RNA-seq data and associated clinical data for GC were obtained from the publicly accessible The Cancer Genome Atlas-Stomach Adenocarcinoma (TCGA-STAD) project (https://portal.gdc.cancer.gov). This study included 448 samples in total, consisting of 36 control samples and 412 tumor samples. Only protein-coding genes were retained, and the expression of duplicates was averaged and normalization performed.

The edgeR package in R (v 4.3.0, Walter and Eliza Hall Institute of Medical Research, Melbourne, VIC, Australia) was employed for conducting differential expression analysis on the TCGA-STAD samples, with $|$logFC$|$ $>$1 & adjust. p value 0.05. The EnhancedVolcano package (v 1.28.2, R package, St. Jude Children’s Research Hospital, Memphis, TN, USA) was applied to generate volcano plots, and heatmaps displaying the top 20 up- and down-regulated genes were drawn with the pheatmap package (v1.0.13, R package, Raivo Kolde, Tartu, Estonia).

The clusterProfiler (v 4.18.4, R package, Guangchuang Yu, Southern Medical University, Guangzhou, Guangdong, China) and org.Hs.eg.db (v 3.22, R package, Marc Carlson, Fred Hutchinson Cancer Center, Seattle, WA, USA) packages were applied for performing Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses, with the significance threshold set at p 0.05. The corresponding analysis results were represented through bubble plots.

EMT score calculation was performed using the GSVA package (v 3.22, R package, Robert Castelo, Universitat Pompeu Fabra, Barcelona, Catalonia, Spain) employing the HALLMARK_EPITHELIAL_MESENCHYMAL_TRANSITION gene set from the Molecular Signatures Database.

DEGs were subjected to weighted gene co-expression network analysis (WGCNA) via the WGCNA package (v 1.72-5, R package, Peter Langfelder and Steve Horvath, University of California, Los Angeles, CA, USA). Samples were clustered and analyzed for independence and average connectivity. Subsequently, a weighted co-expression network was constructed, and the green module (cor = 0.8, p 0.01) that showed the highest relevance to the EMT score was selected as the key module.

Key module genes were identified using criteria of Gene Significance (GS) $>$0.6 & Module Membership (MM) $>$0.6. Functional enrichment analysis of these genes was subsequently conducted, followed by visualization of GO and KEGG enrichment results using bubble maps.

The identified key module genes were subjected to survival analysis via the survival package (v 3.7, R package, Terry M. Therneau, Mayo Clinic, Rochester, MN, USA) in R. Kaplan–Meier curves were plotted with the ggsurvplot function. Genes showing significant association with survival time and rate were screened using the log-rank test, with a p value threshold of 0.01.

Analysis of gene expression was carried out based on the TCGA-STAD dataset. CHST1 expression-EMT score correlation was analyzed using the ggpubr package (v 0.6.0, R package, Alboukadel Kassambara, University of Debrecen, Debrecen, Hungary). Scatter plots were generated with the ggscatter function, incorporating a regression line, the Pearson correlation coefficient, along with its corresponding p-value. The sign and magnitude of the coefficient represent the direction and strength of the correlation between CHST1 expression and EMT, respectively.

Wilcoxon rank sum test was employed to analyze the correlation between the CHST1 gene and relevant clinical characteristics in TCGA-STAD using the ggpubr package. In addition, based on the GSE84437 dataset from the Gene Expression Omnibus database, the correlation between CHST1 expression and clinical characteristics was further assessed, and survival analysis was conducted to examine the link between CHST1 expression and the overall survival in GC patients.

CHST1 expression was validated in tumor tissue samples obtained from patients with GC. The GC tissue samples were provided by the Department of Gastroenterology, The First Affiliated Hospital of Harbin Medical University. Ethical approval for the current study was granted by the Ethics Committee of The First Affiliated Hospital of Harbin Medical University (Approval No.: 2024314). All study participants voluntarily provided informed consent.

Tumor sections embedded in paraffin underwent deparaffinization and rehydration, followed by antigen retrieval via microwave heating. Subsequently, the activity of endogenous peroxidase was blocked with hydrogen peroxide, and non-specific binding was reduced by incubation with normal goat serum. Subsequently, the sections were incubated sequentially with a primary anti-CHST1 antibody (1/200, 44627) overnight at 4 °C and a corresponding HRP-conjugated secondary antibody (Rabbit, 8114) at room temperature. Visualization was performed using DAB (P0202, Beyotime, China), with the reaction monitored under a microscope until distinct brownish-yellow granular staining was observed. Finally, the nuclei were counterstained with hematoxylin (C0107-100 mL, Beyotime, China). After dehydration and clearing, the sections were mounted with neutral resin and examined under a microscope (MoticAE 2000, MOTIC, China). Both antibodies for this experiment were acquired from Cell Signaling Technology, Inc. (CST, Danvers, MA, USA).

Approval for establishing the subcutaneous graft tumor model was granted by the Experimental Animal Welfare Ethics Committee of the First Affiliated Hospital of Harbin Medical University (IACUC NO. 2023068). 4–5-week-old male BALB/c nude mice (16–18 g) used in this experiment were supplied by Beijing Vital River Laboratory Animal Technology Co. Ltd. (China). These experimental mice were kept in pairs under standardized housing conditions and provided with autoclaved food and sterile water. The animal facility maintained controlled environmental parameters: temperature of 22 $\pm$ 2 °C, relative humidity of 55 $\pm$ 5%, and a 12-hour light/dark cycle. Following a five-day acclimation period, all mice were randomly assigned to three groups (AGS, AGS+shRNA, and AGS+sh-CHST1), each containing six mice. Specifically, mice were stratified by body weight and then distributed randomly into three groups (1:1:1) using block randomization generated by R software (v 4.22, R Foundation for Statistical Computing, Vienna, Austria) (set.seed = 123). The allocation scheme was concealed in sealed envelopes maintained by a technician not involved in the experiments, and investigators were blinded to group assignments during all experimental procedures and outcome assessment. AGS cells, shRNA-transfected AGS cells, and CHST1-shRNA (sh-CHST1)-transfected AGS cells were digested, resuspended with PBS, and injected into the left axillary region of nude mice according to group assignment, with 5 $\times$ 10⁶ cells per injection. During the experimental period, mice had unrestricted access to both food and drinking water. Every four days, the tumors at the anterior axillary region were measured for length and width using calipers, and their volume was calculated using the following-listed formula: V = length $\times$ width²/2. After 14 days, mice were anesthetized via intraperitoneal administration of Zoletil 50 (tiletamine-zolazepam, 50 mg/kg, 10 mg/mL) and subsequently humanely euthanized by the method of cervical dislocation. Next, tumor tissues were collected, photographed, measured, and weighed. The volume of the tumor was calculated, ensuring that no tumor exceeded 1500 mm³ as permitted by the Institutional Animal Care and Use Committee. Tumor weights were recorded, and no deaths occurred throughout the entire experiment. The sequence of treatments and measurements was determined by cage position.

Human GC cell lines (MKN28, HGC27, and AGS) and the human gastric mucosal epithelial cell line GES-1 were supplied by Wuhan Procell Life Sciences Co., Ltd. (China). All these cell lines underwent STR profiling for authentication and were tested negative for mycoplasma. They were cultured in DMEM (C11995500BT, Thermo Fisher Scientific, Waltham, MA, USA) containing 10% fetal bovine serum (NFBS-2500A, Noverse, USA), 100 U/mL penicillin (ST488-1, Beyotime, China), and 100 µg/mL streptomycin (ST488-2, Beyotime, China), and maintained at 37 °C in an incubator with 5% CO₂.

96-well plates were seeded with logarithmically growing HGC27 cells (5000 cells/well) and AGS cells (3000 cells/per well) in 100 µL of culture medium. After 24, 48, or 72 h of incubation, 10 µL of CCK-8 solution (CK04, Solarbio, China) was added to each well, and the plates were further incubated at 37 °C for 2 h. MK-3 Microplate Reader (Thermo Fisher Scientific, USA) was employed to detect absorbance at 450 nm.

Lentivirus expressing sh-CHST1 or shRNA (negative control), along with CHST1 overexpression plasmid (OE-CHST1) or the corresponding empty plasmid (OE-NC), was obtained from OBiO Technology (Shanghai) Corp., Ltd. (China). Lentivirus transfection was performed to achieve CHST1 knockdown in AGS and HGC27 cells. Accordingly, AGS and HGC27 cells were categorized into two groups: a negative control group (AGS+shRNA and HGC27+shRNA); and a CHST1 intervention group (AGS+sh-CHST1, HGC27+sh-CHST1). CHST1 was overexpressed in AGS cells using Lipofectamine™ 3000 transfection reagent (L3000015, Thermo Fisher Scientific, USA) with OE-CHST1. AGS cells were subsequently allocated into the following experimental groups: Control, SB203580, and SB203580+OE-CHST1. Both AGS cells and CHST1-overexpressing AGS cells were treated with 20 µmol/L SB203580 (a p38 MAPK inhibitor, HY-10256, MedChemExpress, USA).

The extracted total proteins from tissue or cell lysates in each group were quantified using a BCA assay kit (P0011, Beyotime, China). The collected protein samples were separated by electrophoresis using SDS-PAGE gels (12%) (P1200-25T, Solarbio, China) and next transferred onto polyvinylidene fluoride membranes (ISEQ00010, Millipore, Burlington, MA, USA). After being blocked in 5% skim milk, the membranes were subjected to an overnight incubation with primary antibodies (Table 1) at 4 °C, and subsequently with 1 h of incubation with the corresponding secondary antibody at room temperature. Target proteins were then detected using the chemiluminescent detection substrate (WBULS0500, Millipore, USA).

GC cells of each strain (MKN28, HGC27, and AGS) and human gastric mucosal epithelial cells (GES-1) were inoculated onto anti-degradation slides. Cells were treated with 4% paraformaldehyde (BL539A, Biosharp, Hefei, Anhui, China) for fixation, followed by permeabilization with Triton X-100 (BS084-100 mL, Biosharp, China). After being blocked with 10% goat serum, the cells underwent an overnight incubation with primary antibodies against CHST1 (1:200, CST, USA), E-cadherin (1:200, CST, USA), vimentin (1:100, CST, USA), N-cadherin (1:400, CST, USA), and snail (1:200, CST, USA) at 4 °C, and then 1 h of incubation with fluorescently coupled secondary antibodies (4412, 1:1000, Alexa Fluor® 488, CST, USA) at room temperature. Subsequently, the cell nuclei were stained with DAPI (D9542, Millipore, USA) and photographed using a fluorescence microscope (Axio Observer. A1, Zeiss, Germany).

RNA of the control and CHST1-intervention groups were extracted using Trizol solution (15596026, Thermo Fisher Scientific, USA) and then reverse-transcribed into cDNA (RR037Q, Takara, China). qPCR (CN830S, Takara, China) was performed, with PCR reaction conditions as follows: 95 °C for 10 min, followed by 40 cycles of 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min. The 2^{-Δ⁢Δ⁢Ct} method was applied for calculating the relative level of CHST1 expression. The primer sequences were as follows: CHST1-F: GGTGGGTCTGGGTCTAGGAT; CHST1-R: GGGAGGAAGCACAGCAGATT; $\beta$-actin-F: CAGATGTGGATCAGCAAGCAGGAG; and $\beta$-actin-R: CGCAACTAAGTCATAGTCCGCCTAG.

During the logarithmic growth phase, AGS and HGC27 cells were inoculated in 6-well plates (1 $\times$ 10³ cells/well). After two weeks of culture, the cell cloning in each well was observed. The cells were then washed with PBS (three times, 5 min/time), pre-cooled with methanol fixation at 4 °C for 20 min, stained with crystal violet for 1 h, rinsed with ultrapure water, dried, and photographed with an inverted microscope.

Cells were seeded into 6-well plates and grown to approximately 90% confluence. A scratch perpendicular to the labeled line was created using a 100 µL sterile pipette. Images of the scratch were captured at the same observation point. The width of the scratch was measured at 0 and 24 h, followed by calculation of the wound healing rate.

Flow cytometry (FACS Calibur, BD, USA) was used to quantify apoptosis. After resuspension in 300 µL of binding buffer, cells were treated with Annexin V-FITC (5 µL) and incubated for 10 min. Subsequently, propidium iodide (5 µL) staining was performed for 5 min in the dark, and FlowJo V10 software (BD Biosciences, Ashland, OR, USA) was applied to analyze apoptosis.

R software was employed for performing bioinformatics analysis. No pre-established exclusion criteria were defined for animals or data points in this experiment. All analyses were performed using the original complete dataset. For statistical purposes, each group included three independent cell data or the data of six mice. During the experimental procedures, the first author was aware of the group allocations. The data were subjected to statistical analysis using SPSS 22.0 (IBM Corp., Armonk, NY, USA). The data were reported as mean $\pm$ standard deviation (SD). All statistical data were evaluated for normality and homogeneity of variance. The Student’s t-test was applied for pairwise comparisons, while comparisons among three or more groups were carried out using one-way ANOVA followed by Sidak’s multiple comparisons test. p 0.05 denotes a statistically significant difference.

Comparative analysis between tumor and normal samples revealed 4765 DEGs in total. Among these DEGs, there were 2542 up-regulated and 2223 down-regulated genes (Fig. 1A). Heatmap showing the top 20 up- and down-regulated genes is presented in Fig. 1B. Functional enrichment analyses were subsequently conducted on these identified DEGs. GO enrichment analysis revealed that, the DEGs were mainly enriched in ECM and extracellular structure organization (BP category); collagen-containing ECM (CC category); and receptor ligand activity and signaling receptor activator activity (MF category) (Fig. 1C). KEGG pathway analysis indicated enrichment of these DEGs in the neuroactive ligand-receptor interaction (Fig. 1D).

Fig. 1.

Screening and enrichment analysis of DEGs in GC. (A) Volcano plot of DEGs. (B) Heatmap of the top 20 up- and down-regulated DEGs. (C) GO enrichment analysis of DEGs. (D) KEGG pathway enrichment analysis of DEGs. DEGs, Differentially Expressed Genes; GC, gastric cancer; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

The EMT score was calculated for all samples, revealing a marked elevation in tumor group relative to normal group (p 0.05) (Fig. 2A). WGCNA was subsequently performed on the 4,765 DEGs to generate a heat map of sample clustering and clinical features (Fig. 2B). A power value of 8 was selected (Fig. 2C), corresponding to a R² of 0.85 and a slope of –1.48 (Fig. 2D).

Fig. 2.

EMT score and WGCNA. (A) Comparison of EMT scores between tumor and normal tissues. (B–D) WGCNA of DEGs in GC. ^*p 0.05. EMT, epithelial-mesenchymal transition; WGCNA, weighted gene co-expression network analysis.

A weighted gene co-expression network (Fig. 3A) was constructed, and the transcriptome overview map plots (Fig. 3B) and correlation heatmaps between modules (Fig. 3C) were generated. Subsequently, correlation analysis between sample characteristics (EMT score, tumor) and the modules was performed (Fig. 3D). The results showed that, among all modules, the green module (cor = 0.8, p 0.01) exhibited the strongest correlation and was thus defined as the key module, comprising 169 genes.

Fig. 3.

Screening of EMT-related modules. (A–C) Weighted gene co-expression network and heatmap of correlation between modules. (D) Correlation analysis of each module with EMT score and tumor.

GS vs. MM analysis (Fig. 4A) was performed on the key module, and 50 key module genes were obtained by filtering with thresholds of GS $>$0.6 & MM $>$0.6. Functional enrichment analysis was subsequently conducted on the key module genes. GO enrichment analysis revealed that the key module genes were primarily enriched in extracellular structure organization, ECM organization, and external encapsulating structure organization (BP category); collagen-containing ECM (CC category); and ECM structural constituent (MF category) (Fig. 4B). KEGG pathway analysis indicated significant enrichment in protein digestion and absorption (Fig. 4C).

Fig. 4.

Screening of EMT-related genes and functional enrichment analysis. (A) Screening of EMT-related genes. (B) GO enrichment analysis of EMT-related genes. (C) KEGG pathway enrichment analysis of EMT-related genes.

Survival analysis of the 50 EMT-related genes (Fig. 5) revealed seven genes significantly associated with survival: CHST1, GPR176, OLFML2B, P4HA3, PDGFRB, SPARC, and VCAN, which were defined as hub genes. Among these, CHST1 demonstrates high expression in GC tissues; nevertheless, the underlying mechanism remains unclear. As such, CHST1 was selected for subsequent analysis in this study.

Fig. 5.

Prognostic analysis of EMT-related genes. CHST1, carbohydrate sulfotransferase 1; GPR176, G protein-coupled receptor 176; OLFML2B, olfactomedin-like 2B; P4HA3, prolyl 4-hydroxylase subunit alpha 3; PDGFRB, platelet-derived growth factor receptor beta; SPARC, secreted protein acidic and cysteine rich; VCAN, versican.

Significant expression of CHST1 was found in the tumor group in comparison to the normal group (p 0.001) (Fig. 6A,B). Furthermore, a notable and statistically significant positive correlation was observed between CHST1 expression and the EMT score (R = 0.6, p 0.001) (Fig. 6C).

Fig. 6.

Correlation analysis between CHST1 and EMT score. (A,B) Comparison of CHST1 expression between the tumor and normal groups. (C) Correlation analysis of CHST1 expression with EMT Score. CHST1, carbohydrate sulfotransferase 1; EMT, epithelial-mesenchymal transition.

Clinical correlation analysis demonstrated a significant correlation between CHST1 expression and T-stage (p 0.001), whereas no marked correlation was observed between CHST1 expression and other clinical features, including gender, age, tumor grade, M-stage, N-stage, or stage (p $>$ 0.05) (Fig. 7).

Fig. 7.

Clinical correlation analysis of CHST1. ns: p $>$ 0.05, ^*⁣**p 0.001.

Subsequently, the GSE84437 dataset (comprising 433 GC samples) was analyzed to evaluate the association between CHST1 expression and clinical characteristics/patient survival. It was revealed through analysis of clinical features that CHST1 expression was markedly associated with T-stage but not with age, gender, or N-stage (p 0.05, Fig. 8A). Survival analysis demonstrated that patients with high CHST1 expression exhibited poorer overall survival and shorter survival versus those with low CHST1 expression (p 0.05, Fig. 8B). These findings suggest that high CHST1 expression is significantly related to the malignant progression and the unfavorable prognosis in GC.

Fig. 8.

Clinical correlation analysis and survival analysis of CHST1. (A) Correlation analysis of CHST1 expression with age, gender, N-stage, and T-stage in GC patients. (B) Survival analysis of CHST1. ns: p $>$ 0.05, ^*p 0.05.

CHST1 expression in patient-derived GC tissues was validated through IHC (Fig. 9A) and Western blot analysis (Fig. 9B). It was revealed that CHST1 protein expression was markedly increased in GC tissues versus normal tissues (p 0.01).

Fig. 9.

Validation of CHST1 expression. (A) IHC validation of CHST1 expression in GC tissues. Scale bar = 100 µm. (B) WB analysis of CHST1 expression in GC tissues. Values are reported as mean $\pm$ SD (n = 3 independent replicates). ^**p 0.01 vs. normal. Between-group statistical comparisons were carried out using Student’s t-test. IHC, immunohistochemistry; WB, western blot; SD, standard deviation.

We found high CHST1 expression in GC cell lines (MKN28, HGC27, and AGS) relative to the gastric epithelial cell line GES-1 (p 0.05), with the highest expression observed in AGS cells (Fig. 10A,B). Consistent results were obtained by IF analysis (Fig. 10C).

Fig. 10.

Analysis of CHST1 expression in GC cell lines. (A) WB detection of CHST1 expression. (B) qRT-PCR detection of CHST1 expression. Values are reported as mean $\pm$ SD (n = 3 independent replicates). ^**p 0.01, ^*⁣**p 0.001 vs. GES-1. One-way ANOVA combined with Sidak’s post hoc test. (C) IF detection of CHST1 expression. CHST1 is indicated by the red fluorescence signal, while cell nuclei were counterstained with DAPI (blue). Scale bar = 100 µm. CHST1, carbohydrate sulfotransferase 1; IF, immunofluorescence.

Given the highest CHST1 expression observed in AGS and HGC27 cells, these two cell lines were selected for carrying out subsequent experiments. Our findings showed that CHST1 protein expression was markedly lower in AGS and HGC27 cells in the sh-CHST1 group relative to the shRNA group (p 0.05) (Fig. 11A,B). Consistently, low expression of CHST1 was found to markedly reduce proliferation of both AGS and HGC27 cells (p 0.05) (Fig. 11C,D) but elevated the apoptosis rate (Fig. 11E). Additionally, decreased CHST1 expression led to a reduction in cell migration (Fig. 11F). These results suggest that inhibition of CHST1 suppresses GC cell proliferation and migration while enhancing apoptosis.

Fig. 11.

Effects of CHST1 on the biofunctionality of AGS and HGC27 cells. (A) WB analysis to determine CHST1 expression. Values are reported as the mean $\pm$ SD (n = 3 independent replicates). ^*⁣**p 0.001 vs. AGS+shRNA group, ^#⁢#⁢#p 0.001 vs. HGC27+shRNA group. Between-group statistical comparisons were carried out using Student’s t-test. (B) IF to detect CHST1 expression. CHST1 is indicated by the red fluorescence signal, while cell nuclei were counterstained with DAPI (blue). Scale bar = 100 µm. (C) CCK-8 assay to determine the proliferation of AGS and HGC27 cells. Values are reported as mean $\pm$ SD (n = 3 independent replicates). ^*p 0.05. Between-group statistical comparisons were carried out using Student’s t-test. (D) Plate cloning assay to detect cell proliferation. (E) Flow cytometry to determine cell apoptosis. (F) Scratch assay to determine cell migration. Scale bar = 100 µm. CHST1, carbohydrate sulfotransferase 1; sh-CHST1, CHST1 shRNA; CCK-8, Cell Counting Kit-8.

Enhanced expression of the EMT-related protein E-cadherin and attenuated expression of N-cadherin, vimentin, and snail were found in the sh-CHST1 group relative to the shRNA group (p 0.05) (Fig. 12A). Consistent results were observed by IF (Fig. 12B). It is suggested that CHST1 may be closely related to the EMT of AGS and HGC27 cells.

Fig. 12.

Effects of CHST1 on the EMT of AGS and HGC27 cells. (A) WB analysis to determine the expression levels of EMT-related proteins. Values are reported as mean $\pm$ SD (n = 3 independent replicates). ^**p 0.01, ^*⁣**p 0.001. Between-group statistical comparisons were carried out using Student’s t-test. (B) Detection of EMT-related proteins using IF. Cell nuclei were counterstained with DAPI (blue). Scale bar = 100 µm. CHST1, carbohydrate sulfotransferase 1; sh-CHST1, CHST1 shRNA.

Low expression of CHST1 suppressed the phosphorylation of p38 MAPK and ERK proteins in AGS cells (p 0.001, Fig. 13A), whereas overexpression of CHST1 promoted the phosphorylation of these proteins (p 0.001). Relative to the Control group, the SB203580 group exhibited significantly reduced cell proliferation (p 0.001), decreased colony formation, increased apoptosis rate, and lower cell migration (Fig. 13B–E). Compared with the SB203580 group, the SB203580+OE-CHST1 group demonstrated markedly enhanced proliferation (p 0.001), increased colony formation, decreased apoptosis rate, and improved cell migration. By contrast with the Control group, the SB203580 group had elevated E-cadherin protein expression (p 0.001), but decreased expression of N-cadherin, vimentin, and snail proteins (Fig. 13F, p 0.001). Relative to the SB203580 group, the SB203580+OE-CHST1 group showed decreased E-cadherin protein expression (p 0.001), but elevated expression of N-cadherin, vimentin, and snail proteins (p 0.05). Based on these results, it was indicated that CHST1 may promote EMT and malignant phenotypes in GC cells through the MAPK/ERK signaling pathway.

Fig. 13.

CHST1 modulates EMT and malignant phenotypes in AGS cells via regulation of the MAPK/ERK signaling pathway. (A) WB analysis of CHST1’s effect on the MAPK/ERK signaling pathway. Values are reported as mean $\pm$ SD (n = 3 independent replicates). Between-group statistical comparisons were carried out using Student’s t-test. (B,C) Cell proliferation evaluated using CCK-8 and plate colony formation assays. (D) Apoptosis rate measured by flow cytometry. (E) Cell migration evaluated using the scratch assay. Scale bar = 100 µm. (F) WB analysis of EMT-related protein expression. Values are reported as mean $\pm$ SD (n = 3 independent replicates). ^*⁣**p 0.001 vs. shRNA or Control group; ^#p 0.05, ^#⁢#p 0.01, ^#⁢#⁢#p 0.001 vs. OE-NC or SB203580 group. One-way ANOVA combined with Sidak’s post hoc test. CHST1, carbohydrate sulfotransferase 1; sh-CHST1, CHST1 shRNA; OE, CHST1 overexpression plasmid; OE-NC, overexpression empty plasmid; MAPK, mitogen-activated protein kinase; ERK, extracellular signal-regulated kinase.

Subcutaneous tumor formation experiments in nude mice showed (Fig. 14A) that both the tumor volume and weight were higher in the AGS+shRNA group relative to the AGS+sh-CHST1 group (p 0.001, Fig. 14B,C). In contrast, no marked difference was observed in tumor volume or weight between the AGS and the AGS+shRNA groups (p $>$ 0.05). WB analysis revealed a significant decrease in the phosphorylation levels of p38 MAPK and ERK proteins in the AGS+sh-CHST1 group relative to the AGS+shRNA control group (p 0.01, Fig. 14D). In contrast, no statistically significant difference was observed between the AGS parental and AGS+shRNA control groups (p $>$ 0.05). These results suggest that CHST1 can significantly promote the growth of GC and enhance the activity of the MAPK/ERK signaling pathway.

Fig. 14.

Subcutaneous tumor formation in nude mice. (A) Morphology of subcutaneous tumor formation. (B) Comparison of the weight of subcutaneous tumor formation. Values are reported as mean $\pm$ SD (n = 6, six mice). (C) Comparison of the volume of subcutaneous tumor formation. Values are reported as mean $\pm$ SD (n = 6, six mice). (D) WB analysis of CHST1’s effect on the MAPK/ERK signaling pathway. Values are reported as mean $\pm$ SD (n = 3 independent replicates). ns: p $>$ 0.05 vs. AGS group, ^**p 0.01, ^*⁣**p 0.001 vs. AGS+shRNA group. One-way ANOVA combined with Sidak’s post hoc test. CHST1, carbohydrate sulfotransferase 1; sh-CHST1, CHST1 shRNA.