Introduction
Creutzfeldt-Jacob Disease (CJD), first characterized in 1921 by Alfons Jacob (Jacob, 1921), remains an enigmatic disease. CJD, along with diseases as kuru, Gerstmann-Straussler-Scheinker syndrome (GSS) and Fatal Familial Insomnia (FFI), are "prional" diseases caused by large aggregations of protein in the brain (1). These proteins cause macroscopic vacuolation of brain tissue and as a result the diseases are known as "spongiform encephalopathies."

Prional diseases uniquely do not appear to rely on nucleic acids for their transmission (Griffith, 1967). Diseases such as CJD apparently are the result of a polymerization reaction in with the disease isoform of a protein (designated PrPsc, for prion protein scrapie) causes a conversion of the normal host-encoded protein, PrPc (cellular), to the PrPsc isoform through an autocatalytic process (Prusiner, 1991 and Kocisko, 1994). PrPsc protein does not appear to have its amino-acid sequence modified although its glycosylation pattern almost certainly varies from PrPc. Therefore, the conversion is not a covalent altration but rather a conformational change. Since there are phenotypically distinguishable strains of CJD, it is hypothesized that each strain converts the protein in a unique manner.

Interest in Creutzfeldt-Jacob Disease recently has reached new levels due largely to an unusual outbreak of CJD in Britain, which has been associated with a epizootic of another prion disease, bovine spongiform encephalopathy (BSE). The epizootic was first discovered in 1986: since then, there have been over 164 000 cases of BSE reported in the United Kingdom (Coursens, 1997), affecting almost one-third of all cattle herds. BSE is unusually species-nonspecific for a prion disease, and has been shown to cause encephalopathies in many species, including cats and ungulates, sheep and goats (Bruce, 1994). The disease has been shown, surprisingly, to be transmitted by ingestion of BSE-infected brain homogenates.

Since the beginning of 1985, 26 individuals in Britain between the ages of 19 and 45 have been diagnosed with CJD (2). However, these 26 individuals exhibit features not normally associated with CJD, such as florid plaques rather than synaptic deposits of protein, unusual behavioral changes and ataxia, and a prolonged disease course. Most importantly, these 26 individuals have a median age of 28, compared to a normal CJD-onset disease of roughly 60. As a result, this strain of CJD has been called 'new variant' CJD (nvCJD). The coincidence of this new epidemic of CJD merely ten years after the discovery of BSE has sounded public-health alarms throughout the world, and led to much speculation in both the popular media and scientific circles that the etiological agent of BSE may be responsible for nvCJD. Two years after the first case of nvCJD was described, scientific data are beginning to provide support for this hypothesis.

Is "new variant" CJD really a new disease?
There are three distinct types of traditionally defined Creutzfeldt-Jacob disease. The most common kind, with an occurrence rate of about one in 106, is sporadic and apparently attributable to an unlucky mutation in the PRNP gene or happenstance development of the disease isoform. There is, as well, a genetically transmitted CJD which accounts for approximately fifteen percent of all cases (Windl, 1996) and is due to a coding mutation in PRNP. There have also been numerous cases of iatrogenic exposure, in which patients undergoing dura matter or corneal grafts, receiving human cadaveric pituitary-derived growth hormone or gonadotropin as treatment, or exposed to unclean instruments have acquired CJD (Weller, 1989).

The appearance of the new variant CJD in Britain has caused some to question whether this is in fact a new disease or rather a newly discovered, but historically occurring disease. The evidence points towards the conclusion that nvCJD is in fact, an emergent disease, and not one that has been hidden. While there is always a possibility that this disease has been undetected and historically classified as one of the other three forms of CJD, there are several data that make the chance of such a misclassification remote.

First, as mentioned above, the neuropathologies associated with nvCJD are different from the regular form of the disease. Second, there have been no previous reports of neuropathologies similar to that of nvCJD in the medical literature (Will, 1996) and the similarities between the neuropathologies of the ten cases of nvCJD makes them virtually indistinguishable, unlike the variations seen in traditional CJD. Finally, countries other than the UK with CJD surveillance (including the US and European Union) have not reported an increase in the incidence of CJD among those aged below 45 years. All of these data point to a disease focused in the UK with a new etiological agent.

Mad cows, mad humans? The evidence that BSE may cause nvCJD
Because of the limited number of cases of nvCJD and the relatively recent initial outbreak of the disease, data are only now beginning to shed light on a possible direct link between nvCJD and BSE. Support for BSE as the etiological agent of nvCJD comes form three lines of evidence: first, the epidemiology of nvCJD; second, the relative transmissibility of BSE in other mammals; and third, molecular evidence, published recently. Together, these three lines of evidence provide a compelling case for a link between BSE and nvCJD.

Epidemiology of the outbreak
In 1994, initial speculation that BSE might be the causative agent of CJD was based solely on the correlation between seven surprising cases of CJD and the mad-cow disease outbreak in Britain. Without any hard evidence, the media -- in particular, tabloid newspapers -- implicated BSE as the etiological agent behind the new cases of CJD. The connection, although considered somewhat specious at the time (3), has been substantiated better by other data that have come to light in the ensuing two years.

The first piece of the epidemiological jigsaw is that Britain is the only country in the world with an extensive BSE problem. Save for one occurrence of nvCJD in France, Britain is also the only nation to have reported the new disease. (The French case can be explained by the low rate of BSE found in French herds or alternatively may be attributed to British beef imported into France, before the European Union banned British beef exports in 1995.) It is unlikely that many CJD cases are not reported. The problem of BSE underreporting in other countries may be cause for some debate, as farmers try to avoid having their entire herd culled, but most human cases of CJD are pathologically determined. Because autopsies are performed at relatively high rates in Western countries, it is a fair assumption that most cases of CJD, at lest in the developed world, are reported.

Second, and more compelling, there have been no reports of nvCJD in countries putatively free of BSE, such as the United States. If the disease's agent were found in both the US and the UK, the US should have had some sixty cases of nvCJD reported between 1994 and 1996. Indeed, even on continental Europe, with a population slightly higher than the US and an extremely low rate of BSE infection, there has only been the one case of nvCJD in France. The lack of any case of nvCJD in a country without BSE provides a solid correlation between presence of BSE in cattle and nvCJD in humans, and is the most striking epidemiological evidence.

Pathogenicity of BSE upon ingestion
A controlled study of the infectivity of BSE in humans could never be performed, and its is virtually impossible to determine retrospectively whether any of the victims actually consumed BSE-infected beef. Evidence does, however, show that BSE may be infective across the species boundary. These data are surprising because, based on the prion model of inheritance, the lack of high homology between different species' prional proteins theoretically ought to be an effective barrier against infection. Bruce and colleagues (1994) have demonstrated that many species, including sheep, goats, and some ungulates can be infected with BSE through an oral route of transmission and maintain infectiousness in subsequent experiments that assayed transmissibility to mice -- although the mice were subsequently inoculated intracerebrally.

Other data are not so convincing. Evidence presented by Collinge and colleagues (1996) in a paper that supports the possibility that BSE causes nvCJD may argue exactly the opposite: that the BSE agent cannot be transmitted to humans. While BSE can be transmitted to wild-type FVB and C57BL mice easily, it has not been transmitted to mice, designated HuPrP^+/+ Prn-p^0/0, transgenic for the human prional protein. While Dr. Collinge argues that this may be explained by the low frequency of BSE infection in humans or a long incubation period, the pathogenicity of BSE in these transgenic mice -- which lack a species barrier to human prional diseases such as kuru -- is undoubtedly lower than in nontransgenic mice.

Finally, Ridley and colleagues, reporting on the inefficiency of BSE transmission to macaques (Ridley, 1996), claim that despite exposure to the BSE agent, more than 100 of their macaques have failed to develop a spongiform encephalopathy. These data are particularly unreliable, because they are the result of observation and not a controlled experiment. Nevertheless, given that the BSE agent has been found in the meal fed to these macaques, this may be evidence for a low rate of transmission in higher mammals. Unfortunately, animal models such as this one, and those employed by Collinge and Bruce (Collinge, 1996 and Bruce, 1994), make for very poor models of transmission to humans by the oral route because of physiological, immune, and prion-protein (except for the case of transgenic mice) differences.

Molecular analysis
While the other results outlined above convinced many, including policy makers, that BSE could cause nvCJD, molecular evidence that the etiological agents of nvCJD and BSE are identical is necessary. This is ultimately very difficult to do, because the proteins that cause disease are host encoded and therefore the amino-acid sequences of the proteins will vary as usual between organisms. Crystallographic data would provide the most solid evidence of identical protein conformations among affected humans and cattle, but none has yet been produced.

Strains of prional disease are hypothesized to be based on differences in conformation, and possibly glycosylation, of the host protein (Bessen and Marsh 1994). This evidence is based on differences in migration patterns on an SDS-polyacrylamide gel of two types of mink spongiform encephalopathy, "hyper" (Hy) and "drowsy" (Dr), after limited treatment with proteinase K. It implicates the conformation forced on PrPc by PrPsc as the cause of the difference in neuropathology associated with different forms of the disease. The disease progression for Hy as the name implies is rather rapid, and the Dr strain causes neuropathology more slowly.

Recent experiments (Collinge, 1996) at the molecular level give credence to the widely accepted BSE-nvCJD link, though the data are far from proof. This research entailed comparing the four different types of CJD (genetic, sporadic, iatrogenic, and "new variant") to each other as well as to BSE.

The authors of this paper analyzed the prional proteins from 43 neuropathologically confirmed cases of CJD, including ten nvCJD cases. The authors discovered that, among the traditionally seen cases of CJD, there were three distinct banding patterns on Western analysis after treatment with proteinase K. Types I and II were found to correspond to either the genetic, sporadic, or iatrogenic CJD. Type III was exclusively found in iatrogenic cases. These first three types of CJD differed only in the molecular weights of three fragments of the PrPsc, but had virtually identical band intensities. The type IV was seen solely in cases of nvCJD and showed a startlingly different banding pattern. Other evidence showed a unique glycosylation pattern to nvCJD, further demonstrating that the disease's agent is different that those of the other three types of CJD. Interestingly, PrP from all ten cases of nvCJD was heterozygous for valine at residue 129, implying that there may be a genetic predisposition for contracting nvCJD.

This evidence is confined by the relatively small sample size, particularly for some of the data on iatrogenic cases. Controls are nonexistent for many experiments, in particular, for the study of the valine-methionine polymorphism at amino acid 129. Also, the authors fail to show glycosylation patterns for all three types of traditional CJD, making visual confirmation of their claim impossible.

Further evidence takes the form of comparison studies between the banding patterns, again on Western analysis, of nvCJD to the BSE from infected mice, macaques, cats and, of course, cows. The glycoform signature of new variant CJD was statistically not different to the pattern seen in all of the other cases. This is in agreement with the idea that transmission of BSE to humans causes nvCJD. These data are completely subjective, however, as even the authors admit. There is no basis for normalization of the results, and given the lack of transmission to transgenic mice, these data basically only serve to confirm the findings (Bruce, 1994), that BSE can be transmitted to non-bovine species.

Assessing the evidence
Each of the three lines of evidence, taken individually, is subject to criticism. The epidemiological data only show a correlation, not causation. Studies in other animals may have no relation to what happens in humans. Finally, the molecular evidence is largely circumstantial in that the glycosylation patterns appear similar, but this does not implicate for certain BSE as the causative agent for nvCJD.

Taken together, however, the data are more compelling. Although all three lines of evidence are each not sufficient to prove that BSE causes nvCJD, as a whole there is no singe way to refute the hypothesis, based on the evidence at hand. Interestingly, there are no studies that compellingly argue against the claim that BSE causes new variant CJD. This lack of any meaningful data to refute the BSE-CJD link is another critical element in support of the hypothesis. Importantly, the data also make it almost completely impossible that pure chance is responsible for the development of nvCJD.

Arguments that agents other than prions may be at work, for instance, are inherently dubious because they do not necessarily change any interpretation of the data. For instance, if a nucleic-acid based infectious agent, such as a virus or some smaller, unidentified DNA or RNA particle, caused the abnormal protein to form, it would not change the fact of the evidence presented above. As an example, if a virus were responsible for the pathology associated with both BSE and nvCJD, it would not change the interpretation of any of the data above, though it might lead to new strategies for prevention.

A recent article in Science (Lasmezas, 1997), for instance, claimed to show that the BSE agent could be transmitted without any detectable PrPsc. This study was flawed for several reasons, including the lack of PrP^0/0 mice, too-high detection threshold, and a lack of screening for microscopic vacuolation. Indeed, the best support for the protein-only hypothesis comes from the inability to initiate prional diseases in animals lacking the PrP gene. Furthermore, protein polymerization is often a favored cascade thermodynamically (4) that is kept in check by relatively low body temperatures and rapid protein digestion.

Policy Implications
On March 27, 1996, the European Union decided to ban British exports of beef, live bovine animals or embryos, any mammalian-derived meat and bone meal, and "any products liable to enter the human food chain" that were derived from cows. This action was made in response to publications at the time (Will, 1996 and Tabrizi, 1996) that pointed clearly to a link between the then still-mysterious cases of CJD and BSE. Subsequently, the European Union forged a scathing critique of British policies on the prevention of BSE in an inquiry into the BSE outbreak (European Parliament, 1996).

The British had banned specified offal from the food chain in 1989 in response to the BSE epidemic, which had developed during the mid 1980s, but failed to ensure that the ban on the feeding of meat and bone meal to ruminants was effective. The British, for instance, did not have any legal penalties for the storage or administration of such feed. There was no surveillance mechanism for herds in which BSE had been detected until 1994. Finally, the British did not adequately monitor or maintain the embargo on beef products (European Parliament, 1996).

Nevertheless, the UK now has the world's most stringent rules for the control of BSE, with all cattle in a diseased herd culled or at a minimum prevented from market. The policy failures of the past have been replaced with strict pan-European guidelines on the production and distribution of beef and bovine products.

If BSE is the etiological agent for nvCJD, and assuming a ninety percent effective ban on bovine offal, the CJD epidemic is likely to continue for a number of years. Mathematical models of the epidemic (Coursens, 1997) predict that, if the mean incubation is short for CJD, the number of cases will be around 100. If the mean incubation time is longer, say 25 years, with a low deviation from the mean, the UK could be in for a catastrophic epidemic of some 80 000 individuals eventually contracting new variant CJD.

North America has, for the moment, been spared such a prospect, but conditions are ripe in the US, in particular, for a repeat of the British experience with BSE and CJD. The US Department of Agriculture (USDA) currently places few effective restrictions on the use of animal scraps in feed for ruminants (Kluger, 1997) -- the same mechanism which is posited to have caused to the British BSE outbreak. Given that BSE can occur spontaneously, unmitigated good fortune is all that has kept US herds from becoming infected with BSE. On the other hand, given that BSE is known to have a relatively long incubation time of four years, detection of BSE in one cattle herd would almost surely indicate a widespread problem. The simple solution for the US is to adopt European-like restriction on the use of offal. The cost of a BSE epidemic in terms of human life and lost revenue due to cattle culling would surely outweigh the some $100 million it would take to institute such new regulations on the US cattle industry.

Conclusions
The recent outbreak of a new variant of Creutzfeldt-Jacob disease in Britain is most likely related to the epizootic of Bovine Spongiform Encephalopathy that swept British cattle herds beginning in the mid-1980s. The evidence in support of this conclusion comes from epidemiological, animal, and molecular evidence, which together show that the etiological agent of BSE is identical at the molecular level to that of nvCJD and also that the prion can be acquired, at however low of a rate, by an oral route of transmission, across species boundaries. While exposure to BSE is highly unlikely in Britain anymore, there is a considerable chance that past exposure to BSE will cause the epidemic of nvCJD to continue, at least until the turn of the century.

1. Prion diseases are not limited to humans and cows.  The sheep disease scrapie, for example, is a prion disease that is enzootic in many countries because of husbandry methods.  Prion-like agents have also been found in yeast, the simplest eukaryotes.  The Sup35 protein can be altered in vivo to aggregate.  The determinant that causes the prion-like state, known as [PSI⁺] is cytoplasmically inherited and can be propagated in cell-free systems (Paushkin, 1997). Another yeast prion-like agent, URE3, has also been isolated.

2. My data are from the CJD Surveillance Unit in Edinburg, Scotland. As of 30 June 1997, the CJDSU reports 20 confirmed deaths from nvCJD and one case still alive, a total of 21. The other five cases are reported over the ProMED-mail network and are to the best of my knowedge true.

3. Interestingly, initial reports of the BSE-CJD link caused the UK's minister of agriculture to appear on the BBC with his son, hamburgers in hand, to deny that eating British beef posed any health threat whatsoever.

4. By this I mean that proteins will, in the absence of some agent to prevent them from doing so (such as a proteinase), tend to form aggregates; this is why x-ray crystallography is possible. In particular, those proteins that are largely hydrophobic will precipitate from aqueous solution at relatively low concentrations. The crystallization of proteins is not, of course, favored from an entropy standpoint.

Windl, O. et al. "Genetic basis of Creutzfeldt-Jakob disease in the United Kingdom: a systematic analysis of predisposing mutations and allelic variation in the PRNP gene" Human Genetics September 1996 (98) 259-264.