 |
 |
Researchers use 'virtual patients' to study effects of experimental drugs WASHINGTON (March 11, 2002 4:30 p.m. EST) - Bill and Allen both have asthma, but scientists ran into a problem when testing a highly touted experimental drug on them: The drug flunked. It didn't help Allen breathe better, and Bill had a strange reaction that called into question the drug's entire foundation. That was good news, actually. Bill and Allen are "virtual patients," sophisticated computer programs that mimic disease. That the experimental drug failed on them meant maker Aventis Pharma didn't have to watch it fail in tests using real people - as a competitor did. You've heard of "in vitro" and "in vivo" research, studies in test tubes and living organisms. Welcome to "in silico" research, where companies like Entelos Inc. and Physiome Sciences put reams of biological data into supercomputers to create intricate models of human cells, organs, even multiorgan systems that mimic complex diseases. Just as simulators help build better jumbo jets, the hope is that biosimulation will save pharmaceutical companies time and money in creating new drugs - by helping them pick the most promising chemicals to test in people and skip dead-ends. "This is an incredible tool for drug discovery," says Richard Kahn, chief medical officer of the American Diabetes Association. "It is absolutely astounding what this technology can do." Dismayed that "the pipeline is about empty" of promising new diabetes drugs, the nonprofit group just began an unprecedented collaboration with Entelos, hoping to sign up pharmaceutical companies in the hunt for therapies using the California company's new diabetes simulator. Many scientists fear biosimulators are over-hyped, particularly their role in very complex disorders like diabetes and heart disease where much of the basic biology is still a mystery. "I don't think we're anywhere near being able to predict from a few pathways in the human cell ... things that take place in the whole body," says Dr. Jan Breslow, a Rockefeller University specialist on the pathophysiology of heart disease. But simulators can be updated easily, says Jeremy Levin, whose New Jersey-based Physiome Sciences boasts a heart model that goes into irregular beats and customizable simulators. Indeed, some drug giants are spending millions for access to simulators in hopes they'll improve a bad statistic: A third of potential drugs that pass the years of research needed to begin final-stage testing in people still ultimately fail. Scientists' brains simply can't hold all the discoveries about how healthy and diseased cells work, much less envision all the chain reactions sparked by tweaking even one link in the process. But a supercomputer can. Consider Entelos' model of asthma. Company scientists converted findings from thousands of asthma studies into mathematical formulas that a computer can read. The data link the known pathways of the respiratory inflammation of an asthma attack. Then Entelos tested a prominent theory that hindering production of a protein called interleukin-5 would in turn lower airway-blocking cells called eosinophils during an asthma attack. Aventis had created an anti-IL5 drug that worked in animals. Now enter Bill and Allen, the virtual patients. The drug almost eliminated eosinophils in Allen's airways but he didn't breathe any easier. Bill's asthma attack, slightly different because Entelos mimicked Aventis' animal discoveries, went haywire, casting further doubt. Shortly after the computer experiment, competing scientists reported anti-IL5 failed in a human test, too. No one yet knows how well Entelos' new simulator for diabetes, a much more complex disease, will work. But the diabetes association's Kahn says for the first time scientists can predict "if I change X in the liver, what happens to Y in muscle," before tests in people. Even their staunchest proponents don't expect biosimulators to ever replace human testing of drugs, required by the Food and Drug Administration. Computers can't predict what they aren't asked. Physiome, for example, once predicted the blood pressure drug Posicor wouldn't cause a certain irregular heartbeat - but the drug later was yanked off the market for deadly interactions with other medicines. "There is no virtual human," Levin cautions. But with more simulators being developed, the next step is integrating them so if a drug for one organ hurts another, there might be a hint. "It's a wonderful moment for this area." by Lauran Neergaard, AP Medical Writer who covers health and medicine for The Associated Press in Washington. Copyright 2002 AP Online http://www.nando.net/healthscience/v-text/story/297506p-2613625c.html janet paterson: an akinetic rigid subtype, albeit perky, parky pd: 55/41/37 cd: 55/44/43 tel: 613 256 8340 email: [log in to unmask] smail: 375 Country Street, Almonte, Ontario, Canada, K0A 1A0 a new voice: http://www.geocities.com/janet313/ ---------------------------------------------------------------------- To sign-off Parkinsn send a message to: mailto:[log in to unmask] In the body of the message put: signoff parkinsn