We investigated immune-related long non-coding RNAs (lncRNAs) that may be exploited as potential therapeutic targets in anaplastic gliomas. We obtained 572 lncRNAs and 317 immune genes from the Chinese Glioma Genome Atlas microarray and constructed immune-related lncRNAs co-expression networks to identify immune-related lncRNAs. Two additional datasets (GSE16011, REMBRANDT) were used for validation. Gene set enrichment analysis and principal component analysis were used for functional annotation. Immune-lncRNAs co-expression networks were constructed. Nine immune-related lncRNAs (SNHG8, PGM5-AS1, ST20-AS1, LINC00937, AGAP2-AS1, MIR155HG, TUG1, MAPKAPK5-AS1, and HCG18) signature was identified in patients with anaplastic gliomas. Patients in the low-risk group showed longer overall survival (OS) and progression-free survival than those in the high-risk group (P < 0.0001; P < 0.0001). Additionally, patients in the high-risk group displayed no-deletion of chromosomal arms 1p and/or 19q, isocitrate dehydrogenase wild-type, classical and mesenchymal TCGA subtype, G3 CGGA subtype, and lower Karnofsky performance score (KPS). Moreover, the signature was an independent factor and was significantly associated with the OS (P = 0.000, hazard ratio (HR) = 1.434). These findings were further validated in two additional datasets (GSE16011, REMBRANDT). Low-risk and high-risk groups displayed different immune status based on principal components analysis. Our results showed that the nine immune-related lncRNAs signature has prognostic value for anaplastic gliomas.

The Glioblastoma (GBM) immune microenvironment plays a critical role in tumor development, progression, and prognosis. A comprehensive understanding of the intricate milieu and its interactions remains unclear, and single-cell analysis is crucially needed. Leveraging mass cytometry (CyTOF), we analyzed immunocytes from 13 initial and three recurrent GBM samples and their matched peripheral blood mononuclear cells (pPBMCs). Using a panel of 30 markers, we provide a high-dimensional view of the complex GBM immune microenvironment. Hematoxylin and eosin staining and polychromatic immunofluorescence were used for verification of the key findings. In the initial and recurrent GBMs, glioma-associated microglia/macrophages (GAMs) constituted 59.05 and 27.87% of the immunocytes, respectively; programmed cell death-ligand 1 (PD-L1), T cell immunoglobulin domain and mucin domain-3 (TIM-3), lymphocyte activation gene-3 (LAG-3), interleukin-10 (IL-10) and transforming growth factor-β (TGFβ) demonstrated different expression levels in the GAMs among the patients. GAMs could be subdivided into different subgroups with different phenotypes. Both the exhausted T cell and regulatory T (Treg) cell percentages were significantly higher in tumors than in pPBMCs. The natural killer (NK) cells that infiltrated into the tumor lesions expressed higher levels of CXC chemokine receptor 3 (CXCR3), as these cells expressed lower levels of interferon-γ (IFNγ). The immune microenvironment in the initial and recurrent GBMs displayed similar suppressive changes. Our study confirmed that GAMs, as the dominant infiltrating immunocytes, present great inter-and intratumoral heterogeneity and that GAMs, increased exhausted T cells, infiltrating Tregs, and nonfunctional NK cells contribute to local immune suppressive characteristics. Recurrent GBMs share similar immune signatures with the initial GBMs except the proportion of GAMs decreases.

Anaplastic glioma is Grade III and the median overall survival is about 37.6 months. However, there are still other factors that affect the prognosis for anaplastic glioma patients due to variable overall survival. So we screened four-lncRNA signature (AGAP2-AS1, TPT1-AS1, LINC01198 and MIR155HG) from the lncRNA expression profile from the GSE16011, CGGA and REMBRANDT datasets. The patients in low risk group had longer overall survival than high risk group (median OS 2208.25 vs. 591.30 days; P < 0.0001). Moreover, patients in the low risk group showed similar overall survival to Grade II patients (P = 0.1669), while the high risk group showed significant different to Grade IV (P = 0.0005) with similar trend. So based on the four-lncRNA, the anaplastic gliomas could be divided into grade II-like and grade IV-like groups. On the multivariate analysis, it showed the signature was an independent prognostic factor (P = 0.000). The expression of four lncRNAs in different grades showed that AGAP2-AS1, LINC01198 and MIR155HG were increased with tumor grade, while TPT1-AS1 was decreased. Knockdown of AGAP2-AS1 can inhibit the cell proliferation, migration and invasion, while increase the apoptosis cell rates in vitro. In conclusion, our results showed that the four-lncRNA signature has prognostic value for anaplastic glioma. Moreover, clinicians should conduct corresponding therapies to achieve best treatment with less side effects for two groups patients.