Philip Bourne is associate director of UCSD's Protein Data Bank project. Photo: Sam Hodgson
Monday, July 20, 2009 | University of California, San Diego researchers are harnessing the San Diego Supercomputer Center’s raw-number crunching power for a novel technique in drug testing, potentially slashing costs and seeking new uses for old prescription drugs at a time when the struggling biotechnology industry desperately needs to do both.
Retooling an existing drug for a different purpose as scientists at UCSD’s Skaggs School of Pharmacy did recently isn’t a new idea — it’s known as drug repurposing. A famous example is the drug Viagra, which Pfizer originally developed to treat angina, but was repurposed after clinical trials showed its other benefit.
But the banks of computers at the Supercomputer Center are allowing scientists to super-charge that process — rapidly screening drugs against more than 50,000 possible proteins at once. This research could end up crucial to an industry that is faced with a business model that has in many ways crumbled under the weight of the financial crisis.
The days when a biotech could survive for decades on the prospects of one or two home-run drugs are gone. Now companies have to find multiple revenue sources in shorter time frames. This new paradigm is especially important in San Diego, home to a biotech cluster that employs around 44,000 people and pumps upwards of $9 billion annually into the local economy.
“The notion of what we know as drug discovery — looking at one receptor, one drug, one disease — is out the window,” says Philip Bourne, a UCSD pharmacologist and the leader of the research team. Bourne’s lab takes a broader approach by exploring how a drug might interact with other proteins, instead of focusing only on the one it’s intended to target.
Recently Bourne’s team used this technique to find a possible new application for Comtan and Tasmar, drugs used to treat Parkinson’s disease. Their computer model predicted the two medications would bind to a protein the tuberculosis bacterium uses to repair its cell wall. Researchers later verified that both drugs work against tuberculosis in the test tube as predicted.
And because both drugs have already been approved by the Food and Drug Administration once, they would have a shorter and less expensive path to approval for a second use. Many companies own numerous compounds that were never marketed for one reason or another, but could be resurrected for further study.
All of this could be a welcome boon to the industry, said Joe Panetta, the CEO of Biocom, a regional trade group of Southern California life sciences companies.
“If you look back, in the past the process of developing these drugs didn’t really include a rigorous way of looking at repositioning,” Panetta says. “There’s opportunity in these molecules that may not have been explored because researchers at the time lacked the data or the tools to do so.”
Bourne’s lab performs its screenings on a library of 3-D molecular models called the Protein Data Bank, an international databank currently hosted by the Supercomputer Center. The databank, which began in 1971, has grown rapidly in recent years and now stores information on more than 50,000 proteins.
Beyond finding new uses for old drugs, researchers are using the ever-expanding databank and computing power of the Supercomputer Center to ferret out possible side effects of drugs. Finding side effects in the virtual world before they reach real-world clinical trials could save companies hundreds of millions of dollars, Bourne said.
The researchers make their discoveries by looking at things from an “evolutionary standpoint,” Bourne said. Since all life on earth is related, many proteins in different life forms feature the same kinds of characteristics, and often proteins in the same organism — even those that perform different tasks — share subtle similarities.
Drugs work by attaching to specific proteins in cells. But some drugs will latch onto multiple types of proteins, causing unwanted side effects. Through the Supercomputer Center, Bourne’s team can check a drug against tens of thousands of proteins at once.
Through this process they can identify other proteins the drug binds with. These are called “off-target” proteins, and drugs that hook onto numerous proteins are called “promiscuous” or “dirty” drugs.
The ability to predict the off-target effects of a drug is crucial. Pfizer’s anti-cholesterol drug torcetrapib, for example, cost an estimated $800 million to develop — only to be withdrawn when late-stage clinical trials revealed deadly side effects.
“This is very promising with regards to addressing the major bottleneck — the largest cost factor in bringing most drugs to market,” says John Wooley, associate vice chancellor for research at UCSD. “There’s lots more drugs out there that would work if they didn’t have a side effect.”
By quickly identifying adverse effects, Wooley believes, pharmaceutical companies could slash costs and streamline development. And as the amount of available data increases these algorithms will become even more powerful. “The bigger the data set, the better we can address these questions,” Wooley says.
Repurposing is especially important at a time when pharmaceutical companies are short on promising new medicines. “Drug companies are very conservative … at the same time, their pipelines are pretty empty and they’re in pretty desperate straits,” Bourne says.
Computer-based methods do have their limitations, however; while Bourne’s models can make predictions, those predictions still have to be validated in the lab.
“We’re not going to get away from a lab. You have to have an experimental component,” Wooley says. But this type of computer-assisted process, Wooley believes, will become invaluable in the future. “Less and less drug discovery will happen without a computational component. It’s far, far faster, it’s cheaper, it’s more efficient.”
And the ongoing work at Supercomputer Center could be not only useful to the drug industry but revealing as well. As Bourne has explored the complex impacts of drugs in the biochemical jungle of human cells, his findings have given him some unsettling insights.
Many drugs interact with multiple targets in ways that are still incompletely understood, Bourne said, meaning they may have as yet unknown effects. For Bourne this isn’t very reassuring.
“It’s pretty scary,” he says. “We know very little about the drugs we take.”
Jonathan Parkinson is a San Diego-based freelance writer. Please contact him directly at firstname.lastname@example.org with your thoughts, ideas, personal stories or tips. Or set the tone of the debate with a letter to the editor.
Value investigative reporting? Support it. Donate Now.