Show Me The Data: A Nobel Lesson In The Process Of Science

The recent award of the Nobel Prize in physiology or medicine to Stanley Prusiner, a professor of neurology, virology, and biochemistry at the University of California, San Francisco, for his work on prions sharpened the focus on the concept of data-driven ideas in science. In contrast is the challenge to the well-accepted idea that HIV causes AIDS by Peter Duesberg, a professor of molecular and cellular biology at the University of California, Berkeley. The two hypotheses are iconoclastic, ear

Dec 8, 1997
Barry Palevitz

The recent award of the Nobel Prize in physiology or medicine to Stanley Prusiner, a professor of neurology, virology, and biochemistry at the University of California, San Francisco, for his work on prions sharpened the focus on the concept of data-driven ideas in science. In contrast is the challenge to the well-accepted idea that HIV causes AIDS by Peter Duesberg, a professor of molecular and cellular biology at the University of California, Berkeley. The two hypotheses are iconoclastic, earning their proponents the title of "renegade," yet there the similarities end. One hypothesis, the one that earned the Nobel, succeeded based on accumulated positive evidence; the other appears to have waned, for the opposite reason.

Prusiner's pursuit of an unpopular hypothesis-that a type of protein, or prion, without a nucleic acid, could transmit a disease that essentially destroys the brains of certain mammals-elegantly illustrates the cycle of inquiry at the heart of the scientific method. Observations lead to hypotheses that pose testable predictions. Experiments test those predictions and raise yet other questions, prompting new hypotheses and experiments.

Having lost a patient in 1972 with Creutzfeld-Jakob disease (CJD), a spongiform encephalopathy, Prusiner set about the task of discovering its cause. Sparked by the work of Tikvah Alper's group at Hammersmith Hospital in London, which demonstrated that a nucleic acid was not implicated, Prusiner and colleagues purified from animal brains infectious protein particles he named prions. Years later, Prusiner's many experiments using transgenic mice would confirm that the protein indeed devastates the brain.

If the idea of an infectious protein wasn't bizarre enough, Prusiner discovered that the prion protein has the same amino acid sequence in healthy as well as sick individuals. Could a conformational difference explain the apparent paradox of two types of identical prion protein? Experiments showing that the deranged protein is relatively protease-resistant compared with its normal counterpart suggested such an explanation. But just as important, could the "evil twin" prion spread damage, converting-or subverting-its "good twin" into the pathogenic form, as indicated by injections of affected brain matter into healthy animals, and by in vitro studies? And how can prions be both inherited and infectious, as encephalopathies such as CJD in humans appeared to be? Prusiner's group addressed these questions, and data they obtained convinced others to join the fray.

Despite convincing recent evidence that bovine spongiform encephalopathy and the new variant CJD are one and the same (A.F. Hill, Nature, 389:448-50, 1997; M.E. Bruce, Nature, 389:498-501, 1997), a possibility that Prusiner discussed in a January 1995 article (S. Prusiner, Scientific American, pp. 48-57, 1995), many questions remain. For example, we still do not know precisely how a good prion protein goes bad, although molecular modeling suggests a possible conversion of alpha helices to beta pleated sheets in a crucial part of the protein's conformation. And not everyone in the scientific community is convinced that prions are the sole cause of the spongiform encephalopathies. As such, a report in Science (N. Williams, 278:31, 1997) referred to the "so-called prion protein." But Prusiner is undaunted. When a reporter asked if his winning the Nobel would squelch criticism of his not-yet-complete work, he replied, "I don't think prizes stop criticism. Data does."

The saga of Peter Duesberg differs from that of Stanley Prusiner. Ten years ago, Duesberg, a respected retrovirologist and member of the United States National Academy of Sciences, offered what seemed like an incredible hypothesis: HIV does not cause AIDS (P. Duesberg, Science, 241:514-16, 1988). Instead, he argued, the virus is a marker, a harmless hitchhiker along for the ride in immunosuppressed individuals. Duesberg further claimed that the disorders manifested as AIDS are fomented by lifestyle-related factors, including abuse of intravenous drugs and nitrite inhalants and even using AZT, a drug that is given to treat AIDS.

Clearly, Duesberg was bucking conventional opinion. The scientific community had by this time almost unanimously agreed that HIV causes AIDS after extensive studies in epidemiology, virology, immunology, and a host of other disciplines. Federal grants were beginning to fill the coffers of AIDS researchers. It came as no surprise that, as was the case with Prusiner, the biomedical community gave Duesberg's hypothesis a rather chilly reception. And because of the idea's controversial nature and its proponent's highly respected background as the discoverer of oncogenes, Duesberg's view attracted a lot of attention from the scientific and popular media.

No stranger to controversy, Duesberg, like Prusiner, persevered. Science magazine devoted eight pages to "the Duesberg phenomenon" (J. Cohen, Science, 266:1642-9, 1994), while The Scientist provided further coverage the following year (P. Duesberg, The Scientist, March 20, 1995, page 12). An entire issue of the journal Genetica (95[1-3]:1-193, 1995) addressed alternative hypotheses regarding the cause of AIDS, with Duesberg a major contributor. Letters to the editor and editorials in Science and Nature flew fast and furious. Duesberg received support in the scientific community, some if only to defend his right to be heard. The Sunday Times of London rallied to his defense. And as late as 1996, Duesberg authored a book on the subject (Inventing the AIDS Virus, Washington, D.C., Regnery Publishing, 1996), which was prominently reviewed in the New York Times (J. Osborn, New York Times Book Review, April 7, 1996) and here.

But what of Duesberg's hypothesis today? It appears that the uproar has died down, attention redirected. Why? While many scientists regarded Duesberg's idea as untenable and therefore simply ignored it, the real reason for the silence lies here: Few if any data support it.

Unlike the case with prions, predictions based on the Duesberg hypothesis-such as demonstrating a non-HIV cause of AIDS in people with hemophilia who received contaminated blood products-did not pan out. This particular population helped nail HIV as the cause of AIDS early on because people suffering from hemophilia generally do not have the behavioral risk factors, such as drug abuse, that Duesberg thinks contribute to or cause AIDS. Duesberg argued that non-HIV contaminants of factor VIII pooled from donated blood could cause AIDS when used to treat hemophilia. But this idea fell flat as donated blood began to be screened for HIV and AIDS cases among people with hemophilia plummeted.

Duesberg's hypothesis was perhaps more plausible a decade ago. But as new information on the nature of HIV has accumulated, the idea has not stood the test of time. Why would protease inhibitors work if HIV were not crucial in AIDS pathogenesis? The discovery of HIV coreceptors reveals a new chapter in the evolving epidemic-why and how some individuals resist infection (M. Dean et al., Science, 273:1856-62, 1996; M. Samson et al., Nature, 382:1722-5, 1996; R. Liu et al., Cell, 86:367-77, 1996). Might coreceptor variants be the cofactors some investigators have sought to explain previously mysterious aspects of AIDS etiology (B. Goodman, The Scientist, March 20, 1995, page 1)? And replacement of HIV antibody testing with the more direct detection of viral load (HIV RNA) may have countered Duesberg's long-standing criticism of the limitations and significance of antibody tests (D. Ho, Science, 272:1124-5, 1996). All in all, mounting evidence has strengthened the links between the virus and crippled immunity. As of now, without supporting data to the contrary, HIV is still the prime suspect in causing AIDS.

Barry A. PalewitzRicki Lewis
Barry A. Palevitz is a research professor of biological sciences at the University of Georgia in Athens (palevitz@dogwood. Ricki Lewis is a textbook author and geneticist based in Scotia, N.Y. (
The contemporary tales of Prusiner and Duesberg offer meaningful lessons in the process of science. Hypotheses and theories, even revolutionary ones, stand or fall on the weight of data. True, it may take a lot of time for the final verdict to come in-witness the eventual acceptance of plate tectonics and recent discoveries confirming some of the more counterintuitive predictions of quantum mechanics. Both the Prusiner and Duesberg hypotheses were widely disseminated and discussed, and their predictions addressed by scientific methods. In Prusiner's case, efforts to find a nucleic-acid-based alternative failed, while data supporting an all-protein-based disease entity accumulated. In Duesberg's case, the reverse occurred: Positive evidence for his hypothesis was absent or unconvincing, while data in favor of HIV grew. Duesberg steadfastly maintains that he cannot do the critical experiments for want of grant support. But others in the community have not been idle. Definitive evidence that HIV does not cause AIDS and that something does remains elusive.

The hypotheses of Prusiner and Duesberg each generated intense interest, and even hostility. Some could argue that such hostility is a good example of the subjective, biased nature of science that seeks to protect vested interests. Yet, one hypothesis prevailed and the other withered because, in the end, to borrow from the film Jerry McGuire, scientists insist, "Show me the data."