(Reprinted from the December 1986 issue of Discover magazine with permission.)

The Game of the Name is Fame. But is it Science?

Stanley Prusiner, "discoverer" of prions

By Gary Taubes

Prion is a terrific word. It's snappy. It's easy to pronounce. People like it. It isn't easy to come up with a good word in biology. One hell of a lot of bad words people introduce get thrown away.

--Dr. Stanley Prusiner, "discoverer" of the prion

Naming something before you discover it is a risky business. For starters, you have to couch your definition with great care, so that it will fit whatever it is that you find, when and if you actually find it. In public, you have to deal confidently with your fellow scientists, who are likely to take umbrage at your nomenclatorial presumption. In private, you have to hope that as the data come in they fit your definition, or at least that your definition can be made to fit the data.

Yet if you play your public relations right, the press will make you famous for discovering something you haven't yet found. Fame, in turn, can bring in money that might have gone to your competitors, whom you can now outspend in the race for the actual discovery. In the long run, you may be given credit for the discovery--even if it's made by someone else.

So it was that the word prion entered the scientific lexicon in April 1982, courtesy of Stanley Prusiner, a neurologist at the University of California at San Francisco (UCSF). Prusiner, then 39, was working on some esoteric diseases that were known as slow virus diseases and were thought to be caused by viruses.

But Prusiner told the world they were caused by prions (pronounced PREE-ons), short for proteinaceous infectious particles. (Logically, the word should have been proin, but Prusiner, with melodious intent, transposed the o and the i to make it more terrific.) He hadn't simply put forth what he might then refer to as his "prion model" or his "prion hypothesis" (in the way that in physics, for example, the fundamental particles known as quarks entered the world modestly as "the quark model" and stayed that way for a decade until experimental proof was obtained). Prusiner had unequivocally and pre-emptively renamed the very subject of his research, as well as that of a few dozen other scientists. From then on, he was no longer examining the nature of slow viruses; his work was about prions.

The prion, as Prusiner loosely defined it, might indeed be a virus in which he hadn't been able to find the genes that every virus must have. Or it might be some kind of rogue protein--hence, prion--that goes about its nefarious business without benefit of a gene. If the latter turned out to be the case, the prion would be the most remarkable form of life on the planet, the only one that could multiply without genes. This would require a drastic rewriting of the dogma of molecular biology, a redefinition of the meaning of life itself.

Such heresy is often written off as quackery. But the slow virus diseases are an odd bunch and tough to get a handle on. They're not a threat of the magnitude of cancer or AIDS, but the damage they do looks remarkably similar to that done by Alzheimer's, which afflicts 2.5 million Americans. Prusiner's prion represented a rare, potentially fruitful lead to what might be going on in both slow virus diseases and Alzheimer's.

In the years since he named it, Prusiner has published, alone or with collaborators, more than 20 papers reporting experimental results in support of his prion. He has said that prions or prion-like agents may be the cause of Alzheimer's, as well as a range of other afflictions, from rheumatoid arthritis to Parkinson's disease. He has also said that the prion may represent a critical link between viruses and genetic disorders.

In the twelve years since he began his research, Prusiner has gone from physician in a hospital to full professor at UCSF, perhaps the best medical school in the country. He now heads a laboratory that receives well over a million dollars a year in federal grants. With this money, he's supporting a dozen scientific collaborations, some of them with researchers who are among the most respected in medicine.

Prusiner and his unorthodox prion have taken a field of research that seemed unlikely ever to extend beyond the concerns of the victims of slow virus diseases--sheep, cannibals, and the one in a million people stricken with Creutzfeldt-Jakob disease (CJD)--and turned it into a big-money, high-profile enterprise. He and his research have been written up in everything from the New York Times (seven times), Scientific American (three times), and DISCOVER (three times) to the Reader's Digest (headline: "Killer Diseases from the Dawn of Time").

In recognition of Prusiner's work and the potential medical importance of the prion, his lab won a $4 million congressional award in 1985. It would be used, according to a UCSF press release, "to determine the structure of prions and how they cause disease."

Yet the prion remains a mystery in more ways than one--perhaps the most controversial being why the government gave $4 million to a scientist whose work is disputed by virtually every other researcher in his field save his immediate collaborators.

From the day Prusiner announced his "discovery," those researchers have suggested that his heresy wasn't in his prion, but in his premature claim that such a bizarre creature was needed to explain slow virus diseases. Since then, they've only grown more incredulous in the face of Prusiner's apparent tour de force. As one of his own former researchers puts it, "The whole thing is so infuriating from a scientific standpoint. The interpretation of the data has gotten so far from that data, it's hard to believe how it could've happened." Although Prusiner's critics have repeatedly contested his science, they've been virtually ignored. It's as if the decision had already been made: the prion had been hailed as a finding of Nobel Prize caliber, and that was that.

The story of the prion isn't a story of science as a layman might expect it to function. It's not so much a story of rogue proteins and obscure diseases as of how a single set of results can be interpreted by different scientists in antithetical ways, and how one scientist with a flair for public relations can overcome seemingly valid criticism and dominate his field. It's also a story of how the traditional checks and balances of science--peer review and experimental replication--can be rendered impotent in the name of progress. This isn't a new phenomenon in science; the story of the prion is merely an extreme example.

At the heart of the matter are the agents of the slow virus diseases, which have proved so elusive that the bottom line on them is, as one biologist puts it, that there is no bottom line. The diseases lead to dementia and death. They have no cures, and they aren't your everyday afflictions, unless you happen to be a sheep. Scrapie, identified in the seventeenth century, is so insidious that until recently if a single sheep was afflicted the entire flock had to be killed.

Kuru, the first human malady to be labeled a slow virus disease, came to the attention of Western doctors only in the 1950s; it was the primary cause of death among the Fore, a cannibalistic New Guinea tribe. By 1966 National Institutes of Health (NIH) researchers Carleton Gajdusek (who later won the Nobel Prize for his work) and Joe Gibbs had proved that, like scrapie, kuru was infectious: chimps inoculated with brain tissue of human victims took sick with kuru-like symptoms and died. The spread of the disease was attributed to cannibalism.

During the 1960s Gibbs and Gajdusek inoculated tens of thousands of lab animals with tissue from victims of amyotrophic lateral sclerosis (Lou Gehrig's disease), multiple sclerosis, Alzheimer's, and forty-odd other neurological disorders on the chance that they, too, might prove infectious. The only one that took was CJD, which kills about 225 Americans a year, fewer than traffic accidents do in two days. Its symptoms are nearly identical to those of kuru and scrapie.

Hard facts about these diseases are few and bizarre. First, whatever the agent is, it seems almost impossible to kill. "God in the guise of a virus" is how Paul Brown, an NIH virologist, describes it. Sterilization procedures that would kill any other virus several times over, from an hour in an autoclave to a decade in formaldehyde, seemed to affect the slow viruses not at all.

And whatever the agent is, it replicates, and it does so just as a virus would. A mouse injected with what is thought to be a few infectious particles of scrapie will provide enough infectious material in one year to kill a million of its kin. But if the agent is a virus, it takes its time doing its work and apparently leaves the immune system unalerted. Regardless of how or when CJD enters a human victim, symptoms won't usually appear until the sixth or seventh decade of life (hence the word slow).

Finally, tracking down the agent has been an expensive and seemingly endless task. The only way to certify the presence of the disease in man or animal is to wait for the victim to die, "puree" its brain, and inject it into a lab animal to see if it dies. Preliminary experiments to isolate the culprit, which in a routine virus might take days or weeks, could take years in the case of slow virus diseases.

Slow virus disease research was in the doldrums from the time of Gibbs and Gajdusek's successes to Prusiner's appearance on the scene in 1974. Researchers never got close to pinning the agent down. For all they knew, it could have been something in the air, it could have been phlogiston, or kryptonite, or even God's will. When Prusiner arrived, as Brown says, "if for no other reason than his prickly and aggressive nature, he leavened the field the way you leaven bread."

Prusiner grew up in Cincinnati and went to medical school at the University of Pennsylvania, where he spent his spare time doing biochemistry research. By the time he got his M.D., he says, "I was cocky enough to think I could do research."

In 1972, Prusiner began a residency at UCSF in neurology--the central nervous system, he says, was the "last great frontier of medicine." When one of his patients died of CJD, he talked to a neurologist about slow virus diseases, and the man told him what was known, which was little, and what was not, which was virtually everything. That was the kind of challenge any cocksure researcher would jump at.

"I started reading about scrapie," says Prusiner, "and it became clear that this was a wonderful problem for a chemist. It had been attacked by pathologists, physicians, veterinarians. Those who tried to unravel the chemistry of the disease hadn't taken a very careful approach. I spent much of my time thinking about how I was going to do this problem. When I finished, I set up a lab here. I got some money from the neurology department, but not a lot. This subject needed a lot of money from the very beginning, hundreds of thousands of dollars."

When Prusiner went to NIH for money, he came up empty-handed. As he tells it, "They said 'Who the hell are you? You know something about enzyme chemistry, but you know nothing about virology, and nothing about scrapie, and you never trained with anybody.' " So he took a course on virology, to counter the NIH's first complaint. For the latter, he collaborated with William Hadlow, a pathologist at the NIH's Rocky Mountain Laboratories, who had first pointed out the similarities between kuru and scrapie 15 years earlier.

With Hadlow and virologist Carl Ecklund, Prusiner set about with almost monomaniacal zeal to isolate the scrapie agent--and ran head on into the reality that had beaten his predecessors. Whatever the agent, isolating it would require, as Prusiner later estimated, as many as 250,000 lab mice and a few lifetimes. To purify or isolate the agent they had to cull it from all the proteins, DNA, and RNA normally present in the brain. They'd have to divide infected brain tissue from a mouse afflicted with scrapie into fractions, inject the fractions into mice, and wait a year to see which ones caused the mice to die. Then they'd have to subdivide the most virulent of those fractions into a dozen smaller ones to see which one of those contained the agent. Wait another year to find out. And so on, until they were down to virtually nothing but the agent in their test tubes. "We rapidly went through our ten thousand mice," says Prusiner, "and even if we were handed money on a silver platter, we couldn't go on like that."

In 1978 the NIH decided to "reorient" slow virus research at Rocky Mountain Labs. They cut Prusiner and company off. Joe Gibbs says they wanted new blood involved in the work and to force Hadlow, Ecklund, and Prusiner to publish whatever data they had.

It turned out to be Prusiner's big break--a classic story of triumph through adversity. Prusiner came up with some better ideas for doing scrapie research. He switched from mice to hamsters, in which the onset of the illness is twice as fast. He began using a test called an incubation-time assay, which had been used for scrapie research by British scientists since the 1960s. Prusiner's refined incubation-time assays could be scored after a couple of months, when the animals took sick, rather than after a year, when they died. With these two innovations, Prusiner increased the efficiency of his experiments 90-fold. "Over the next two or three years," he says, "we did more experiments on the biochemistry of scrapie than everyone else in the history of scrapie combined."

By 1981 Prusiner claimed that he'd achieved 100-fold purification of the scrapie agent--meaning he had derived a preparation that had as much infectivity as a diseased brain but 99 per cent less of the extraneous junk. Then he threw in enzymes that would digest any proteins, and found that they killed the infectivity of the sample, rendering it harmless. A protein, he concluded, must be an important part of the agent. When he tossed in enzymes that specifically go after nucleic acids, the molecules in genes, he found no change in infectivity.

"We used at least five different techniques to show that a protein was necessary for infectivity," says Prusiner. "Then we used five different techniques to look for a nucleic acid. We couldn't find any. So we introduced the term prion. Then the modern era of all this begins."

Prions, as Prusiner defined them in his April 1982 Science article, are "proteinaceous infectious particles which are resistant to inactivation by most procedures that modify nucleic acids." That didn't rule out a virus, but it leaned strongly toward an infectious protein, the heresy. The press took to it.

The first article appeared in the San Francisco Chronicle nearly two months before Prusiner's scholarly paper in Science. The headline read "Tiny Life Form Found." As Prusiner tells it, "They put my picture and prions in the upper left-hand corner of the front page on Friday. Reagan was on the right. And everyone in the world played catch-up on Saturday. The New York Times was very upset. So was everybody else. That kind of thing did more than anything I could ever do. The prion became a household word among biologists immediately. They didn't even have to read Science."

In contradiction to Prusiner's claim that the press was responsible for the prion's sudden fame, Paul Bendheim, a post-doc with Prusiner at the time, says that Prusiner "rammed that word down the throats of everybody in that laboratory and in the world." Bendheim's view is shared by Dave Bolton, who also did research for Prusiner. (They each spent three years in Prusiner's lab and came up with critical discoveries for him.) Both now work for the competition, the Institute for Basic Research in Developmental Disabilities (IBR) on Staten Island.

Bendheim and Bolton speculate that Prusiner coined the word prion as much to raise money for his research as for scientific reasons. Although Prusiner denies the charge, they also say, as do other former colleagues, that he hired fund-raising experts--something almost unheard of in academic science--to help convince private foundations and the public of the importance of his research. "Stan discussed this with us," says Bolton. "He said 'Look, this whole area's getting lost in a muddle of slow virus this and unconventional that and a whole bunch of other things. If we coin a new term for it, and go out and tell people of the potential link to Alzheimer's, we're going to draw people's attention to this. And we're going to get money.' " (A link between Alzheimer's and slow virus diseases had been considered for decades but never proved.)

One other factor may have been behind Prusiner's prion: the slow virus field had already yielded one Nobel (Gajdusek's), and it's a given that another will go to the discoverer of the agent itself. "It didn't escape Stan at all that this thing had Nobel Prize-winning possibilities and could shower him with personal glory," says another of Prusiner's former post-docs.

Frank Masiarz, a biochemist and Prusiner's second-in-command between 1978 and '82, says he quit after Prusiner decided to publish the prion article. "We were co-writing that," says Masiarz. "I wanted it to be a very critical overview in terms of the possibilities for the structure of the agent. By creating the name prion, he clearly wanted to push the entire interpretation in the direction of a protein-only agent. I said there's no point in creating a name for something that we don't even know exists yet. But Stan's ego is unusually important in the way he interprets the data. He tends to jump to conclusions, to give less credence to facts that would discount his preferred interpretation of the results."

Prusiner's competitors knew nothing of the prion politics. They did know that Prusiner had taken an idea that had been around for years and made it front-page news. They weren't happy. That there might not be a gene in slow viruses had been raised as early as 1967 by British researchers. Prusiner had actually based part of his claim on the British experiments, which had used radiation to show that the agent was much smaller than other viruses and too small for a gene of any consequence.

His critics indignantly pointed out that Prusiner had made no discovery: he'd found no gene, but that didn't mean there wasn't one. As Masiarz says, "I've dealt with a lot of negative results in my day. You don't know if negative results mean that the effect doesn't exist, or that you've simply been unable to measure it in this situation." In fact, they had equally compelling circumstantial evidence that there was a gene (i.e., the agent multiplied), as well as precedents among other known viruses, to suspect that Prusiner was simply missing the agent, just like everybody else.

Among the most convincing evidence for a gene was that produced in the '70s by Alan Dickinson, George Outram, and Richard Kimberlin, of the British Neuropathogenesis Unit in Edinburgh. They had isolated different strains of the scrapie agent that created different forms of the disease when injected into animals. Apparently, something in the agent was controlling what happened in the animal, and doing it with great precision. In traditional molecular biology, only a gene could do that.

Dickinson, Outram, and Kimberlin had even proposed their own slow virus model, the virino: a small piece of nucleic acid protected by a protein made by genes in the host, rather than genes in the agent, as was the case with conventional viruses. Their virino model seemed to fit Prusiner's data as well as his prion did, and the rest of the data in the field even better. "The biology of scrapie tells us the damn thing has a genome," says Kimberlin. "To try to create models of an infectious protein with which you can encode strain variation is damn difficult. And Stan Prusiner, bless his heart, never even tried--which is wise, because you can't do it." ("Of course I try," Prusiner responds. "But there's no point in writing voluminous amounts about what you think may or may not be happening.")

Not surprisingly, editorials in the British journals Nature (written by Kimberlin) and the Lancet came down hard on Prusiner. They likened him to biologists in the 1930s who had wrongly claimed that viruses were only proteins--"easily seen as the type of half-truth that puts the cart before the horse," said the Lancet. Both publications insisted that Prusiner had elucidated nothing with his prion. He had only confused matters.

One anonymous researcher even took some measure of revenge on Prusiner's premature re-appellation by writing a sequence of satirical limericks, which made the rounds of the field. A sampling:

There was a young turk named StanWho embarked on a devious plan. "If I simply rename it, I'm sure I can claim it," Said Stan as he pondered his scam.

"Eureka!" cried Stan, "I have found it. Well . . . maybe not actually found it. But I talked to the press Of the slow virus mess And invented a name to confound it!"

In December 1982 Prusiner reported that he'd pushed the purification of the scrapie agent even further. Although he'd been unable to identify a discrete protein or gene in his purified preparation, when he threw in a detergent enzyme known as proteinase k, it cleaned up all the extraneous odds and ends and left single protein intact. Prusiner called it the prion protein, or PrP. He said it seemed to be related to the agent. It might even be the main part of the agent. Maybe even all of it.

The result impressed medical researchers in general, but had a lesser effect on Prusiner's colleagues in his field. They pointed out that to identify his prion protein, Prusiner had first had to denature it--a process that required pouring detergents into the preparation and boiling it--and while doing so, had rendered the preparation harmless. Since the stuff was no longer infectious, he had no way of telling what else he had killed in the process--such as the virus itself.

In December 1983, Prusiner was in the press again. This time he did what no one else had ever done, although many had speculated and tried. He linked scrapie, and all slow virus research, firmly and publicly with Alzheimer's. With George Glenner, an Alzheimer's expert at the University of California at San Diego, he reported that his PrP molecules aggregated to form infinitesimal rods. Perhaps more important, clusters of these rods appeared to be almost identical to amyloid plaques, which appear in the brains of Alzheimer's victims. Prusiner said that if prions cause scrapie, and if prion rods and amyloid are the same, then perhaps Alzheimer's researchers should consider that amyloids aren't a trivial waste product of the disease but are composed of the agent itself--which might well be a prion.

"It is an astounding finding," Prusiner told the New York Times, "because we never would have dreamed that amyloid and prions are the same. The implications of the findings may be enormous." He then suggested to the Times that the discovery could eventually lead to the prevention and treatment of Alzheimer's.

This time it was Glenner who was left to tell the press that they had no proof that the proteins in Alzheimer's amyloids and those in scrapie were the same. (They later proved not to be.) He told one reporter at the time, "I have the greatest respect for Stanley, but he wanted to get in the press fast. I don't think it should have been released at all." Some researchers speculated that Prusiner seized the opportunity to go to the medical and lay press with what was probably a trivial link between the two diseases to promote the potential of his research for Alzheimer's, which had recently become the target of massive federal spending.

In fact, once again Prusiner had apparently slighted orthodox medical research in favor of his prions. Amyloid plaques are found in a variety of diseases, and in many were known to be composed of the body's own proteins. Prusiner could have interpreted this to mean that the protein in his prion rods was also produced naturally by the body (which indeed turned out to be the case) and was therefore unlikely to be part of the scrapie agent. He preferred the racier scenario.

Prusiner's hypothesis also indirectly took care of his toughest competitor--Pat Merz, a researcher at IBR, a mother of two and part-time housewife with a bachelor's degree in chemistry and no more. "I'm way out on a limb in this field as a woman without a Ph.D.," she says. She's also, in the opinion of University of North Carolina biophysicist Bob Rohwer, "one very good electronmicroscopist."

In 1978, Merz had identified minute fibrils in the brains of mice with scrapie that couldn't be found in healthy mice. By 1980 she had found similar fibrils in a variety of animals infected with a variety of strains of the disease, yet never in healthy animals. In 1981 she and her collaborators published the finding of what she called scrapie associated fibrils (SAF)--the first report of a structure specific to the disease. It predated Prusiner's prion paper, in which it wasn't mentioned, by a year, and his claim of the discovery of prion rods, which seemed to be identical to SAF, by nearly two.

Most scientists studying slow virus diseases credit Merz with the first breakthrough in the modern era of research. In their papers they refer to SAF and cite Merz's work. Prusiner refers to prion rods and cites his own. He steadfastly refuses to admit that prion rods and SAF are the same. "Other workers have suggested that these prion rods are related to longer fibrils," he wrote in a March 8, 1984 New England Journal of Medicine article that otherwise ignored Merz's work. Actually, other workers suggest that the two are one and the same. Says Brown, for instance, "I've always been tempted to put up pictures taken from Pat's work and from Stan's, and to ask Stan to identify the prion rod. He couldn't do it. Nobody could."

But Prusiner also says that Merz stated plainly that SAF aren't amyloid. (In fact, her papers say that SAF are amyloid-like, although apparently not Alzheimer's amyloid plaques.) And because he proved that prion rods are amyloid, they therefore can't be SAF. Paradoxically, he then insists that neither SAF nor prion rods per se exist in the brain, but are created in the process of isolating the scrapie agent. "The prion rods are an artifact of the detergent extraction," he says, "but they faithfully reproduce what you see in the tissue, which is the amyloid filament."

Even Prusiner's collaborators have trouble with this rationale. "The conditions under which one forms these structures have a lot to do with what they look like," says Dave Kingsbury, a collaborator of Prusiner's who now works at the National Science Foundation. "It's always very dangerous to try to extrapolate in vitro manipulations to in vivo observations."

Bolton criticizes Prusiner's interpretation on different grounds: "You can sum it up quickly. He doesn't want Pat's discovery of SAF to be significant. I think he's so consumed with this idea that he's now gone beyond what he can base in fact." (When Prusiner was told that both his former researchers, Bolton as well as Bendheim, disagreed with him on the SAF-prion rods issue, he simply replied, "Students are sometimes rebellious. That's fine. They're lovely people.")

While Prusiner continued to release findings supporting the heresy of the scrapie agent, the anti-prion forces were busy trying to disprove it. At the forefront was Bob Rohwer, who in 1976 had been enticed by Gajdusek into doing slow virus work at the NIH. "At that time the scrapie problem offered the single most incredible challenge to the central paradigms of molecular biology," says Rohwer, who's now at the University of North Carolina. "The molecular biology story was unfolding in a very understandable way, but this stuff didn't seem to fit the pattern. Being trained at Caltech as a skeptic, I thought I'd apply skepticism to the problem."

In a series of papers, Rohwer professed to show that Prusiner's evidence that the agent was bizarre, immortal, too small to contain genes, and apparently just protein was full of loopholes. He first repeated 30-year-old studies that had seemed to indicate that the elusive agent was virtually immortal, and found that it wasn't quite so hardy as they'd assumed. "In fact," he says, "99.9999 per cent of the virus is actually killed almost instantly upon exposure to conventional autoclave temperatures, and the vast majority is destroyed by boiling, Clorox, formaldehyde, and a few other things."

He then took on the 1967 radiation studies that were Prusiner's strongest evidence that the agent was too small to contain a gene of any consequence. Rohwer re-analyzed the old data, comparing them to data on other viruses that had been similarly measured but subsequently sized exactly with molecular biology techniques. His result: not only was the agent not that small compared to known viruses, but it also had to be at least as large as some. "That gives a very, very, very unprion-like picture of the agent," he says.

Rohwer then went after Prusiner's methods. He reiterated one of the few undisputed facts about the scrapie agent: it's sticky. It sticks to cell membranes (which is probably why it's so difficult to isolate from the brain), and it sticks to itself. No one could accurately estimate the amount of infectivity in a brain if what was assumed to be one infectious unit was actually 1,000 stuck together.

Furthermore, although Prusiner's incubation-time assay may have been quicker and cheaper than traditional methods, it was also "fantastically less accurate," says Rohwer. The British researchers claim to have known this since the 1960s, and give it as the reason for their much more cautious use of the assay. The supposed purifications of the agent could be off by a factor of from 100 to 1,000. What Prusiner called 3,000-fold purified, for instance, might actually be only 30-fold. In which case, says Rohwer, "you don't have any purification at all. You've just deluded yourself."

Prusiner does admit that Rohwer has a point--"I'm the one who has continuously said, look, these are the limits of precision of the assay." But then he adds, "The proof of my approach is that we've made an enormous amount of progress here."

If Rohwer was right--and his work was published in reputable journals, just as Prusiner's was--the amount of progress hadn't been enormous. The scrapie agent might simply be a virus, or a virino. And if Prusiner's preparations were much less pure than advertised, no one had any way to prove that it wasn't. Prusiner's PrP could just be some harmless piece of brain cell that tagged along with the agent whenever he tried to purify it.

These possibilities didn't seem to slow Prusiner. He continued to reap laurels, money, and headlines with every paper. "People outside the field read Prusiner," says Brown. "They don't read the twenty years of literature. And they say everyone inside the field screaming at Stan is biased because he's this successful young Turk. It's like McCarthy saying if you don't wave an American flag, you're a communist."

For example, when Laura Manuelidis, a neuropathologist at Yale, wrote to the New England Journal of Medicine criticizing Prusiner's scientific interpretations, the reaction of some researchers was to question Manuelidis, not Prusiner. "Her letter was vicious," said one Alzheimer's expert at the Albert Einstein College of Medicine. "It's typical of these people. They hate him. They're watching him run away with what they spent their life trying to do--and failing to do."

The strong support for Prusiner can be linked to two factors. First, Prusiner's talks at conferences, which some researchers find to be near religious experiences. "They're a repetition of the notion that the prion is the infectious agent and therefore all data must fit, and any data that don't fit better be ignored," says Ashley Haase, a former Prusiner collaborator now at the University of Minnesota. "What happens to people who have to listen to this stuff is that they come away with the impression that the slow virus problem has been solved."

Second, any substantial peer review in the field ceased with Prusiner's first prion paper. From then on, when Prusiner submitted articles to journals, he recommended they not be refereed by his competitors--and few were. His competitors, in turn, suggested that their work not be refereed by Prusiner. One journal, Cell, for whatever reason, published only papers from Prusiner and his colleagues, whereas Nature, for instance, published mostly papers from the competition.

Although reviewers' comments are sent to authors unsigned, Bolton, Bendheim, Brown, Rohwer, and others claim they've received negative reviews so vehement that they could only have come from Prusiner or someone in his lab. "I've reviewed a lot of papers from scientists at Staten Island, Yale, and the rest," says Rohwer, "but we all recommend each other as reviewers. Frankly, I've had papers reviewed by Prusiner--or at least I think I have. You can tell his reviews right away because they start out by saying you didn't use the word prion for this, that, and the other thing. Then they say you didn't cite this Prusiner paper and that one. And the papers are always recommended for rejection. That's apparently the experience of everyone else as well."

Prusiner doesn't question the need for peer review, but he no longer seems to consider his competitors his peers. He claims that because he publishes in "the best journals in the world," his papers get better scrutiny from scientists in related fields than they possibly could by those in his own. "Why should I have to put up with their preconceived notions about what's happening?" he says. "These people don't publish in those journals, because their papers get rejected, and they don't get rejected by me." (In fact, his competitors' papers have been published in Nature, Science, and the New England Journal of Medicine, among others.)

On one recent occasion when Prusiner did referee a competitor's paper, he not only recommended that it be rejected but tried to supplant it with work of his own. In January 1985 he and his collaborators suggested in the New England Journal that prion proteins might be used as a diagnostic test for Creutzfeldt-Jakob, while reporting that they had found the protein in the brains of two CJD victims but not in the brain of a patient with another disorder. (Two years earlier, Merz and collaborators had reported in Nature that they had found SAF in CJD brains. Still, in a UCSF press release entitled "Prions Shown to Cause Rare Human Senile Dementia," Prusiner said, "This is extremely important, because we've taken ten years of work on scrapie and shown that it applies to humans.") In April 1985, Brown and his NIH collaborators sent a paper to the Journal extending Prusiner's work to 21 CJD brains and 30 controls, giving Prusiner full credit for the idea.

The Journal asked Prusiner to referee the paper. He suggested rejecting it on the grounds that it was confusing and not well controlled. And, as he wrote to the Journal, "a question of originality must be addressed. Clearly, the paper does nothing but increase the number of cases from 2 to 21 that were reported earlier this year. ... " (Prusiner denies having reviewed the paper, although the New England Journal of Medicine says that indeed he did.)

Prusiner then wrote up a paper based on his own work, in which he had had a graduate student search for the prion protein in 15 CJD brains, six fewer than Brown had used to gain what Prusiner had considered no additional scientific value. Less than two months after recommending that Brown's paper not be published, Prusiner submitted his own, for which he claimed better results--although even Kingsbury, his collaborator, has described it as essentially the same science as Brown's.

The New England Journal rejected Prusiner's paper. "If a referee submits a paper to us on the same subject that he's being asked to review," says Dr. Arnold Relman, the Journal's editor, "we don't consider it until we make a decision on the manuscript under review." Brown's paper was eventually accepted with minor revisions, but only after the angry intervention of Gajdusek.

By mid-1984 Prusiner's science and public relations had begun paying off. Not only could his competitors no longer duplicate his work because none of them had as many animals as he did, but they couldn't attract the high-powered collaborators that he could because they didn't have the money.

Once he had the prion protein, Prusiner went to Leroy Hood of Caltech and Charles Weissmann of the University of Zurich for help in elucidating its structure. The two are known among medical students as "gods of molecular biology." "Stan needed to have his prion sequenced," says Hood. "He had precious little of it, and we can do it better than anybody in the world."

Hood deduced a small sequence of the amino acids in the prion protein and from that worked backward to make a copy of the DNA that would create that sequence. Weissmann then used the DNA to fish for the gene that coded for the prion protein from a "library" of genetic material from scrapie-infected brain tissue.

To this point, the scenario played out as well as Prusiner could have hoped. Then the plot twisted on him. In tests at UCSF, he confirmed that the RNA that coded for his prion protein was in the brains of hamsters with scrapie, but it also appeared in equal amounts in the brains of healthy hamsters--exactly what should not have been the case if the protein was the agent.

Prusiner's best evidence that the scrapie agent was an infectious protein had now been contradicted. He admits that he was astounded by this turn of events. Rather than publish immediately, he sat on the findings for six months while he tried to make sense of this self-inflicted apparent disproof of his theory. By the end of that time he had discovered that the prion proteins in healthy and infected animals--what he called the cellular protein and the scrapie protein, respectively--were different. The scrapie protein aggregated into rods; the cellular didn't. And the scrapie protein survived proteinase k; the cellular protein disintegrated when exposed to it. With this information, Prusiner published.

The paper appeared in the April 1985 issue of Cell, signed by Hood, Weissmann, and Prusiner and his colleagues. It seemed to concede that the infectious protein theory was teetering on the abyss: "Our finding that PrP . . . is encoded by the host genome and that a closely related, if not identical, protein is expressed in uninfected, normal cells would seem to support the argument for a scrapie-specific nucleic acid." That meant the agent was either a regular virus or a virino that had escaped detection. Only if a scrapie protein could somehow modify a cellular protein into a copy of itself, which would set off a chain reaction converting other cellular proteins into scrapie proteins, the paper said, could the authors "envision a scrapie prion devoid of nucleic acid."

Prusiner's competitors embraced this report. If the scrapie prion wasn't devoid of nucleic acid, and the prion protein existed in healthy animals, then the protein wasn't part of the agent. As Nature put it, the search for the scrapie agent was "back with a vengeance."

Prusiner disagreed. He held tenaciously to the difference between cellular and scrapie proteins, and to the idea that somehow the scrapie proteins altered the cellular ones into "infectious" versions. This hypothesis, still a protein-only one, could be made to fit some of the seemingly genetic aspects of the diseases.

When Prusiner spoke of his findings, either to reporters or in conferences, he played up the angles that seemed the most advantageous to salvaging the prion. He reported, for instance, at a meeting of the American Federation for Clinical Research, that the virus hypothesis had now been ruled out. He then said, "The discovery of cellular PrP seems to explain one of the most perplexing aspects of scrapie infection. That is, how can the infectious process which devastates the brain and leads to death go on in the absence of any response by the immune system?"

The UCSF press release on the work--approved and authorized by Prusiner--claimed with equal brashness that they had "accumulated important new evidence suggesting that scrapie is caused by a radically new type of infectious agent." This was somewhat confusing, since that agent now seemed more likely than ever to be a virus or a virino, but since those had always been included within the definition of the prion, the release wasn't semantically incorrect. And, as one pathologist put it, "when there are megabucks involved in the research, you don't want to split hairs."

Soon molecular biologists punched more holes in the prion theory. For instance, after Prusiner's paper came out, Bruce Chesebro, a virologist at Rocky Mountain Labs, and Haase reported that when mice were infected with scrapie, within two months the amount of infectivity in their spleens shot up to a million infectious units, yet the RNA that would have to be there to produce PrP remained virtually undetectable. "It suggests," Chesebro concludes, "that the protein and the infectious agent may have nothing to do with one another."

Prusiner, on the other hand, reported that he had found the RNA in the spleen of hamsters (although still not in proportion to the amount of infectivity), and said that Haase and Chesebro must have missed it. And in case his competitors were right, he again invoked the broad definition of the prion. Thus he could say that prions are found in the spleen, but what's not found is the prion protein, which isn't necessarily part of the prion. When it was pointed out that this seemed confusing, he reiterated that the prion was only a synonym for the infectious agent of the disease, and that infectivity, hence prions, was found in the spleen. QED.

To clear up the confusion, all Prusiner has to do is use bacteria to clone the prion protein. "If we put the protein into an animal," he says, "and it causes scrapie, we know we've got the agent. This is the key experiment. I'm optimistic." The catch is that if the protein doesn't cause scrapie, it proves nothing one way or the other. "If you make this protein, and you put it into animals, will they get sick?" says Chesebro. "And if they don't, which I firmly believe they won't, then why not? Is it because we've got the wrong gene? Or is it because the bacteria didn't make the protein the way we expected them to? The controversy will just go on and on."

This next year should be crucial. Just about everybody in the field is sitting on results that they claim will prove something significant. Gibbs says he has found the natural form of the prion protein in ever tissue in the body: "The implication is that the prion protein isn't the agent." Manuelidis says she has compelling evidence that the prion protein isn't infectious. Heino Diringer, a biochemist at the Robert Koch Institute in Berlin, says he has similar data. And Prusiner says he has new results showing that the protein is the agent, isn't a virus, has no genes, and can justifiably be called a prion. All these revelations are as yet unpublished.

Any final answer is as likely as not to come from Prusiner's lab. After his molecular biology paper came out in Cell, Prusiner won that multimillion-dollar congressional award that pushed the level of his support way beyond that of his competition. Many of his competitors are funded at the poverty level, or even working on scrapie with money that has fallen through the cracks from other research.

Prusiner has the animals, the money, the high-powered help, and the skills to track down the agent. The question is whether he's so obsessed with sticking to his prion story to the end that, as Bolton says, "even if the scrapie agent was a virus and all the data was staring him in the face, my bet would be that he'd never see it." Indeed, if the agent does turn out to be a virus, and Prusiner's lab comes up with it first, it could still end up going into the history books as a prion. Says Prusiner, "I never said it's only an infectious protein. I've never said that in one paper. You'll not find it. I've been very, very careful."