Comprehensive analysis of matrix metalloproteinases and their inhibitors in head and neck squamous cell carcinoma
DOI:
https://doi.org/10.1080/0284186X.2021.2009564Keywords:
Head and neck squamous cell carcinoma, matrix metalloproteinase, tissue inhibitors of metalloproteinases, bio-markers, gene signatureAbstract
Objective: This study aimed to explore the association of matrix metalloproteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) with cancer progression and prognosis in head and neck squamous cell carcinoma (HNSCC).
Methods: Differentially expressed genes (DEGs) were identified by LIMMA package using R software. The correlation between the expression levels of MMPs and TIMPs in HNSCC cancer samples and adjacent normal tissue samples was performed using Pearson correlation analysis. The Kruskal-Wallis test (H-test) was used to determine the association between the expression level of MMPs/TIMPs and HNSCC clinical stage. The survival result was expressed as a KM curve, and the log-rank test was used for statistical analysis. Lasso regression and multivariate Cox regression analyses were used to examine whether the gene signature based on MMPs and TIMPs was an independent prognostic factor in patients with HNSCC.
Results: Among the top 10 most up-regulated genes in HNSCC cancer tissues when compared with normal tissues, six genes belonged to the MMPs. Spearman correlation analysis revealed that only MMP11 and MMP23B were positively correlated with tumor stage. Survival analysis showed that patients with a high expression of MMP14, MMP20, TIMP1, and TIMP4 had a worse prognosis than low expression patients. Additionally, a novel five-gene (MMP3, MMP17, MMP19, MMP24, and TIMP1) signature was constructed and significantly associated with prognosis as an independent prognostic signature.
Conclusions: Our data show that the accuracy of a single gene of MMP or TIMP as predictors of progression and prognosis of HNSCC is limited, although some studies have proposed that MMPs act as driving factors for cancer progression. The prediction performance of the five-gene signature prediction model was much better than that of the gene signatures based on every single gene in prognosis prediction.
