TAMI-11. INCREASED M1 MACROPHAGE INFILTRATION CORRELATED WITH POOR PROGNOSIS OF WHO IV GLIOMAS

Zhaoming Zhou,Mingyao Lai,Jiangfen Zhou,Qingjun Hu,Ruyu Ai,Shaoqun Li,Cheng Zhou,Linbo Cai
DOI: https://doi.org/10.1093/neuonc/noab196.795
2021-01-01
Neuro-Oncology
Abstract:Abstract BACKGROUND Gliomas are the malignancy with a poor prognosis. Our previous database mining study demonstrated that M1 macrophage infiltration predicted the survival of GBM patients. Here in this study, we further explored the findings. METHODS RNA-seq was performed on 90 WHO IV glioma tissue samples. The sequencing data was investigated with xCell for the cell infiltration levels, and the M1 macrophage infiltration was further analyzed for the prognostic prediction effect with overall survival (OS) data. Differentially expressed genes (DEGs) were calculated between groups and the hub genes were determined by the MCC models in Cytoscape. The survival risk score (SRS) calculating models were established by several machine learning methods, including the least absolute shrinkage and selection operator (LASSO), generalized linear model (GLM), and linear discriminant analysis (LDA). RESULTS Compared with M1 macrophages none infiltration, WHO IV gliomas with M1 macrophages infiltration was associated with poor prognosis, and this result remained significant in multivariate analyses (hazard ratio [HR], 0.219; 95% CI, 0.047–0.723; P = 0.035). Protein-to-protein (PPI) network analysis of top 200 up-regulated DEGs determined 10 hub genes (P4HB, PDIA6, LAMB1, PRKCSH, CSF1, LAMB2, LGALS1, RCN1, CALU, and TNC). Further analysis determined that the 10 hub genes were enriched in the ECM-receptor interaction signaling pathway, and six out of the ten gene expressions were confirmed by immunohistochemistry staining. Based on the 6 genes, a survival risk score (SRS) was established by machine learning methods. SRS was able to distinguish the high-risk and low-risk WHO IV gliomas with an AUC = 0.80 [95% CI: 0.74 – 0.86, P < 0.01]. CONCLUSIONS M1 macrophage infiltration was an unfavorable prognostic biomarker for WHO IV gliomas. ECM-receptor interaction signaling pathway was involved in M1 macrophage infiltration. Hub genes in the signaling pathway could be the potential therapeutic targets for WHO IV gliomas.
What problem does this paper attempt to address?