Integrating bioinformatics and machine learning to analyze drug resistance-associated biomarkers and immune cell infiltration characteristics in colorectal cancer

Chemotherapy resistance remains a major challenge in colorectal cancer (CRC) treatment. The aim of this study is to identify biomarkers associated with Selumetinib resistance in CRC. Two gene expression datasets were available from the Gene Expression Omnibus collection. The differentially expressed genes (DEGs) were found using the R program. Then GO, KEGG and GSEA enrichment analysis were conducted, followed by PPI network construction with the STRING online database. The signature genes were identified by use of logistic regression and Random Forest. The ROC curve and nomogram model were used to verify the discrimination and efficacy of CALU and MRPS33. Besides, CIBERSORT analysis was used to explore the relationship between CALU/MRPS33 and infiltrating immune cells. For in vitro experiments, western blot was used to detect MRPS33 protein expression. CCK-8 and transwell assays were performed to assess CRC cell proliferation, migration, and invasion during Selumetinib/Trametinib treatment. We analyzed the GSE120993 dataset to identify 335 DEGs between sensitive and resistant CRC cells, including 198 up-regulated and 137 down-regulated genes. These DEGs were mainly enriched in ‘Metabolic pathways’, ‘Human T-cell leukemia virus 1 infection’, ‘Human papillomavirus infection pathway’ and ‘Prion disease’. CALU and MRPS33 were identified as potential biomarkers, which showed different expression in the Selumetinib/Trametinib-resistant group compared to the Selumetinib/Trametinib-sensitive group. Additionally, MRPS33 was positively correlated with Neutrophils. CALU was positively correlated with Macrophages M2 and T cells CD8, and was negatively correlated with Mast cells activated and T cells CD4 naive. There were significant correlations between MRPS33/CALU and Trametinib/Selumetinib-related genes. More importantly, MRPS33 was verified to show high protein expression in Selumetinib/Trametinib-resistant HT29 cells and HCT116 cells. Silencing of MRPS33 repressed Selumetinib/Trametinib resistance in HT29 cells and HCT116 cells. MRPS33 was recognized as potential diagnostic biomarkers for Selumetinib/Trametinib resistance in CRC, which may contribute to personalized treatment strategies.