and M stage \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eQ, R\u003cb\u003e)\u003c/b\u003e. Functional experiments were conducted using the MKN-45 cell line, which was infected with lentivirus carrying AP000695.2-shRNA. The efficiency of the lentiviral vectors was confirmed by qPCR \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eS\u003cb\u003e)\u003c/b\u003e. The MTT assay demonstrated that, knockdown of AP000695.2 significantly inhibited the proliferation of MKN-45 cells \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eT\u003cb\u003e)\u003c/b\u003e. This was further validated \u003cem\u003ein vivo\u003c/em\u003e using a subcutaneous xenograft nude mouse model, where the growth of tumors was weakened by knocking down AP000695.2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eU, V\u003cb\u003e)\u003c/b\u003e. Based on analysis of the five lncRNAs in the prognostic model, AP000695.2 emerged as the most significant gene related to the prognosis and is a promising potential target for GC treatment.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eThe lncRNA-miRNA-mRNA regulatory axes of AP000695.2\u003c/h2\u003e \u003cp\u003eThe Sankey diagram demonstrated the lncRNA-miRNA-mRNA axes associated with AP000695.2 in the ceRNA network, and obtained a total of 13 potential ceRNA-mRNAs \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, Supplemental Table\u0026nbsp;3\u003cb\u003e)\u003c/b\u003e. K-M survival curves of the TCGA-STAD cohort revealed that patients with low expression of CDH11 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.046) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e, COL12A1 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.025) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e, COL5A2 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.009) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG\u003cb\u003e)\u003c/b\u003e, and VCAN (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.004) \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eN\u003cb\u003e)\u003c/b\u003e had a significant survival advantage. Conversely, the expression levels of COL11A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e, COL1A2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE\u003cb\u003e)\u003c/b\u003e, COL5A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF\u003cb\u003e)\u003c/b\u003e, LOXL2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e, PRRX1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eI\u003cb\u003e)\u003c/b\u003e, STC2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eJ\u003cb\u003e)\u003c/b\u003e, THBS2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eK\u003cb\u003e)\u003c/b\u003e, TIMP3 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eL\u003cb\u003e)\u003c/b\u003e, and TNFSF11 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eM\u003cb\u003e)\u003c/b\u003e did not appear to significantly affect the prognosis of GC patients (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.05). To validate the lncRNA-miRNA-mRNA axes illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, MKN-45 cells were transfected with miR-144-3p-mimic and miR-7-5p-mimic. In AP000695.2 knockdown cells, the down-regulation capacity of miR-144-3p-mimic on CDH11 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eO\u003cb\u003e)\u003c/b\u003e, COL5A2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eP\u003cb\u003e)\u003c/b\u003e, COL12A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eQ\u003cb\u003e)\u003c/b\u003e and VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eR\u003cb\u003e)\u003c/b\u003e expression was diminished (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01). Similarly, the knockdown of AP000695.2 also reduced the down-regulation effect of miR-7-5p-mimic on COL5A2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eS\u003cb\u003e)\u003c/b\u003e, COL12A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eT\u003cb\u003e)\u003c/b\u003e and VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eU\u003cb\u003e)\u003c/b\u003e expression (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.01) but showed no similar function on expression of CDH11 (Supplementary Fig.\u0026nbsp;1). To elucidate the molecular mechanisms related to lncRNA-miRNA-mRNA axes based on CDH11, COL12A1, COL5A2 and VCAN in GC patients, GSEA analysis was conducted on high- and low-expression groups of these genes. GSEA enrichment analysis indicated that the CDH11 low-expression group was enriched in base excision repair, DNA replication, and mismatch repair pathways, whereas the high-expression group was primarily associated with the calcium signaling pathway, extracellular matrix (ECM) receptor interaction, and TGF-β signaling pathway \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eV\u003cb\u003e)\u003c/b\u003e. The high-expression groups of COL5A2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eW\u003cb\u003e)\u003c/b\u003e, COL12A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eX\u003cb\u003e)\u003c/b\u003e and VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eY\u003cb\u003e)\u003c/b\u003e were mainly enriched in ECM receptor interaction, JAK-STAT signaling pathway, MAPK signaling pathway, TGF-β signaling pathway, and Toll-like receptor signaling pathway.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe aforementioned signaling pathways have been well elucidated in their role in regulating the tumor immune microenvironment\u003csup\u003e[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e, making it reasonable to hypothesize that the AP000695.2-based lncRNA-miRNA-mRNA axis may be involved in tumor immune evasion.\u003c/p\u003e \u003cp\u003e \u003cb\u003eEstimation of immune response and immune cell infiltration using the key genes of lncRNA-miRNA-mRNA axes related to AP000695.2\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo explore potential indicators for immunotherapy within the AP000695.2-related lncRNA-miRNA-mRNA axes, we utilized RNA-Seq data from the immunotherapy cohort PRJEB25780 available at 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 anti-programmed cell death protein 1 (PD-1) therapy, grouping them by clinical outcomes into complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD)\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. We compared the expression of 13 mRNAs related to the AP000695.2 axis between the Pembrolizumab treatment-sensitive (CR/PR) group and the Pembrolizumab treatment-resistant (SD/PD) group. The expression levels of COL12A1 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), COL1A2 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), COL5A1 (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) and VCAN (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) in the CR/PR group were lower \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA\u003cb\u003e)\u003c/b\u003e. The ROC curves for COL12A1 (AUC\u0026thinsp;=\u0026thinsp;0.172), COL1A2 (AUC\u0026thinsp;=\u0026thinsp;0.720), COL5A1 (AUC\u0026thinsp;=\u0026thinsp;0.745) and VCAN (AUC\u0026thinsp;=\u0026thinsp;0.816) demonstrated their predictive effect on immunotherapy response \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB-E\u003cb\u003e)\u003c/b\u003e. Next, we calculated the proportions of infiltrating immune cells in the TCGA-STAD using the CIBERSORT method. We compared the proportions of each immune cell between the differential expression groups of COL12A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF\u003cb\u003e)\u003c/b\u003e, COL1A2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH\u003cb\u003e)\u003c/b\u003e, COL5A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eJ\u003cb\u003e)\u003c/b\u003e and VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eL\u003cb\u003e)\u003c/b\u003e. Some immune cells, including naive B cells, memory B cells and Plasma cells, were more enriched in the low-expression groups, whereas M2 macrophages were more enriched in the high-expression groups. Notably, CD8\u003csup\u003e+\u003c/sup\u003e T cells were more enriched only in the low-expression groups of COL12A1 and COL5A1. Proportions of infiltrating fibroblasts in the TCGA-STAD cohort were calculated using the XCELL method, revealing low fibroblasts infiltration in the low-expression groups of COL12A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG\u003cb\u003e)\u003c/b\u003e, COL1A2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eI\u003cb\u003e)\u003c/b\u003e, COL5A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eK\u003cb\u003e)\u003c/b\u003e and VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eM\u003cb\u003e)\u003c/b\u003e. Additionally, the low-expression groups of COL12A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eN\u003cb\u003e)\u003c/b\u003e, COL1A2 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eO\u003cb\u003e)\u003c/b\u003e, COL5A1 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eP\u003cb\u003e)\u003c/b\u003e and VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eQ\u003cb\u003e)\u003c/b\u003e preserved lower expression of immune checkpoints.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003eValidation of AP000695.2/miR-144-3p/VCAN axis\u003c/h2\u003e \u003cp\u003eAs it showed in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eU, both AP000695.2 and miR-144-3p affect the expression of VCAN. To further investigate the regulatory network of AP000695.2, miR-144-3p and VCAN, we validated the AP000695.2/miR-144-3p/VCAN axis by dual-luciferase reporter experiments. The online database starbase (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://starbase.sysu.edu.cn\u003c/span\u003e\u003cspan address=\"http://starbase.sysu.edu.cn\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) predicted the binding sites of miR-144-3p to AP000695.2, The mutant plasmid vectors were constructed and specific binding sites are shown \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA\u003cb\u003e).\u003c/b\u003e The dual-luciferase reporter assay displayed that MKN-45 \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB\u003cb\u003e)\u003c/b\u003e. knockdown of AP000695.2 significantly inhibited the proliferation of MKN-45 cells, transfection of miR-144-3p mimic in MKN-45 cells that had knocked down AP000695.2 resulted in further diminution of the proliferative capacity of MKN-45 cells cells \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC\u003cb\u003e)\u003c/b\u003e. Next, by targetscan database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.targetscan.org/vert_80/\u003c/span\u003e\u003cspan address=\"https://www.targetscan.org/vert_80/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), we predicted the binding site of miR-144-3p to the 3'UTR of VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD\u003cb\u003e)\u003c/b\u003e. The luciferase report assay was applied to verify miR-144-3p directly targeted VCAN in MKN-45 cells. The results showed the luciferase activity was obviously declined when VCAN-wt and miR-144-3p mimic were transfected into MKN-45 cells, while VCAN-mut had no effect \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE\u003cb\u003e)\u003c/b\u003e. Overexpression of VCAN promoted cell proliferation, which was impaired by transfection of miR-144-3p mimic \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eF\u003cb\u003e)\u003c/b\u003e. Based on the above results, The regulation of AP000695.2, miR-144-3p and VCAN interacted with each other to affect cell proliferation. In addition, the difference in VCAN expression levels between Pembrolizumab treatment-sensitive group and Pembrolizumab treatment-resistant group in the immunotherapy queue PRJEB25780 is most significant, also the expression of VCAN exhibiting the highest predictive power for immunotherapy response. CD8\u003csup\u003e+\u003c/sup\u003e T cells are crucial anti-tumor immune cells within the tumor microenvironment and are the primary immune cells in tumor targeting therapy. PD-1 ligand (PD-L1) can be targeted to alleviate the exhaustion of CD8\u003csup\u003e+\u003c/sup\u003e T cells\u003csup\u003e[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e. Therefore, we used immunohistochemistry to detect the CD8\u003csup\u003e+\u003c/sup\u003e T cell infiltration in tumor tissues of clinical gastric cancer patients with different levels of VCAN expression. The clinical characteristics of 67 gastric cancer patients and the IHC score of VCAN in tumor tissue were demonstrated (Supplementary Table\u0026nbsp;4). In tissues with higher VCAN expression, the infiltration of CD8\u003csup\u003e+\u003c/sup\u003e T cell is comparatively lower \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eG, H\u003cb\u003e)\u003c/b\u003e K-M survival curves of the clinical gastric cancer patients showed Progression-Free-Survival (PFS) stratified on high and low expression of VCAN \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eI\u003cb\u003e)\u003c/b\u003e, which was consistent with the conclusion in the K-M survival curve of VCAN in the TCGA-STAD cohort in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eN.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eCurrently, numerous studies have focused on ceRNA networks in cancer. LncRNAs and miRNAs have been shown to play regulatory role in various biological processes and tumor-related signaling pathways. For example, researchers have found that the lncRNA-ATB can be activated by TGF-β, competitively binding to the miR-200 family, thereby inducing epithelial-mesenchymal transition (EMT)\u003csup\u003e[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. The lncRNA-MALAT1 stimulates PI3K/Akt axis to accelerate proliferation and invasion of tumor cells and to stimulate chemoresistance\u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. In gastric cancer, the lncRNA-THAP7-AS1 inhibits the transcription of miR-22-3p and miR-320a, playing a carcinogenic role\u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e, and lncRNA-NKX2-1-AS1 activates the VEGFR-2 signal pathway by regulating miR-145-5p, promoting GC progression and angiogenesis\u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e. In these processes, both lncRNAs and miRNAs act as ceRNAs.\u003c/p\u003e \u003cp\u003eTherefore, studies on the ceRNA network in GC are expected to provide new insights the mechanism of GC occurrence and development and simultaneously guide the identification of new therapeutic targets for GC. Weighted correlation network analysis (WGCNA) can be used to find clusters (modules) of highly correlated genes and associate these gene modules with biological traits. Hence, we constructed a prognostic model for GC based on the ceRNA network using WGCNA. Firstly, we screened differentially expressed RNAs and identified prognosis-related RNA modules by WGCNA. The selected RNAs were used to construct the ceRNA network. Then, we developed a ceRNA-related prognostic model composed of 5 lncRNAs (AC010333.1, LINC01579, AP000695.2, LINC00922, AL121772.1) using Cox regression analysis and Lasso regression analysis. The accuracy of the prognostic model was predicted and validated by a validation set. In addition, univariate and multivariate Cox regression analysis suggested that the risk score could be serve as an independent prognostic factor for GC patients.\u003c/p\u003e \u003cp\u003eRegarding the roles of the 5 lncRNAs (AC010333.1, LINC01579, AP000695.2, LINC00922, AL121772.1) in tumors within the ceRNA-related prognostic model, LINC00922 has been extensively studied. Ji et al. found that LINC00922 accelerated GC progression through the miR-204-5p/HMGA2 axis\u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e. Additionally, LINC00922 promotes the proliferation, migration and invasion of lung cancer through the miRNA-204/CXCR4 axis\u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e. In ovarian cancer, LINC00922 can competitively bind miR-361-3p to promote CLDN1 expression and activate the Wnt/β-catenin signaling pathway, accelerating ovarian cancer progression\u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/sup\u003e. Moreover, in hepatocellular carcinoma, LINC00922 promotes cell proliferation, migration, invasion, and EMT by regulating miR-424-5p/ARK5 axis\u003csup\u003e[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]\u003c/sup\u003e. In addition, some researchers applied an APEX-seq method to identify plasma membrane-associated RNAs and found that RNA-AL121772.1 has a considerable affinity for sphingomyelin (SM)\u003csup\u003e[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. Previous studies have confirmed that LINC01579 competitively binds to miR-139-5p and regulates cell proliferation and apoptosis in glioblastoma by modulating EIF4G2 expression\u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSubsequently, we plotted Kaplan-Meier (K-M) survival curves for the five lncRNAs in the ceRNA-related prognostic model in GC. The results indicated that the expression level of AP000695.2 was significantly negatively correlated with the prognosis of GC patients. Moreover, the expression level of AP000695.2 was significantly higher in GC tissues than in normal tissues and correlated with the T stage. Therefore, we speculated that AP000695.2 may promote the development of GC and serve as a potential target for GC treatment. Functional experiments with GC cells showed that the reducing AP000695.2 expression suppressed the proliferative ability of GC cells, which was also verified \u003cem\u003ein vivo\u003c/em\u003e using the subcutaneous xenograft nude mouse model.\u003c/p\u003e \u003cp\u003eBesides, we identified a ceRNA network comprising AP000695.2, two miRNAs (miR-144-3p, miR-7-5p), and thirteen mRNAs (CDH11, COL12A1, STC2, THBS2, COL1A2, COL5A1, COL5A2, LOXL2, PRRX1, TIMP3, COL11A1, TNFSF11, VCAN). The K-M survival curves showed that low expression levels of CDH11, COL5A2, COL12A1, and VCAN were associated with improved prognosis. CDH11 has been implicated in the development and progression of various cancers, including breast cancer, pancreatic cancer and gastric cancer\u003csup\u003e[\u003cspan additionalcitationids=\"CR28\" citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e. Both COL5A2 and COL12A1, members of the collagen family, have been associated with the progression of gastric cancer\u003csup\u003e[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]\u003c/sup\u003e. VCAN exhibits pro-carcinogenic properties in several malignancies, such as breast cancer, ovarian cancer and bladder cancer\u003csup\u003e[\u003cspan additionalcitationids=\"CR32\" citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]\u003c/sup\u003e. In gastric cancer, VCAN is reportedly regulated by miRNA or cirRNA-miRNA axis to facilitate gastric cancer progression\u003csup\u003e[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/sup\u003e, and high VCAN expression is associated with resistant to immunotherapy\u003csup\u003e[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]\u003c/sup\u003e. Through miR-mimic and real-time PCR, we revealed that AP000695.2 competitively binds miR-144-3p with CDH11, COL12A1, VCAN and miR-7-5p with COL5A2, COL12A1. The GSEA enrichment suggested that CDH11, COL5A2, COL12A1, VCAN may regulate downstream signaling pathways, including Hedgehog signaling pathway, JAK-STAT signaling pathway, MAPK signaling pathway, TGF-β signaling pathway, Wnt signaling pathway. Based on the above studies, we hypothesized that AP000695.2 enhances the expression of CDH11, COL5A2, COL12A1 and VCAN through competitive binding miR-144-3p and miR-7-5p. This interaction likely promotes the proliferation of GC cells through the tumor-associated signaling pathway, resulting in poorer patient survival.\u003c/p\u003e \u003cp\u003eCancer immunotherapy exhibits promising antitumor efficacy across various malignancies, particularly through immune checkpoint blockade with antibodies, demonstrating highly durable response rates in phase I studies involving patients with advanced melanoma, non-small-cell lung cancer, and other solid tumors\u003csup\u003e[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/sup\u003e. The TME is crucial in cancer progression and therapeutic outcome\u003csup\u003e[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]\u003c/sup\u003e. According to CIBERSORT and XCELL algorithms, we predicted the abundance of infiltrating immune cells in patients with GC from the TCGA datasets. Low expression groups of COL12A1, COL1A2, COL5A1 and VCAN presented a favorable immune microenvironment and lower expressions of immune checkpoints. Increased expressions of immune checkpoints can directly inhibit specific immune cells activation and contribute to immune escape, resulting in unfavorable outcomes\u003csup\u003e[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]\u003c/sup\u003e. This observation elucidates why GC patients in the low expression groups of COL12A1 and VCAN exhibited better prognosis. Although there was no significant difference in PDCD1 expression between high- and low-expression groups of COL12A1, COL1A2, COL5A1 and VCAN, the low expression groups of COL12A1, COL1A2, COL5A1 and VCAN in the immunotherapy cohort PRJEB25780 from ENA indicated a promising response to anti-PD-1 immunotherapy, especially the VCAN. This may be attributed to the limited significance of PD-1 as a solitary evaluation indicator for the therapeutic response of immune checkpoint inhibitors\u003csup\u003e[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/sup\u003e. Moreover, Due to the significant impact of VCAN expression on the outcomes of immunotherapy and its improved predictive value for immune responses, we demonstrated that elevated VCAN expression leads to reduced infiltration of CD8\u003csup\u003e+\u003c/sup\u003e T cells in gastric cancer tissue, as well as a less favorable prognosis in clinical gastric cancer patients, as evidenced by IHC analysis.\u003c/p\u003e \u003cp\u003eIn summary, we employed WGCNA bioinformatic analysis and ceRNA network construct to develop a prognostic model. Further analysis identified AP000695.2 as a promoter of GC cell proliferation and elucidate its regulatory ceRNA network. We propose that AP000695.2 forms the AP000695.2/miR-144-3p/VCAN axis by competitively binding to miR-144-3p with VCAN, and forms the AP000695.2/miR-144-3p, miR-7-5p/COL12A1 axis by competitively binding to miR-144-3p and miR-7-5p with COL12A1, thereby influencing the expression of VCAN and COL12A1. Changes in VCAN and COL12A1 expression affects tumor-related signaling pathways such as MAPK signaling pathway, TGF-β signaling pathway, Wnt signaling pathway, etc., which in turn impact the prognosis of GC patients. In addition, analysis of immune response and TME revealed that the expression of VCAN affects the immune microenvironment and exhibits a high predictive ability for immunotherapy response. Therefore, we further validated the binding between RNAs in the AP000695.2/miR-144-3p/VCAN axis. Subsequently, we demonstrated in clinical gastric cancer patients that VCAN expression inhibits the infiltration of CD8\u0026thinsp;+\u0026thinsp;T cells within the TME and influences patient prognosis. However, the specific regulatory mechanisms of the AP000695.2/miR-144-3p/VCAN axis as well as the pathways they regulate, require further verified. Whether these mechanisms can be exploited to enhance anti-tumor immunity and improve immunotherapy responses remains to be elucidated.\u003c/p\u003e "},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eL.X., W.L. and G.H. designed the research. Y.A., X.L. and J.L. performed the majority of experiments and analyzed data. D.W., W.Y. and W.Y. contributed to clinical samples and reagents. All authors reviewed the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eWe are grateful for participation and cooperation from gastric cancer patients. This work was supported in part by Suzhou Medical Engineering Collaborative Innovation Research Project (SLJ2022001).\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eTranscriptome data and clinicopathologic information were downloaded from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). Data of immunotherapy cohort PRJEB25780 were obtained from the European Nucleotide Archive (ENA) (https://www.ebi.ac.uk/ena/browser/home). Protein interaction networks were constructed from the Search Tool for the Retrieval of Interaction Gene/Proteins (STRING) database (https://string-db.org/). Further details of all data generated or analyzed for this study can be obtained from the corresponding author upon reasonable request.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eEthical Compliance Statement:\u003c/strong\u003e All animal experiments were performed in accordance with the institutional guidelines of the Laboratory Animal Center of Soochow University and were approved by the Animal Ethics Committee of Soochow University.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eARRIVE Guidelines Statement:\u003c/strong\u003e This study strictly adhered to the ARRIVE guidelines 2.0 (https://arriveguidelines.org). Experimental procedures included a priori sample size calculation, randomized allocation of animals into treatment groups, and blinded assessment of outcomes to ensure scientific rigor and reproducibility.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMachlowska J, Baj J, Sitarz M, et al. Gastric Cancer: Epidemiology, Risk Factors, Classification, Genomic Characteristics and Treatment Strategies[J]. International Journal of Molecular Sciences, 2020, 21(11): 4012.\u003c/li\u003e\n\u003cli\u003eSexton R E, Al Hallak M N, Diab M, et al. Gastric cancer: a comprehensive review of current and future treatment strategies[J]. Cancer Metastasis Reviews, 2020, 39(4): 1179-1203.\u003c/li\u003e\n\u003cli\u003eZhou J, Ma X, Bi F, et al. Clinical significance of circulating tumor cells in gastric cancer patients[J]. Oncotarget, 2017, 8(15): 25713-25720.\u003c/li\u003e\n\u003cli\u003eDjebali S, Davis C A, Merkel A, et al. Landscape of transcription in human cells[J]. Nature, 2012, 489(7414): 101-108.\u003c/li\u003e\n\u003cli\u003eXie S S, Jin J, Xu X, et al. Emerging roles of non-coding RNAs in gastric cancer: Pathogenesis and clinical implications[J]. World Journal of Gastroenterology, 2016, 22(3): 1213-1223.\u003c/li\u003e\n\u003cli\u003eTaft R J, Glazov E A, Lassmann T, et al. Small RNAs derived from snoRNAs[J]. RNA (New York, N.Y.), 2009, 15(7): 1233-1240.\u003c/li\u003e\n\u003cli\u003eBartel D P. MicroRNAs: target recognition and regulatory functions[J]. Cell, 2009, 136(2): 215-233.\u003c/li\u003e\n\u003cli\u003eMcCabe E M, Rasmussen T P. lncRNA involvement in cancer stem cell function and epithelial-mesenchymal transitions[J]. Seminars in Cancer Biology, 2021, 75: 38-48.\u003c/li\u003e\n\u003cli\u003eZhu Y, Zhou B, Hu X, et al. LncRNA LINC00942 promotes chemoresistance in gastric cancer by suppressing MSI2 degradation to enhance c-Myc mRNA stability[J]. Clinical and Translational Medicine, 2022, 12(1): e703.\u003c/li\u003e\n\u003cli\u003eSalmena L, Poliseno L, Tay Y, et al. A ceRNA hypothesis: the Rosetta Stone of a hidden RNA language?[J]. Cell, 2011, 146(3): 353-358.\u003c/li\u003e\n\u003cli\u003eTh\u0026eacute;ry C. Exosomes: secreted vesicles and intercellular communications[J]. F1000 Biology Reports, 2011, 3: 15.\u003c/li\u003e\n\u003cli\u003eBinnewies M, Roberts E W, Kersten K, et al. Understanding the tumor immune microenvironment (TIME) for effective therapy[J]. Nature Medicine, 2018, 24(5): 541-550.\u003c/li\u003e\n\u003cli\u003eKim S T, Cristescu R, Bass A J, et al. Comprehensive molecular characterization of clinical responses to PD-1 inhibition in metastatic gastric cancer[J]. Nature Medicine, 2018, 24(9): 1449-1458.\u003c/li\u003e\n\u003cli\u003eFang Z, Meng Q, Xu J, et al. Signaling pathways in cancer-associated fibroblasts: recent advances and future perspectives[J]. Cancer Communications (London, England), 2023, 43(1): 3-41.\u003c/li\u003e\n\u003cli\u003eFitzgerald K A, Kagan J C. Toll-like Receptors and the Control of Immunity[J]. Cell, 2020, 180(6): 1044-1066.\u003c/li\u003e\n\u003cli\u003eFarhood B, Najafi M, Mortezaee K. CD8+ cytotoxic T lymphocytes in cancer immunotherapy: A review[J]. Journal of Cellular Physiology, 2019, 234(6): 8509-8521.\u003c/li\u003e\n\u003cli\u003eYuan J hang, Yang F, Wang F, et al. A long noncoding RNA activated by TGF-\u0026beta; promotes the invasion-metastasis cascade in hepatocellular carcinoma[J]. Cancer Cell, 2014, 25(5): 666-681.\u003c/li\u003e\n\u003cli\u003eMao T L, Fan M H, Dlamini N, et al. LncRNA MALAT1 Facilitates Ovarian Cancer Progression through Promoting Chemoresistance and Invasiveness in the Tumor Microenvironment[J]. International Journal of Molecular Sciences, 2021, 22(19): 10201.\u003c/li\u003e\n\u003cli\u003eLiu H T, Zou Y X, Zhu W J, et al. lncRNA THAP7-AS1, transcriptionally activated by SP1 and post-transcriptionally stabilized by METTL3-mediated m6A modification, exerts oncogenic properties by improving CUL4B entry into the nucleus[J]. Cell Death and Differentiation, 2022, 29(3): 627-641.\u003c/li\u003e\n\u003cli\u003eTeng F, Zhang J X, Chen Y, et al. LncRNA NKX2-1-AS1 promotes tumor progression and angiogenesis via upregulation of SERPINE1 expression and activation of the VEGFR-2 signaling pathway in gastric cancer[J]. Molecular Oncology, 2021, 15(4): 1234-1255.\u003c/li\u003e\n\u003cli\u003eJi Z, Qiu Y, Cai Q, et al. LINC00922 acts as a novel oncogene in gastric cancer[J]. World Journal of Surgical Oncology, 2022, 20(1): 121.\u003c/li\u003e\n\u003cli\u003eLiang T, Wang B, Li J, et al. LINC00922 Accelerates the Proliferation, Migration and Invasion of Lung Cancer Via the miRNA-204/CXCR4 Axis[J]. Medical Science Monitor: International Medical Journal of Experimental and Clinical Research, 2019, 25: 5075-5086.\u003c/li\u003e\n\u003cli\u003eWang L, Ren C, Xu Y, et al. The LINC00922 aggravates ovarian cancer progression via sponging miR-361-3p[J]. Journal of Ovarian Research, 2021, 14(1): 77.\u003c/li\u003e\n\u003cli\u003eYe Z, He Q, Wang Q, et al. LINC00922 promotes the proliferation, migration, invasion and EMT process of liver cancer cells by regulating miR-424-5p/ARK5[J]. Molecular and Cellular Biochemistry, 2021, 476(10): 3757-3769.\u003c/li\u003e\n\u003cli\u003eWu E, Guo X, Teng X, et al. Discovery of Plasma Membrane-Associated RNAs through APEX-seq[J]. Cell Biochemistry and Biophysics, 2021, 79(4): 905-917.\u003c/li\u003e\n\u003cli\u003eChai Y, Xie M. LINC01579 promotes cell proliferation by acting as a ceRNA of miR-139-5p to upregulate EIF4G2 expression in glioblastoma[J]. Journal of Cellular Physiology, 2019, 234(12): 23658-23666.\u003c/li\u003e\n\u003cli\u003ePeran I, Dakshanamurthy S, McCoy M D, et al. Cadherin 11 Promotes Immunosuppression and Extracellular Matrix Deposition to Support Growth of Pancreatic Tumors and Resistance to Gemcitabine in Mice[J]. Gastroenterology, 2021, 160(4): 1359-1372.e13.\u003c/li\u003e\n\u003cli\u003eYang Z, Yan C, Yu Z, et al. Downregulation of CDH11 Promotes Metastasis and Resistance to Paclitaxel in Gastric Cancer Cells[J]. Journal of Cancer, 2021, 12(1): 65-75.\u003c/li\u003e\n\u003cli\u003eChen J H, Huang W C, Bamodu O A, et al. Monospecific antibody targeting of CDH11 inhibits epithelial-to-mesenchymal transition and represses cancer stem cell-like phenotype by up-regulating miR-335 in metastatic breast cancer, in vitro and in vivo[J]. BMC cancer, 2019, 19(1): 634.\u003c/li\u003e\n\u003cli\u003eChivu-Economescu M, Necula L G, Matei L, et al. Collagen Family and Other Matrix Remodeling Proteins Identified by Bioinformatics Analysis as Hub Genes Involved in Gastric Cancer Progression and Prognosis[J]. International Journal of Molecular Sciences, 2022, 23(6): 3214.\u003c/li\u003e\n\u003cli\u003eLi R, Hou S, Zou M, et al. miR-543 impairs cell proliferation, migration, and invasion in breast cancer by suppressing VCAN[J]. Biochemical and Biophysical Research Communications, 2021, 570: 191-198.\u003c/li\u003e\n\u003cli\u003ePing Q, Wang C, Cheng X, et al. TGF-\u0026beta;1 dominates stromal fibroblast-mediated EMT via the FAP/VCAN axis in bladder cancer cells[J]. Journal of Translational Medicine, 2023, 21(1): 475.\u003c/li\u003e\n\u003cli\u003eSalem M, O\u0026rsquo;Brien J A, Bernaudo S, et al. miR-590-3p Promotes Ovarian Cancer Growth and Metastasis via a Novel FOXA2-Versican Pathway[J]. Cancer Research, 2018, 78(15): 4175-4190.\u003c/li\u003e\n\u003cli\u003eYang L, Zhou Y N, Zeng M M, et al. Circular RNA Circ-0002570 Accelerates Cancer Progression by Regulating VCAN via MiR-587 in Gastric Cancer[J]. Frontiers in Oncology, 2021, 11: 733745.\u003c/li\u003e\n\u003cli\u003eSong J, Wei R, Huo S, et al. Versican enrichment predicts poor prognosis and response to adjuvant therapy and immunotherapy in gastric cancer[J]. Frontiers in Immunology, 2022, 13: 960570.\u003c/li\u003e\n\u003cli\u003eTopalian S L, Sznol M, McDermott D F, et al. Survival, durable tumor remission, and long-term safety in patients with advanced melanoma receiving nivolumab[J]. Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology, 2014, 32(10): 1020-1030.\u003c/li\u003e\n\u003cli\u003eTopalian S L, Hodi F S, Brahmer J R, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer[J]. The New England Journal of Medicine, 2012, 366(26): 2443-2454.\u003c/li\u003e\n\u003cli\u003eZhang Y, Zhang Z. The history and advances in cancer immunotherapy: understanding the characteristics of tumor-infiltrating immune cells and their therapeutic implications[J]. Cellular \u0026amp; Molecular Immunology, 2020, 17(8): 807-821.\u003c/li\u003e\n\u003cli\u003eZhou A L, 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[J]. Pathology, Research and Practice, 2020, 216(9): 153093.\u003c/li\u003e\n\u003cli\u003eLu S, Stein J E, Rimm D L, et al. Comparison of Biomarker Modalities for Predicting Response to PD-1/PD-L1 Checkpoint Blockade: A Systematic Review and Meta-analysis[J]. JAMA oncology, 2019, 5(8): 1195-1204.\u003c/li\u003e\n\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":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"gastric cancer, lncRNA, WGCNA, prognostic model, tumor microenvironment, immunotherapy","lastPublishedDoi":"10.21203/rs.3.rs-4989662/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4989662/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eNumerous studies have demonstrated that long non-coding RNA (lncRNA) play critical roles in regulating physiological processes and contributing to pathological diseases. This study aimed to develop lncRNA-based signatures to predict the prognostic risk of gastric cancer (GC) patients and provide therapeutic guidance.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eGene expression profiles and clinical information were obtained from The Cancer Genome Atlas (TCGA) database. Differentially expressed RNAs, including lncRNA, miRNA, and mRNA, in cancerous and adjacent non-cancerous tissues were analyzed using Weighted correlation network analysis (WGCNA) and construction of a lncRNA-miRNA-mRNA competing endogenous RNA (ceRNA) network. Then, a lncRNA-based risk model was constructed by Cox regression and Lasso regression analyses.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eA ceRNA network comprising 235 lncRNAs, 60 miRNAs, and 52 mRNAs was identified. Based on the expression of five lncRNAs (including AC010333.1, LINC01579, AP000695.2, LINC00922 and AL121772.1) screened from the ceRNA network, a lncRNA-based risk model was developed, which effectively predict the prognosis of GC patients. The expression of AP000695.2 was significantly associated with poor prognosis and higher T stage. The knockdown of AP000695.2 inhibited the growth of GC cells both \u003cem\u003ein vitro\u003c/em\u003e and \u003cem\u003ein vivo\u003c/em\u003e. Transfection with miR-144-3p and miR-7-5p mimics attenuate the up-regulation of targets genes, including CDH11, COL5A2, COL12A1, and VCAN, which was induced by AP000695.2, suggesting a ceRNA mechanism. Additionally, elevated VCAN expression was correlated with poorer survival and a reduced response to anti-PD-1 immune checkpoint inhibitor treatment of GC.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eThis study established a lncRNA-based risk model for predicting the prognosis of GC patients and identified a ceRNA mechanism involving AP000695.2-miR-144-3p-VCAN, presenting novel biomarkers and therapeutic targets for GC treatment.\u003c/p\u003e","manuscriptTitle":"Identification of a lncRNA based ceRNA network signature to establish a prognostic model and explore potential therapeutic targets in gastric cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-03 06:45:44","doi":"10.21203/rs.3.rs-4989662/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-04-30T11:23:27+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-28T04:30:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"56877339995143031367395734227711219242","date":"2025-04-10T03:07:59+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-07T00:08:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"30779308696454809362580095625687275145","date":"2025-04-01T16:15:41+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-01T15:59:59+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-04-01T04:43:36+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-03-22T16:27:29+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1b8572da-c514-4c66-835a-4632740c9cdf","owner":[],"postedDate":"April 3rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":46533325,"name":"Biological sciences/Cancer"},{"id":46533326,"name":"Health sciences/Biomarkers"}],"tags":[],"updatedAt":"2025-07-07T16:11:37+00:00","versionOfRecord":{"articleIdentity":"rs-4989662","link":"https://doi.org/10.1038/s41598-025-05105-x","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-07-01 15:58:02","publishedOnDateReadable":"July 1st, 2025"},"versionCreatedAt":"2025-04-03 06:45:44","video":"","vorDoi":"10.1038/s41598-025-05105-x","vorDoiUrl":"https://doi.org/10.1038/s41598-025-05105-x","workflowStages":[]},"version":"v1","identity":"rs-4989662","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4989662","identity":"rs-4989662","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}