Identification of a lipid homeostasis-related gene signature for predicting prognosis, immunity, and chemotherapeutic effect in patients with gastric cancer

Abstract

AbstractGastric cancer (GC) is one of the most common and deadliest cancers worldwide. Lipid homeostasis is essential for tumour development because lipid metabolism is one of the most important metabolic reprogramming pathways within tumours. Elucidating the mechanism of lipid homeostasis in GC might significantly improve treatment strategies and patient prognosis. GSE62254 was applied to construct a lipid homeostasis-related gene signature score (HGSscore) by multiple bioinformatic algorithms including weighted gene coexpression network analysis (WGCNA) and LASSO-Cox regression. A nomogram based on HGSscore and relevant clinical characteristics was constructed to predict the survival of patients with GC. TIMER and xCell were used to evaluate immune and stromal cell infiltration in the tumour microenvironment. Correlations between lipid homeostasis-related genes and chemotherapeutic efficacy were analysed in GSCAlite. RT‒qPCR and cell viability assays were applied to verify the findings in this study. HGSscore was constructed based on eighteen lipid homeostasis-related genes that were selected by WGCNA and LASSO-Cox regression. HGSscore was strongly associated with advanced TNM stage and showed satisfactory value in predicting GC prognosis in three independent cohorts. Furthermore, we found that HGSscore was associated with the tumour mutation burden (TMB) and immune/stromal cell infiltration, which are related to GC prognosis, indicating that lipid homeostasis impacts the formation of the tumour microenvironment (TME). With respect to the GSCAlite platform, PLOD2 and TGFB2 were shown to be positively related to chemotherapeutic resistance, while SLC10A7 was a favourable factor for chemotherapy efficacy. Cell viability assays showed that disrupted lipid homeostasis could attenuate GC cell viability. Moreover, RT‒qPCR revealed that lipid homeostasis could influence expression of specific genes. We identified a lipid homeostasis-related gene signature that correlated with survival, clinical characteristics, the TME, and chemotherapeutic efficacy in GC patients. This research provides a new perspective for improving prognosis and guiding individualized chemotherapy for patients with GC.