New Drugs: The Haystack Gets Smaller

Developing new treatments from protein "needles" will be easier

Recombinant insulin, EPO, TPA. Together, these big-sellers and a half dozen other biotech drugs accounted for $5 billion in revenues last year. They also share a common biological trait: They are all genetically engineered versions of proteins naturally secreted into the body by human cells. Secreted proteins are the workhorses of the body, conveying messages from one cell to another and playing key roles in countless bodily functions. They are also among the most highly sought-after candidates for new drugs.

Researchers once spent lifetimes looking for new secreted proteins. But advances in genetic technology make it likely that nearly all of the body's secreted proteins will be found within two years. Using computer-based techniques, scientists from such gene-hunting powerhouses as Human Genome Sciences (HGS), Millennium, and InCyte are amassing databases containing thousands of previously unknown proteins. On Sept. 25, Genetics Institute Inc. (GI) in Cambridge, Mass., announced it had perfected a biological technique to isolate virtually all secreted proteins. Finding them "used to be like looking for a needle in a haystack," says Kenneth A. Jacobs, the GI senior scientist who discovered the technique. "Now, we have a single magnet that will find all the needles."

GI says it has found about 5,000 genes that code for secreted proteins so far. Jacobs predicts that in the next 24 months, the company will have identified nearly all the genes thought to code for secreted proteins--about 10,000 of the human body's 100,000 genes.

The new technology builds on the discovery a decade ago that all secreted proteins have a characteristic "tail" that differentiates them from other proteins. Gene hunters were the first to take advantage of that. Companies such as HGS and InCyte use computers to scan vast genetic databases, looking for DNA fragments that correspond to this tail. Using this technique, HGS has identified 5,000 secreted proteins and identified six that have potential use as therapeutics for major diseases, says Chief Executive William A. Haseltine. The company has filed for scores of patents and received two--including one for a substance that blocks formation of the white cells responsible for inflammation in multiple sclerosis and rheumatoid arthritis.

TARGETS. GI hopes Jacobs' technology will help it find secreted proteins that the gene hunters might miss. The technique involves using specially modified strains of yeast that will grow only when exposed to the gene fragments that code for secreted proteins' tails. Researchers add random human gene fragments, taken from assorted tissues, to the modified yeast in a petri dish. If the yeast grows, the fragments must be part of a gene for a secreted protein. Researchers can then produce the protein for study as a possible drug target.

GI says its technology is generating far more information than one company can digest. So it is sharing its catalog of secreted proteins with major rivals at a modest up-front cost. The fee buys companies samples of GI's proteins. They then agree to share half the profits with GI if any of the proteins lead to drugs. HGS and other genomics companies, by contrast, have kept most of their data confidential--or sold it to select companies willing to pay steep prices.

GI's yeast technology is impressive, but there's no guarantee the company will profit from its secreted proteins. For one, competition from the likes of HGS is stiff. "For many of the molecules they are going to find, [GI] will discover that HGS is already there with a full-length patent," says Haseltine.

Also, some scientists believe that most of the useful secreted proteins have already been discovered, says Joshua Boger, chairman of Vertex Pharmaceuticals. Remaining proteins may have very specific--and limited--activity, or they could be so general that side effects would outweigh benefits. In either case, they would be poor candidates for drug development.

GI is willing to take that chance. Already, the company has signed agreements with Genentech Inc. and Chiron Corp.--two of the biotech industry's leaders. Genetics Institute Chairman Gabriel Schmergel says he is in talks with 16 other companies, including Amgen Inc. "Genetics Institute's approach offers the power of numbers," says David Ebersam, vice-president for business development at Genentech. "If you have lots of proteins, you increase the chances you get a `hit"' or find a protein that can lead to a drug.

At the least, says Nobel prize winner Phillip A. Sharp, head of the biology department at Massachusetts Institute of Technology, GI's library will turn up proteins that give researchers valuable clues about how diseases work. In recent years, geneticists have identified a number of disease-related genes that code for proteins, but the proteins themselves are not necessarily useful as drugs. Leptin, a highly touted protein associated with obesity, may or may not make a useful drug. But its discovery has pointed researchers toward other possible anti-obesity drugs.

Schmergel says he has not yet worked out how GI will be paid by its partners if its proteins turn out to be useful only as research tools and not as drugs. But GI, Haseltine, and others hope advances in technology will trigger a renaissance in the use of proteins as therapies. "Some people thought the era of therapeutic proteins was over," says Haseltine. But with a flood of new proteins hitting some of the best labs in the country, biotech might see another string of best-sellers.

    Before it's here, it's on the Bloomberg Terminal.