Genentech's Gamble

(page 4 of 4)

To better match patients with drugs, Levinson took the odd step of hiring away five autoimmune specialists from a single lab at the University of Minnesota last year. They are experts in searching for "biomarkers," molecular signals that reveal the unique makeup of each patient's disease. While the pool of patients for each personalized drug is small, such treatments can be lucrative. For example, only 30% of breast cancer patients carry the markers that qualify them for Genentech's drug Herceptin, yet the treatment is worth $1 billion a year. But autoimmune diseases are more challenging, says rheumatologist Timothy W. Behrens, who was the lead biomarker expert scouted from the University of Minnesota. Lupus, for one, "can involve skin, brain, lungs, heart, kidneys," he says. "Every lupus patient I've ever treated is unique."

The economics of a personalized approach are fragile. Suppose a new lupus drug works for only 20% of patients with that condition. Can Genentech use the small patient pool to justify charging sky-high prices? The answer should be yes, insists Levinson. But Genentech has come under fire in Congress for overpricing its drugs. By nature a private person, Levinson loathes lobbying, yet he recently made several trips to Washington to lecture lawmakers on the high costs of developing biotech drugs. He worries that Congress may make "stupid decisions" to cap drug prices "based on an incorrect understanding of the economics of the industry."

Identifying biomarkers might also help Genentech screen out patients who are susceptible to side effects. In February, 2005, while Genentech was still developing Rituxan for MS, its partner, Biogen, released its own heralded MS drug, Tysabri. It looked like a surefire success—right up the point when patients started to die. The culprit was a rare brain virus that may have been able to infect patients because their immune systems were weakened by Tysabri. "It was so unexpected," recalls Dr. Michael Panzara, chief medical officer of Biogen's neurology business. "We think we're so smart about autoimmune disease. Then something like this happens and we have to rethink everything." Genentech promptly redesigned its Rituxan trials, building in extra patient checkups in hopes that physicians would spot dangerous side effects fast. (Biogen is on the block and is expected to fetch more than $20 billion, largely because of its stakes in Tysabri and Rituxan.)

Early Success

It is still early days for Rituxan in autoimmune disease, but there have been some important milestones. On a Saturday in the summer of 2006, the MS research team gathered in a conference room to review early results of a key Rituxan trial. As they studied MRI scans of patients' brains, they were amazed to discover that the signs of the nerve-ravaging disease had dropped by 91%. "I can forever remember sitting in this room and watching a story unfold that we really hadn't seen before," recalls Craig Smith, a medical director at Genentech.

Rituxan was approved to treat rheumatoid arthritis in February, 2006, and has captured 10% of the market. For patients such as Nancy Kowalski, the sign-language interpreter, the drug has been a lifesaver. She got her first infusion in March, 2007. While the needle was still in her arm she glanced at her hands and couldn't believe what she was seeing. "The swelling was going down." Kowalski has gone back to work.

Genentech will announce key Rituxan data in lupus and MS next year, and it's working on some completely new approaches to autoimmune disease. The mandate, as always, is to consider ideas others might overlook. "There's this tremendous herd instinct out there," Levinson says. "That's a great opportunity, because often the crowd is wrong."

Weintraub is a senior writer for BusinessWeek's science and technology department.