Spatially Defining the ADPKD Transcriptome

Background: PKD is a common inheritable disease of limited treatment. Robust characterization of human disease may define the faithfulness and utility of animal and cell-based models, focusing efforts on clinically relevant pathomechanisms to improve translation and drug discovery. By dissociating human samples, ADPKD kidney has been transcriptionally characterized with unprecedented depth, yet dissociation eliminates cyst-defining morphology. An absence of well-characterized anchor genes that specify cystic epithelia, and only cystic epithelia, limits a retrograde single cell analysis. Here spatial transcriptomics (ST) morphologically and transcriptionally defines cysts, delineates cystic anchor genes, and discerns ligand-receptor engagement of cystic epithelia. Methods: 28yo with ADPKD due to PKD1 mutation (p.Thr3223_Glu3231del) underwent nephrectomy with a cortical FFPE block sectioned, deparaffinized, and 10X Genomics Visium CytAssist ST performed. 5mm x 5mm of cortex was mounted onto a Visium V4 Slide divided into Spots, each targeting 18,085 genes in triplicate with 53,504 probes (Visium Human Transcriptome Probe Set v2.0). Sequencing by Spaceranger 3.1.2, normalization by Seurat’s SingleCellTransform pipeline, DEGs by Wilcoxon Rank Sum Test, cell ID by Seurat anchor-based label transfer, pathways by GSEA, ligand-receptors by Cell chat, visualization by Loupe browser. Results: In ADPKD, 4435 Spots had mean 38364 reads, and ARPKD had 4578 Spots with mean 38621 reads. In ADPKD, anchor-based label transfer segregated 8 clusters, validated by H&E superimposition to visualize tubules, glomeruli, interstitium, vasculature, and fibrotic areas. PC/CNT was enriched in cyst lining and had tope 5 DEGs including AQP2, SCNN1G, CALB1. Ligand-receptor analysis associates cystic epithelia as the sender of autocrine and paracrine WNT9B-LRP5/6 signaling and of COL4A5, COL4A6, and COL9A2 binding with localized integrins. PC/CNT subclustering to 5 clusters (3.1 – 3.5), shows 3.3 highly localized to cysts and has top DEGs of GLIS2, MSC, SMIM5, NR4A3, CDK2AP2, HMGCS2, and VTCN1 with MSC and SMIM5 as leading putative anchor genes. Conclusion: Spatial transcriptomics of ADPKD kidney morphologically defines cystic epithelia to identify MSC and SMIM5 as putative anchor genes. Autocrine and paracrine signaling of WNT9B and the localized attachment of COL4A5, COL4A6, and COL9A2 of cystic epithelia may be localized druggable targets.