Identification of 3 subpopulations of tumor-infiltrating immune cells for malignant transformation of low-grade glioma

BACKGROUND: Tumor-infiltrating immune cells (TIICs) are highly relevant to clinical outcome of glioma. However, previous studies cannot account for the diverse functions that make up the immune response in malignant transformation (MT) from low-grade glioma (LGG) to high-grade glioma (HGG). METHODS: Transcriptome level, genomic profiles and its relationship with clinical practice were obtained from TCGA and CGGA database. The “Cell type Identification By Estimating Relative Subsets Of RNA Transcripts (CIBERSORT)” algorithm was used to estimate the fraction of 22 immune cell types. We divided the TCGA and CGGA set into an experiment set (n = 174) and a validation set (n = 74) by random number table method. Univariate and multivariate analyses were performed to evaluate the 22 TIICs’ value for MT in LGG. ROC curve was plotted to calculate area under curve (AUC) and cut-off value. RESULTS: Heterogeneity between TIICs exists in both intra- and inter-groups. Several TIICs are notably associated with tumor grade, molecular subtypes and survival. T follicular helper (TFH) cells, activated NK Cells and M0 macrophages were screened out to be independent predictors for MT in LGG and formed an immune risk score (IRS) (AUC = 0.732, p < 0.001, 95% CI 0.657–0.808 cut-off value = 0.191). In addition, the IRS model was validated by validation group, Immunohistochemistry (IHC) and functional enrichment analyses. CONCLUSIONS: The proposed IRS model provides promising novel signatures for predicting MT from LGG to HGG and may bring a better design of glioma immunotherapy studies in years to come.