P122 Digital spatial profiling reveals the predictive characteristics in patients with immune checkpoint inhibitor-induced colitis

Abstract Background The increasing use of immune checkpoint inhibitors (ICIs) against cancers is inevitably accompanied with immune-related adverse events (irAEs), and irAE colitis is one of the common types. The overactivation of CD8+ T cells plays a central role in irAEs. However, the local mechanisms related to irAE colitis remain unclear. Hence, we aimed to explore the transcriptional characteristics related to CD8+ T cell responses of irAE colitis. Methods Colon biopsy samples were collected from 6 irAE colitis patients and 5 non-irAE controls after ICIs treatment. Leveraging digital spatial profiling, 23 regions of interest (ROI) enriched with CD8+ T cells were selected and divided into 3 groups, which were non-irAE, pre-treatment irAE and post-treatment irAE. Differentially expressed genes (DEGs), protein-protein interaction (PPI) and weighted gene co-expression network analysis (WGCNA) were performed to identify the hub genes, and they were further verified by external single-cell RNA-seq data. Also, the immune infiltration analysis was performed (Figure 1A). Results Principal component analysis (PCA) indicated similar distribution among the 3 groups (Figure 1B). The comparisons of non-irAE vs. pre-treatment irAE, and pre-treatment irAE vs. post-treatment irAE yielded 105 and 210 DEGs respectively. WGCNA resulted in 14 clustering modules, with MEturquoise and MEmagenta exhibiting the highest correlation scores (Figure 1C). Importantly, 174 key genes were identified and prominently associated with various mitochondrial processes e.g. oxidative phosphorylation (Figure 1D). PPI analysis revealed a sub-module consisting of 22 nodes in MCODE. Further overlap analysis of CytoHubba and DEGs pinpointed 9 hub genes. Remarkably, COX5B was found at the intersection of both groups (Figure 1E). External single-cell RNA-seq data verified that all the hub genes were significantly highly expressed in the irAE colitis group except COX6A1 (Figure 1F). Besides, immune infiltration analyses revealed the higher levels of endothelial cells (P = 0.043) and fibroblasts (P <0.01) by SpatialDecon, and activated dendritic cells (P <0.001) by xCell in pre-treatment irAE, compared to non-irAE (Figure 1G). Further, immune cells correlation analysis demonstrated highly similar patterns of endothelial cells, fibroblasts, pDCs, and macrophages (Figure 1H). Conclusion COX5B related to mitochondrial processes characterized irAE colitis. Expansion of dendritic cells and stromal compartments including endothelial cells and fibroblasts may contribute to irAE development.