OBJECTIVE This study aims to identify immune-related subtypes, prognostic biomarkers and potential therapeutic targets in gastric cancer (GC) through integrative multi-omics analysis, with a particular focus on the immunoregulatory mechanisms mediated by the active components of the traditional Chinese herbal formula Wei-Liu-An Mixture (Wei-Liu-An-He-Ji, abbreviated as Wei-Liu-An).

METHODS Based on immune infiltration profiles, unsupervised clustering was performed to stratify GC patients into three immune phenotypes: an immune-desert subtype (Cluster 1) , an immune-excluded subtype (Cluster 2) and an immune-inflamed subtype (Cluster 3). Differentially expressed genes (DEGs) among immune subtypes were identified and intersected with metabolite-related targets of Wei-Liu-An Mixture obtained by ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Single-cell RNA sequencing (scRNA-seq) data were leveraged to delineate the cell-type-specific expression patterns and putative localization features of prognostic genes across distinct tumour immune microenvironment (TIME) compartments, with particular emphasis on functional enrichment analyses of crystallin alpha B (CRYAB)-positive fibroblasts. Finally, we developed an immune-inflamed subtype-derived 9-gene prognostic risk score model, termed IIP-9GS (Immune-Inflamed phenotype-derived 9-Gene Signature Risk Score), and externally validated its predictive performance for survival risk stratification in two independent GC cohorts.

RESULTS The IIP-9GS prognostic risk score model effectively stratified patients with GC into high- and low-risk groups, demonstrating robust and stable discriminative performance across cohorts.Among the signature genes, interferon regulatory factor 1 (IRF1) and interferon-gamma (IFNG) were identified as potential targets of Wei-Liu-An, and their high expression was significantly associated with favorable prognosis. In contrast, CRYAB was downregulated in the immune-inflamed Cluster 3, whereas previous analyses indicated that its high expression correlated with poor survival, suggesting that CRYAB may represent a key high-risk gene in GC. Notably, CRYAB was highly enriched in fibroblasts, smooth muscle cells and endothelial cells, implying that it may exert complex biological effects by modulating the stromal compartment of the TIME.

CONCLUSION The study performs multi-omics integrative bioinformatics analyses based on immune infiltration characteristics in gastric cancer (GC) and establishes the IIP-9GS prognostic risk score model. At the single-cell level, it suggests that CRYAB⁺ fibroblasts may represent a pivotal stromal regulatory node within the TIME. Integrating target prediction results for the active constituents of the Wei-Liu-An, we propose that Wei-Liu-An may contribute to TIME remodeling by modulating immune-related pathways and stromal cell states, thereby alleviating immunosuppression and mitigating stroma-associated resistance. Collectively, CRYAB and its associated fibroblast subpopulations may serve as potential prognostic biomarkers and therapeutic targets for overcoming immunosuppression and stroma-driven resistance, providing new clues and a theoretical basis for individualized integrative treatment strategies in GC.