Stem Cell Landscape Aids in Tumor Microenvironment Identification and Selection of Therapeutic Agents in Gastric Cancer.
Chao He,Yongfeng Ding,Yan Yang,Gang Che,Fei Teng,Haohao Wang,Jing Zhang,Donghui Zhou,Yanyan Chen,Zhan Zhou,Haiyong Wang,Lisong Teng
DOI: https://doi.org/10.1016/j.cellsig.2023.110965
IF: 4.85
2024-01-01
Cellular Signalling
Abstract:Gastric cancer stem cells (GCSCs) are strongly associated with the refractory characteristics of gastric cancer, including drug resistance, recurrence, and metastasis. The prognosis for advanced gastric cancer patients treated with multimodal therapy after surgery remains discouraging. GCSCs hold promise as therapeutic targets for GC patients. We obtained 26 sets of stem cell-related genes from the StemChecker database. The Consensus clustering algorithm was employed to discern three distinct stemness subtypes. Prognostic outcomes, components of the tumor microenvironment (TME), and responses to therapies were compared among these subtypes. Following this, a stemness-risk model was formulated using weighted gene correlation network analysis (WGCNA), alongside Cox regression and random survival forest analyses. The C2 subtype predominantly showed enrichment in negative prognostic CSC gene sets and demonstrated an immunosuppressive TME. This specific subtype exhibited minimal responsiveness to immunotherapies and demonstrated reduced sensitivity to drugs. Four pivotal genes were integrated into the construction of the stemness model. Gastric cancer patients with higher stemness-risk scores demonstrated poorer prognoses, a greater presence of immunosuppressive components in TME, and lower rates of treatment response. Subset analysis indicated that only the low-stemness risk subtype derives benefit from 5-fluorouracil-based adjuvant chemotherapy. The model's effectiveness in immunotherapeutic prediction was further validated in the PRJEB25780 cohort. Our study categorized gastric cancer patients into three stemness subtypes, each demonstrating distinct prognoses, components of TME infiltration, and varying sensitivity or resistance to standard chemotherapy or targeted therapy. We propose that the stemness risk model may help the development of well-grounded treatment recommendations and prognostic assessments.