Abstract 773: High-throughput in situ multiomic profiling of kidney, lung, colon, and breast cancer FFPE samples for profiling of tumor-immune interactions

Abstract Molecular level spatial profiling of tissue provides crucial insights into tumor biology and immunotherapeutic responses. The G4X Spatial Sequencer, our innovative platform, facilitates extensive translational studies by simultaneously detecting RNA transcripts, proteins, and H&E staining in FFPE tissues. Capable of imaging up to 40 cm2 of tissue in a single run, it significantly enhances throughput compared to traditional methods and enables 3D reconstruction from serial sections in one run. We designed ∼300-gene transcript panels for each of four tissue types: kidney, lung, colon, and breast, with an emphasis on immuno-oncology relevant targets. The panels share a common set of 150 immuno-oncology and 50 stromal genes. Each panel also includes a set of ∼100 tissue-specific genes to profile cell types, cancer pathways and tumor microenvironment. We imaged gene and protein expression in 5-μm thick sections of human kidney, lung, colon, and breast, from both healthy and cancer FFPE tissue samples. Using a novel tissue transfer process, we arrayed multiple tissue sections in two formats: a large sample format (10 mm x 10 mm) with 10 sections/slide, and a small sample format (4.5 mm x 4.5 mm) with 32 sections/slide. Each slide was analyzed in three modalities (RNA, proteins, fluorescent H&E). Following antibody staining, probe binding and amplification of targets, images were acquired using 4-color SBS readout, followed by nuclear and cytoplasmic staining to generate a fluorescence-based H&E stain. In FFPE cancer samples and healthy controls, we detected an average of 50-200 transcripts/cell and 20-50 unique genes/cell, with the dynamic range extending to > 500 transcripts/cell. Using 6 rounds of priming, we achieved high transcript density profiling while maintaining a false discovery rate of ∼1%. Strong correlation between single-cell transcript and protein expression profiles confirmed specificity of each modality. Comparison across serial sections showed strong concordance in expression profiles for both modalities, demonstrating robust reproducibility. We also show the assembly of serial sections to reconstruct a 3D tumor volume. Finally, we identified distinct tumor-associated cellular phenotypes and spatial neighborhoods using a novel multi-modal graph-based clustering method leveraging RNA, protein, morphology, and spatial information. Citation Format: Kenneth H. Gouin, Michael Lawson, Andrew Pawlowski, Sabrina Shore, Yuji Ishitsuka, Daan Witters, Eli N. Glezer. High-throughput in situ multiomic profiling of kidney, lung, colon, and breast cancer FFPE samples for profiling of tumor-immune interactions [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 773.