Amino acid metabolism-related genes as potential biomarkers and the role of MATN3 in stomach adenocarcinoma: A bioinformatics, mendelian randomization and experimental validation study.

BACKGROUND

Stomach adenocarcinoma (STAD) is a major contributor to cancer-related mortality worldwide. Alterations in amino acid metabolism, which is integral to protein synthesis, have been observed across various tumor types. However, the prognostic significance of amino acid metabolism-related genes in STAD remains underexplored.

METHODS

Transcriptomic gene expression and clinical data for STAD patients were obtained from the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. Amino acid metabolism-related gene sets were sourced from the Gene Set Enrichment Analysis (GSEA) database. A prognostic model was built using LASSO Cox regression based on the TCGA cohort and validated with GEO datasets (GSE84433, GSE84437, GSE84426). Kaplan-Meier analysis compared overall survival (OS) between high- and low-risk groups, and ROC curves assessed model accuracy. A nomogram predicted 1-, 3-, and 5-year survival. Copy number variations (CNVs) in model genes were visualized using data from the Xena platform, and mutation profiles were analyzed with "maftools" to create a waterfall plot. KEGG and GO enrichment analyses were performed to explore biological mechanisms. Immune infiltration and related functions were evaluated via ssGSEA, and Spearman correlation analyzed associations between risk scores and immune components. The TIDE database predicted immunotherapy efficacy, while FDA-approved drug sensitivity was assessed through CellMiner database. The role of MATN3 in STAD was further examined in vitro and in vivo, including amino acid-targeted metabolomic sequencing to assess its impact on metabolism. Finally, Mendelian randomization (MR) analysis evaluated the causal relationship between the model genes and gastric cancer.

RESULTS

In this study, we developed a prognostic risk model for STAD based on three amino acid metabolism-related genes (SERPINE1, NRP1, MATN3) using LASSO regression analysis. CNV amplification was common in SERPINE1 and NRP1, while CNV deletion frequently occurred in MATN3. STAD patients were classified into high- and low-risk groups based on the median risk score, with the high-risk group showing worse prognosis. A nomogram incorporating the risk score and clinical factors was created to estimate 1-, 3-, and 5-year survival rates. Distinct mutation profiles were observed between risk groups, with KEGG pathway analysis showing immune-related pathways enriched in the high-risk group. High-risk scores were significantly associated with the C6 (TGF-β dominant) subtype, while low-risk scores correlated with the C4 (lymphocyte-depleted) subtype. Higher risk scores also indicated increased immune infiltration, enhanced immune functions, lower tumor purity, and poorer immunotherapy response. Model genes were linked to anticancer drug sensitivity. Manipulating MATN3 expression showed that it promoted STAD cell proliferation and migration in vitro and tumor growth in vivo. Metabolomic sequencing revealed that MATN3 knockdown elevated levels of 30 amino acid metabolites, including alpha-aminobutyric acid, glycine, and aspartic acid, while reducing (S)-β-Aminoisobutyric acid and argininosuccinic acid. MR analysis found a significant causal effect of NRP1 on gastric cancer, but no causal relationship for MATN3 or SERPINE1.

CONCLUSION

In conclusion, the amino acid metabolism-related prognostic model shows promise as a valuable biomarker for predicting the clinical prognosis, selecting immunotherapy and drug treatment for STAD patients. Furthermore, our study has shed light on the potential value of the MATN3 as a promising strategy for combating the progression of STAD.