Prion diseases are transmissible neurodegenerative disorders which affect a range of mammalian species. In humans they can be inherited and sporadic as well as acquired by exposure to human prions. Prions appear to be composed principally of a conformational isomer of host-encoded prion protein and propagate by recruitment of cellular prion protein. Recent evidence argues that prion protein can also encode disease phenotypes by differences in its conformation and glycosylation. Such molecular analysis of prion strains suggests that new variant Creutzfeldt-Jakob disease is caused by BSE exposure. The novel biology of prion propagation may not be unique to these rare degenerative brain diseases.
The human prion diseases or transmissible spongiform encephalopathies have been traditionally classified into Creutzfeldt-Jakob disease (CJD), Gerstmann-Sträussler-Scheinker disease (GSS) and kuru. They are rare neurodegenerative disorders affecting ~1 person per million world-wide per annum. Despite their rarity, remarkable attention has been recently focused on these diseases. This is because of both their unique biology and concerns that the epidemic of a newly recognised bovine prion disease, bovine spongiform encephalopathy (BSE), could pose a threat to public health through dietary exposure to infected tissues.
Transmissibility of the human diseases to experimental animals was demonstrated with the transmission, by intracerebral inoculation with brain homogenates into chimpanzees, of first kuru and then CJD in 1966 and 1968 respectively (1 ,2 ). Transmission of GSS followed in 1981 (3 ). The prototypic prion disease, scrapie, is a naturally occurring disease of sheep and goats, and has been recognised for >200 years (4 ). Scrapie was demonstrated to be transmissible by inoculation in 1936 (5 ) and it was the recognition that histopathologically kuru, and then CJD, resembled scrapie that led to suggestions that these diseases may also be transmitted by inoculation (6 ,7 ). Kuru is a neurodegenerative condition characterised principally by a progressive cerebellar ataxia which reached epidemic proportions amongst the Fore linguistic group in the Eastern Highlands of Papua New Guinea and which was transmitted by cannibalism. Since the cessation of cannibalism in the 1950s the disease has declined but a few cases still occur as a result of the long incubation periods in this condition (8 ). Transmission of CJD from case to case has occurred by a number of routes involving accidental inoculation with human prions as a result of medical procedures. Such iatrogenic routes include the use of inadequately sterilised neurosurgical instruments, dura mater and corneal grafting, and use of human cadaveric pituitary derived growth hormone or gonadotrophin.
The epidemiology of the human prion diseases encompasses the three aetiological types of prion disease: inherited, sporadic and acquired (Table 1 ). Inherited cases constitute ~15% of cases, acquired cases (kuru and iatrogenic CJD) are rare. Most human cases are sporadic, and their precise aetiology is still unclear. Both human and animal prion diseases share common histopathological features. The classical triad of spongiform vacuolation (affecting any part of the cerebral grey matter), astrocytic proliferation and neuronal loss, may be accompanied by the deposition of amyloid plaques (9 ). These plaques are composed principally of an abnormal, partially protease resistant isoform of a host encoded sialoglycoprotein, prion protein (PrP). PrP is expressed widely but at much higher levels in neuronal cells as a glycosylphosphatidyl inositol-anchored cell surface protein (10 ).
The transmissible agent, or prion, seems to consist principally of an abnormal isoform of the prion protein (PrP), designated PrPSc (11 ). PrPSc is known to be derived from the cellular isoform, PrPC, by a post-translational mechanism (12 ,13 ) and evidence is mounting that this change is conformational rather than covalent (14 ,15 ).While PrPC is fully sensitive to proteolysis, PrPSc, which accumulates in the brain during disease, is partially protease resistant. The human PrP gene (PRNP) is a single copy gene located on the short arm of chromosome 20 (16 ); it spans 16 kb and contains two exons (17 ,18 ). The complete open reading frame of 759 nucleotides is contained within the larger second exon which comprises the majority of the 2.4 kb mRNA. A key advance in the understanding of prion biology was the identification of mutations in PRNP in the familial forms of the human diseases. The first mutation to be identified in PRNP was in a family with CJD and constituted a 144 bp insertion into the coding sequence (19 ). A second mutation was reported in two families with GSS and genetic linkage was confirmed between this missense variant at codon 102 and GSS, confirming that GSS was an autosomal dominant Mendelian disorder (20 ). This genetic variant was not present in the normal population but was rapidly identified in numerous other unrelated GSS kindreds. These diseases are therefore both inherited in the germline and transmissible by inoculation, and are biologically unique in this regard.
The identification of one of the pathogenic PrP gene mutations in a case with neurodegenerative disease allows not only molecular diagnosis of an inherited prion disease (21 ) [and pre-symptomatic testing of at risk individuals (22 )] but also its sub-classification according to mutation. Pathogenic mutations reported to date in the human PrP gene consist of two groups: (i) point mutations within coding sequence resulting in amino acid substitutions in PrP or in one case production of a stop codon resulting in expression of a truncated PrP; (ii) insertions encoding additional integral copies of an octapeptide repeat present in a tandem array of five copies in the normal protein (Fig. 1 ). The availability of direct gene markers for these diseases has enabled identification of highly atypical cases and has widened the known phenotypic range of these disorders (23 ,24 ). Remarkable phenotypic variability may be present in the same family, with phenotypes ranging from classical CJD to atypical dementias with extremely long duration and lacking the typical histological features of spongiform encephalopathy (25 ). Clinical and pathological syndromes of the human inherited and other prion diseases are reviewed in detail elsewhere (26 ).
Genetic susceptibility is also relevant to both the sporadic and acquired prion diseases. There is a common polymorphism of the human prion protein, with either methionine or valine present at residue 129. In Caucasians, ~38% are homozygous for the more frequent methionine allele, 51% are heterozygotes and 11% are homozygous for valine (27 ). The large majority of sporadic CJD occurs in individuals homozygous for this polymorphism (28 ) and most pituitary hormone related iatrogenic CJD cases are homozygotes, with a particular excess of valine homozygotes (29 ). This protective effect of PRNP heterozygosity is also seen in some inherited prion diseases where the age at onset of disease is 1-2 decades later in heterozygotes at PRNP codon 129 as compared with homozygotes (30 ,31 ).
Although prion diseases can be transmitted between mammalian species by inoculation, in practice it is difficult to do so, and typically transmission occurs in only a small proportion of inoculated animals and then only after prolonged incubation periods on primary passage. On second and subsequent passage of infectivity into animals of the same species, incubation periods are typically much shortened, relatively synchronous as compared with primary passage, and all inoculated animals succumb to disease. A key component of this so-called `species barrier' is the degree of homology between the PrP in the donor and recipient species. Mice transgenic for hamster PrP lose their species barrier to hamster prions (32 ). When such transgenic mice were challenged with mouse derived prions, the infectivity they produced was fully pathogenic for mice but not hamsters, while on challenge with hamster derived prions they produced prions fully pathogenic to hamsters but not mice. The interpretation of this finding is that a direct interaction between PrP molecules occurs at some stage in the process of prion propagation and that such interaction occurs most easily if the interacting PrP molecules are identical. Such work led to the idea that replication of prions may occur by PrPSc interacting directly with PrPC and catalysing the conversion of PrPC to PrPSc (32 ,33 ). Such an effect could then lead to a chain reaction of conversion leading to the progressive conversion of increasing amounts of PrPC to PrPSc. The pathogenic mutations in the prion protein may result in the production of a protein that converts spontaneously to PrPSc in individuals with inherited prion diseases. Such a model provides an explanation of how a disease can be simultaneously inherited and transmissible. The finding that nearly all sporadic CJD occurs in homozygotes with respect to a common and apparently innocent protein polymorphism lends strong support to such a mechanism (28 ). Again prion protein interaction would occur most favourably in individuals with two identical copies of the prion protein. Heterozygotes, producing two different proteins, would be somewhat protected, as if by an internal `species barrier'. It is possible that the occasional individuals heterozygous at codon 129 who do develop sporadic CJD have a more prolonged illness although more detailed studies are required to investigate this further (34 ). PrP valine 129 and PrP methionine 129 would be expected to differ slightly in their propensity to conversion to PrPSc. The excess of PrP valine 129 homozygotes amongst human pituitary hormone related cases suggests that PrP valine 129 may be the more susceptible (29 ).
A high resolution structure of full length PrP has not yet been reported, but an NMR structure determination of a mouse PrP fragment spanning residues 121-231 represents an important step forward (37 ). However, this structure lacks a highly conserved hydrophobic region (106-126), which contains several of the known pathogenic human mutations, as well as the N-terminal octapeptide repeat region, expansion of which by integral numbers of repeat elements results in a group of inherited prion diseases. The conformation of this repeat region has been studied by circular dichroism and appears to form an extended, flexible domain with properties similar to the poly-L-proline type II helix (38 ). It has been hypothesised that this may form a low specificity binding domain.
Despite the wealth of evidence indicating the central role of PrP in these diseases, its normal cellular function remains unclear. Mice homozygous for PrP null alleles appear to develop and behave normally (39 ). Such mice are completely resistant to developing prion disease following inoculation and do not propagate infectivity (40 ,41 ). However, electrophysiological studies have demonstrated impaired GABAA mediated synaptic inhibition in hippocampal brain slices maintained in vitro and also reduced long-term potentiation (42 ), an abnormality independently confirmed in a different PrP knockout line (43 ). However, others have failed to replicate this finding, albeit using different techniques and experimental conditions (44 ). That this phenotype is rescued by expression of human PrP in such mice confirmed its specificity for PrP (45 ). These abnormalities of synaptic inhibition are reminiscent of the neurophysiological abnormalities seen in patients with CJD and in scrapie infected mice (46 ), raising the possibility that prion neurodegeneration may be, at least in part, due to loss of PrP function rather than to a deleterious effect of PrPSc (42 ). The relative normality of PrP null mice, which do not develop progressive neurodegeneration, could result from effective adaptive changes during neurodevelopment. It is possible that the sequestration of PrPC into PrPSc during prion disease in a developed nervous system lacking such plasticity has more severe consequences. Certainly, PrPSc does not appear to be toxic to cells lacking PrPC (47 ) and PrPSc is difficult to detect in some cases of the inherited human prion disease fatal familial insomnia (FFI) (48 ), and in transgenic mice infected experimentally with FFI (49 ), suggesting that, at least in these cases, neurodegeneration may not be the result of PrPSc toxicity. According to such a model, the gain of function implied by the autosomal dominant mode of inheritance could reflect the formation of PrPSc which then functionally depletes PrPC by a dominant negative effect.
Two additional phenotypes have now been described in PrP null mice, abnormalities of circadian rhythms (50 ) and cerebellar Purkinje cell degeneration (51 ). Further neurophysiological abnormalities in the hippocampus of PrP null mice, including disrupted Ca2+-activated K+ currents and abnormal intrinsic properties of CA1 pyramidal cells have recently been reported (52 ,53 ).
Extensive epidemiological studies of CJD have been performed in a number of countries and all obtained broadly similar results (35 ). The overall annual incidence in most studies was ~1 case per million. No significant case clustering is present other than with respect to familial clusters. Cases are distributed apparently at random with a frequency related only to local population density. In particular, there is no evidence for case-to-case spread or association with local scrapie prevalence. For instance, CJD is as common in Australia and New Zealand, which have been scrapie free for many years, as in the UK where scrapie is endemic. Numerous case control studies have not shown consistent associations with particular occupational groups or dietary components.
The core clinical syndrome of CJD is of a rapidly progressive dementia usually with myoclonus. The onset is usually in the 45-75 year age group with peak onset between 60-65. Sporadic CJD is exceedingly rare in individuals under age 30. The clinical progression is typically over weeks progressing to akinetic mutism and death often in 2-3 months. Around 70% of cases die in under 6 months.
However, in late 1995 two cases of sporadic CJD were reported in the UK in teenagers (54 ,55 ). Only four cases of sporadic CJD had previously been recorded in the literature in teenagers, and none of these cases occurred in the UK. In addition, both cases were unusual in having kuru-type plaques, a finding seen in only ~5% of CJD cases. Soon afterwards a third very young sporadic CJD case occurred (56 ). These cases caused considerable concern and the possibility was raised that they might suggest a link with BSE. By March 1996, further extremely young onset cases were apparent and review of the histology of these cases showed a remarkably consistent and unique pattern (57 ). These cases were named `new variant' CJD although it was clear that they were also rather atypical in their clinical presentation; in fact most cases did not meet the accepted clinical diagnostic criteria for probable CJD (57 ) and, in some respects at least, clinically more closely resembled kuru (58 ). Extensive studies of archival cases of CJD or other prion diseases failed to show this picture and it seemed that this did represent the arrival of a new form of prion disease in the UK. The statistical probability of such cases occurring by chance was vanishingly small and ascertainment bias seemed most unlikely as an explanation. It was clear that a new risk factor for CJD had emerged and appeared to be specific to the UK. The UK Government advisory committee on spongiform encephalopathy (SEAC) concluded that, while there was no direct evidence for a link with BSE, exposure to specified bovine offal (SBO) prior to the ban on its inclusion in human foodstuffs in 1989, was the most likely explanation. A case of the new variant was reported in France soon after (59 ). PRNP analysis showed that all cases available for study were homozygous, for methionine at codon 129, and that no known or novel pathogenic mutations were found in PRNP coding sequence (60 ). Recently, transmission of BSE to three macaques has been reported, with production of histopathological appearances that are closely similar to the very unusual pattern of pathology seen in new variant CJD (61 ), providing further evidence that these cases may result from BSE transmission. However, detailed comparisons with transmissions of typical CJD cases to macaques will be needed to fully evaluate the significance of this finding.
If these cases do represent transmission of BSE to humans it is unclear why none had a pattern of unusual occupational or dietary exposure to BSE; it is possible that they represent a genetically susceptible sub-population. In addition, it is unknown why this age group should be particularly affected. However, little is known about which foodstuffs contained high-titre bovine offal. It is possible that certain foods containing particularly high titres were eaten predominately by younger people. It is also possible that children have an inherently higher intrinsic susceptibility or shorter incubation period following dietary exposure to the BSE agent. Possibilities might include, for example, differences in gut permeability or lymphoreticular system or higher PrP expression levels.
Direct experimental evidence that new variant CJD may be caused by BSE was provided by molecular analysis of human prion strains (PrPSc typing) (see below). The identification of a biochemical marker that distinguished new variant CJD from previously recognised forms of sporadic and iatrogenic CJD confirmed that this was a novel variant and implied that it was caused by exposure to a single novel prion strain type (62 ). The lack of any history of iatrogenic exposure to humans in these patients suggested that this was a novel animal prion. The molecular strain characteristics of new variant CJD closely resembled those seen in BSE itself and BSE transmitted to mice, domestic cat and macaque, consistent with BSE being the origin of new variant CJD (62 ). Such PrPSc markers are detectable in tonsil (63 ), which may allow pre-mortem diagnosis of new variant CJD without brain biopsy, and may potentially allow earlier diagnosis.
Experimental transmission studies of the human prion diseases have, until recently, largely involved transmission to laboratory primates, in particular chimpanzees and squirrel monkeys. The extensive experience of the NIH group, reporting >300 successful transmissions has recently been summarised (64 ). However, transmission studies in primates are severely limited by the expense of such studies and by ethical concerns as to the use of primates in such work. Attempts by most laboratories to transmit human prions to rodents have been fairly unsuccessful, with only occasional transmissions occurring and then at very prolonged incubation periods close to the natural lifespan of the mice. Some groups have, however, reported more frequent transmissions (65 ). Recently transgenic mice have become available which have increased susceptibility to human prions. Mice expressing a chimaeric human/mouse PrP were susceptible to three CJD isolates with short incubation periods (66 ). It has now become clear that transgenic mice expressing wild-type human PrP, but not mouse PrP, are highly susceptible to CJD, with all inoculated mice succumbing at short incubation periods usually in the range of 180-220 days (67 ,68 ). Such mice appear to lack a species barrier to human prions and can now be used for extensive studies of the transmission characteristics of human prion diseases and to bioassay human prions (68 ).
A major problem for the `protein-only' hypothesis of prion propagation has been how to explain the existence of multiple isolates or strains of prions. Such strains are distinguished by their biological properties: they produce distinct incubation periods and patterns of neuropathological targeting in inbred mouse lines. As they can be serially propagated in inbred mice with the same Prn-p genotype they cannot be encoded by differences in PrP primary structure. Furthermore, strains can be re-isolated in mice after passage in intermediate species with different PrP primary structures (69 ). Understanding how a protein-only infectious agent could encode such phenotypic information has been of considerable biological interest.
Support for the contention that strain specificity is encoded by PrP alone is provided by study of two distinct strains of transmissible mink encephalopathy prions which can be serially propagated in hamsters, designated hyper (HY) and drowsy (DY) (70 ). These strains can be distinguished by differing physiochemical properties of the accumulated PrPSc in the brains of affected hamsters (71 ). Following limited proteolysis, strain specific migration patterns of PrPSc on polyacrylamide gels can be seen. DY PrPSc appears to be more protease sensitive than HY PrPSc producing a different banding pattern of PrPSc on Western blots following proteinase K treatment. This relates to different N-terminal ends of HY and DY PrPSc following protease treatment and implies differing conformations of HY and DY PrPSc (72 ). Furthermore, the demonstration that these strain specific physiochemical properties can be maintained during in vitro production of protease resistant PrP, when PrPC is mixed with HY or DY hamster PrPSc, further supports the concept that prion strains involve different PrP conformers (73 ).
Recently, several human PrPSc types have been identified which are associated with different phenotypes of CJD (62 ,74 ). The different fragment sizes seen on Western blots following treatment with proteinase K suggests that there are several different human PrPSc conformations. However, to fulfil the criteria of strains, these patterns must be transmissible to animals both in same and in different species. Remarkably, this is the case, with both PrPSc fragment sizes and the ratios of the three PrP glycoforms (diglycosylated, monoglycosylated and unglycosylated PrP) maintained on passage in transgenic mice expressing human PrP (62 ). Transmission of human prions and bovine prions to wild-type mice results in murine PrPSc with fragment sizes and glycoform ratios which correspond to the original inoculum (62 ). New variant CJD is associated with PrPSc glycoform ratios which are distinct from those seen in classical CJD. Similar ratios are seen in BSE and BSE when transmitted to several other species (62 ). These data strongly support the `protein only' hypothesis of infectivity and suggest that strain variation is encoded by a combination of PrP conformation and glycosylation. Transmission of PrPSc fragment sizes from two different sub-types of inherited prion disease to transgenic mice expressing a chimaeric human mouse PrP has also been reported (75 ). As PrP glycosylation is thought to be a co-translational process, the different glycoform ratios may represent selection of particular PrPC glycoforms by PrPSc of different conformations. According to such a hypothesis, PrP conformation would be the primary determinant of strain type with glycosylation being involved as a secondary process. However, since it is known that different cell types may glycosylate proteins differently, PrPSc glycosylation patterns may provide a substrate for the neuropathological targeting that distinguishes different prion strains. Particular PrPSc glycoforms may replicate most favourably in neuronal populations with a similar PrP glycoform expressed on the cell surface. Such targeting could also explain the different incubation periods which also discriminate strains, targeting of more critical brain regions, or regions with higher levels of PrP expression, producing shorter incubation periods.
Such studies may allow a new molecular classification of human prion diseases; it is likely that additional PrPSc types or strains will be identified. This may well open new avenues of epidemiological investigation and offer insights into causes of `sporadic' CJD. In addition, PrPSc typing can be applied to other species; it is already apparent that PrP glycoform analysis alone can distinguish a number of mouse passaged scrapie strains (76 ). The combination of fragment size and glycoform analysis should allow better resolution and might be applied, for instance, to study if BSE has transmitted to other species involved in the human diet such as sheep. It will be necessary to study the full range of sheep strains to determine whether these can all be distinguished at a molecular level from BSE, it is possible that more refined molecular techniques may be necessary (77 ). Classical strain typing, involving mouse transmissions, takes 1-2 years and can only be applied to limited numbers of cases. Molecular strain typing takes days and allows large scale screening.
The ability of a protein to encode a disease phenotype has important implications in biology, as it represents a non-Mendelian form of transmission. It would be surprising, and also itself intriguing, if evolution had not used this mechanism for other proteins in a range of species. The recent identification of prion-like mechanisms in yeast is particularly interesting in this regard (78 ,79 ).
As BSE appears to have transmitted to humans, a key issue is the extent of the bovine to human species barrier. Clearly, this cannot be measured directly since this would require inoculation of humans with BSE. However, transgenic models may offer a way to address this issue, at least in part. The principal determinants of the `species barrier' are the degree of homology between PrP molecules in the host and inoculum (32 ) and strain of agent; however, BSE appears to be caused by a single prion strain (69 ). Transgenic mice expressing human PrP, which are competent to produce human PrPSc and `human' prions on CJD challenge, offer an opportunity to address the issue as to whether it is possible for bovine prions to induce production of human PrPSc. To date results are reassuring. Incubation periods to BSE were unaltered in mice expressing human PrP in addition to mouse PrP, and only mouse PrPSc appeared to be produced (68 ). A potentially more revealing experiment is to challenge mice expressing only human PrP with BSE (68 ). No transmission occurred at incubation periods well beyond those of CJD in these mice (68 ); it remains to be seen if transmission will occur at longer incubation periods. These mice express valine at polymorphic codon 129 of PRNP and these studies are being repeated with mice transgenic for human PrP methionine 129. This is of particular importance given that all new variant CJD cases seen to date are methionine 129 homozygotes.
However, even the presence of a highly effective species barrier between cattle and humans does not exclude some degree of BSE transmission to humans, given the very large numbers of people that have been exposed. Genetic susceptibility may well be important in this regard and in particular, PRNP codon 129 homozygotes would be expected to be at considerably higher risk than heterozygotes, although it is possible that heterozygotes may simply have longer incubation periods.
Human Molecular Genetics
Pages
Introduction
Molecular Genetics Of The Human Prion Diseases And A Model For Prion Propagation
Structure And Function OF PrP
Creutzfeldt-Jakob disease and the `new variant' putatively linked to BSE
Transmission studies
Molecular basis of prion strains
Bovine to human species barrier
References
REFERENCES
This page is maintained by OUP admin. Last updated Fri Sep 12 18:09:23 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1997