Prognostic role of tumor microenvironment and immune- and autophagy-related genes in colorectal adenocarcinoma.

Abstract

Background: Colorectal adenocarcinoma (COADREAD) is the second most common cause of cancer-associated deaths. Immunity and autophagy play a key role in the development and progression of COADREAD, but the specific mechanisms have not been fully elucidated. We aimed to explore immune- and autophagy-related genes (IARGs) to establish prognostic risk assessment and clinical prediction models and to understand the molecular basis of COADREAD.

Methods: Transcriptomic and clinical data from colon (COAD) and rectal cancers (READ) were obtained from TCGA and GEO databases, including 460 COADREAD cases and validation cohorts (GSE161158/GSE17536). Immune-related (IRGs) and autophagy-related genes (ARGs) were integrated to identify 22 immune-autophagy-related genes (IARGs). Molecular subtypes were constructed via consensus clustering of IARGs, followed by gene set variation analysis (GSVA) to explore pathway activity. Differentially expressed immune-autophagy-related genes (IARDEGs) were identified using limma and subjected to functional enrichment [Gene Ontology (GO)/Kyoto Encyclopedia of Genes and Genomes (KEGG)]. A prognostic risk model was developed via LASSO and Cox regression, validated in external cohorts. Immune infiltration was assessed using ssGSEA, and a nomogram integrating clinicopathological features was established. Statistical analyses were performed in R (v4.2.2), with significance at P<0.05.

Results: The IARG data could be used to distinguish between cancerous and normal specimens of COADREAD. VEGFA, BIRC5, and BID genes were highly expressed in COADREAD, while the expression of TNFSF10 was low. Most of the IARGs were positively correlated with COADREAD. The GSVA results of four classes of C4 verified that the clustering effect was best. More IARGs, such as CXCR4, CCL2, and CTSB, were in the C4 class than the C1 class. In the risk model, the T cell and B cell receptor pathways were substantially upregulated in patients in the low-risk group. The risk score greatly differed with the different expression levels of key immune checkpoints and immune cell infiltration, and the levels of immune cells were higher in the low-risk group.

Conclusions: In this study, bioinformatic analysis proved that immune-a1-related genes could be used to distinguish between normal and COADREAD specimens and that immunity and autophagy are associated with low-risk COADREAD; therefore, these genes have the potential to improve clinical predictions of COADREAD risk.