Gerald Koelsch, who is at the Oklahoma City site of the biotech company CoMentis, Inc., surprised the audience with his presentation of human data on the company’s β-secretase inhibitor CTS-21166. In a first proof-of-concept human study, the compound appeared safe and reduced plasma amyloid-β levels substantially for an extended period of time. Many pharmaceutical companies are known to pursue BACE inhibitors, and some are rumored to have begun initial human testing, including Merck (see ARF related news story), Eli Lilly and Co., and Takeda in Japan, but this is the first time clinical results of a BACE inhibitor have been reported at a large scientific conference in the U.S.

Scientists at CoMentis have pursued BACE as a drug target since two of its scientific founders, Jordan Tang and Arun Ghosh, published a crystal structure with an inhibitor caught inside it (Hong et al., 2000). That inhibitor was based on the APP sequence, and since then, the company has focused on medicinal chemistry to develop a suitable small-molecule drug. CTS-21166 is a transition-state analog that is the company’s lead candidate at present, Koelsch said. He would not reveal the molecule’s structure but recounted its properties on the list of requirements a drug must meet, from pharmacokinetics to pharmacodynamics to oral availability, and more.

Koelsch noted that the compound’s potency and selectivity passed muster when it proved to bind BACE with a potency of 1.2 to 3.6 nanomolar in cellular assays, but not to bind some 60 other enzymes, channels, and receptors. In response to a question about non-specific plasma protein binding, which has posed problems for some other candidate BACE inhibitors in the past, Koelsch noted that this has not been an issue with CTS-21166.

When injected intra-peritoneally for six weeks at a dose of 4 mg/kg into an aggressive transgenic mouse model that expresses both the Swedish and London APP mutations (Rockenstein et al., 2001), CTS-21166 reduced levels of brain Aβ40 and 42 by 38 and 35 percent, respectively. It also reduced plaque load in the hippocampus and cortex by some 40 percent. Koelsch said he assumed this plaque drawdown happened indirectly through an equilibrium shift, whereby a drop in new Aβ production allows existing Aβ to come off plaques and be degraded (see DeMattos story). Pharmacokinetic measurements in rats and mice indicated that CTS-21166 displays good brain penetration at a concentration of 8 ng/ml relative to that in plasma of 18 ng/ml at a single time point (4 mg/kg over six weeks). Previously, it has been difficult to find BACE inhibitors that cross the blood-brain barrier.

In these experiments, the drug did not damage the myelin sheath in the sciatic nerve of this mouse model, indicating that the BACE enzyme might aid myelination primarily during development (Willem et al., 2006). They will continue to monitor for demyelination, Koelsch said.

With that data, the scientists undertook a Phase 1 study in humans. Single intravenous injections of doses ranging from 7.5 to 225 mg were given to six volunteers plus two placebo recipients per group. In this small trial, the drug was well tolerated at all doses, Koelsch said. As in prior studies, a dose-dependent response was seen in plasma pharmacokinetics. The volunteers varied little in their response to the drug. They cleared it slowly, indicating the drug could possibly be given once a day. (Though this human study used intravenous administration, the inhibitor has been shown to be orally available in mice, rats, dogs, and monkeys, Koelsch said.)

Koelsch showed data suggesting that the drug reduced plasma Aβ levels by up to 80 percent in the volunteers at the highest dose, reaching its trough three hours after injection and gradually returning to baseline over the course of several days. Interestingly, Aβ40 levels did not overshoot baseline after the drug had washed out. Such a “rebound” effect has puzzled scientists with LY450139 treatment (Siemers et al., 2007. This γ-secretase inhibitor has entered a Phase 3 clinical trial called IDENTITY).

The CoMentis study did not include spinal taps; hence, no CSF data are available. Another major question that remains open at this point is by how much brain Aβ levels need to come down for a drug to be clinically meaningful to AD patients. This question will find its answer in the clinic.—Gabrielle Strobel.