169P - Tumour microenvironment in glioblastoma: A single-cell perspective on macrophage dynamics and T cell exhaustion

Background

The tumour microenvironment (TME) in glioblastoma includes tumour-associated macrophages (TAMs), with microglia-derived (Mig-TAMs) being pro-inflammatory and monocyte-derived (Mon-TAMs) being tolerogenic (Atunes et al, 2021). This study uses single-cell RNA sequencing (scRNAseq) to explore TAM dynamics and T cell exhaustion in newly diagnosed (ndGBM) and recurrent GBM (rGBM), aiming to uncover distinct TME profiles.

Methods

We analysed scRNA datasets from two studies in the EMBL-EBI database, encompassing ndGBM and rGBM. TAM clusters were reproduced by the Louvain algorithm and annotation by Seurat, followed by manual curation of TAM subsets. We compared Mig-TAMs and Mon-TAMs by their phenotypes and proportions of the overall TAM cluster (using normalised mean cell counts). We correlated the TAM subsets with T cell exhaustion and interferon gene signature.

Results

Our analysis included seven patients (ndGBM=3, rGBM=4), examining 47,965 cells and 23,944 genes. We identified 19 Louvain clusters. The TAM cluster comprised two subsets: Mon-TAMs (markers: LGALS1+, S100A10+, MIF1+) and Mig-TAMs (markers: CD74+, C1QB+, C1QC+). In ndGBM, the mean cell count ratio of Mon-TAMs to total TAM was 0.45 (95% CI, 0.42–0.87), and Mig-TAMs was 0.54 (95% CI, 0.12–0.95, t(2)=4.73, p=0.042). In rGBM, Mon-TAMs had a ratio of 0.65 (95% CI, 0.16–1.19), and Mig-TAMs was 0.34 (95% CI, -0.19–0.88, t(3)=3.87, p=0.03). In the recurrent GBM samples (rGBM), T cell exhaustion markers (IL2RA, ICOS, CTLA4, TIGIT) were significantly more expressed, chemokine and interferon gene signatures were significantly reduced and we observed a predominance of the Mon-TAM cluster compared to the Mig-TAM cluster. Interestingly, in the treatment naïve group (ndGBM), we observed no signals of T cell exhaustion and high Mig-TAM cluster prevalence compared to the Mon-TAM cluster. In concordance, we observed increased interferon gamma gene signature and increased CXCL10 expression (≥50%), which are known predictors for immunotherapy response.

Conclusions

The immunoactive Mig-TAM populations and gene expression signatures in treatment-naive GBM could facilitate translational hypothesis and patient enrichment for future immunotherapy studies in GBM.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.