UNLABELLED: COX-2 is a key enzyme involved in the response of bone to loading. However, using mice with a null mutation of the COX-2 gene, we found that a functional COX-2 gene is not required for mechanotransduction. This paradoxical finding may have resulted, in part, from mechanically induced COX-1 activity. INTRODUCTION: Cyclooxygenase-2 (COX-2) is an important mediator in the response of bone to mechanical loading, with pharmacological inhibition of COX-2 effectively eliminating or reducing mechanically induced bone formation. In this study, we further investigated the role of COX-2 in skeletal mechanotransduction using a genetic approach. The aim was to compare the skeletal responsiveness of COX-2 homozygous mutant (COX-2(-/-)) and wildtype control (COX-2(+/+)) mice to investigate whether a functional COX-2 gene is necessary for mechanotransduction. MATERIALS AND METHODS: Adult female COX-2(+/+) and COX-2(-/-) mice on a C57BL/6x129/ola background were studied using the ulna axial loading model. The response to 2 days of loading for 120 cycles/day at 2 Hz was measured histomorphometrically. Phenotypic characterization of the femurs in these mice was also performed. In a separate group of animals, the expression of the remaining COX isozyme, COX-1, was assessed using real-time RT-PCR 4 h after one bout of 120 loading cycles. RESULTS: Null mutation of the COX-2 gene resulted in a consistent femoral phenotype of reduced bone mass, altered architecture, and inferior mechanical properties. Many of these differences were nullified after adjustment for body weight. Nevertheless, body weight-corrected values showed a consistent trend of reduced mechanical properties in COX-2(-/-) mice. Genotype did not influence the response to mechanical loading, with no histomorphometric differences being found between COX-2(+/+) and COX-2(-/-) mice. Real-time RT-PCR showed COX-2(-/-) mice to express significantly greater COX-1 expression in loaded ulnas than in loaded ulnas in COX-2(+/+) mice. There were no differences in COX-1 expression in nonloaded ulnas. CONCLUSIONS: A functional COX-2 gene was not found to be required for skeletal mechanotransduction. This is in contrast to previous pharmacological studies showing that COX-2 is critical to the response of bone to loading. Investigating a potential reason for the absence of a genotype difference in this study, we found that mice with a null mutation in the COX-2 gene possess inductive skeletal COX-1 expression.