Abstract 758: Spatial profiling of matched primary and recurrent glioblastoma samples reveals changes in cellular states and the immune microenvironment

Abstract Background: Glioblastoma (GBM), the most aggressive brain tumor, is treated with surgery, chemotherapy, and radiation, yet 90% recur within two years. PD-1 blockade improves outcomes in other cancers, but shows limited effect in GBM, likely due to tumor heterogeneity and an immunosuppressive microenvironment. This study aims to investigate post-treatment transcriptional changes in tumor cells and tumor-associated macrophages (TAMs) in matched primary and recurrent GBM, focusing on neoadjuvant PD-1 blockade. Methods: We used the GeoMx Digital Spatial Profiler to spatially select and transcriptomically profile tumor cells (SOX2+) and TAMs (IBA1+). Gene expression changes between primary and recurrent GBMs were analyzed in 30 patients (26 IDHwt, 4 IDHm) with 20 paired primary samples. All patients received standard treatment, and 20 also received second-line nivolumab one week before recurrence surgery. Tumor purity was confirmed using InferCNV and cell signatures. Subsequently, differential expression and pathway enrichment analyses were performed to compare recurrent tumors treated with nivolumab to those untreated, as well as primary vs. recurrent tumors. We also evaluated Neftel signatures in matched tumors to assess treatment- and recurrence-associated malignant state changes. Results: Principal component analysis of all samples revealed distinct transcriptional profiles between tumor cells and TAMs. Separate analyses showed clear differences in tumor cell and TAM segments based on tumor status (primary/recurrent), though no separation related to nivolumab treatment. Similarly, comparing nivolumab-treated recurrent tumors with untreated controls showed no significant changes in malignant states or gene expression in tumor cells or TAMs, consistent with the lack of clinical survival benefit.To investigate overall transcriptional changes induced by standard treatment, differential expression analysis comparing tumor cell segments in paired primary and recurrent tumors revealed few significant genes, likely reflecting high inter-patient variability driven by tumor heterogeneity. Unsupervised clustering of tumor cell segments identified a distinct cluster specific to recurrent tumors, with DEGs highlighting pathways linked to this particular tumor cell phenotype. In TAM segments, increased expression of checkpoint-related genes in recurrent tumors suggests TAM phenotype changes that may influence immune dynamics in the microenvironment. Conclusion: Spatial profiling of matched primary and recurrent GBM samples reveals distinct transcriptional changes in both tumor cells and TAMs, reflecting the complex microenvironment of recurrent GBM. However, our findings point to a limited impact of neoadjuvant PD-1 blockade on transcriptional activity, highlighting ongoing challenges in targeting GBM with immunotherapy. Citation Format: Sara Artzi, Marc N. Klausen, Dylan S. Harwood, Signe R. Michaelsen, Simone B. Maarup, Alessio Locallo, Nicolai S. Bager, Nadine M. Hammouda, Benedikte Hasselbalch, Ulrik Lassen, Joachim Weischenfeldt, Bjarne W. Kristensen. Spatial profiling of matched primary and recurrent glioblastoma samples reveals changes in cellular states and the immune microenvironment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 758.