Cathepsin K Inhibitors for Osteoporosis: Biology, Potential Clinical Utility, and Lessons Learned

Abstract

Cathepsin K is a cysteine protease member of the cathepsin lysosomal protease family. Although cathepsin K is highly expressed in osteoclasts, lower levels of cathepsin K are also found in a variety of other tissues. Secretion of cathepsin K from the osteoclast into the sealed osteoclast-bone cell interface results in efficient degradation of type I collagen. The absence of cathepsin K activity in humans results in pycnodysostosis, characterized by increased bone mineral density and fractures. Pharmacologic cathepsin K inhibition leads to continuous increases in bone mineral density for ≤5 years of treatment and improves bone strength at the spine and hip. Compared with other antiresorptive agents, cathepsin K inhibition is nearly equally efficacious for reducing biochemical markers of bone resorption but comparatively less active for reducing bone formation markers. Despite multiple efforts to develop cathepsin K inhibitors, potential concerns related to off-target effects of the inhibitors against other cathepsins and cathepsin K inhibition at nonbone sites, including skin and perhaps cardiovascular and cerebrovascular sites, prolonged the regulatory approval process. A large multinational randomized, double-blind phase III study of odanacatib in postmenopausal women with osteoporosis was recently completed. Although that study demonstrated clinically relevant reductions in fractures at multiple sites, odanacatib was ultimately withdrawn from the regulatory approval process after it was found to be associated with an increased risk of cerebrovascular accidents. Nonetheless, the underlying biology and clinical effects of cathepsin K inhibition remain of considerable interest and could guide future therapeutic approaches for osteoporosis.

Figure 1.

Bone cell morphology by light microscopy on 5-μm-thick undecalcified bone biopsy sections from patients with pycnodysostosis. (a, b, d, e) The top two rows correspond to patient A, and (c, f) the bottom row corresponds to patient B. (Left column) Goldner’s trichrome stain showing osteoclasts. (Right column) Giemsa staining showing lining cells and osteoclasts. (a–c) Defective multinucleated osteoclasts adjacent to demineralized collagen fringes (pink) and mineralized bone matrix (green). Note the comparatively deeper resorption lacunae in (a, b) patient A vs (c) patient B. (d–f) Bone-lining cells (dark blue) in resorption area (light pink, arrows) after osteoclast detachment from (d) bone surface or (e, f) in orphan resorption pits. The mineralized bone surface is dark pink. Reproduced from Fratzl-Zelman et al. (83) with permission.

Figure 2.

Changes in BMD. The mean percentage of change from baseline over 24 months in BMD at the (a) lumbar spine (LS), (b) total hip, (c) femoral neck (FN), and (d) distal one-third radius in subjects treated once weekly with either placebo (open circles) or odanacatib (ODN) 50 mg (solid circles). Data from Bone et al. (131) with permission. Weighted LS mean, weighted least squares mean.

Figure 3.

Changes in biochemical markers of bone resorption and formation. Mean percentage of change from baseline over 24 months for markers of bone resorption [(a) urinary NTX (NTx)/creatinine (Cr) and (b) serum CTX (CTx)] and bone formation [(c) serum bone-specific alkaline phosphatase (BSAP) and (d) serum P1NP] in subjects treated once weekly with either placebo (open circles) or odanacatib (ODN) 50 mg (solid circles). Data from Bone et al. (131) with permission. LS, least squares.

Figure 4.

Changes in biochemical markers of bone resorption and formation. Mean percentage of change from baseline over 36 months for markers of bone resorption [(a) urinary NTX (NTx)/creatinine (Cr) and (b) serum CTX (CTx)] and bone formation [(c) serum bone-specific alkaline phosphatase (BSAP) and (d) serum P1NP] in subjects treated once weekly with placebo/placebo (PBO/PBO; dark squares), odanacatib 50 mg/placebo (ODN/PBO; open circles), or odanacatib 50 mg/odanacatib 50 mg (ODN/ODN; solid circles). Reproduced from Eisman et al. (132) with permission. SE, standard error.

Figure 5.

Changes in BMD. Mean percentage of change from baseline over 60 months in BMD at the (a) lumbar spine and (b) total hip in subjects treated once weekly with placebo/placebo (PBO/PBO; dark squares), odanacatib 50 mg/placebo/placebo (50 mg/PBO/PBO; open circles), or odanacatib 50 mg/odanacatib 50 mg/odanacatib 50 mg (50 mg/50 mg/50 mg; solid circles). Reproduced from Langdahl et al. (133) with permission.

Figure 6.

(a–c) Histologic slides (magnification ×40) and (d) composite schematic of the BRC, which comprises the cells constituting the BMU—specifically osteoclasts (OCs), osteoblasts (OBs), and osteocytes—and canopy of bone-lining cells and associated capillary. (a) A BRC in trabecular bone, demonstrating the location of the OBs along the bone-forming surface. The osteocytes are shown embedded in the bone matrix and the canopy of cells consists of bone-lining cells. (b) A BRC in cortical bone (outer demarcation shown by broken line) filled with erythrocyte ghosts (EGs) and OBs; a few OCs can also be seen. (c) A BRC stained with an antibody specific for CD34, demonstrating staining of endothelial cells in the marrow capillary adjacent to the BRC. (d) A composite schematic of the BRC, showing connections between the osteocyte network, surface bone-lining cells, and the BRC. All cells in this network are connected with gap junctions, which might provide a pathway (block arrows) by which signals generated by osteocytes deep within the bone reach the surface and elicit remodeling events by OCs and OBs. Note also the potential direct physical contact between OCs and OBs, which would allow for signaling between these cells. Reproduced, with permission, from Khosla et al. (168). CV, central vessel of the Haversian system, which forms the basic structural unit in cortical bone.

Figure 7.

(a) Working model for mechanisms by which osteoclasts regulate osteoblasts and bone formation. (b) Proposed changes in osteoclast–osteoblast coupling after treatment with conventional antiresorptive agents, including bisphosphonates and denosumab. (c) More complex changes in proposed osteoclast–osteoblast coupling after treatment with cathepsin K inhibitors. Reproduced from Khosla (226) with permission.