SCRAPING THE BARREL OF RESEARCH

When young people were seen to be dying of CJD in 1996 the media was happy to blame it on the prion, "discovery" of US neurologist Stanley Prusiner. But, as Emily Green reports, it is only now that Scottish scientists are coming near to proving the link between the deaths and BSE in cattle

WE HAVE learnt a new word. Prion. It sounds so modern, so menacing, and it is, the media has assured us, the rogue protein behind the BSE epidemic, and the infectious agent that caused 15 deaths of young people from a sinister new variant of Creutzfeldt-Jakob disease.

The prion as killer is a recent invention, a snappy label coined by an enterprising American in 1982. It entered common British usage shortly after the mini-BSE scare of December 1995, which was prompted by Sir Bernard Tomlinson, a distinguished neuropathologist in Newcastle, announcing on Radio 4 that he would not eat bovine offal. Shortly after newspapers erupted with controversy over the safety of beef, the press office of the Wellcome Trust piped up with prions, adding the helpful note to editors: "BSE is a member of a group of diseases known as transmissible spongiform encephalopathies, or prion diseases."

How very scientific-sounding, particularly coming from Professor John Collinge of the Prion Diseases Group at the Imperial College of Medicine at St Mary's Hospital in London. (Collinge's unit is supported by the Wellcome Trust, along with the two leading research councils, and the David and Frederick Barclay Foundation, a charity set up by the identical twin owners of the Ritz Hotel.) Within a fortnight of Tomlinson's remark, Wellcome informed the nation that Collinge already had an experiment underway in which transgenic mice expressing human prion proteins had been dosed with BSE. It was, basically, the next best thing to testing BSE on humans.

On 20 March 1996, when health secretary Stephen Dorrell linked BSE and a new variant of CJD, it dawned on a shocked country that we humans were already being tested. Every reporter in the land wanted a culprit, and they wanted one fast. As disease agent, the prion did nicely. The media also wanted a saviour. Collinge did nicely. He, like the new orthodoxy, was untainted by the decade of political blundering that had led to the crisis. Moreover, our lives seemed to hang in the balance of whether or not his transgenic mice got CJD from BSE. Our very prions depended on it.

We all have prions. Not nasty ones, but nice ones, too. We just didn't know it. Nor, in 1982, did Stanley Prusiner, the neurologist at the University of California at San Francisco who unveiled the word in an article in the US magazine Science. Though he had no real evidence, the prion was, he hypothesised, the infectious agent behind CJD and a school of mystifying neurological diseases.

It stands for "proteinaceous infectious particle", which should be abbreviated as "proin", but Prusiner inverted the O and I to improve the ring. "Prion is a terrific word," he told Discover magazine in 1986." 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." Or mispronounced. Prusiner pronounces his word "pree- on". If a scientist in the field pronounces it "pry-on" this is the subtlest note of dissent. If they say "protein-only hypothesis", it is safe to assume they are sceptics.

CLUES FROM THE CANNIBALS

THERE IS nothing new about transmissible spongiform encephalopathies (TSEs) - the school of ailments supposedly caused by prions. Records show that ovine TSE, known as scrapie, was spotted in East Anglian sheep in 1732. From 1913 to 1921, cases of what turned out to be human analogues of scrapie - CJD - were spotted in Germany.

Over the years, various diseases joined the TSE camp: transmissible mink encephalopathy (TME) and chronic wasting disease (CWD), a disease of captive mule deer and elk; and, in humans Gerstmann-Straussler-Scheinker Syndrome (GSS), fatal familial insomnia (FFI) and, most famously, kuru, discovered among the Fore-speaking tribe in Papua New Guinea in the Fifties. In 1976, evidence that kuru was transmitted by ritual cannibalism, and that, along with CJD, it was transmissible at all gained the American paediatrician, Daniel Carleton Gajdusek, a Nobel Prize. Gong in hand, he loftily crowned TSEs "slow virus infections".

It seems unlikely now that TSEs are caused by viruses, at least in the classical sense of a disease agent consisting of DNA or RNA with its own protein coat. That said, even in the Fifties, the infectious entity already had a long-established name, "the scrapie agent". The importance of scrapie research, and sheep, to the field is so fundamental that, though it took Gajdusek decades to admit it, it would never have occurred to him to look at kuru as a TSE had an American veterinarian working in Britain, William Hadlow, not pointed out its similarity to scrapie.

It took some audacity for Prusiner to rename the scrapie agent a prion. It is a time-honoured phrase, and a British one. During most of this century it has been the UK, not the US, that has been the traditional home of TSE research. Britain, after all, has a lot of sheep - at least one for every five humans. Scotland is the epicentre of UK expertise due to that country's scientific tradition and an agricultural disaster of 1935, when a seemingly routine vaccine for a common sheep disease, "louping-ill", became accidentally contaminated with the scrapie agent. It killed 1,200 of the 18,000 animals immunised (about 7 per cent).

Investigations that were initiated to examine the vaccination fiasco led to the discovery throughout the Fifties, Sixties and Seventies, of markedly different strains of scrapie. These were led by a geneticist, Alan Dickinson, and a pathologist, Hugh Fraser, who employed a process known as a bioassay. In this, a homogenate of infected tissue, usually brain, is passaged through generations of mice and its effects noted. While rough on the mice, it allows scientists to study closely the havoc wreaked by the scrapie agent.

These bioassays led, in turn, to the 1968 discovery of a gene in mice controlling the progress of scrapie. It was called "sinc" (for scrapie incubation). "The way it controls the disease is so accurate," Dickinson said at the time, "it looks more like civil engineering than biology." Soon they found a similar gene in sheep, which they called "sip" (scrapie incubation period). Eventually they defined more than 20 strains of scrapie. The discovery of sinc and sip and the existence of many strains, implied the scrapie agent was getting genetic information from somewhere. No one, anywhere, has been able to isolate a nucleic acid, or any sort of informational molecule, which enables the scrapie agent to go forth and cause brain rot. Today there is no end of speculation about the identity of the scrapie agent. The government BSE co-ordinator, Ray Bradley, reckons he has heard "more than 30 theories", including suggestions that it is a spiroplasma and "nemavirus". The most plausible explanation, a virino, is hinted at in a 1995 paper, from no less than Stanley Prusiner.

BAD NEWS IN A PROTEIN COAT

THE VIRINO, however, is anything but Prusiner's idea. It received its international debut in Tokyo in August 1984 and came from Alan Dickinson, the most senior of the Edinburgh scrapie researchers. The scrapie agent, he suggested, was a hybrid structure, which borrowed the protein from the host, but contained an unidentified molecule, possibly a nucleic acid, that gave it strain-like properties. This molecule was transmitted, for example, when sheep grazed on placentas left behind after lambing. The virino hypothesis amounted to a new spin on an old quip "a virus is bad news in a protein coat". Effectively, the hypothetical virino is terrible news in a borrowed protein coat.

But it does not have the ring of prion. When Prusiner first unveiled the prion, he suggested the scrapie agent was not remotely viral, but a rogue protein, and something of an acrobat with it, capable of flipping spontaneously and setting off a chain reaction that destroys the brain. At first he suffered ridicule ("PRION is ion followed by PR"). But if Prusiner was right, he had every chance of scooping the second Nobel Prize almost certainly destined for the person who cracks the structure of TSE agents. Proving the existence of an infectious agent based only on protein, without RNA or DNA, but capable of transmitting strain differences, would redefine the limits of biology.

A heretical notion, yes. Original, no. British researchers had been toying with the same sort of speculation since the Sixties. In 1967 Tikvah Alper, a radiologist at Hammersmith Hospital, observed that the scrapie agent survived radiation which suggested it had no DNA or RNA. Gibbons and Hunter, a pair of scrapie researchers from what is now the Institute for Animal Health at Compton, Berkshire, promoted the same concept around the same time, and a physicist from Bedford College, London, pushed the idea in Nature. The Oxford physiologist, Colin Blakemore, is good-humoured about Prusiner's adoption of the theory. "He had the wit to develop the idea," he said last summer.

Prusiner's American colleagues were less sanguine, and scrapie researchers in North Carolina, Connecticut, Maryland, New York and elsewhere came out against him in force in a damning 1986 article in Discover magazine. Most seriously, they suggested that he had used other peoples' ideas.

As evidence, Discover offered the example of "prion rods". Some diseases of the TSEs, but not all, leave pronounced plaque deposits in diseased brains. In 1983, Prusiner posited that infectivity was found in remnants from plaque, in structures he called "prion rods". It was, at best, a case of seriously belated synchronicity, which redefined a 1981 discovery by a New York electron microscopist, Pat Merz, and her collaborator, a Scottish biochemist Robert Somerville. A year before Prusiner unveiled his "rods", Merz and Somerville had published results finding essentially identical "scrapie associated fibrils" (SAFs) unique to all TSE brain samples, and not found in healthy controls.

PRUSINER PULLS IT OFF

BY 1986, Prusiner was redefining the entire history of scrapie research by claiming he had, in three years, conducted "more experiments on the bio-chemistry of scrapie than everyone in the history of scrapie combined." This sort of braggadocio had attracted a $4m Congressional award - money that enabled Prusiner to go out and discover the prion which he had, somewhat prematurely, named four years earlier.

Working with Leroy Hood and Charles Weissmann, world-class scientists from the California Institute of Technology and University of Zurich, he organised the sequencing of the amino acid chain of the protein he had, effectively, named before its discovery. It took cutting-edge techniques of molecular biology to identify the structure, and then backtrack to deduce its DNA. Following their work, mutations were found at various links in the protein's chain, which seemed to relate to susceptibility of various host genotypes to TSEs. Even those who question Prusiner's tactics call it "a heroic" discovery.

It is a shame that he couldn't resist playing the renaming game. This amino acid chain, complete with its individual mutations, was almost certainly the replication site for TSE agents. This is to say, whatever agent causes the disease makes its first chemical bonding here. These replication sites amounted exactly to what Dickinson and Fraser had already named the sip and sinc genes. No one denies this. Yet, while traditionally terminology of the first discoverer enters common usage, sip and sinc got bumped for prion-speak. Today one of the very few scientists courtly, stubborn and brave enough to still employ the original terminology is Professor Heino Diringer of the Robert Koch Institute in Berlin.

If name changes make the scrapie literature difficult to penetrate, the prion also comes to us with its own special collection of tortuous special clauses. In 1985, prion proteins were found in abundance in the brains of healthy control animals, and later they were found throughout their bodies - nervous system, muscle, the works. That is to say, we all have "prions" aplenty, though nobody is quite sure why we have them, or what it is they do. But rather than admit his disease agent wasn't a disease agent, Prusiner set about redefining the prion. The protein in its nominal form (cellular) became PrP (c) and this was differentiated from its rogue form PrP (Sc).

Other hurdles were not so easily cleared. When Scottish and American researchers objected that a protein could not account for more than 20 strains of scrapie, Prusiner had no answers. Infuriated by his intransigence, Laura Maneulidis, a neurologist from Yale University, observed that all conventional understanding had suddenly been dismissed as heretical, and all heretical notions as conventional. A British counterpart is still flabbergasted. "He simply ignored us. It was as if work in the Thirties, Forties, Fifties and Sixties never existed. It was as if strain typing hadn't happened."

It took some stalling, but the prion camp finally devised a response: biochemical conformation. The notion is, roughly, that different mutations in the prion gene can unexpectedly send the protein from a corkscrew shape to a wave patterns, and the brains of affected animals consequently become diseased. The fashion in which certain proteins flip is determined by the host genotype. Mice, for example, have two sorts of mutations in their prion protein sequences. If mutation differences were accents, it would go like this: one prion says pry-on, the other says pree-on, so their respective proteins turn into like-voiced pry-ons and pree-ons. Hold on, shout the sceptics. Two mutations produce more than 20 accents? There isn't an explanation to this challenge that leaves the prion theory intact. It would need 18 more pronunciations of prion, so to speak.

Bruised by the Discover article, Prusiner stopped talking to the press a decade ago. Yet he is still busy rewriting the language he thinks we should use. By the early Nineties his team was suggesting that classic descriptions such as Creutzfeldt-Jakob disease and Gerstmann-Straussler-Scheinker syndrome should be renamed as prion diseases, followed by specific- molecular notations. The idea was politely rebuffed.

A LACK OF PROOF

THE CONTROVERSY over prions is such in the US that the specialist press, even prestigious Science, now tiptoes around the subject. In fact, left to the Americans, it seems likely the prion would have been put to bed with a consoling pat on the head in the late Eighties had it not been rescued by several tantalising advances.

The first was in the field of gene-splicing. In 1989, Prusiner's laboratory produced transgenic mice which were capable of producing species-specific prion proteins. This laid open the prospect of testing for human susceptibility to "prion diseases" in mice. Then in 1993, Adriano Aguzzi, a neuropathologist at the University of Zurich, reported in Cell that mice lacking the prion protein gene, known as "null" or "knockout" mice, did not get sick when infected with scrapie. Was this proof that the prion was involved? Yes, and more evidence that the prion protein is what Dickinson called the scrapie agent's "replication site" in the virino hypothesis.

Most important to the survival of an increasingly divergent number of the protein-only hypotheses was work at the Rocky Mountain Laboratories in Montana. Here, researchers including Byron Caughey and Richard Bessen managed to trigger instantaneous prion conformation (or flipping) in a test-tube. This was exciting work, which really got interesting when they made one sort of normal prion aggregate in two different ways, following on from the conformation of two distinct sets of PrP (Sc) residues. Finally, there was a hint at strain specific behaviour free of DNA or RNA.

If the protein-only hypothesis has a leg to stand on, it is this work. It begins to show us how spent proteins might accumulate. Yet those who argue that aggregations formed in test tubes illustrate how protein behaves in a living brain are jumping from A to Z. Caughey is the first to stress that his work is cell-free chemistry, not an experiment in a living animal. Secondly, aggregation does not identify the prion protein as a disease agent. They have no way of knowing if they reproduced infectivity in the aggregated protein. They may simply have a recipe for protein residue found in plaque. "Until we or someone else can measure new infectivity, the proof just isn't there," Caughey told Science last summer.

LOOKING FOR PROTEIN X

MEANWHILE, Prusiner's own argument is evolving beyond all recognition. In October 1995, an article by his team appeared in Cell amounting to the admission it was by no means clear that the rogue prion is the scrapie agent. It may well need another protein - "protein X". Or maybe not a protein. Some sort of "molecular chaperone". Something that makes it the scrapie agent. The protein-only hypothesis now strikes Laura Maneulidis as so tortured that she describes it as "medieval". What is worse, this prion with knobs on is uncomfortably similar to the original virino hypothesis.

Back in Britain, John Collinge's much publicised batch of transgenic mice never got CJD from BSE. However, in the summer of 1996, while they were on the verge of dying of old age, he was quick to puncture any false optimism by appearing on Newsnight, to say, in effect, good news is no news. His laboratory may have been using the wrong sort of mice. The human victims were all one genotype. The genes of the mice had been spliced with another gene, the wrong one.

In October 1996, Collinge announced he had, using electrophoresis, discovered a specific isotope molecular marker for new variant CJD, and that this was the closest proof to date that it was related to BSE. Moreover, he claimed, it put a diagnostic test in view. It might even come in handy as a test to determine whether BSE was masquerading as scrapie in sheep and goats. This prompted the father of scrapie strain-typing, Alan Dickinson, to question Collinge's controls in a letter to the Daily Telegraph. As it turns out, Collinge had no scrapie controls. He had suggested he could tell X (BSE) from Y (scrapie) without identifying Y.

Dickinson's rebuke was followed only weeks, ago by international objections. In late March, a letter appeared in Nature signed by a team of researchers from the US, Japan, France and Germany, who reckoned Collinge's electrochemical analysis was well off the mark. Collinge replied in the same issue: "We did not claim the 'identity' of the prion strain causing BSE and CJD."

Oh no? The Wellcome Trust press releases, which, according to the publicists, he approved, certainly led reporters to think so. As a lesson for the future, the media may wish to study the fine art of scientific weasel wording. The press release announced Collinge had "identified a molecular marker which distinguishes 'new variant' CJD from other forms of CJD. This characteristic molecular 'signature' is also seen in BSE". All Collinge had said was that, in the confines of his own work, he had found something that looked like something else. Yet more objections to his lack of scrapie controls appear in the current edition of Nature, along with a reply from Collinge to the effect that while there are still bugs in the system, his molecular techniques will be fast once they, ahem, work.

Claim has followed claim from Collinge. Four months ago, he was back in the mainstream media as the public face of yet another putative discovery: a diagnostic test using tonsil scrapings, from which scientists would then isolate his marker (eventually). As every national newspaper jumped on the story, colleagues could not credit the fuss. "They only had one tonsil," said one. "And it was from an autopsy."

This was not only a premature suggestion, but also hardly original. Scottish scientists isolated infectivity in the lymphoreticular system in the Sixties and Seventies, and, six months before Collinge and his collaborators suggested tonsil scrapings in the Lancet, Dutch veterinarians had made a nearly identical suggestion in Nature.

The press, and prions, have been good to Collinge. Last June, the Wellcome Trust announced he had become one of its Senior Research Fellows in Clinical Science, which provided his group with an additional pounds 1.6m. Late last month, he was prominent among scientists supported by pounds 8m worth of new grants from the two leading research councils (BBSRC and MRC). Like Prusiner before him, his success needn't hinge on the proof in the prion. His Prion Diseases Group has recently been renamed the "Neurogenetics Unit".

Ironically, the first conclusive indication that new variant CJD is, or is not, related to BSE is likely to come not from the publicity wizards and their prions, but from the besieged old-guard scrapie specialists in Edinburgh. The first blow to the prion camp came in 1993 at a Royal Society meeting, when Moira Bruce, a Scottish biologist with 20 years scrapie research experience, outlined the results of a series of bioassays with BSE material in mice, done to determine the relative incubation periods, and gather pathological evidence. So far, the "lesion profiles" resulting from this work have proved uncannily similar, whether the mouse inoculate was made from tissue taken from a cow with BSE, or a sheep, goat, cat or antelope. All these animals have distinctly different prion protein genotypes. Material taken from diseased tissue from each should not, according to prion orthodoxy, produce the same pathology when passaged through mice. But it does.

The earliest indications from the lesion profiles from mice infected with tissue from humans who died of new variant CJD may be out sometime this summer. A match with BSE samples would define very soundly that new variant CJD is caused by the same agent as BSE. It may also show irrefutably that there is something more than species-specific host protein mutations at work, something with spookily consistent pathological fingerprints, something pointing to genetic information in the scrapie agent.

Moira Bruce's work is the first of a one-two punch. One of her colleagues, Nora Hunter, recently published results showing that while common sheep genotypes in the Antipodes, which are all descended from British stock, do not get scrapie, their British cousins do. It will take some clever explaining to justify why proteins spontaneously flip in sheep flocks here, whereas they don't down under. Perhaps it's because their prions are upside-down already. !

THE SEARCH FOR SCRAPIE

THE HUNT for the scrapie agent began in earnest in Scotland in the Thirties. The keepers of this tradition insist the scrapie agent can't, like the "prion", be only protein. To do such distinctive damage, it must carry genetic information - RNA or DNA. In the opposite corner is the prion camp, from California, Zurich and London, who dismiss the Scots and their allies as flat-earthers. It is, they insist, protein alone that has long plagued European sheep with scrapie and, since 1986, visited the BSE epidemic along with new variant CJD on Britain.

1961-69: Alan Dickinson, a geneticist, and pathologist Hugh Fraser, working from the Moredun Institute near Edinburgh, discover a gene controlling the development of scrapie in mice. They name it sinc. By 1969, they discover sip, a corresponding gene in sheep.

1967-68: Picking up on Moredun work done in the Forties, three groups of British researchers independently suggest that scrapie does not have a nucleic acid (DNA or RNA) and might be a revolutionary, protein-only agent.

1971: Dickinson et al say of sip and sinc: "If we could find exactly how this gene worked at the molecular level in an uninfected mouse, it may be possible to deduce the molecular structure of the infective agent."

1978: At a conference in Montana, Dickinson calls the scrapie agent a "virino" - a hybrid of host protein and some sort of transmissible informational molecule. The meeting's hosts include San Francisco neurologist, Stanley Prusiner.

1981: An American electron microscopist, Pat Merz, and Scottish biochemist Robert Somerville, publish in Acta Neuropathologica the discovery of particles unique to the brains of mice with scrapie and name them scrapie associated fibrils (SAFs). Alan Dickinson, Hugh Fraser and Richard Kimberlin form the Neuropathogenesis Unit in Edinburgh, the first dedicated research outpost for study of TSEs.

1982 April: In an article in Science, Prusiner re-names the scrapie agent a "prion", short for "protinaceous infectious particle".

May: Dickinson puts the virino hypothesis as a retort to prions in the Lancet, as does Kimberlin in Nature.

1983: Prusiner's team report in Cell the discovery of "prion rods". These prove indistinguishable from Merz and Somerville's "SAFs".

1984: Dickinson formally launches the virino hypothesis at a TSE conference in Tokyo.

1985-1990: Working with Charles Weissmann in Zurich and Leroy Hood in California, Prusiner finally discovers the prion, sequences its amino acid chain and deduces its DNA. "Protinaceous infectious particles" are found in healthy tissue. Prusiner renames the disease agent a rogue prion - PrP (Sc). By 1990, his laboratory has introduced transgenic mice expressing human prion proteins into research.

1993: Adriano Aguzzi, a neuropathologist at the University of Zurich, reports in Cell that mice lacking the prion protein gene do not get sick when infected with scrapie. It becomes clearer that the prion is, or contains, what Dickinson and Fraser named sip and sinc.

Dickinson's colleagues from the Neuropathogenesis Unit, Moira Bruce and Hugh Fraser, have passaged tissue from cattle with BSE, and from antelope, goats and cats with BSE-like spongiform encephalopathies, through inbred mice. They find a consistent pathological fingerprint, implying the infectious agent for BSE carries its own genetic coding. Moira Bruce presents the results of the "lesion profiles" in London. Weissmann and Prusiner argue over the significance of the BSE fingerprint.

1995 June: A team trusted by both sides of the argument, Richard Bessen and Byron Caughey in Montana, achieve strain-specific conformation of prion proteins in test-tube experiments. They caution the public that they do not know if they created infectious materials. If they do this, the protein-only hypothesis will be vindicated.

October: Prusiner's team publish in Cell that the prion may need "protein X", or "chaperone X", to qualify as the scrapie agent.

December: The Wellcome Trust press office describes John Collinge's work with BSE and transgenic mice at St Mary's, London as work on a "prion disease".

1996 March: At the height of the BSE scare, in an article in the Lancet, Collinge refers to prion "strains", and footnotes Bruce's work.

June: A team of Dutch scientists publish a letter in Nature suggesting pre-clinical BSE and CJD could be detected from tonsil scrapings.

October: Collinge claims in Nature that he has found a specific isotype molecular marker for new variant CJD and BSE, which puts a diagnostic test in view.

1997 January: A Lancet article by Collinge, suggests tonsil scrapings as a diagnostic test for CJD.

March: Nature publishes a challenge to the accuracy of Collinge's marker from researchers in the US, France, Germany and Japan.

April: Researchers from Edinburgh and an associated lab in Berkshire write in Nature that Collinge's molecular marker is present in strains of scrapie. Collinge replies: "it is unsurprising that not all of these strains can be completely differentiated from BSE."