RESUMO
Emerging evidence suggested that ferroptosis and immune activation, as well as their interactions, played a crucial role in the occurrence and progression of Parkinson's disease (PD). However, whether this interaction could serve as the basis for a hematological diagnosis of PD remained poorly understood. This study aimed to construct a novel hematological model for PD diagnosis based on the ferroptosis-related immune genes. The brain imaging of PD patients was obtained from the Affiliated Hospital of Nantong University. We used least absolute shrinkage and selection operator (LASSO) to identify the optimal signature ferroptosis-related immune genes based on six gene expression profile datasets of substantia nigra (SN) and peripheral blood of PD patients. Then we used the support vector machine (SVM) classifier to construct the hematological diagnostic model named Ferr.Sig for PD. Gene set enrichment analysis was utilized to execute gene functional annotation. The brain imaging and functional annotation analysis revealed prominent iron deposition and immune activation in the SN region of PD patients. We identified a total of 17 signature ferroptosis-related immune genes using LASSO method and imported them to SVM classifier. The Ferr.Sig model exhibited a high diagnostic accuracy, and its area under the curve (AUC) for distinguishing PD patients from healthy controls in the training and internal validation cohort reached 0.856 and 0.704, respectively. We also used the Ferr.Sig into other external validation cohorts, and a comparable AUC with the internal cohort was obtained, with the AUC of 0.727 in Scherzer's cohort, 0.745 in Roncagli's cohort, and 0.778 in Meiklejohn's cohort. Furthermore, the diagnostic performance of Ferr.Sig was not interfered by the other neurodegenerative diseases. This study revealed the value of ferroptosis-related immune genes in PD diagnosis, which may provide a novel direction and strategy for the development of novel biomarkers with less invasiveness, low cost, and high accuracy for PD screening and diagnosis.
RESUMO
BACKGROUND: Multiple molecular subtypes with distinct clinical outcomes in gastric cancer have been identified. Nonetheless, the immunogenomic subtypes of gastric cancer and its mediated tumor microenvironment (TME) characterizations have not been fully understood. METHODS: Six gastric cancer cohorts with 1506 samples were obtained. Unsupervised methods were used to perform immunogenomic phenotype clustering. The least absolute shrinkage and selection operator regression method was used to construct immunogenomic characterization score (IGCS). RESULTS: Three distinct immunogenomic phenotypes were determined. We observed a prominent survival difference between three phenotypes. The TME cell-infiltrating characteristics under these three phenotypes were highly consistent with three immune subtypes of tumors. Cluster 1, was characterized by the "immune-desert" phenotype, with relatively lower cell infiltration level (type 1 "cold tumor"); Cluster 2, characterized by "immune-inflamed" phenotype, with abundant innate and adaptive immune cell infiltration ("hot tumor"); Cluster 3, characterized by "immune-excluded" phenotype, with significant stromal activation and inactivated immune cell infiltration (type 2 "cold tumor"). We demonstrated IGCS signature was significantly correlated with TME inflammation and stroma activity, molecular subtypes, genetic variation, microsatellite instability, immune checkpoint molecules, and patient prognosis. High IGCS subtype, with poorer survival and enhanced stromal activity, presented an immune-exclusion and non-inflamed TME characterization. Low IGCS, related to increased mutation/neoantigen load and microsatellite instability, showed enhanced responses to anti-checkpoint immunotherapy. Four immunotherapy cohorts confirmed patients with low IGCS exhibited prominently enhanced clinical responses and treatment advantages. CONCLUSIONS: This study demonstrated the immunogenomic characterizations could play a crucial role in shaping the complexity and diversity of tumor microenvironment. Targeting tumor immunogenomic characteristic in order for changing adverse phenotypes may contribute to exploiting the novel immunotherapy combination strategies or novel immunotherapeutic drugs, and promoting the advance of tumor personalized immunotherapy.