A novel signature of aging-related genes associated with lymphatic metastasis for survival prediction in patients with bladder cancer

Abstract

Background: The predominant and most prevalent form of metastatic bladder cancer (BCa) is lymphatic metastasis, which is associated with a highly dismal prognosis for patients. Aging-related genes (ARGs) are believed to contribute significantly to tumor development. However, the effect of ARGs on lymphatic metastasis of BCa is unclear. This research sought to establish a prognosis model based on ARGs associated with lymphatic metastasis in BCa.

Methods: We downloaded BCa data from the TCGA and GEO databases and ARGs from the Aging Atlas database. The least absolute shrinkage and selection operator (LASSO) approach was applied to obtain the characteristic ARGs of risk signature in the TCGA cohort. Verification was done using the GSE13507 dataset. The R package 'ConsensusClusterPlus' was employed to identify the molecular subtypes based on the characteristic ARGs. Protein-Protein interaction network, MCODE analysis, enrichment analysis (KEGG, GO, GSEA), and immune infiltration analysis were performed to investigate underlying mechanisms. EdU, migration and invasion assays, wound healing assays, immunofluorescence staining, and quantitative polymerase chain reaction were conducted to evaluate the impact of ELN on the proliferative, migratory, and invasive capacities of BCa cells.

Results: We identified 20 differently expressed ARGs. A four ARGs risk signature (EFEMP1, UCHL1, TP63, ELN) was constructed in the TCGA cohort. The high-risk group (category) recorded a reduced overall survival (OS) rate relative to the low-risk category (hazard ratio, 2.15; P <0.001). The risk score could predict lymphatic metastasis in TCGA cohort (AUC=0.67). The GSE13507 dataset was employed to verify the validity of this risk score. Based on the four ARGs, two distinct aging profiles (Cluster 1 and Cluster 2) were discovered utilizing the ConsensusClusterPlus, and Cluster 2 possessed a favorable OS in contrast with Cluster 1 (hazard ratio, 0.69; P =0.02). Classical tumor signaling pathways, ECM-associated signaling pathways, and immune-related signaling pathways participate in BCa progression. ELN recombinant protein affected the expression of collagen and increased migration and invasiveness in BCa cells.

Conclusion: We constructed a four-ARG risk signature and identified two aging molecular subtypes. This signature could serve as an effective survival predictor for patients with BCa.

Keywords: aging; bladder cancer; extracellular matrix remodeling; lymphatic metastasis; molecular subtypes; prognostic signature.