COX2-independent and COX2-dependent effects of naproxen on bone quality, osteocytes, and fatigue fracture healing in male and female mice

Numerous pre-clinical and clinical studies link regular non-steroidal anti-inflammatory drug (NSAID) use to an increased risk of skeletal stress fracture, but the causative mechanisms remain poorly understood. Here, we used a mouse model of cyclooxygenase 2 (COX2) inhibition, the Ptgs2-Y385F mouse, to identify effects of the NSAID naproxen on bone that are independent of or dependent on COX2 activity. We hypothesized that naproxen would decrease strain adaptive bone remodeling in a COX2-dependent manner, but would decrease bone toughness and fracture resistance through COX2-independent effects. We first used microCT analysis and mechanical testing to compare bones from Ptgs2-Y385F mice to those of wild type (WT) littermates, revealing few skeletal differences due to the genetic mutation. Next, WT and Ptgs2-Y385F mice were subjected to six bouts of non-damaging uniaxial forelimb compression over two weeks while receiving vehicle or naproxen drinking water. Here, naproxen decreased load-induced bone formation in WT, but not Ptgs2-Y385F mice, indicating a COX2-dependent effect. Contrastingly, naproxen reduced bone toughness and post-yield deformation across genotypes compared to vehicle treatment, supporting a COX2-independent mechanism influencing fracture resistance. The effects of naproxen were consistent across different doses and were not dose-dependent within the administered range. Additionally, histological analysis showed that naproxen, compared to vehicle treatment, increased empty osteocyte lacunae number across genotypes while increasing osteoblast number, total osteocyte lacunae number, and fluorophore perilacunar labeling in WT mice only. These findings suggest that naproxen acts on bone cells through both COX2-independent and -dependent mechanisms. Separately, mice were also pre-treated with vehicle or naproxen for 15 days before fatigue fracture induction by axial forelimb compression, then received further treatment for 15 days as limbs healed. Naproxen treatment resulted in mild, but sex-specific, effects on fracture formation and healing in both genotypes compared to vehicle. Further analysis of osteocyte dendritic networks in uninjured femurs from these same mice also revealed sexually dimorphic changes in dendrite number and density in naproxen treated mice as compared to vehicle, correlating with previously observed bone toughness differences. Together, these results demonstrate that naproxen influences bone through both COX2-dependent and -independent mechanisms and displays sex-dependent differences. These findings also suggest that regular COX2-selective NSAID use can have skeletal consequences, thus safe pain management methods for individuals with higher fracture risk must be identified.