Cathepsin K is believed to have an eminent role in the pathologic resorption of bone. However, several studies have shown that other proteinases also participate in this process. In order to clarify the contribution of cathepsin K to the destruction of arthritic bone, we applied the human tumor necrosis factor (hTNF)–transgenic mouse model, in which severe polyarthritis characterized by strong osteoclast‐mediated bone destruction develops spontaneously.

Arthritis was evaluated in hTNF‐transgenic mice that were either wild‐type for cathepsin K (CK^+/+), heterozygous for cathepsin K (CK^+/−), or deficient in cathepsin K (CK^−/−). Arthritic knee joints were prepared for standard histologic assessment aimed at establishing a semiquantitative score for joint destruction and quantification of the area of bone erosion. Additionally, microfocal computed tomography was performed to visualize bone destruction in mice with the different CK genotypes. CK^+/+ and CK^−/− osteoclasts were generated in vitro, and their proteinase expression profiles were compared by complementary DNA array analysis, real‐time polymerase chain reaction, and activity assays.

Although the area of bone erosion was significantly reduced in hTNF‐transgenic CK^−/− mice, the absence of cathepsin K did not completely protect against inflammatory bone lesions. Several matrix metalloproteinases (MMPs) and cathepsins were expressed by in vitro–generated CK^−/− osteoclasts, without marked differences from CK^+/+ osteoclasts. MMP activity was detected in CK^−/− osteoclasts, and MMP‐14 was localized by immunohistochemistry in inflammatory bone erosions in hTNF‐transgenic CK^−/− mice, suggesting MMPs as potential contributors to bone destruction. Additionally, we detected a reduction in osteoclast formation in cathepsin K–deficient mice, both in vitro and in vivo.

The results of our experiments raise doubts about a crucial role of cathepsin K in arthritic bone destruction.