Swim against the tide in medical research, and your colleagues may
regard you with suspicion. But challenge one of its central orthodoxies,
and you may end up branded as either a genius or a charla-tan. Stanley Prusiner,
a biochemist at the University of California at San Francisco, has been
cast in both roles. His heresy was to challenge the received wisdom that
all infectious agents must carry genetic material in the form of DNA or
RNA. Now, more than a decade later, this idea is slowly being absorbed into
mainstream thinking, helping researchers to understand fatal brain diseases
such as scrapie in sheep, BSE in cattle and Creutzfeldt-Jakob (CJD) disease
in humans.
The story goes back to a controversial paper published in the journal
Science in 1982. In it, Prusiner proposed that scrapie and CJD are caused
by an infectious agent containing protein molecules – but no genes. At
that point, these diseases were a biological enigma which four decades of
research had failed to crack. It was impossible to say in detail what the
infectious agent was. However, in experiments on infectious extracts from
brains riddled with scrapie, Prusiner had detected high concentrations of
protein. Moreover, the infec-tiveness of these extracts was killed off
by treatments that destroyed proteins, but not by ones that destroyed nucleic
acids.
Whatever the agent was, it could survive conditions that would kill
all known viruses and bacteria. Prusiner, a flamboyant character who knew
the value of presentation, immediately christened it ‘prion’, shorthand
for ‘proteinaceous infectious particle’. As a label for scrapie-like diseases,
it had an irresistible advantage over ‘transmissible spongiform encephalopathies’.
Advertisement
Evangelical zeal
Within months came a second Science paper. Prusiner’s team had identified
a protein that appeared to be unique to scrapie-infected brains, and which
was unusually resistant to digestion by enzymes. This so-called protease-resistant
protein, or PrP, provided the strongest lead for decades in the quest to
discover the nature of the elusive scrapie agent. Prusiner had a hunch that
the agent, the thing he had christened prion, and the PrP protein were
one and the same. And he pursued this idea with almost evangelical zeal
in scientific papers and conferences.
Initial reactions ranged from scepticism to outrage. To suggest that
the scrapie agent could cause infection without the help of genetic material
was one thing. Indeed a British researcher, Tikvah Alper of the Medical
Research Council’s Radiopathology Research Unit at Hammersmith Hospital
in London, had speculated as much in a letter to Nature as long ago as
1967. But to invent, and go as far as to name, a completely new type of
infectious particle on what looked like the slenderest of evidence was something
quite different. To many researchers, particularly in Britain, it smacked
of opportunism, of a brazen attempt to appropriate an entire area of research
with a smart buzzword – particularly as Prusiner hadn’t yet tackled the
imp- ortant question: how an agent devoid of genetic information, a humble
protein, could infect tissue and spread inside it. It was hard to see how
infection could occur without replication and how replication could occur
without genetic information.
Matters came to a head in 1986, when the American science magazine Discover
published a damning article headed ‘The Name of the Game is Fame: But is
it Science?’. Prusiner had, it claimed, achieved his ends by manipulating
the media and suppressing opposing views. From then on, Prusiner kept his
head down, concentrated on laboratory research and vowed never again to
talk to journalists.
Today that decision seems to be paying off. Prusiner and his team at
the University of California, working with colleagues all over the world,
have accumulated some persuasive evidence to support the idea that prions
are the major, perhaps the only, component of the infectious agent in scrapie-like
diseases. A fascinating picture is beginning to form of how infectious agents
could spread in tissue without conventional replication taking place. The
term ‘prion’ is routinely used by journals such as Science and Nature, and
in September 1993 the Royal Society in London held a meeting entitled ‘The
Molecular Biology of Prion Diseases’.
Yet old prejudices die hard. The term prion still sticks in the throats
of opponents of the ‘protein-only hypothesis’. Researchers are fond of calling
it the ‘P-word’, and only half in jest. Part of the problem, say Prusiner’s
critics, is that his original definition of the prion was so vague that
it will be able to assume whatever identity the disease-causing agent turns
out to have. ‘It could include all conventional viruses,’ says Moira Bruce
of the Institute for Animal Health in Edinburgh. ‘Now it’s become associated
with the protein-only hypothesis.’
Slippery definitions are not the only reason why the prion is still
battling for respectability. Important questions remain unanswered about
the kind of molecule the prion protein is and how it generates illnesses
such as CJD and BSE. One of the biggest remaining threats to the protein-only
hypothesis concerns strain variation. It has long been known that there
are many different strains of scrapie, each strain producing a slightly
different pattern of symptoms. Now the same appears to be true of BSE. And
here lies the rub, for disease strains normally reflect genetic variation
in an infectious agent. Nobody has yet shown how an agent consisting solely
of protein could produce different strains of disease.
The challenge of explaining strain variation comes at a crucial time
for Prusiner and other proponents of the protein-only hypothesis. Evidence
linking the PrP protein to scrapie-like diseases has grown so dramatically
in recent years that few researchers today would doubt that it plays some
part in infection. The question is, what part. The gene encoding the PrP
protein – known, confusingly, as Prn-P in animals and PRNP in people – is
found in some form in all mammals. According to researchers in Germany,
it now seems that this gene is an essential accomplice of the infectious
agent. In 1992, Charles Weissmann and his colleagues at the University of
Zurich found that mice lacking the Prn-P gene are perfectly normal in every
respect – except that they are not susceptible to infection with scrapie.
Since then, researchers everywhere have been trying to tie this observation
in with an earlier, but perhaps even more fundamental, discovery – namely,
that the PrP protein can exist in two forms. In terms of their amino-acid
sequences, these molecules seem to be identical. Yet they act differently
in cells. Researchers suspect this is because they adopt different molecular
shapes. PrPC, the normal form of the protein, is found in all cells, especially
on the surfaces of neurons, and can be broken down by enzymes. PrPSc, the
abnormal form of the protein, is resistant to attack by enzymes and is only
found in diseased brain extracts.
Perfect solution
For Prusiner and others in the ‘prion camp’, the existence of these
two forms of PrP offers a perfect solution to the central mystery: how a
protein-based agent could spread in tissue without ever replicating in the
conventional way. Their argument is based on the idea that normal prion
proteins can be converted, irreversibly, into abnormal prion proteins, and
that this might happen when an abnormal protein comes into contact with
a normal one. According to this theory, the abnormal protein acts as the
proverbial bad apple, corrupting its neighbours one by one: the protein
does not synthesise new copies of itself as it would if it were replicating
in a conventional sense. Confusingly, however, many researchers refer to
this process as ‘prion replication’.
The challenge now is to discover what this corruption amounts to in
a physical sense. Prusiner and other researchers suspect it involves the
normal protein flipping from a ‘healthy’ molecular shape to a ‘disease-causing’
one. Recently, Prusiner’s team has discovered that the abnormal form of
the prion protein adopts a shape in which part of its amino-acid chain folds
tightly into a molecular structure known as a beta-sheet. By contrast, the
normal form of the protein contains very little of this structure. Instead,
its amino-acid chain folds tightly into structures known as alpha-helices.
Vulnerable protein
These structural differences hold the secret to understanding how normal
prion proteins become corrupted by abnormal ones, argues Prusiner. The corruption
may occur, he suggests, when normal prion proteins unfold into looser configurations
inside cells. At this point, normal prion proteins would be relatively vulnerable.
Abnormal prion proteins – should the cell contain any – could conceivably
interact with them, causing the proteins to switch into the beta-sheet structure.
A chain reaction would result, with the abnormal molecular structure spreading
inside cells.
If all this turns out to be true, it will mean something quite revolutionary:
that the biological information needed to cause an infection can be encoded
by the shape of a protein as well as by DNA or RNA. Supporters of this heterodox
theory turn to genetics for further evidence.
People afflicted with inherited forms of illnesses like CJD carry mutated
versions of the PRNP gene. Twenty or more such mutations have been identified
so far, and different mutations can produce slightly different versions
of disease. The mutated genes code for a less stable form of the prion protein
which is much more likely to flip into the abnormal form. People who produce
such protein are, in effect, carrying their own ‘bad apples’. They will
develop disease in the absence of any contact with outside infection.
Or so say proponents of the protein-only hypothesis. John Collinge,
who heads the Prion Research Group at St Mary’s Hospital Medical School
in London, believes that this model of infection can be extended to people
who suffer from noninherited forms of CJD, which are much more common. These
people could be victims of spontaneous mutations corrupting the PRNP gene
within single cells, he argues. Such cells would die, releasing abnormal
prion proteins and setting off a chain reaction throughout the brain.
None of this surprises Carleton Gajdusek, a virologist at the US National
Institutes of Health (NIH) in Bethesda, Maryland, and one of the most influential
scientists investigating scrapie-like diseases. He won a Nobel prize for
his work on the degenerative disease kuru in Papua New Guinea during the
1950s and 1960s, and subsequently showed that CJD could be transmitted to
laboratory animals. Gajdusek’s response to Prusiner is not that his ideas
are heretical, but that they are nothing new to anyone who knows anything
about amyloid, an insoluble deposit of protein seen in a number of brain
diseases, including scrapie and Alzheimer’s disease.
A variety of different proteins can form amyloid. The protein content
of the amyloid implicated in Alzheimer’s is, for instance, different from
that found in brains infected with scrapie. But in both cases, explains
Gajdusek, the protein in amyloid deposits en-courages yet more amyloid to
form. The bad apple effect comes into play – just as it does in the prion
model of scrapie-like diseases. Indeed, the very term prion is redundant,
says Gajdusek: in his view its coinage was ‘political’, designed to distance
Prusiner’s work from that of others in the same field.
Back in the early 1980s, when Prusiner burst on the scene, Gajdusek
was convinced the scrapie agent was an ‘unconventional virus’. He still
is. Taking Pasteur’s definition of a virus as a ‘small obligate parasite
requiring energy and information from a host organism’, he sees no conflict
in continuing to call the scrapie agent a virus even though he, like Prusiner,
is convinced that it carries no nucleic acid. ‘I use the term virus facetiously
– Pasteur would have applauded.’
Two worlds
Strip away the disputes about terminology, and most theories about scrapie-like
diseases begin to look very similar. In 1992, Michael Alpers of the Papua
New Guinea Institute in Goroka, one of Gajdusek’s colleagues, wrote wryly
in a symposium report: ‘The ideas converge, but no one would get that impression
from reading papers from the two groups; it is as if they were operating
in two different worlds.’
But a contentious difference remains: whether the PrP protein acts
as a self-contained agent or as something that is necessary for disease
but insufficient to cause it. ‘The abnormal form of the prion protein itself
is the central component of the infectious agent,’ says Collinge. ‘Whether
it’s the sole component remains to be seen.’
‘It probably is the protein – everything is pointing to that,’ says
Paul Brown of the NIH. ‘But so far not a single observation is inconsistent
with the view that the protein is a receptor for an environmental agent,
although no one has found such an agent, or any disease-specific foreign
nucleic acid.’ In Edinburgh, Bruce is equally adamant that the case remains
open. ‘Weissmann’s experiment shows that host PrP plays a key role and is
necessary for the disease to develop, but it does not prove that the protein
is sufficient for disease transmission.’
Of all Prusiner’s critics, the Edinburgh researchers are the least convinced.
Like most researchers, they have no problem with the idea that the PrP protein
is central to the disease process. But they believe that the infectious
agent must also contain an ‘informational molecule’, such as DNA or RNA.
In their scheme of things, this molecule – possibly a tiny length of nucleic
acid too small to be broken down – would be an essential accomplice of
the PrP protein. Packaged in a protective shield of abnormal PrP protein,
it would enter cells, shed its packaging and then go on to catalyse the
conversion of normal PrP into abnormal PrP. In Prusiner’s protein-only hypothesis,
the abnormal form of the PrP protein is the corrupting influence. But in
the Edinburgh model, this role is played by the informational molecule.
Until there is enough evidence to distinguish these two possibilities,
Bruce and her colleagues will continue to refer simply to ‘the agent’. But
Collinge takes a more practical approach. ‘I don’t think that calling it
the prion protein necessarily means that you’ve accepted that protein is
all there is to the infectious agent,’ he says. ‘I certainly haven’t reached
that point yet – I still think there is room for co-factors, for example
in explaining strain variation.’
The number of different strains in scrapie exposes the biggest weakness
in ‘protein-only’ thinking. Since 1965, Alan Dickinson and his colleagues
at the Neuropathogenesis Unit in Edinburgh, run jointly by the Medical Research
Council and the Agricultural and Food Research Council, have isolated some
twenty different strains of scrapie. Mice respond differently to infection
with these different strains. The incubation time before symptoms appear
varies from strain to strain, for example, as does the pattern of damage
that appears in their brains. Most importantly, the characteristics of
these strains remain consistent as the disease is transmitted from one
animal to another, or even, in some instances, beween different species
which have slightly different versions of the Prn-P gene.
Recently, Bruce has come up with further evidence showing that the different
strains are somehow imprinted with their own identity. She infected mice
with BSE from a wide variety of different animals, including naturally infected
cows, cats, two species of captive antelope and experimentally infected
sheep and goats. For each breed of mouse used in the experiment, the disease
developed in the same way, regardless of the source of the infection. In
each case the disease profile differed from that of natural sheep scrapie.
‘The fact that the BSE agent has maintained its identity on passage through
six different species adds to previous evidence that scrapie-like agents
have an informational molecule which determines disease characteristics,’
Bruce wrote in her report of this work to the Royal Society meeting last
September.
‘It’s difficult to explain this finding in terms of the protein-only
model,’ she says. ‘There isn’t an established mechanism by which the protein
could replicate and carry information.’
The physical nature of the informational molecule she and her colleagues
have proposed as an alternative remains elusive, however. ‘It’s only recently,
since the identification of the PrP protein, that we’ve been able to ask
questions like that,’ says Bruce. She herself has not given up on the idea
that some form of viral nucleic acid may be involved, probably in close
association with PrP. This is despite work from Prusiner’s group apparently
showing that purified PrPSc is infective on its own.
Brown is ‘disinclined to attach much importance to strains’. For Collinge,
however, the data on strains ‘are a key component of what we know about
these diseases. If you are constructing an explanation of prion replication,
you have to explain the phenomenon of strains. We don’t know if there are
human strains yet – they’ve only been documented in the context of scrapie
– but there may be human strains, and they may be relevant to the pattern
of disease in humans.’ His own hunch is that the properties of the strains
are intrinsic to the protein molecule, possibly something to do with its
pattern of glycosylation (the process by which sugars are attached to the
protein) or its three-dimensional shape.
Single spectrum
With the current level of knowledge, there seems to be nothing to choose
between various explanations of these variations in strain. But the latest
techniques in molecular biology are providing new ways of tackling the whole
question of what happens at the cellular and molecular level. Brown, Collinge
and other groups in Japan and elsewhere have amassed detailed knowledge
of the mutations in the PRNP gene that predispose people to CJD and related
diseases. It is now widely accepted that all these diseases should be regarded
as part of a single spectrum. Collinge supports the view that they should
now properly be known as the prion diseases. Whatever you call them, the
challenge is to find out how the mutations cause such devastation in the
brain.
The approach Collinge and others have chosen is to breed laboratory
mice that carry the mutant human gene – a feat that has become possible
only in the past two or three years. ‘Ultimately we hope to put in multiple
mutations to produce a protein that spontaneously adopts the abnormal form
in a test tube, and which would produce infectivity,’ says Collinge. Doing
that, he adds, would demonstrate that protein alone is sufficient for infectiveness.
‘But no one has yet produced prion replication outside the cell. It may
be that co-factors in the cell membrane are necessary.’
The other reason for developing the transgenic mouse lines is that even
with today’s imperfect level of knowledge, therapy for ‘prion diseases’
is at last a real possibility. Two approaches seem promising: to design
drugs that interfere with the abnormal protein, preventing it from accumulating,
or to use gene therapy to switch off production of the protein. ‘I’m optimistic
that over the next five to ten years we’ll have agents that will interfere
with these diseases,’ says Collinge. ‘It’ll take some time after that before
they reach clinical trials.’
Would we have got this far sooner without the posturing and backbiting
that went on in the 1980s? As with virtually all scientific controversies
born of intense competition, opinions vary. Bruce says: ‘In some ways it’s
been a problem because big clashes of personality get in the way of proper
research.’ But Collinge is not so sure. ‘If anything, the competition has
stimulated progress. I think it’s contributed to making people even more
keen to go back and do the experiments to prove the other side wrong.’
Prusiner’s own view of his role in the history of the subject is well
illustrated by the quotation, from the Belgian poet and dramatist Maurice
Maeterlinck, he chose to preface a collection of conference papers published
in 1992: ‘At every crossroads on the road that leads to the future, tradition
has placed against us ten thousand men to guard the past.’
Georgina Ferry is a freelance science writer and broadcaster.
* * *
Of mice, men and cattle
The first of the ‘prion diseases’ to be widely studied was scrapie,
which is endemic in the sheep flocks of Western Europe and has been known
for over two centuries. The brains of the affected animals suffer from massive
degeneration, they develop a staggering gait and other behavioural abnormalities
before they die. While scrapie never exactly grabbed the public imagination,
it is essentially the same disease as the much better-known BSE, or mad
cow disease, first identified in the mid-1980s. Indeed, the commission of
inquiry set up to investigate the outbreak of BSE attributed it to the practice
of including the rendered remains of slaughtered sheep, some of which may
have been infected with scrapie, in the high-protein diets fed to calves
and milk cows. This family of diseases, however, is much more than a veterinary
curiosity.
Several human diseases present a very similar picture, and more may
yet be identified. Creutzfeldt-Jakob disease, kuru, Gerstmann-Straussler-Scheinker
syndrome (GSS) and fatal familial insomnia make up the current tally. All
are rare, degenerative diseases of the brain. GSS, FFI and some cases of
CJD have been traced to defective genes; kuru was transmitted through handling
and possibly eating tissue from infected corpses during funeral rites in
Papua New Guinea; CJD can be passed on through medical procedures involving
the use of infected tissue or contaminated instruments, but most cases
occur sporadically, and without any obvious cause. The area of the brain
where most damage occurs, with consequent effects on behaviour, also varies
from one disease to another.
Recent research in molecular biology has shown that these diseases and
the animal diseases they resemble all have the same agent in common, and
has suggested a way of transmitting disease that is unlike anything in
the textbooks.
But do the latest laboratory findings have any implications for the
question of whether or not BSE has been, or could be, transmitted to humans?
According to John Wilesmith, an epidemiologist at the Central Veterinary
Laboratory in Weybridge, Surrey, the answer is not really. ‘The discovery
of the molecular structure of the agent could lead to the development of
a diagnostic test in live animals which could be useful in man,’ he says.
‘But as in the case of HIV, the molecular structure does not necessarily
provide all the answers about transmissibility.’
The focus of Wilesmith’s research is to try to establish how much of
the BSE epidemic was due to tissue from BSE-infected cattle, rather than
sheep, being included in cattle diets, as happened for a few years before
the practice was stopped. ‘The cannibalistic element may be important,’
he says. This ties in with laboratory research showing that it is much harder
to transmit prion diseases from one species to another than within the same
species.
However, the fact remains that at least one cow may well have developed
the infection after eating sheep products infected with scrapie, and that
is enough to cause widespread concern that the disease could pass from cattle
to humans. The only evidence to counter such concern is epidemiological,
and no such link has ever been demonstrated between the incidence of scrapie
or BSE and human brain disease. Australia and New Zealand, for example have
large populations of scrapie-free sheep; yet the incidence of CJD is the
same there as it is elsewhere. And according to Wilesmith, a new survey
funded by the European Union finds that cases of CJD occur in Britain at
about the same rate as in other European countries, where there is no BSE.
But critics rightly point out that epidemiology only tells you about
the risk of animal-to-human transmission in the population as a whole.
It cannot tell you how or where specific individuals contracted their diseases.