Background
EMT is an important biological process in the mechanism of tumor invasion and metastasis. However, there are still many unknowns about the specific mechanism of EMT in tumor. At present, a comprehensive analysis of EMT-related genes in colorectal cancer (CRC) is still lacking.
Methods
All the data were downloaded from public databases including TCGA database (488 tumor samples and 52 normal samples) as the training set and the GEO database (GSE40967 including 566 tumor samples and 19 normal samples, GSE12945 including 62 tumor samples, GSE17536 including 177 tumor samples, GSE17537 including 55 tumor samples) as the validation sets. One hundred and sixty-six EMT-related genes (EMT-RDGs) were selected from the Molecular Signatures Database. Bioinformatics methods were used to analyze the correlation between EMT-RDGs and CRC prognosis, metastasis, drug efficacy, and immunity.
Results
We finally obtained nine prognostic-related EMT-RDGs (FGF8, NOG, PHLDB2, SIX2, SNAI1, TBX5, TIAM1, TWIST1, TCF15) through differential expression analysis, Unicox and Lasso regression analysis, and then constructed a risk prognosis model. There were significant differences in clinical characteristics, 22 immune cells, and immune functions between the high-risk and low-risk groups and the different states of the nine prognostic-related EMT-RDGs. The methylation level and mutation status of nine prognostic-related EMT-RDGs all affect their regulation of EMT. The Cox proportional hazards regression model was also constructed by the methylation sites of nine prognostic-related EMT-RDGs. In addition, the expression of FGF8, PHLDB2, SIX2, and SNAIL was higher and the expression level of NOG and TWIST1 was lower in the non-metastasis CRC group. Nine prognostic-related EMT-RDGs also affected the drug treatment response of CRC.
Conclusions
Targeting these nine prognostic-related EMT-RDGs can regulate CRC metastasis and immune, which is beneficial for the prognosis of CRC patients, improve drug sensitivity in CRC patients.