K. Guion1, B. Rocque1, P. Singh1, S. Bangerth1, L. Pickard1, J. Bhattacharjee2, S. Eguizabal1, C. Weaver3, S. Chopra4, S. Zhou5, R. Kohli2, L. Sher1, B. Ekser6, J. Emamaullee1,3,7 1University Of Southern California, Division Of Abdominal Organ Transplantation And Hepatobiliary Surgery, Department Of Surgery, Keck School Of Medicine, Los Angeles, CA, USA 2Children’s Hospital Los Angeles, Division Of Gastroenterology, Hepatology And Nutrition, Los Angeles, CA, USA 3Children’s Hospital Los Angeles, Division Of Abdominal Organ Transplantation, Los Angeles, CA, USA 4University Of Southern California, Department Of Pathology, Keck School Of Medicine, Los Angeles, CA, USA 5Children’s Hospital Los Angeles, Department Of Pathology And Laboratory Medicine, Los Angeles, CA, USA 6Indiana University School Of Medicine, Division Of Transplant Surgery, Department Of Surgery, Indianapolis, IN, USA 7University Of Southern California, Department Of Molecular Microbiology And Immunology, Keck School Of Medicine, Los Angeles, CA, USA
Introduction:
While policy changes have reduced waitlist mortality and improved post-transplant survival in pediatric acute liver failure, Biliary Atresia (BA) remains the primary cause of childhood end-stage liver disease. Epithelial injury, flawed embryogenesis, environmental toxins, and viruses, all may contribute to BA's complex immune landscape that is not yet fully understood. To tackle this, we employed spatial transcriptomics to compare healthy and fibrotic liver tissue. Our goal was to uncover gene expression patterns specific to BA, while also demonstrating a robust data analysis technique in the emerging field of spatial transcriptomics.
Methods:
Liver specimens from patients with BA with advanced fibrosis and normal liver were obtained. Spatial transcriptomics and single nucleus RNA sequencing (snRNASeq) were performed. Cell-cell interactions were predicted from Ligand-Receptor Analysis from snRNASeq. Spatial transcriptomic data was then deconvoluted with snRNASeq to generate single-cell resolution spatial data. Differential gene expression pathways and cell-cell interactions associated with BA were derived from atlas comparison.
Results:
Deconvolution of the spatial transcriptome using paired snRNASeq data generated a spatially resolved, single cell dataset of 24 unique liver cell phenotypes. BA tissue had increased cholangiocytes, liver sinusoidal endothelial cells (LSEC) zone 1, T/NK cells, B/Plasma cells, hepatic stellate cells (HSC)/fibroblasts, and portal endothelial cells when compared to normal tissue. When compared to microanatomy annotations, the increased immune populations in BA were concentrated in the periportal regions, while these same populations were more spatially diffuse in the normal liver. BA tissue also had decreased epithelial progenitor cells, LSEC zones 2&3, and hepatocyte progenitor cells. Further analysis determined that BA was characterized by a loss in cell-cell interaction complexity and dysregulation of FGF23-FGFR1 signaling, and loss of LAMC3- CD44 interactions between hepatic stellate cells and Kupffer cells.
Conclusion:
BA with advanced fibrosis has a quantitative increase in immune cells that are spatially grouped near the portal triad. LSEC zones 2&3 had a decrease in progenitor cells and dysregulation of cellular signaling. This was characterized by a loss in cell-cell interaction complexity and dysregulation of FGF23-FGFR1 signaling, and a loss of LAMC3-CD44 interactions between hepatic stellate cells and Kupffer cells. In addition, deconvolution of spatial transcriptomic data with matched snRNASeq data enables creation of a spatially resolved, single cell atlas to study gene expression and cell-cell interactions in biobanked clinical samples with advanced liver disease.
