A multi-component view of the glioma transcriptome identifies unique immune infiltration patterns in primary glioblastomas and patients with inferior prognosis.

Patients diagnosed with glioblastomas continue to have a dismal prognosis, highlighting the need for a better characterization of these tumors in order to identify more efficient therapies. In this study, we generated genome-wide expression data from normal brain and glioma samples, representing major glioma subtypes, using exon-level microarrays (n=70). A multi-component approach was used to characterize molecular pathways and tumor infiltrate, with particular focus on the poor prognosis primary glioblastomas (pGBM). Two independent, publically available datasets were used for validation of survival data. Grade II glioma and secondary GBM were enriched in cytotoxic lymphocytes, while pGBM exhibited a heterogeneous cell-related immune infiltrate with high infiltration of monocytic cells. Infiltration by cells of monocytic origin predicted poorer prognosis, regardless of the tumor subtype. High TSPYL2 expression was found to be associated with a poor prognosis in pGBM and to correlate with genes involved in tumor invasion. The multi-component transcriptome phenotypes of pGBM introduced in this study add insights into the main pathways associated with aggressiveness, and identified infiltration by cells of monocytic origin as a universal prognostic marker for all glioma subtypes.

70