Causal association and mechanistic validation of the USP8‐DLK1 axis in osteoarthritis: A genetic meta‐analysis integrated with in vitro and in vivo investigations
Sayed Abdulla Jami, Md Ariful Haque, Ernest Dapaah, Vidmi Taolam Martin, Md Tofazzale HosainAbstract
Background/Objective
Osteoarthritis (OA) affects more than 500 million people globally and remains without effective disease‐modifying therapy, owing in part to an incomplete understanding of the molecular mechanisms driving chondrocyte degeneration. Although genetic studies have implicated ubiquitin‐specific peptidase 8 (USP8) and delta‐like 1 (DLK1) in OA susceptibility, their causal relevance and precise mechanistic relationship—particularly through the ferroptosis pathway—have not been established.
Methods
We performed two‐sample Mendelian randomization (MR) using protein quantitative trait locus (pQTL) instruments from deCODE Genetics and expression QTL (eQTL) instruments from eQTLGen, with genome‐wide association summary statistics from the Genetics of Osteoarthritis (GO) Consortium ( n = 826 690) as outcomes. Causal inference was supplemented by Bayesian colocalization (coloc v5.2.3) and Summary‐data‐based Mendelian Randomization / Heterogeneity in Dependent Instruments (SMR/HEIDI) analyses. Functional validation employed IL‐1β‐ or erastin‐stimulated primary human chondrocytes with USP8 overexpression or siRNA‐mediated knockdown, and in vivo validation used the destabilization of the medial meniscus (DMM) mouse model with intra‐articular delivery of Adeno‐Associated Virus (AAV)‐USP8 or USP8‐siRNA.
Results
Genetically predicted USP8 protein levels were associated with lower knee OA risk (inverse variance‐weighted odds ratio [OR] = 0.83, 95% CI: 0.76–0.91, p = 1.2 × 10 − 5 ) and hip OA risk (OR = 0.87, 95% CI: 0.81–0.93, p = 3.5 × 10 − 5 ). Bayesian colocalization supported a shared causal variant for USP8 and knee OA at 11p15.1 (PP.H4 = 0.82). In chondrocytes, USP8 physically interacted with and deubiquitinated DLK1, prolonging its half‐life from approximately 4 h to more than 8 h. USP8 overexpression restored glutathione peroxidase 4 (GPX4), reduced 4‐hydroxynonenal (4‐HNE) and MMP13, and recovered COL2A1 under inflammatory and ferroptotic stress; USP8 knockdown produced directionally opposite effects. In vivo, intra‐articular AAV‐USP8 reduced OARSI scores from 4.2 ± 0.6 to 1.5 ± 0.4 ( p < .001), preserved cartilage thickness, and attenuated ferroptosis‐associated 4‐HNE accumulation, while USP8‐siRNA worsened all structural and molecular endpoints.
Conclusion
USP8 protects against OA‐associated chondrocyte injury through a DLK1–GPX4–ferroptosis suppression axis. Early intra‐articular AAV‐USP8 delivery preserves cartilage structure in a preclinical model, but late administration does not reverse established cartilage loss, positioning USP8 as an early disease‐modifying rather than regenerative target.