Spatial transcriptomics landscape of livers from patients with metabolic dysfunction-associated steatohepatitis

Metabolic dysfunction-associated steatohepatitis (MASH) progression involves heterogeneous responses and coordinated interactions among parenchymal and nonparenchymal cells. Single-cell transcriptomics has advanced our understanding of cellular heterogeneity, while spatial omics maps molecular signatures within tissue architecture, enabling inference of cell–cell and ligand–receptor interactions, particularly those involving macrophages in MASH. Here, we aimed to characterize spatially resolved transcriptomic alterations in hepatocyte-, hepatic stellate cell (HSC)-, and immune cell-enriched regions in individuals with obesity. We further sought to validate these signatures using public single-cell level RNA sequencing datasets and protein analyses and assess their clinical associations and changes following bariatric surgery. We performed spatial transcriptomic profiling of human liver tissues using the GeoMx Digital Spatial Profiler in 16 patients with MASH across fibrosis stages and eight NonMASH controls. Regions of interest (ROIs) enriched for hepatocytes, immune cells, and HSCs were defined by immunofluorescence-guided pathological annotation. In addition to various bioinformatic analyses on spatial transcriptomic data, we integrated our data with several other public datasets. Finally, the study findings were further validated at the protein level. Each hepatic ROI exhibited distinct molecular alterations during MASH progression. Cytochrome P450 genes were significantly downregulated in hepatocyte-, HSC- and immune cell-enriched ROIs, and negatively correlated with the expression of nuclear receptor and pro-inflammatory genes. With advanced fibrosis in MASH, nonparenchymal ROIs showed marked mitochondrial dysfunction. Ligand‒receptor analysis revealed disrupted inter-regional communication (e.g. loss of protective signals and enhanced pro-fibrotic interactions). Notably, IL32 was consistently upregulated across all three ROI types in MASH versus NonMASH. In addition to validation using public single-cell level datasets, we identified 19 MASH-associated genes, whose expression was correlated with clinical and histologic parameters and reversed following bariatric surgery. Among the key genes identified in our analysis, we selectively validated the hepatic expression of CYP1A2, IL32, and S100A8/9 at the protein level in patients with MASH. Our results reveal compartment-specific metabolic and inflammatory reprogramming in human MASH, highlighting CYP450 suppression, immune activation, disrupted intercellular signaling, and reversibility after bariatric surgery, thereby providing therapeutic implications.