Thick bones, big drug

It was 1994, and Scott Simonet of Amgen's molecular genetics department in Thousand Oaks, Calif., was looking at some strange X-rays. He had engineered five transgenic mouse lines to overexpress a mysterious secreted protein. The mice looked and behaved normally, but that ordinariness was only skin-deep. "On the X-rays, it was pretty obvious that the long bones had higher bone mineral density," says

Elie Dolgin
Aug 1, 2008

It was 1994, and Scott Simonet of Amgen's molecular genetics department in Thousand Oaks, Calif., was looking at some strange X-rays. He had engineered five transgenic mouse lines to overexpress a mysterious secreted protein. The mice looked and behaved normally, but that ordinariness was only skin-deep. "On the X-rays, it was pretty obvious that the long bones had higher bone mineral density," says Simonet. The pelvic and vertebral bones also appeared thicker than usual. It was clear that the mysterious protein played a role in building bone.

Simonet had picked the protein because its DNA sequence partially matched that of a family of cytokine receptors involved in cell death, known as tumor necrosis factor (TNF) receptors. But unlike other TNF receptors, this protein lacked a transmembrane-spanning sequence; instead, it showed features of being secreted. As a protein that is released and flows throughout the body, Simonet suspected it might serve...

Bone is in a constant state of flux, continuously being reformed to remove parts worn out by daily wear and tear. Two basic types of bone-regulating cells control this remodeling: Osteoclasts act to dissolve and resorb older bone, leaving behind tiny cavities, which the second cell type, osteoblasts, fill with new bone. A delicate balance between the two cell types maintains normal bone density. When osteoclasts outnumber osteoblasts, however, bones become brittle and prone to fracture - the hallmarks of osteoporosis.

It was clear that the mysterious protein played a role in building bone.

Simonet's secreted protein appeared to limit the number of osteoclasts and protect the skeleton from bone resorption, so he named the protein osteoprotegerin (OPG) - Latin for "the bone protector." Simonet also showed that giving extra OPG to cell lines or to normal mice blunts osteoclast formation and leads to denser bones (Cell, 89:309-19, 1997).

The Amgen team had stumbled upon a very important protein - but so did a number of other researchers around the globe. On the other side of the Pacific, researchers at the Snow Brand Milk Products Company in Japan independently identified OPG in 1998; later that year, both groups also discovered OPG's target ligand. As it turned out, the OPG ligand was identical to the ligand of another TNF receptor called RANK, which researchers at Immunex in Seattle discovered in 1997.

Researchers now know that this ligand, called RANKL, binds to RANK to promote osteoclast formation. OPG, though, acts as a "decoy receptor," interacting with RANK to prevent its association with RANKL. The balance between RANKL and OPG controls the amount of bone resorption, says Simonet. "Almost all discoveries associated with bone loss are involved in this pathway."

In 1999, Amgen entered recombinant OPG in Phase I trials of postmenopausal women with osteoporosis. Two years later, the protein was also tested against cancer-related bone lesions. In 2001, Amgen even teamed up with NASA to test OPG's effect in zero gravity on mice aboard the space shuttle Endeavor.

OPG showed promise in all these studies, the results of which were published in Cancer and The Journal of Bone and Mineral Research, but the clinical trials didn't progress much further. That's because, in the meantime, Amgen had developed another product to target the same pathway, denosumab. Whereas OPG prevents RANKL from binding RANK, denosumab works by destroying RANKL directly. Denosumab is a human monoclonal antibody, and, as such, it's "more potent and more selective in its binding capacity" than OPG, Simonet says. He adds that denosumab would never have been developed without first discovering OPG and the RANK pathway.

So far, the biotech company has instigated 25 clinical trials with the drug involving around 20,000 patients, including two Phase III head-to-head trials showing that denosumab outperforms the current leading osteoporosis drug, Fosamax, for some bone traits.

Switching from OPG to denosumab was "certainly the right decision," says Michael McClung, an endocrinologist at the Oregon Osteoporosis Center in Portland, who oversaw some of the denosumab clinical trials and serves on Amgen's global advisory board. With denosumab, "the osteoclasts are essentially in hibernation, rather than made unhappy" as they would be with an OPG-based drug. This makes denosumab treatment fully reversible and much safer, he says.