Abstract
Purpose :
Exploring the relationship between matrix metalloproteinases(MMPs) and the prognosis of uveal melanoma(UM) through high-throughput sequencing
Methods :
We collected transcriptome sequencing data from 17 normal choroidal tissues and 53 UM tumor tissues. We analyzed differentially expressed mRNAs and enriched their functions with GO and KEGG databases. We used LASSO regression to select MMPs associated with prognosis based on clinical information and follow-up results of UM patients. Then used a risk score formula to construct a prognostic risk scoring system based on MMP2 and MMP28, and integrated pathological results to build a nomogram for predicting prognosis. Additionally, "estimate" and "cibersort" packages were performed on analyzing tumor microenvironment (TME) and immune cell infiltration from different risk groups.
Results :
The differentially expressed mRNA functions between normal choroidal and UM tumor tissues are mostly enriched in the extracellular matrix. Among the differentially expressed members of the MMPs family, MMP-2,-9,-12,-15,-16,-17,-19 and-28 were identified. A risk scoring system was constructed based on MMP-2 and -28, with the formula Risk Score = (MMP2 * 0.14643730) + (MMP28 * -0.03730257). With this system, 53 UM patients were divided into high-risk group (HRG) and low-risk group (LRG). Analysis revealed a significantly worse prognosis in the HRG compared to the LRG in both our own dataset and the TCGA validation set. The accuracy of the prognostic model based on the risk score was approximately 80% for 1, 3, and 5 years. After incorporating other clinical features, both risk score and pathology were identified as independent prognostic factors. A nomogram model was constructed based on these two factors, and the results indicated good concordance with actual outcomes at 1, 3, and 5 years. TME analysis showed HRG had higher ESTIMATE, immune, and stromal scores and more macrophages and Treg cells than LRG.
Conclusions :
This study is based on the largest comparative analysis to date between normal choroidal and UM tissues worldwide. The predictive model based on MMP-2 and -28 can accurately predict the prognosis of UM patients. The enrichment of macrophages and Treg cells in UM tissues may be associated with an adverse prognosis in the TME.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.