Discovery and functional analysis of osteoarthritis susceptibility genes using families from a large population-based cohort

      Purpose: The main obstacle to the development of disease-modifying therapies for OA is poor understanding of the disease process: we do not know the cell types or molecular pathways that normally function to limit susceptibility to OA. Our goal is to discover molecular pathways that are vulnerability points for the development of OA and generate mouse models using human disease alleles. We discover these pathways by identifying rare gene mutations in coding sequences that have a strong effect on susceptibility to OA in families identified from a unique statewide population-based cohort, the Utah Population Database. From the family studies, we identify gene variants that we view as candidates for having a strong effect on OA susceptibility. To rapidly identify those candidate variants that measurably alter gene function, we quantitatively assay WT and variant gene functions in zebrafish embryos or tissue culture. We then determine whether and how the functionally altered genes confer susceptibility to OA in genetically modified mouse models.
      Methods: We take an approach used infrequently in the OA field. We study many unrelated families with clear-cut inherited forms of OA to identify susceptibility alleles that have strong determinate effects. To discover pathways that when modified lead to strong and unambiguous susceptibility to OA, we analyzed the exomes of >50 unrelated families from the Utah Population Database that had a significant enrichment of dominantly inherited shoulder OA, distal and proximal interphalangeal joint hand OA, erosive hand OA, and 1st metatarsophalangeal (MTP) joint OA. We use loss- and gain-of-function analyses in zebrafish to rapidly screen for gene variants that have a functional impact on protein activity. Finally, Crispr/Cas9 technology is used to introduce OA-associated disease alleles into the mouse where we analyze the how these mutations affect onset and progression of OA.
      Results: Our results indicate that we can use families to identify genes and pathways that confer strong susceptibility to the early stages of OA. Our approach has led to the discovery of candidate gene variants in families with different forms of OA that affect common pathways. These pathways include genes that regulate 1) the proinflammatory response, 2) mechanotransduction/mechanosensation, 3) the TGF-b and Notch signaling pathways, 3) transcription modulators, and 4) ECM components and ECM modification. We have discovered variants that are worth pursuing experimentally as they coincide with loci detected in GWAS, affect genes of related biological function, represent multiple alleles of a candidate gene, and identify genes not previously linked to OA.
      We have identified mutations in the NOD-RIPK2 pathway in families with interphalangeal joint, erosive hand, and 1st MTP joint OA. The NOD-RIPK2 pathway is a central regulator of the proinflammatory response. We have shown the OA-associated variants have increased proinflammatory activity relative to the wildtype protein. Based on these findings, we propose the gene variants act dominantly as a gain-of-function allele to over-stimulate the inflammatory response to naturally occurring joint damage leading to OA. To test this hypothesis we introduced a single amino acid change, encoded by the RIPK2 disease allele, into an isogenic mouse. Current results indicate mice harboring the disease allele exhibit increased severity of OA, an increased systemic inflammatory response, and altered gene expression when compared to controls. We have generated a conditional disease allele to test whether expression of the RIPK2 disease allele in innate immune cells or chondrocytes is sufficient to confer increased susceptibility to experimentally induced or spontaneous OA in the mouse.
      Erosive hand OA is characterized radiographically by centrally located interphalangeal joint erosions, and clinically by its rapid onset and progression. Genomic analysis of one family identified a rare coding variant in the procollagen galactosyltransferase COLGALT1. Loss- and gain-of-function analyses in the zebrafish indicate that colgalt1a/1b are necessary for cartilage formation. We are also examining if collagen structure is disrupted in articular cartilage of mice with a Colgalt1 mutation and if Colgalt1 mutant mice are more susceptible to injury induced OA.
      Conclusions: Our studies demonstrate our innovative approach of combining genetic analysis of families with OA and functional variant analysis in the zebrafish to identify well-supported OA-associated alleles and pathways. Our approach has led to the discovery of 1) compelling novel candidate genes, 2) multiple genes that contribute to a common pathway, 3) and identification of common pathways altered in multiple forms of OA, 4) and to the generation of new mouse models of OA. Our genetic studies of families with OA and functional analyses in the mouse support our hypothesis that variation in the inflammatory response may be a key factor in susceptibility to OA. Our work is significant in that we have taken a novel approach to identify genes that when mutated have a significant impact on the onset or progression of OA. Developing new animal models with these susceptibility genes will be useful for discovery of therapeutic drugs.