Metabolic reprogramming of synovial fibroblasts in osteoarthritis by inhibition of pathologically overexpressed pyruvate dehydrogenase kinases

Osteoarthritis (OA) is the most common degenerative joint disease and a major cause of age-related disability worldwide, mainly due to pain, the disease's main symptom. Although OA was initially classified as a non-inflammatory joint disease, recent attention has been drawn to the importance of synovitis and fibroblast-like synoviocytes (FLS) in the pathogenesis of OA. FLS can be divided into two major populations: thymus cell antigen 1 (THY1)- FLS are currently classified as quiescent cells and assumed to destroy bone and cartilage, whereas THY1+ FLS are invasively proliferative cells that drive synovitis. Both THY1- and THY1+ FLS share many characteristics with fibroblast-like progenitors - mesenchymal stromal cells (MSC). However, it remains unclear whether synovitis-induced metabolic changes exist in FLS from OA patients and whether metabolic differences may provide a mechanistic basis for the identification of approaches to precisely convert the pathologically proliferative synovitis-driven FLS phenotype into a healthy one. To identify novel pathological mechanisms of the perpetuation and manifestation of OA, we analyzed metabolic, proteomic, and functional characteristics of THY1+ FLS from patients with OA. Proteome data and pathway analysis revealed that an elevated expression of pyruvate dehydrogenase kinase (PDK) 3 was characteristic of proliferative THY1+ FLS from patients with OA. These FLS also had the highest podoplanin (PDPN) expression and localized to the sublining but also the lining layer in OA synovium in contrast to the synovium of ligament trauma patients. Inhibition of PDKs reprogrammed metabolism from glycolysis towards oxidative phosphorylation and reduced FLS proliferation and inflammatory cytokine secretion. This study provides new mechanistic insights into the importance of FLS metabolism in the pathogenesis of OA. Given the selective overexpression of PDK3 in OA synovium and its restricted distribution in synovial tissue from ligament trauma patients and MSC, PDKs may represent attractive selective metabolic targets for OA treatment. Moreover, targeting PDKs does not affect cells in a homeostatic, oxidative state. Our data provide an evidence-based rationale for the idea that inhibition of PDKs could restore the healthy THY1+ FLS phenotype. This approach may mitigate the progression of OA and thereby fundamentally change the clinical management of OA from the treatment of symptoms to addressing causes.