The immunosuppressive microenvironment in glioblastoma (GBM) prevents an efficient antitumoral immune response and enables tumor formation and growth. Although an understanding of the nature of immunosuppression is still largely lacking, it is important for successful cancer treatment through immune system modulation. To gain insight into immunosuppression in GBM, we performed a computational analysis to model relative immune cell content and type of immune response in each GBM tumor sample from The Cancer Genome Atlas RNA-seq data set. We uncovered high variability in immune system-related responses and in the composition of the microenvironment across the cohort, suggesting immunologic diversity. Immune cell compositions were associated with typical alterations such as IDH mutation or inactivating NF1 mutation/deletion. Furthermore, our analysis identified three GBM subgroups presenting different adaptive immune responses: negative, humoral, and cellular-like. These subgroups were linked to transcriptional GBM subtypes and typical genetic alterations. All G-CIMP and IDH-mutated samples were in the negative group, which was also enriched by cases with focal amplification of CDK4 and MARCH9. IDH1-mutated samples showed lower expression and higher DNA methylation of MHC-I-type HLA genes. Overall, our analysis reveals heterogeneity in the immune microenvironment of GBM and identifies new markers for immunosuppression. Characterization of diverse immune responses will facilitate patient stratification and improve personalized immunotherapy in the future. This study utilizes a computational approach to characterize the immune environments in glioblastoma and shows that glioblastoma immune microenvironments can be classified into three major subgroups, which are linked to typical glioblastoma alterations such as IDH mutation, NF1 inactivation, and CDK4-MARCH9 locus amplification. http://cancerres.aacrjournals.org/content/canres/78/19/5574/F1.large.jpg .

No abstract

<p>Tables contain statistically significant associations between cluster activities and mutations or copy number variations in genes.</p>

<p>Table contains immune response related gene clusters identified in clustering analysis and used in other analyses.</p>

<p>Table contains copy number variation of genes adjacent to CDK4 (AGAP2, MARCH9, TSPAN31) in samples with CDK4 amplification.</p>