Development of a novel lipid metabolism-related model predicting the prognosis of gastric cancer and exploration the role of NPR3 in gastric cancer metastasis

Aim To establish a novel lipid metabolism-related (LMR) prognostic model for gastric cancer (GC) and explore the potential mechanism of natriuretic peptide receptor-3 (NPR3) in the process of GC metastasis. Method LMR genes were identified from the Gene Set Enrichment Analysis (GSEA) and mRNA expression profile were download from The Cancer Genome Atlas (TCGA) database. We used the R package “limma” to obtain the LMR differentially expressed genes (DEGs) between GC and adjacent tissues. Consensus clustering was then performed based on the expression of LMR DEGs using the R package “ConsensusClusterPlus”. We adopted the weighted correlation network analysis (WGCNA) to obtain the best module related to metabolic subtypes. A prognostic model based on 6 LMR genes (FBLN7, NPY1R, VTN, NPR3, EPHB3 and AUH) was established through least absolute shrinkage and selection operator (LASSO) penalized Cox regression analysis based on progression-free interval (PFI). In addition, we verified the NPR3 expression in several GC cell lines by quantitative Real-time PCR and Western Blotting, and explored the effect of NPR3 on GC cell migration using the wound healing assay and transwell test. We performed immunohistochemistry (IHC), H&E and collagen staining on 42 GC tissues to clarify the clinical significance of NPR3 in gastric cancer. Results 2 LMR subtypes (C1 and C2) were confirmed using consensus clustering of 153 LMR-DEGs. Compared with C1, C2 was associated with a worse prognosis, especially in terms of PFI (HR: 1.64, 95%CI: 1.15–2.33, P < 0.001). Using WGCNA and univariate cox regression, 558 genes were screened out to build and optimize the model. Finally, a novel predictive formula system based on 6 genes (FBLN7, NPY1R, VTN, NPR3, EPHB3 and AUH) were constructed and the time-dependent area under the receiver operating characteristic curve (time-ROC, 1/3/5 years) was 0.79/0.77/0.71 and 0.73/0.68/0.64 in the training set (N = 214) and validation set (N = 141), respectively. In addition, we found that NPR3 over-expression could promote the migration of GC cells. And its expression was higher in tumor tissues than in paracancerous tissues and patients with high expression of NPR3 were more likely have the vascular invasion (OR: 5.056, 95%CI: 1.159–22.060, p = 0.031) and higher stage (OR: 5.100, 95%CI: 1.336–19.470, p = 0.017). Conclusion We established a novel LMR prognostic model predicting the prognosis of gastric cancer, and found that NPR3 can promote the tumor migration and vascular invasion of gastric cancer.