By Randall Osborne

BRUSSELS, Belgium — Calling pharmacogenomics proof of the industry's "imagination to overcome what was once thought to be impossible," Glenn Miller, associate scientific director of Genzyme Corp., told European biotechnology leaders the flood of clues to potential targets is a problem full of opportunity.

"You don't want to have to make 'go/no-go' decisions on hundreds of thousands of potential targets," he said. "You want to make those sorts of decisions on a very small number of highly qualified targets."

That's where pharmacogenomics comes in, Miller said, noting that working with genotype and phenotype data is nothing new but is moving into previously unexplored areas.

Miller spoke as part of a panel on new technology frontiers at EuropaBio '98, the second annual European Biotechnology Congress, focused on the unfolding biotechnology industry in Europe.

"There are decades of experience looking at metabolic genes and the difference of drug metabolism [between] individuals," Miller said. "Where it's leading to now is patient population segmentation. You're able to identify patient populations without administering bioactive drugs and waiting to find out how well an individual metabolizes [them]. You're able to do a genoytpe on very large populations, and find out who you can treat and who you can't treat."

Such a strategy is especially important in developing clinical trials, Miller said, and Genzyme has devised comparative expression methods.

"The problem we're attempting to solve here, or at least provide some information on, is the identification of disease and treatment-relevant expressed genes," Miller said. "These are taking a look at two different populations who have been treated or not treated with a particular compound, looking at normal vs. disease groups [to find out] the genes involved in a toxic response to a particular compound, and how you can parse those genes out."

Patterns of expression are important, too, he said, since diseases affecting larger numbers of people are not single-gene disorders.

Cambridge, Mass.-based Genzyme uses its Serial Analysis of Gene Expression (SAGE) technology to "reduce the bulk that you're looking at, from thousands or millions of tags, or hundreds of thousands of genes, down to just a few that you can look at in detail," Miller said.

Although in genomics, "collaborations are already [valued at] over $3 billion, and this is a number that is a year or two old," more partnerships are required to step up the science.

"Pharmacogenomics is really the platform of the future for drug development," he said. "The number of targets that will need to be sifted through is not growing smaller, it's growing larger. There's no single technology here that will be the answer. It will have to be a pool of technologies."

Future Holds 'Black Box' For Tailored Drug Selection

Ultimately, Miller said, companies will take a clinical sample from an individual and put the sample through a "black box," which is "a process, and not a particular piece of equipment."

Included in that process will be information from expressed genes and genotypes — as much information as can be gained, from all the sources available. "What will come out of that process will be a patient-specific pattern allowing you to customize therapy," Miller said. "That's not one drug, one patient. That's one drug that is known to act on a wide variety of individuals, but has perhaps some adverse side effects you want to avoid."

The next couple of years will bring "the identification [of every gene] or at least the knowledge that every gene is being held in a repository somewhere. If you stop there, you're no better off than if you stop today."

The EuropaBio meeting, with about 550 in attendance, continues through Friday. *