Identification and Validation of Amino Acid Metabolism-Related Biomarkers and Investigation of their Potential Mechanisms in Lung Adenocarcinoma

Background: In lung adenocarcinoma (LUAD), the metabolism of amino acids (AAs) plays a crucial role in the growth, infiltration, and metastasis of tumor cells. Nevertheless, the potential of AA metabolism-associated genes (AAMRGs) to serve as prognostic indicators in LUAD remains ambiguous. Thus, this study sought to evaluate the prognostic value of AAMRGs in LUAD patients. Methods: Herein, we extracted LUAD transcriptomic information from two key repositories, namely The Cancer Genome Atlas Program (TCGA) and Gene Expression Omnibus. The non-negative matrix factorization (NMF) clustering technique was used to categorize the LUAD cases based on their AAM profiles before assessing the survival rates and composition of immune cells. Using limma software, shared dysregulated transcripts were identified across subgroups before functional annotation via DAVID, which comprised exploration of gene ontology and the Kyoto Encyclopedia of Genes and Genomes pathway. The prognostic framework was developed using five prognostic indicators through TCGA-derived LUAD specimens. We performed the analysis using singlevariable Cox, least absolute shrinkage and selection operator regression, and multi-factorial Cox regression. Molecular pathways between cohorts were compared with gene set enrichment analysis (GSEA). Real-time quantitative polymerase chain reaction (RT-qPCR) and immunohistochemical (IHC) analysis were utilized to validate the key genetic components of the model. Results: NMF clustering analysis was performed to categorize 497 LUAD patients into three distinct subgroups with obvious variations in the survival rates. The subtypes exhibited substantial disparities in immune cell populations, particularly in monocytes and mast cells. Analysis of 176 shared differentially expressed genes (DEGs) revealed enrichment in T lymphocyte stimulation, immunological reactions, and extra immune-related processes within the subgroups. The prognostic framework was constructed using biomarkers, such as ERO1LB, HPGDS, LOXL2, TMPRSS11E, and SLC34A2. Moreover, GSEA demonstrated a correlation between elevated risk and cell cycle processes, but lower risk was linked with arachidonic acid metabolic pathways. Analysis of 1128 DEGs revealed enrichment in various physiological processes, including cellular division, p53 signaling cascades, immunological responses, and additional pathways upon the comparison of high and low-risk cohorts. The RT-qPCR analysis confirmed elevated expression levels of ERO1LB and TMPRSS11E in LUAD specimens. Consistent with RT-qPCR analysis, the IHC results affirmed that the expression levels of ERO1LB and TMPRSS11E were increased in LUAD specimens. Conclusion: The five identified AAMRGs in LUAD were validated and appropriately utilized to construct a risk assessment model that could potentially act as prognostic biomarkers for LUAD patients.