The pathway to a treatment for Alzheimer’s disease is littered with drugs that looked promising but ultimately failed because of the disease’s complex and poorly understood mechanism. One drug candidate still soldiering on is Merck’s verubecestat, which is currently in Phase III clinical trials.
On Wednesday at the 7th International Chemical Congress of Pacific Basin Societies, or Pacifichem, in Honolulu, Merck scientist Daniel Wyss reported how scientists developed the clinical candidate.
Verubecestat targets β-site amyloid-cleaving enzyme 1, also known as BACE1 or β-secretase. The enzyme is involved in generating amyloid-β, the peptide responsible for the plaques that accumulate in the brains of Alzheimer’s patients. Amyloid-β’s role in Alzheimer’s disease is controversial, and if researchers could prove that blocking its production slowed or halted the disease’s progression, it would be a major breakthrough.
But roughly 10 years ago, scientists considered BACE1 an undruggable target, Wyss tells C&EN. Most groups were using peptidelike compounds to inhibit the enzyme, he says. “They had suboptimal pharmaceutical properties and couldn’t cross the blood-brain barrier.”
To tackle BACE1, researchers at Schering-Plough decided to use fragment-based lead discovery (FBLD)—a relatively new approach to drug discovery at the time. (Schering-Plough later merged with Merck.) In contrast to the classical drug-discovery approach of high-throughput-based screening, in which scientists screen a library of millions of small molecules against a particular target, FBLD tests only thousands of simpler chemical motifs to see if they bind in a target’s active site.
“It’s like a puzzle,” Wyss explains. “You take the building block and you try to gradually expand it into nearby binding pockets. Every interaction that you build into the core contributes to additional binding affinity.”
Wyss’s team initially hit upon an isothiourea as the starting piece for their discovery process. Unfortunately, isothioureas don’t make good drugs because of their instability. So the scientists examined what the isothiourea was doing in BACE1’s active site to find a way to mimic that function with another structure. X-ray crystallography revealed that the isothiourea moiety was making an extended network of hydrogen bonds with functional groups that were key to BACE1’s catalytic activity. “That was completely novel,” Wyss says. “Those kinds of interactions were unknown in the field.”
The researchers determined that iminopyrimidinones could bind in the same manner as the isothiourea moiety. They then modified the molecule further by adding a fluorinated phenyl ring so it could extend into one of the nearby binding regions in BACE1, where hydrophobic interactions were key. After some classical medicinal chemistry, they arrived at the clinical candidate, verubecestat.
“Although more than 30 drug candidates derived from fragment-based methods have entered the clinic, Wyss’s work shows this approach can be effective even on a particularly difficult target,” comments Daniel Erlanson, the symposium organizer and cofounder of Carmot Therapeutics, a biotech company that focuses on FBLD. The approach, Erlanson adds, “has delivered a molecule with sufficiently good properties that Merck will be able to directly test the amyloid hypothesis for Alzheimer’s disease.”