INSIHGT revived: Unified 3D tissue profiling for spatially resolved, multi-marker, quantitative expression scoring at single-cell resolution for fresh and archival samples.

e15098 Background: Biology is three-dimensional, yet we lack the technology to picture the structural and molecular details throughout the tumour tissue volume, undermining our efforts to understand the heterogeneous and complex 3D spatial relationships. Ideally, archived paraffin blocks should also be revivable for comprehensive 3D analysis of multiple molecular markers at single-cell resolution. Methods: We developed a foundational chemistry (termed INSIHGT) that enables highly multiplexed, deeply penetrating immunostaining in human tissue samples. We then developed a tissue revival protocol enabling INSIHGT to be used in archived samples retrieved from paraffin blocks. We further designed and constructed a light-sheet microscope that achieves can image cleared tissues (1 cm x 1 cm x 5 mm) within 6 hours at 0.5 μm isotropic resolution. We also curated a panel of markers and probes, along with deterministic image processing strategy to achieve consistent universal cell boundary detection and segmentation in normal and tumourous tissues. Results: The overall technology platform - INSIHGT-revived - allows non-destructive virtualization of any archived tissues retrived from paraffin wax. With one round of staining and imaging, 4 markers - along with nuclear and cell segmentation panel - can be visualized at the subcellular level in centimetre-sized tissues within 6 days. The staining and retrieval chemistry has been proven to be compatible with >300 off-the-shelf antibodies, including those for epigenetics and post-translational modifications. Using our custom designed panel, accurate single-cell segmentation can be achieved for nearly all epithelial, immune and stromal cell types, improving marker expression scoring by 20%, while 3D imaging also improve machine learning-based segmentation accuracy by ~30% compared to 2D images. Up to 28-plex profiling using multi-round staining can be achieved in 1mm-thick samples preserved only with conventional formalin fixation. Samples processed with our pipeline is indistinguishable from non-processed controls by pathology specialist using H&E, immunohistochemistry, bulk transcriptomics by RNA sequencing, whole genome sequencing, and spatial proteomics profiling by mass spectrometry. As a proof-of-concept, we quantified IO-related, ADC-related, prognostically relevant and biologically interesting markers in lung, colon, and breast cancer specimens. Conclusions: INSIHGT-revived can virtualize a gold mine of data from tissue archives. With extremely large data throughput reflecting ground-truth 3D tumour architectures and spatially resolved single-cell expression profiles, INSIHGT-revived can empower the discovery and clinical translation of biomarkers for patient selection for therapeutics.