·GEPIA (Gene Expression Profiling Interactive Analysis) platform was used to analyze the MMPs expression in 30 common tumor tissues. R software was used to integrate transcriptomic data of 167 normal pancreatic tissues from GTEx (Genotype-Tissue Expression) database with 178 pancreatic cancer tissues and 4 para-tumor tissues from TCGA (The Cancer Genome Atlas) database to compare the differences of MMP14 expression. GEPIA platform was used to perform survival analysis in the patients with different MMP14 expression levels. Perl and R project software were used to analyze the correlation of MMP14 expression and clinical characteristics. R project software and GSEA v3.0 software were used for Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis and Gene Ontology (GO) functional enrichment analysis. CIBERSORT analysis was used to calculate the proportions of immune cell subsets in the tumor immune microenvironment, and analyze correlation of MMP14 expression and the proportion of immune cell subsets. Six fresh surgical samples of pancreatic cancer tissues and paired para-tumor tissues were collected to detect various immune cell subsets by flow cytometry. t-SNE dimensionality reduction and statistical analysis were used to explore the characteristics of tumor immune microenvironment in pancreatic cancer. A tissue microarray was used to conduct multiplex immunohistochemistry (mIHC) to detect MMP14 expression and the proportions of immune cell subsets to verify bioinformatics analysis results.

·In the bioinformatics database, compared with the other members of MMPs, MMP14 was highly expressed in a variety of tumors, and MMP14 expression in pancreatic cancer was higher than that in the other cancers. The expression of MMP14 in pancreatic cancer tissues was significantly higher than that in normal pancreatic tissues and para-tumor tissues (P=0.000). The expression of MMP14 in the TNM stage II tumor was higher than that in the stage I tumor (P=0.012). The expressions of MMP14 in the histological grade 2 (G2) and grade 3 (G3) tumors were higher than those in the grade 1 (G1) tumor (P=0.000). The survival prognosis of the patients in MMP14-low expression group was significantly better than that in high expression group (P=0.033). KEGG and GO analysis results showed that MMP14 was enriched in the pancreatic cancer pathway and immune-related pathways. The expression level of MMP14 was negatively correlated with the percentages of CD8⁺ T cells and monocytes (P<0.05) , and positively correlated with the percentage of macrophages (P=0.000). Flow cytometry analysis suggested that tumor microenvironment of pancreatic cancer was immunosuppressive. The results of mIHC showed that MMP14 expression in the tumor tissues was higher than that in the para-tumor tissues (P=0.000). Tumor tissues with higher TNM stage and histological grade expressed relatively higher MMP14 levels (P<0.05). In the tumor tissues with high MMP14 expression level, the percentage of CD8⁺ T cells decreased (P=0.001) and the percentage of macrophages increased (P=0.000), which were consistent with bioinformatics analysis results.

·Compared with normal pancreatic tissues and para-tumor tissues, the expression of MMP14 significantly increases in pancreatic tumor tissues and the high expression is associated with poor prognosis of the patients; in the pancreatic tumor tissues with high MMP14 expression, the percentage of CD8⁺ T cells decreases and the percentage of macrophages increases; with high MMP14 expression, the tumor microenvironment of pancreatic cancer is highly immunosuppressive.