Science & Technology: CHEMISTRY
SWAPPING A TEST TUBE FOR A SHOTGUN
Back in the mid-1980s, Australian scientist H. Mario Geysen was a prophet without a following. He had figured out a way to rapidly combine molecular building blocks in many different ways--an approach known as combinatorial chemistry. While conventional chemists could synthesize only a few dozen druglike compounds a year, he could churn out thousands of protein fragments called peptides in a few weeks. That offered a possible cure for one of the pharmaceutical industry's biggest headaches: discovering new drug candidates. The industry's response? A resounding yawn. "We couldn't get the message across," recalls Geysen.
Today, nobody is yawning. Geysen is now a drug-discovery chief at Glaxo Wellcome Inc., his former company is a division of biotech powerhouse Chiron Corp., and once neglected combinatorial chemistry is on its way to becoming the drug industry's hottest new tool. "Not since recombinant DNA [gene-splicing] have I seen a technology catch on so fast," marvels Isis Pharmaceuticals Inc. Vice-President David J. Ecker.
Indeed, drug giants such as Merck, Pfizer, and Chiron have set up major in-house efforts, while others rush to buy companies specializing in combinatorial approaches. In the past year, Marion Merrell Dow Inc. snapped up Tucson-based Selectide Corp. for $58 million, Glaxo paid $538 million for Affymax, based in Palo Alto, Calif., and Eli Lilly acquired Sphinx Pharmaceuticals in Durham, N.C., for $72 million. Add in such startups as Pharmacopeia in Princeton, N.J., ArQule in Medford, Mass., and CombiChem in La Jolla, Calif., and there may be more companies doing combinatorial chemistry than Wall Street analysts who understand the new technology, jokes David K. Stone, managing director of Cowen & Co. Says Martin F. Haslanger, president of Lilly's Sphinx subsidiary: "I don't know of any big pharmaceutical company that doesn't have, or isn't scrambling to get, a major combinatorial chemistry effort."
The reason is the sheer difficulty of discovering new drugs. In today's typical approach, a company starts with a key element in a biological pathway--an enzyme involved in heart disease, for example, or a cell receptor that signals the immune system to fight cancer. The trick is finding a chemical that binds to these enzyme or receptor "targets," turning them on or off to treat an illness. Historically, the odds of finding a blockbuster drug weren't much better than those of winning a lottery.
COMBING NATURE. Over decades, big drugmakers have painstakingly synthesized hundreds of thousands of chemicals--one at a time. They are combing the natural world, from rain forests to marine life, for new compounds. And recently they've developed automated methods for rapidly testing thousands of chemicals against molecular targets. Problem is, the drug hunters still aren't coming up with many winning lottery tickets. "We have targets we know are important but haven't been able to find molecules that will interact with them," admits Paul Armond, Pfizer's assistant research and development director. And now that gene sleuths and molecular biologists are finding thousands of new targets, the challenge is greater than ever.
Hence the immense allure of combinatorial chemistry. Glaxo's Geysen estimates that the nascent technology already has produced more new compounds in just a few years than the pharmaceutical industry previously did in its entire history.
What's more, the combinatorial approach isn't limited to drugs. At the University of California at Berkeley, chemist Peter G. Schultz and colleagues are making 10,000 new materials in single experiments. The approach, Schultz predicts, may be the key to discovering everything from new semiconductors to superconductors. Already, he has uncovered a new material with special electromagnetic properties and founded a company, Symyx.
"LIKE BLASPHEMY." The combinatorial idea comes not from chemistry, but from the immune system's ability to churn out many types of antibodies. "To the classic organic chemist, this is like blasphemy," says Scripps Research Institute chemist and molecular biologist Kim D. Janda, founder of CombiChem Inc. While chemists pride themselves on meticulously purifying and analyzing each chemical compound separately, combinatorial chemistry uses a shotgun approach. The basic idea: produce huge numbers of compounds almost at random by shuffling around molecular building blocks. Only when a chemical shows interesting biological activity do scientists go back and figure out the composition. It's such a simple idea that "we look back and say, `My God, why didn't we see this 20 years ago?"' says Walter H. Moos, vice-president for R&D at Chiron.
The main reason for the slow uptake: Combinatorial chemistry's pioneers, such as Geysen and Affymax, initially fashioned only peptides and oligonucleotides, which are compounds made from the building blocks of DNA. Not until the early 1990s, when scientists such as Berkeley chemist Jonathan A. Ellman were able to synthesize collections of small organic molecules--the typical raw material for drugs--did most of the pharmaceutical industry take notice.
Despite the enthusiasm, there are raging debates over the best combinatorial approach. Perhaps the most common tack, used at Pharmacopeia, Chiron, and other companies, is a "pool and split" strategy. Chemists begin by anchoring different molecular building blocks to solid supports (typically plastic beads) in each of several containers (diagram, page 116). They mix the compounds together, then divide up the pool. Chemists then add a second chemical unit--a different one to each container--and pool and split again. Repeating this a number of times can create millions of different compounds in a handful of containers.
If one of the compounds "hits" a biological target, chemists must then go back and figure out the composition of the chemical. One clever solution is using molecular tags, much like a bar code, that keep track of each compound's structure. But many chemists think the pooling and tagging strategy is too complicated. That's why Merck, CombiChem, and others are making thousands of compounds simultaneously but separately, using robots to deliver building blocks to each vial in different order. The most exotic version of this method comes from a collaboration between the David Sarnoff Research Center and SmithKline Beecham. Sarnoff scientists plan to put 10,000 microscopic test tubes, along with a maze of hair-thin channels, pumps and sensors, on chips the size of business cards. Not only could such chips make 10,000 molecules in a few hours, but "we can also use them to screen the molecules against our biological targets," says SmithKline R&D chief George Poste.
NO MIRACLE CURE. Whichever method prevails, many industry experts think the payoff from combinatorial approaches will be huge. Indeed, drugs produced this way are nearing clinical trials. Lilly hopes to start testing a central nervous system drug by yearend. Isis is excited about an anti-AIDS compound.
Combinatorial chemistry, though, is no miracle cure for the ills of drugmakers. Companies have learned that simply making racks of compounds isn't enough, says Columbia University chemist W. Clark Still, founder of Pharmacopeia. "You have to make huge quantities of cleverly chosen molecules," he says. And to become drugs, compounds also must be relatively nontoxic, get into the right cells, and stick around in the body for the right amount of time. That's why drugmakers are also beginning to simultaneously test the new chemicals for such practical considerations as absorption across the stomach wall. "The combinatorial chemistry bandwagon has a lot of people riding on it," warns Geysen. "But few appreciate what it takes to make it practical."
Still, even skeptics agree that it may be the best hope for finding truly novel drugs for previously intractable diseases. And once a potential drug is found, the technology can quickly generate all its possible relatives, increasing the chances of finding the best version--and helping a company stake out a strong patent position.
The bottom line: Combinatorial chemistry could transform drug discovery and new material synthesis, much as gene splicing stood modern biology on its head. "It's a revolutionary technology," says biochemist James Rothman of Memorial Sloan-Kettering Cancer Center. All companies need to do is figure out how best to use it.By John Carey in Washington