Computational Identification of Repurposable Drug Candidates for Psoriatic Arthritis via Single-Cell RNA Sequencing of Synovial Tissue

This provisional patent specification describes a computational drug repurposing pipeline applied to publicly available single-cell RNA sequencing data from human psoriatic arthritis (PsA) and rheumatoid arthritis (RA) synovial tissue biopsies (GSE200815; Floudas et al., Ann Rheum Dis 2022). After quality control filtering, 48,201 cells were analyzed using scVI batch-corrected dimensionality reduction and Leiden clustering, identifying 11 synovial cell types including lining fibroblasts (PRG4+/MMP3+), sublining fibroblasts (THY1+/CXCL12+/DKK3+), SPP1+ macrophages, endothelial cells, pericytes, and lymphoid populations. Wilcoxon differential expression analysis identified 467 genes upregulated in PsA versus RA, prominently including MMP3, FN1, PRG4, VCAM1, HTRA1, AXL, DDR2, and PDGFRA. ChEMBL bioactivity screening identified 11,910 drug candidates; the top 2,000 were assessed for novelty via PubChem, all classified as NOVEL for the PsA indication. Top candidates include CHEMBL5653589 (score 79.24, targeting HSPA5/ER stress and ZMYND8 epigenetic regulation), foretinib (50.03, AXL/DDR2/PDGFRA axis), nintedanib (38.72), prinomastat (30.52, MMP1/MMP3 inhibitor), and dasatinib (29.40, DDR2/PDGFRA). CHEMBL5653589 represents a pan-disease signal independently identified across seven scRNA-seq pipelines spanning neurodegeneration and autoimmune disease. Inventors: Glen Charles Ritschel, Ritschel Research, Tega Cay SC; Claude, Anthropic, San Francisco CA.