Drug Discovery is the lifeblood of the pharmaceutical industry, and companies don't outsource it lightly.? So becoming one of Merck & Co.'s first discovery chemistry outsourcing partners could not have been an easy feat. But that's what ChemBridge Corp. did in 1995 soon after Eugene Vaisberg, its chairman and chief executive officer, visited Merck scientists in Rahway, N.J.
Vaisberg and partner Sergey Altshteyn had started ChemBridge in 1993 with the idea of bringing the fruits of chemistry research in the former Soviet Union to the West. At the time, the high-throughput screening of combinatorial libraries for drug leads was coming into vogue. ChemBridge started creating its own libraries by gathering nonproprietary compounds from researchers around the world.
Within two years, the firm had begun refining the process by harnessing three dimensional pharmacophore design and molecular modeling techniques to fashion quality-controlled libraries out of the sprawling and sometimes low-quality universe of compounds floating around the drug industry.
Calling on Merck in January 1995, Vaisberg met with Arthur A. Patchett, the legendary medicinal chemist who helped discover early blood-pressure- and cholesterol-lowering drugs, and Steven M. Hutchins, who was just launching Merck's discovery chemistry outsourcing effort. Within days, Vaisberg says, ChemBridge had won assignments from both men.
From that beginning, the ChemBridge-Merck relationship has steadily evolved--as has the combinatorial chemistry world around the companies.
In 1997, ChemBridge built chemistry labs in Moscow where it began to synthesize its own compounds with desirable medicinal chemistry properties. At first, these compounds were nonexclusive--available to any buyer. But in 2001, ChemBridge signed a three-year agreement to design and produce libraries that were exclusive to Merck. Some were to be produced in Moscow and some at ChemBridge Research Laboratories (CRL), the San Diego chemistry company that ChemBridge opened in 2000 to focus on more collaborative projects.
According to Vaisberg, the Merck deal--plus a similar one signed with Pfizer later that year--marked the start of a shift in ChemBridge's business mix. In the firm's early years, its stock-in-trade was multisource compound collections. By 2000, sales were split almost evenly between these traditional collections and the ChemBridge-made nonexclusive libraries. Exclusive libraries were just emerging.
Today, about two-thirds of the firm's business is in exclusive discovery chemistry collaborations. The rest is in the traditional multisource collections and ChemBridge-developed libraries.
Merck's participation in combinatorial chemistry has similarly evolved from its start in the early 1990s. As Hutchins explains, the company's compound collection was then made up of historical compounds from past medicinal chemistry projects.
At first, Merck combinatorial/parallel synthesis chemists began adding to this base. In the ensuing years, however, the firm's strategy progressed and it started to buy rather than make its compound libraries--nonexclusive ones at the outset but shifting over time to exclusive ones.
Internal combinatorial chemistry resources were focused instead on designing libraries and turning the screening hits they generate into drug leads. In addition to ChemBridge, Merck struck library deals with firms such as Array Biopharma, Discovery Partners International, Exelixis, WuXi PharmaTech, and Comgenex.
Hutchins says Merck launches a new library by polling its medicinal chemists around the globe for ideas. The firm's molecular modeling scientists then try to determine whether or not the ideas will work and whether they will enhance the company's compound collection.
"We design down to the exact structure," he says, "then work with external companies to help us evaluate our designs, and provide feedback on the chemistry and how it fits with their capabilities. We want to make sure that when a project goes out, we assign it to the company with the best skills."
A typical library might contain 2,000 compounds built around a single molecular scaffold with three points of diversity. Such a library can take anywhere from four months to a year to construct.
Vaisberg and Hutchins agree that the art and science of combinatorial chemistry library creation have come a long way since the early 1990s. "In the early days, it was much more of a numbers game," Hutchins recalls. "We have become much better at designing libraries. The parallel synthesis techniques have improved over the years to allow for more medicinally relevant compounds to be prepared."
LIKEWISE, Vaisberg acknowledges that the early hype surrounding combinatorial chemistry was accompanied by a wave of low-quality libraries. He credits Merck with being one of the first drug companies to build a culture and infrastructure that channels interesting medicinal chemistry into the design of novel libraries. Also forward-looking of Merck, Vaisberg notes, was its recognition of the value of unique library building-block compounds. "Your library is only as good as its input," he says.
Ten years after the rise of combinatorial chemistry, the relationship between drug companies and compound library providers is beginning to shift again. In February, Pfizer broadened its relationships with ChemBridge/CRL and Discovery Partners with new agreements that allow them to provide hit follow-up libraries designed using the results of initial rounds of screening.
Hutchins notes that Merck signed a similar follow-up agreement with WuXi last May. "It's a natural evolution," he says. "If you are working with an outside company to prepare lead generation libraries and you get hits that require follow-up, it makes sense to work with the same company that made the original library."
ChemBridge has yet to strike a hit follow-up agreement with Merck, but Vaisberg says such a step would be a logical progression in their relationship. "It may be the next thing on the horizon," he says.