BoneKEy-Osteovision | Commentary
Function and regulation of cathepsin K in bone
DOI:10.1138/2001036
Adult bone continuously remodels, being removed (via bone resorption) by the osteoclasts and rebuilt (via bone formation) by the osteoblasts. These two major bone resident cell types act in a coordinated fashion, with input from the osteoblast-related osteocytes within the bone and from the lining cells, on the bone surface. Osteoclasts are the hematopoietically-derived specialized multinucleated cells capable of bone resorption. Following tight attachment to the bone surface, osteoclasts secrete protons into a closed extracellular compartment enclosed by a “sealing zone”. The acidic pH (about 4) removes the bone mineral and exposes the underlying matrix. Osteoclasts also secrete the proteases that digest the matrix components. The search for the identity of the protease(s) responsible for degradation of the matrix, which consists of 90% type I collagen, has been the subject of extensive research and numerous publications during the last three decades. Currently, cathepsin K, a member of the papain cysteine protease family, is considered the leading candidate protease responsible for the degradation of most of the bone matrix.
Cathepsin K was discovered in 1994 using differential display of osteoclast and macrophage cDNA libraries from rabbits (). Subsequently the human and mice enzymes were also cloned (). Cathepsin K is abundantly and relatively selectively expressed in osteoclasts, where it is localized in lysosomes, in the ruffled border and in the resorption lacunae on the bone surface (). This cellular and extracellular localization suggests an important role in bone resorption, as does its recently characterized functional activity profile. The ability of cathepsin K to degrade type I collagen both within and outside the helical regions, and to act at an acidic and neutral pH, make it a unique mammalian protease, distinct from the other MMPs (). Interestingly, Kafienah et al. (), also showed that human cathepsin K not only cleaves native type I collagen but also type II collagens at the N-terminal of the triple helix, suggesting that cathepsin K may play a role in cartilage breakdown as well. In fact cathepsin K mRNA expression is upregulated in the rheumatoid arthritis synovium, where its expression was confined to sites of bone destruction (). While, inhibitors of cathepsin K have been shown to effectively inhibit bone resorption in vitro and in vivo () there are no similar reports to date on inhibition of cartilage destruction.
To supplement above biochemical evidence, the importance of cathepsin K in the resorptive process is also supported by a disorder in humans called pyknodysostosis, a rare sclerosing skeletal dysplasia with low bone turnover, which occurs as a result of mutations in the cathepsin K gene (). Furthermore, the deletion of the cathepsin K gene in mice results in osteopetrosis (), an abnormality of bone remodeling caused by decreased bone resorption, manifested as skeletal sclerosis. Osteopetrotic mice have reduced bone marrow cavities, usually filled with trabecular bone.
In addition to mice lacking cathepsin K, which have functionally impaired osteoclasts, there are several other mouse models of osteopetrosis, the study of which has increased the understanding of osteoclast function during bone resorption. Rodent osteopetroses, caused by spontaneous or induced genetic defects, are a heterogeneous group characterized by the absence or dysfunction of osteoclasts. For example, in mice where c-src has been deleted, there are multinucleated osteoclasts but they do not form a ruffled border and resorb bone matrix poorly (), while fos-null mice completely lack osteoclasts and have an increased number of macrophages (). In β3 integrin knockout mice there is actually . an increase in osteoclast number but the osteoclasts are dysfunctional, as evidenced by the inability of isolated osteoclasts to resorb bone in vitro (). The V-ATPase-deficient atp6i mouse (), which carries the same mutation as the spontaneous oc/oc mouse (), also has dysfunctional osteoclasts, unable to acidify the resorption lacuna. Interestingly, disruption of a late endosomal/lysosomal Cl- channel, which may provide the chloride conductance required for efficient proton pumping by the H+-ATPase, yielded a phenotype similar to the atp6i mouse (). An osteopetrotic phenotype was also produced in the double knockout of the nfkb1 and nfkb2 genes, which interfered with osteoclast differentiation (). The op/op mouse, a spontaneous mutation, contains an inactivating mutation in the M-CSF gene, resulting in both osteoclast and macrophage deficiency (). The mi/mi microphthalmia mutant mice have a reduced number of multinucleated osteoclasts, due to a defect in fusion, and lack of ruffled borders, resulting in compromised bone resorption (). The mouse microphthalmia gene encodes a Myc-related basic helix-loop-helix-leucine zipper (b-HLH-ZIP) protein () that belongs to a family of transcription factors which includes Mitf, TFE3, TFEB, and TFEC. Weilbaecher et al. () showed that Mitf and TFE3 are expressed in osteoclasts.
In rare cases the regulation of these osteoclast-essential genes has been elucidated, which brings us back to Cathepsin K. Based on the phenotypic similarity between the Mitfmi/mi mutant mice and cathepsin K null mice (both exhibit osteopetrosis due to non functional osteoclasts), Motyckova et al., () evaluated the status of cathepsin K mRNA and protein in the Mitfmi/mi mice and identified cathepsin K as a protein regulated by the Mitf transcription factor family. Using spleen cells from Mitfmi/mi mice and the osteoclastogenic cytokine RANKL, osteoclasts were generated in culture, but they had diminished TRAP intensity and fewer nuclei per cell. Cathepsin K mRNA and protein expression were reduced (but not absent) in these osteoclasts, compared to wild-type osteoclasts, suggesting that factors other than Mitf are likely to contribute to its expression. Examination of the human cathepsin K promoter revealed four E-boxes that match the binding consensus sequence of the Mitf family members and indeed, gel shift assays proved that endogenous Mitf binds to E-boxes 1-3 but not to E-4. Since there was evidence for other factors interacting with E-box elements, interaction with TFE3, also expressed by osteoclasts, was evaluated. In transient transfection experiments using MCF-7 cells, the wild-type cathepsin K promoter was up-regulated by both Mitf and TFE3. Mutations in E-boxes 1 and 2 abolished and in E-box 3 reduced the regulation of cathepsin K. Adenoviral infection of primary human osteoclasts with Mitf resulted in up-regulation of the expression of cathepsin K as well as TRAP. Although the endogenous cathepsin K promoter appears to be regulated by Mitf, Mitf absence produces no osteoclast defect. A loss of function mutation at the Mitf locus does not affect osteoclast differentiation. However, the dominant negative mutations of Mitf, Mitfmi/mi and Mitfor/or, which contain intact HLH-ZIP motifs that dimerize with wild-type partners but disrupt binding to DNA, resulting in inactive complexes, lead to osteopetrosis. Another allele Mitfce, which contains a mutation that ablates the HLH-ZIP region and produces a protein incapable of either dimerization or DNA binding is not osteopetrotic.
The study by Motyckova et al. () concludes therefore that cathepsin K is a downstream target gene of Mitf. Karsenty () in his review of this article suggests that the milder osteopetrosis induced by cathepsin K inactivation in humans and mice compared to mi/mi mice is consistent with these conclusions since TRAP and presumably other genes such as those responsible for osteoclast fusion are also downstream of Mitf. There are other mutations that cause osteopetrosis highly reminiscent of the cathepsin K phenotype. One such mutation is the grey-lethal (gl) mouse. However, this mutation did not result in lower levels of cathepsin K in osteoclasts ().
Cathepsin K is one of the major players in bone resorption, the deficiency or inhibition of which causes impaired osteoclast activity, and as mentioned above, skeletal abnormalities in patients with pycnodysostosis who lack this protein. There is however a possibility that there are species differences between humans and mice lacking cathepsin K. While humans have denser bones and suffer fractures, the cathepsin K null male mice have less severe osteopetrosis and stronger bones (). In addition, the skeletal abnormalities seem to be restricted to long bones in mice whereas in humans there are abnormalities including unclosed cranial sutures. Everts et al. () suggested the possibility of site specific differences in the skeleton with respect to osteoclastic bone resorption, based on differences in cathepsin K activity between calvarial and long bone osteoclasts in mice. This would be consistent with a lack of a cranial phenotype in the cathepsin K null mouse.
In conclusion, cathepsin K plays a pivotal role in bone resorption and inhibition of this enzyme could be useful in the treatment and prevention of conditions in which bone resorption exceeds bone formation, such as osteoporosis. The mouse cathepsin K knockout has many phenotypic similarities to people who have pyknodysostosis, along with distinct differences. Future studies, like the one reported by Motyckova et al. (), could elucidate the regulation of other genes involved in osteoclast formation and function, as illustrated for cathepsin K in this case, and further advance our understanding of osteoclasts and potential links between mice and men.
This work is licensed under a
Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License.