Construction and validation of a risk model for colorectal cancer prognosis based on N6-methyladenosine modification and immune infiltration

Objective To investigate the prognostic value of N6-methyladenosine (m⁶A) modification related genes and immune infiltration in colorectal cancer (CRC) and construct a risk model for predicting outcome of patients.

Methods The transcriptome data and matched clinical information of CRC patients were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) database. The prognostic value of m⁶A modification related genes and immune infiltration were investigated using the consensus clustering method and single sample gene set enrichment analysis. The weighted gene co-expression network analysis (WGCNA) was used to identify prognostic genes related with m⁶A modification and immune infiltration. Lasso regression analysis was used to construct a multi-gene risk model. The expression differences of prognostic genes identified were further validated through expression differential analysis in the Gene Expression Profiling Interactive Analysis (GEPIA) database. Finally, the Kaplan-Meier was used to evaluate the predicting performance of the model in different subgroups and external validation cohorts.

Results Both the m⁶A modification and immune infiltration phenotype could effectively stratify the prognosis of CRC patients from the TCGA cohort. Most m⁶A modification related genes were significantly correlated with immune infiltration in CRC tissues. Four following prognostic genes were selected using the WGCNA method combined with Lasso regression analysis: intelectin-1, lymphocyte antigen 6 complex locus G6D, atonal homolog 1 and matrix metalloproteinase 28. In colorectal cancer tissues, the expression levels of lymphocyte antigen 6 complex locus G6D and matrix metalloproteinase 28 exhibited significant differences compared to adjacent non-cancerous tissues (P < 0.05). The risk model constructed based on the above prognostic genes can effectively identify the potential risk population with poor prognosis in different clinical subgroups of the TCGA cohort and the GEO validated cohort.

Conclusion A risk model based on m⁶A modification and immune infiltration could effectively predict the clinical outcome of CRC patients, and related prognostic genes have potential to be developed as molecular targets for CRC therapy.