Exploring the potential mechanisms of Yi-Yi-Fu-Zi-Bai-Jiang-San therapy on the immune-inflamed phenotype of colorectal cancer via combined network pharmacology and bioinformatics analyses.

Background: The development and progression of colorectal cancer (CRC) is closely associated with its complex tumor microenvironment (TME). Assessment of the modified pattern of immune cell infiltration (ICI) will help increase knowledge regarding the characteristics of TME infiltration. Yi-Yi-Fu-Zi-Bai-Jiang-San (YYFZBJS) has been shown to have positive effects on the regulation of the immune microenvironment of CRC. However, its pharmacological targets and molecular mechanisms remain to be elucidated.

Methods: Network pharmacological analysis was used to identify the target of YYFZBJS in the TME of CRC. Patients with the immune-inflamed phenotype (IIP) were identified using CRC samples from The Cancer Genome Atlas (TCGA) database. Consensus genes were identified by intersecting YYFZBJS targets, CRC disease targets and differentially expressed genes in the CRC microenvironment. Then, least absolute shrinkage and selection operator (LASSO) Cox analyses were used to identify a prognostic signature from the consensus genes. Cytoscape software was further used to build a unique herb-compound-target network diagram of the important components of YYFZBJS and prognostic gene targets. In addition, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis was performed using the prognostic gene sets to explore the molecular mechanism of the prognostic genes in drug therapy for CRC IIP patients. Finally, single-cell analysis was performed to validate the expression of the prognostic genes in the TME of CRC using the TISCH2 database.

Results: A total of 284 IIP patients were identified from 480 patients with CRC. A total of 35 consensus genes were identified as targets of YYFZBJS in the TME of CRC patients. An eleven-gene prognostic signature, including PIK3CG, C5AR1, PRF1, CAV1, HPGDS, PTGS2, SERPINE1, IDO1, TGFB1, CXCR2 and MMP9, was identified from the consensus genes, with areas under the receiver operating characteristic (ROC) curve (AUCs) values of 0.84 and 0.793 for the training and test cohorts, respectively. In the herb-compound-target network, twenty-four compounds were shown to interact with the 11 prognostic genes, which were significantly enriched in the IL-17 signaling, arachidonic acid metabolism and metabolic pathways. Single-cell analysis of the prognostic genes confirmed that their abnormal expression was associated with the TME of CRC.

Conclusion: This study organically integrated network pharmacology and bioinformatics analyses to identify prognostic genes in CRC IIP patients from the targets of YYFZBJS. Although this data mining work was limited to the study of mechanisms related to prognosis based on the immune microenvironment, the methodology provides new perspectives in the search for novel therapeutic targets of traditional Chinese medicines (TCMs) and accurate diagnostic indicators of cancers targeted by TCMs.