Purpose: Uveal melanoma (UM) is a primary intraocular tumor in adults, with a high percentage of metastases to the liver. Identifying potential key genes may provide information for early detection and prognosis of UM metastasis.
Patients and Methods: Differentially expressed genes (DEGs) were identified using the GSE22138 dataset. Weighted gene co-expression network analysis was used to construct co-expression modules. Functional enrichment analysis was performed for DEGs and genes of key modules. Hub genes were screened by co-expression network and protein–protein interaction network (PPI), and validated by survival analysis in The Cancer Genome Atlas database. Gene set enrichment analysis (GSEA) was used to explore the potential metastasis mechanism of UM. Transient transfection was used to investigate the effect of TIMP1 on the proliferation, migration, and invasion of UM cells.
Results: In total, 552 DEGs were identified between primary and metastatic UM and mainly enriched in extracellular matrix, cellular senescence and focal adhesion pathway. A weighted gene co-expression network was built to identify key gene modules associated with UM metastasis (n=36). The turquoise module is positively correlated with metastasis and genes in this module were mainly enriched in peptidyl-tyrosine autophosphorylation and regulation of organ growth. The hub gene TIMP1  was screened out by co-expression network and PPI analysis. High expression of TIMP1 was related to p53 pathway by GSEA and short overall survival time. Experimental results indicated that overexpression of TIMP1 inhibited the proliferation and migration, while it had no significant effect on invasion of UM cells.
Conclusion: Our study indicates that TIMP1  might be associated with metastasis in UM, which might have important significance for identifying patients with high risk of metastasis and predicting the prognosis of UM.
Keywords: weighted gene co-expression network analysis, WGCNA, uveal melanoma, liver metastases, GO analysis, pathway analysis