Spatial Glyco-Codes Define Human Liver Pathology and Progression

Glycosylation is a relatively underexplored aspect of the central dogma of biology, yet a key determinant of biological function and encodes disease-associated cellular states. However spatial glycomics is limited primarily by workflows that are costly, specialized, or insufficiently informative for conformation-dependent motifs. To address the need for a deep, cost-effective, and portable spatial glycomics technology, we developed spatial-GPT, a multimodal lectin-based platform for the simultaneous profiling of glycans, proteins, and transcripts from same-slide archival FFPE tissues. By combining DBiT-GPT sequencing with CODEX-GP imaging, spatial-GPT provides robust glycan detection in long-stored specimens, cost-effective multiplexing, and ascription of glycan motifs to cellular identity, pathology, and gene-regulatory programs at subcellular spatial resolution. Applied to human liver disease, spatial-GPT resolved glyco-codes of steatosis, fibrosis, cirrhosis, and hepatocellular carcinoma (HCC) subtypes, identified tumor-like glycan remodeling in premalignant regions, revealed conserved HCC-associated glycan programs, and uncovered glycan-defined immune and stromal neighborhoods across tissue microarrays. Spatial-GPT offers a practical extension of pathology, unlocking the glycan dimension on the same tissue section to expose a granularity that may be otherwise obscured by morphology, protein markers, or transcriptomics alone.