An FDA‑approved dicarboxylic acid allosterically targets PNPLA3 to resolve osteoarthritis

Osteoarthritis (OA) lacks disease‑modifying therapies. The infrapatellar fat pad (IPFP) is hypothesized to regulate joint metabolism, but its molecular effectors are unknown. We investigated whether IPFP‑derived metabolites protect cartilage via lipid‑metabolizing enzymes.

Non‑targeted metabolomics, IPFP excision/transplantation, single‑cell RNA‑seq, molecular dynamics, enzyme kinetics, and AAV‑mediated gene silencing were used in patient samples and the DMM mouse model.

Azelaic acid (AA) was the most significantly reduced metabolite in OA IPFP, and its levels correlated inversely with KL grade and WOMAC scores. IPFP excision substantially depletes joint AA (to ~8% of sham levels) and aggravates cartilage damage, while AA supplementation rescued the phenotype. Single‑cell sequencing identified a subset of metabolic chondrocytes (MetCs) that upregulate PNPLA3 upon IPFP loss. AA directly bound PNPLA3 with high affinity (Kd = 4.9μM) at a previously unrecognized allosteric site (>15Å from the catalytic triad), causing non-competitive inhibition (Ki = 12.4μM). Selectivity profiling against a panel of seven lipases, 50 kinases, and 50 GPCRs confirmed high selectivity for PNPLA3 within the tested panels (selectivity index >40). This inhibition was conserved in human primary OA chondrocytes and abolished by K42A mutation. PNPLA3 knockdown by AAV‑shPNPLA3 ameliorated OA, but its effect was lost in IPFP‑deficient mice, establishing an IPFP-AA-PNPLA3 axis.

IPFP‑derived AA protects against OA by allosterically and selectively inhibiting PNPLA3‑driven lipolysis in MetCs. Because AA is already FDA‑approved for topical use, intra‑articular repurposing is immediately testable. Preserving IPFP integrity or delivering AA offers a first‑in‑class metabolism‑targeted therapy for OA.