Biotech Grows Up

(2 of 3)

But outside a handful of laboratories, the process still relies more on brute force than strategy. It begins with what's known in pharmaceutical jargon as "target validation"--checking out a protein to see if it plays a role in disease. If so, the protein is subjected to a battery of molecules until one binds to it and exerts a desired effect--say, shutting it down. That binding molecule is then modified in a variety of ways, and the cycle is repeated until a potential drug emerges. This proto-drug is analyzed to ensure that it can be absorbed by the body and isn't too toxic. Only then does it move to clinical trials on humans.

The costs and inefficiency of this approach are staggering. Today's drugs attack just 500 out of an estimated 5,000 proteins that they could be devised to attack. On average, only 1 in 1,000 molecules that bind to a target makes its way to human trials. Of the survivors, just 1 in 5 reaches the market. If failures are taken into account, pharmaceutical companies on average spend $802 million--and 10 to 15 years--to develop each new drug, according to a report released on Nov. 30 by the Tufts Center for the Study of Drug Development in Boston.

This does not bode well for the pharmaceutical industry--nor for patients waiting for a miracle drug. Large pharmaceutical firms such as GlaxoSmithKline and Aventis will each have to launch up to 45 new drugs--not the variations on existing ones that now account for so much of the industry's output (and advertising)--in the next decade if their sales are to grow at an 8% rate, estimates the consultancy Accenture. By 2010 that will mean producing six new drugs a year. Most companies today grind out just one or two a year. "Productivity is the No. 1 issue going forward in the pharmaceutical industry," says Millennium CEO Mark Levin.

Genomics can help increase productivity by ramping up the identification of target proteins and by increasing the efficiency of drug discovery. Says Ben Conway, principal at Boston venture-capital firm Adams, Harkness & Hill: "Genomics was going to be the Holy Grail. Now people appreciate that it's come down to proteomics"--the large-scale analysis of proteins.

CuraGen of New Haven, Conn., is one of the innovative firms that have refined this art. Its proprietary technology--an adaptation of a decade-old technique--allows CuraGen's scientists to put human genes into yeast cells and effectively "fish" for proteins relevant to drug discovery. "We learned at the seat of the inventor of this technology," boasts technology group leader Bruce Taillon, "and showed him what would happen when CuraGen was set loose on it." The company stunned the biotech world in January, when it announced a 15-year, $1.4 billion deal with Bayer to develop drugs against obesity and diabetes. "CuraGen has a mastery of the genome," says CEO Jonathan Rothberg. "We needed a large drug company like Bayer to help us turn that mastery into a product."

This kind of deal is "a model we're likely to see more of as big pharmas and big biotechs try to incorporate these new technologies into their drug-discovery programs," says Ernst & Young analyst Scott Morrison. And in this case it seems to be working. CuraGen has presented Bayer with 24 targets for drug development.

Not all proteomics involves baker's yeast. Millennium uses mass spectroscopy--a way of producing unique signatures of proteins by tearing them apart and weighing the fragments--to detect drug targets rapidly. Its technology has won the company strong alliances with traditional drug firms, including a $450 million anti-inflammatory drug-development alliance with Aventis and a $465 million research agreement with Bayer. Millennium also has a $250 million partnership with Abbott Laboratories focused on metabolic diseases. Last month the pair announced that they were beginning clinical trials on a candidate drug against obesity.

Human Genome Sciences, like many of these companies, hopes to use proteins themselves as drugs--and has six in clinical trials. CEO William Haseltine says HGS does hundreds of experiments on some 10,000 distinct proteins in its hunt for novel drugs. The firm has invested heavily in bioinformatics to get a grip on the mounds of data this process generates (see box).

That investment has paid off. HGS moved a drug against autoimmune disease into clinical trials within 18 months of its discovery. Haseltine has managed his pharmaceutical aspirations astutely, building commercial manufacturing facilities for proteins years in advance--aware that a lack of capacity has hampered some of the most exciting biotech drugs. "It is my belief that if a skill is critical to your success," says Haseltine, "you must build it and control it yourself." But underscoring the risks of drug discovery, HGS's shares dropped 10% last week when it announced that its leading drug in clinical trials is safe but seemingly ineffective.