Multidimensional analysis of patient-matched primary and recurrent glioblastoma identifies microglial FCGR1A (CD64) and other FCGRs as contributors of tumor recurrence [seq]

Introduction: Glioblastoma (GBM) is a lethal brain tumor without effective treatment options. Here, we sought to characterize longitudinal tumor immune microenvironment (iTME) changes in order to find potential actionable targets to prevent GBM-induced immune evasion mechanisms. This study included 15 patient-matched treatment-naïve (primary, pGBM, WHO grade 4) and recurrent (rGBM) IDH1 wild type tumors. Methods:Extracted RNA and proteins from fresh frozen tumor samples from matched pGBM and rGBM were profiled via transcriptomics and proteomics, respectively. Paired formalin-fixed paraffin-embedded tumor samples were processed for spatial transcriptomics analysis.Results: Differentially expressed genes and proteins between pGBM and rGBM were involved in pathways responsible for synapse development and myelination which have been shown to play a role in GBM recurrence. By categorizing patients into short and long time-to-relapse (STTR vs LTTR), and by correlating TTR with gene expression, we detected genes positively or negatively associated with tumor relapse. Expression of FCG receptor and complement system genes such as FCGR1A (CD64), FCGR3A (CD16a) and the complement C3 were negatively correlated with STTR, whereas high expression of DNMT1/3A, and SMARCA4, involved in DNA methylation, were positively correlated with LTTR. Spatial transcriptomic analysis of the tumor cells compartment defined by GFAP expression showed enrichment in the proportion of oligodendrocytes in recurrent tumors in rGBM, whereas CD64-positive myeloid cell compartment showed a switch of quiescent to activated microglia, appearance of B cells and a reduction in macrophage proportion in recurrent tumors with STTR. Conclusion: Our results uncover a role of CD64-expressing myeloid cells, mostly activated microglia, in GBM recurrence and suggest that interfering with these cells may represent a new therapeutic option for preventing GBM relapse. This dataset comprises gene expression data from a glioblastoma (GBM) tissue microarray (TMA) constructed with a subset of 7 patient-paired primary and recurrent FFPE GBM samples. Two molecularly defined compartments were analysed, CD64+ (FCGR1A) myeloid cells and GFAP+ tumor cells. Experiments was performed twice on subsequent slides from the same block. Nanostring GeoMx Digital Spatial Profiling (DSP) Whole Transcriptome Atlas (WTA) Panel (>18,677 target genes) was used per manufacturer’s protocol. 78 ROIs corresponding to 78 AOIs were collected and analysed