BSE to Humans: official announcement
Nature commentary: Showing the strain
Exclusive!: commentary by Dr. Stephen Dealler
A suspicious signature: commentary by Aguzzi and Weissmann
Nature v-CJD full text strain article

BSE to Humans: the announcement

At a press conference held this morning Wed, 23 Oct 1996 at the Wellcome Foundation, London, it was announced that John Collinge and his team have found a means of using the biochemical properties of the prion protein to trace the likely cause of V-CJD. They have found that the biochemical "signature" of prions in patients who died of V-CJD matches that of prions in mice and macaque monkeys experimentally infected with BSE, but differs from that of prions in human patients with acquired or sporadic CJD. This method has enabled a conclusion to be reached earlier than by waiting for the results of the strain-typing experiments with mice at Edinburgh.

The full account appears in a paper to be published in Nature Thursday (24 October), page 685. A discussion of the work, by Adriano Aguzzi and Charles Weissmann, appears in the same issue on page 666.

Showing the strain

Prion diseases such as bovine spongiform encephalopathy (BSE) in cattle or Creutzfeldt-Jakob disease (CJD) in humans, are characterized by the presence of an abnormal prion protein. Collinge et al. have now investigated the biochemical properties of these proteins. They find that 'new variant' CJD (a form recently recognized in 12 British individuals and thought to stem from consumption of BSE-infected beef) has properties that are sharply distinct from those of other human strains, but which resemble those of BSE transmitted to mice, cats and monkeys from infected cows. The finding greatly strengthens the conclusion that new variant CJD is a consequence of the BSE epidemic in British cattle, and suggests that prion proteins themselves may encode the disease phenotype.

Article analysis by Dr. Stephen Dealler

Commentary kindly by email 10.24.96 by S. Dealler

"What you have there are the abstracts that came with the articles. The actual thing was really quite complex.

An attempt was made to electrophorese purified PrPsc derived from CJD of different types. The electrophoretic pattern that was seen was well divided into 4 kinds. The first was the sporadic CJD that appears from nowhere and no cause is known, the second is the centrally inoculated CJD that comes from another person, the third was the peripherally inoculated CJD that they find coming in thepatients that received contamiated growth hormone. The fourth type, which could be easily separated from the others was that of the new variant CJD. When this was compared with the electrophoretic patterns found when BSE in cats, macaque monkeys, and mice is tested, it was found to be the same. i.e. the CJD type 4 was the same as the other BSEs. This evidence is very strong that BSE gve rise to new variant CJD.

One of the most important findings was that it is the sugar molecules (glycan chains) that are on the prion protein at positions 181 and 197 that decide the strain of disease as it is those molecules that caused the variance in the electrophoretic patterns.

The importance of these findings is not simply because it shows BSE to have probably infected humans, but also means that we may actually be able to treat the disease (see Dealler in Medical Hypotheses 1994;42:69-72) and that we may actually be able to diagnose the disease in asymptomatic people by looking at other tissues, perhaps their blood. This might seem a long way off but it may well not be so. The article by Aguzzi and Weissman was not really perfect in that it made a few errors. Anyway, it attempted to explain the articles. See my latest scientific research page. At the moment there is a furore in the press here about this. One set says that we will all die and the other says that there may be a way out."

Regards,
Steve Dealler

A suspicious signature: commentary by Adriano Aguzzi and Charles Weissmann

                                          Adriano Aguzzi and Charles Weissmann 
                                          A suspicious signature  Nature 383, 666-667 (1994)
All the evidence so far linking the new variant of Creutzfeldt-Jakob (CJD) disease with bovine spongiform encephalopathy (BSE) has been merely circumstantial. Now, laboratory studies comparing the various strains of prion responsible for the diseases have established a physical link between this new disease and BSE. 
Adriano Aguzzi and Charles Weissmann
Prof. Dr. med. Adriano Aguzzi (MRCPath)
Professor of Neuropathology and Head, 
National Reference Center for Human Prion Diseases
Institute of Neuropathology
University Hospital of Z¸rich
Schmelzbergstrasse 12
CH-8091 Z¸rich, Switzerland
Tel. ++41-1-255 2107 or 255 2869
Fax: ++41-1-255 4402
pager: ++41-1-255 1111, ask for 124 482
The nature of the infectious agent, the prion, causing transmissible spongiform encephalopathies (TSEs) is the subject of one of the most passionate controversies of contemporary biology. In addition to its academic aspects, this question now impinges on a grave issue of public health: the possible linkage between bovine spongiform encephalopathy (BSE) and vCJD, a new variant of Creutzfeldt-Jakob disease (CJD) which has primarily affected young individuals in Great Britain1 but also one in France2 (Table 1).

This linkage is plausible since a variety of animals can acquire BSE through oral administration of BSE-infected cattle brain. A paper by Collinge and coworkers3 provides an exciting new approach for tracing the passage of individual prion strains within and between species. TSEs are transmitted by an agent which apparently lacks informational nucleic acids4. This, along with an impressive body of genetic evidence5-7, led to the hypothesis that the prion consists solely of the modified form of a protein, named PrPC, which is encoded by a singular host gene8-10. The conjectured infectious form of PrP has been designated as PrP*11.

An abnormal form of PrPC, called PrPres or PrPSc and characterized by its partial resistance to proteolytic digestion and its tendency to aggregate, accumulates during the advanced stages of most cases of TSEs and has therefore been equated with PrP*. PrPC and PrPres appear to be chemically identical and may therefore represent conformational isomers. PrPres is thought to propagate by interacting with PrPC and causing its conversion into PrPres. As predicted by the "protein only" hypothesis, PrPC is essential for propagation of infectivity12, for clinical disease12 and brain pathology13.

The "protein only" hypothesis accounts readily for all findings but one, namely the occurrence of distinct prion strains which can be propagated indefinitely in hosts homozygous in regard to their PrP genes. Prion strains are defined by the incubation time or by the pattern of brain lesions14. Stable prion strains would be easily explained if the infectious agent were virus-like: strains would then be encoded by the viral genome. If, however, the prion is a conformational isoform of PrPC, then all its phenotypic characteristics should be solely determined by the host PrP genes and any prion strain should, after a few passages, become monomorphic.

How can the existence of strains be explained by the "protein only" hypothesis? Three possibilities have been suggested. (1) PrP* or PrPres is associated with an accessory molecule, such as a small nucleic acid15 which modulates the prion's phenotype without being essential for infectivity. No evidence for a nucleic acid of this kind has been adduced. (2) PrP* undergoes secondary modifications, such as glycosylation, and particular glycans specifically target the molecule to cells which in turn glycosylate PrP in an identical fashion (the "target cell hypothesis", Meyer, personal commun.) (3) PrP* or PrPres exists in various different conformations (at least one for each prion strain), and each type of prion is capable of imparting its own conformation to the PrPC molecule with which it interacts. More than one protein chemist has declared this idea to be unworldly - and yet this is precisely what is implied by a growing number of studies.

Western analysis for PrP from extracts of normal or prion-infected brains reveals three major bands, corresponding to PrP carrying two, one or no N-attached glycans. Protease treatment degrades all PrP in the case of normal brain and leads to increased mobility of the three species in prion-infected brain, due to amino terminal truncation. Presumably the changed conformation of PrPres protects a large part of the molecule against degradation. It was first shown by Bessen and Marsh16 that two strains of prions derived from mink, when propagated in inbred Syrian hamsters, give rise to PrPres molecules with distinct and heritable migration patterns after proteinase K treatment. This mobility difference was shown to be due to cleavage at distinct sites.

The most common form of prion disease in man, Creutzfeldt-Jakob disease (CJD), occurs in four forms: a familial form, linked to mutations in the PrP gene, an acquired or iatrogenic form, due to inadvertent parenteral exposure to CJD-contaminated materials, notably human cadaveric growth hormone, variant CJD, which is attributed to ingestion of BSE-contaminated bovine products and sporadic CJD, the most common form, which is of unknown etiology. Recently, Parchi et al. demonstrated that protease-treated PrPres in sporadic CJD can exhibit two mobility patterns which in conjunction with a Met/Val polymorphism at codon 129 specify four distinct clinical entities17.

Collinge and colleagues have now extended these observations to iatrogenic and vCJD. In addition to the mobility of protease-treated PrPres, they investigate the relative intensity of the three bands that reflect the glycosylation state. On this basis they define 4 "types" of pattern. Protease-digested PrPres from sporadic CJD cases gave rise to type 1 or type 2 patterns and that from iatrogenic cases predominantly to type 3 pattern. Astonishingly, all vCJD samples were uniformly of the type 4 pattern, which was not seen in any other form of CJD. In addition, brains of BSE cattle, as wells as of cats and Kudu who are believed to have acquired BSE, and macaques with experimental BSE showed a type 4 pattern, which is mainly characterized by a high proportion of diglycosylated PrPres, as compared to a low value in the other patterns.

All forms of human spongiform encephalopathies have been transmitted to wild type mice or mice transgenic for the human PrP genes; as summarized in Fig. 1, sporadic or iatrogenic CJDs with type 1, 2 or 3 patterns give rise to type 2 or type 3 patterns in mice carrying human transgenes, while BSE-cattle brain gave rise to a type 4 pattern in wild type mice, again with characteristically high content of diglycosylated species. Thus the conjectured conformational specificity is conserved after transfer between species. The patterns given by vCJD after transmission to the mouse have not yet been reported but are awaited with interest.

How do the different patterns come about? As set forth above, the different mobilities of protease-treated PrPres are ascribed to differing conformations and this is also true for the bands corresponding to non-glycosylated PrP, so that the variability cannot be ascribed to variable glycans. Closer inspection of the Western blots suggests that even within a single pattern type there is variability in the mobility, so that there may be even more different conformations represented among the samples. The unique feature of the type 4 pattern, namely the high ratio of diglycosylated to unglycosylated PrP may reflect increased susceptibility of the diglycosylated form to BSE-mediated conformational change. Alternatively, cells producing preferentially the diglycosylated form may be more readily targeted by the BSE agent.

The findings of Collinge et al. support the proposal that strain specificity is linked to distinct, heritable conformational states, and may be of great significance for the classification of human and animal prion diseases. Moreover, they provide further evidence that the BSE agent has been transmitted to man. This conclusion will be further reinforced if vCJD and BSE impart a similar signature to PrPres when transmitted to mice. One obvious extension of this work will be to compare the PrPres pattern of natural sheep scrapie to that of sheep with experimental BSE. A pattern difference, if present, may be used to determine if and to what extent BSE has been transmitted to sheep by contaminated feed, a question of particular importance for public health. Also interesting will be the search for type 4 PrPres in elderly patients with encephalopathy, since the aged have thus far, and for no obvious reasons, been spared by vCJD.

Comparison of sporadic CJD with new variant of CJD (vCJD)

Sporadic CJD vCJD
typical age of onset 55-70 yr.19-39 (median 28) yr.
Presenting features dementia, myoclonusbehavioral changes, ataxia, dysaesthesias
clinical courserapidly progressiveinsidious onset, prolonged course
codon 129predominantly homozygousMet/Met 100% so far
depositssynaptic deposits, rarely plaquesprominent florid plaques
banding patterntype 1, type 2* type 4 (~ to BSE in mice, macaques, others)
*: type 3 iniatrogenic cases with intramuscular inoculation.

Figure 1: On Western blots, different prion strains exhibits characteristic patterns of PrPres glycoforms, which are inherited even when passaging the infection across different species. Green arrows indicate successful transmission. The major glycoform of PrP is drawn in blue; arrows on western blots indicate increased mobility as compared to type 1. Provokingly, the vCJD pattern is similar to that seen in cows and other species with BSE. * J. Collinge, pers. comm. 

 References
1.      Will, R.G., et al. Lancet 347, 921-925 (1996).
2.      Chazot, G., et al. Lancet 347, 1181 (1996).
3.      Collinge, J., Sidle, K.C.L., Ironside, J., Meads, A. & Hill, A. Nature 383, in press (1996).
4.      Riesner, D., et al. Dev Biol Stand 80, 173-81 (1993).
5.      Hsiao, K., et al. Nature 338, 342-5 (1989).
6.      Medori, R., et al. Neurology 42, 669-70 (1992).
7.      Palmer, M.S., Dryden, A.J., Hughes, J.T. & Collinge, J. Nature 352, 340-342 (1991).
8.      Prusiner, S.B. Science 216, 136-44 (1982).
9.      Oesch, B., et al. Cell 40, 735-46 (1985).
10.     Basler, K., et al. Cell 46, 417-28 (1986).
11.     Weissmann, C. Nature 349, 569-71 (1991).
12.     B¸eler, H., et al. Cell 73, 1339-47 (1993).
13.     Brandner, S., et al. Nature 379, 339-43 (1996).
14.     Bruce, M.E., McBride, P.A., Jeffrey, M. & Scott, J.R. Mol Neurobiol 8, 105-12 (1994).
15.     Weissmann, C. Nature 352, 679-83 (1991).
16.     Bessen, R.A. & Marsh, R.F. J Virol 68, 7859-68 (1994).
17.     Parchi, P., et al. Ann Neurol 38, 21-9 (1995).

Molecular analysis of prion strain variation and the aetiology of 'new variant' CJD

John Collinge, Katie C. L. Sidle, Julie Meads, James Ironside, Andrew F. Hill 
24 Oct 1996  Nature 382, 685-690 (1996) Received 5 September; accepted 10 October 1996.
Strains of transmissible spongiform encephalopathies are distinguished by differing physicochemical properties of PrPSc, the disease-related isoform of prion protein, which can be maintained on transmission to transgenic mice. 'New variant' Creutzfeldt-Jakob disease (CJD) has strain characteristics distinct from other types of CJD and which resemble those of BSE transmitted to mice, domestic cat and macaque, consistent with BSE being the source of this new disease. Strain characteristics revealed here suggest that the prion protein may itself encode disease phenotype.

THE prion diseases or transmissible spongiform encephalopathies are a group of neurodegenerative diseases that affect both humans and animals. They can be transmitted between mammals by inoculation with, or in some cases dietary exposure to, infected tissues. They are all associated with the accumulation in affected brains of an abnormal isoform of a host-encoded glycoprotein, prion protein (PrP), which seems to be the central (and possibly the sole) component of the transmissible agent or prion1. This disease-related isoform, PrPSc, can be distinguished from the normal cellular isoform, PrPC, by its insolubility and partial resistance to proteases. PrPSc is derived from PrPC by a post-translational mechanism2 which seems to involve a conformational, rather than covalent, modification3. Transgenic, human molecular genetic and in vitro conversion studies support a model for prion propagation which involves a direct protein-protein interaction between host PrPC and inoculated PrPSc, with PrPSc acting to promote further conversion of PrPC to PrPSc in an autocatalytic process which proceeds most efficiently when the interacting proteins are of identical primary structure4-7.

In addition to the unique biology of these diseases, interest in them has been intensified because of the epidemic of the prion disease bovine spongiform encephalopathy (BSE)8 in the UK, and now in other countries, and the possibility that this may represent a significant threat to public health through ingestion of BSE-infected tissues. BSE is known to have caused prion disease in a number of other species, including domestic cats (feline spongiform encephalopathy) and captive exotic ungulates (nyala and kudu), presumably as a result of ingestion of BSE-contaminated feed9. The pathogenicity of bovine prions for humans is unknown, although the results of exposing transgenic mice expressing human prion protein, which lack a species barrier to human prions, indicate that induction of human prion production by bovine prions is inefficient10.

Although many converging lines of evidence support the 'protein only' hypothesis for prion propagation1, the existence of several distinct isolates or 'strains' of agent that can be stably passaged in inbred mice of the same prion protein genotype has yet to be explained satisfactorily within this model. Strains can be distinguished by their different incubation periods and patterns of neuropathology when passaged in mice11. A number of distinct strains of natural sheep scrapie are recognized, although BSE seems to be caused by a single strain of agent9.

Support for the idea that strain specificity is encoded by PrP alone is provided by study of two distinct strains of transmissible mink encephalopathy prions, designated hyper (HY) and drowsy (DY), which can be serially propagated in hamsters12. These strains can be distinguished by differing physicochemical properties of the accumulated PrPSc in the brains of affected hamsters13. After limited proteolysis, strain-specific migration patterns of PrPSc on polyacrylamide gels can be seen. DY PrPSc seems to be more protease-sensitive than HY PrPSc, producing a different banding pattern of PrPSc on western blots after proteinase K treatment. This relates to different amino- terminal ends of HY and DY PrPSc after protease treatment and implies differing conformations of HY and DY PrPSc (ref. 14). Furthermore, the demonstration that these strain-specific physicochemical 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 conformers15.

The human prion diseases occur in inherited, acquired and sporadic forms. About 15% are inherited, and are associated with coding mutations in the prion-protein gene (PRNP)16. Acquired prion diseases include kuru and iatrogenic Creutzfeldt- Jakob disease (CJD). Recognized iatrogenic routes of transmission are treatment with human cadaveric pituitary-derived growth hormone or gonadotrophin, dura mater or corneal grafting, and the use of inadequately sterilized neurosurgical instruments17. However, the majority of human prion disease occurs as sporadic CJD, in which pathogenic PRNP mutations and a history of iatrogenic exposure are absent. The majority of sporadic CJD cases are homozygous at polymorphic residue 129, a common protein polymorphism in human PrP that is known to be important in genetic susceptibility to human prion diseases6-16,18-20. Recently, the occurrence of a new form of human prion disease was reported in the United Kingdom, affecting unusually young people and having a highly consistent and unique clinicopathological pattern21. To date, all patients studied are homozygotes (for methionine) at polymorphic residue 129 of PrP and no coding mutations are present22. None has a history of iatrogenic exposure to human prions. This may indicate the arrival of a new risk factor for CJD in the United Kingdom, with dietary exposure to specified bovine offals, before their statutory exclusion from the human diet in 1989, seeming to be the most likely candidate. It is not known how many strains of human prions cause CJD; only two distinct patterns of protease-resistant PrP have been reported to date, associated with different clinicopathological types of sporadic CJD23.

We have investigated a wide range of cases of human prion disease to identify patterns of protease-resistant PrP that might indicate different, naturally occurring, prion strain types. We then studied 'new variant' CJD to determine whether it represents a distinct strain type that can be differentiated by molecular criteria from other forms of CJD. We demonstrate that sporadic and iatrogenic CJD is associated with three distinct patterns of protease-resistant PrP on western blots. Types 1 and 2, as previously described, are seen in sporadic CJD, and also in some iatrogenic CJD cases. A third type is seen in acquired prion diseases that arise from a peripheral route of exposure to prions. New variant CJD is associated with a unique and highly consistent appearance of protease- resistant PrP on western blots, involving a characteristic pattern of glycosylation. Whereas transmission of CJD to inbred mice produces a pattern characteristic of the inoculated CJD, transmission of BSE produces a glycoform ratio pattern closely similar to new variant CJD. Similarly, experimental BSE in macaques and naturally acquired BSE in domestic cats show an indistinguishable glycoform pattern to experimental murine BSE and new variant CJD. Transmission of types 1, 2 and 3 CJD to transgenic mice expressing human PrP reveals persistence or conversion of strain type dependent on PRNP codon-129 genotype, providing supportive evidence for the 'protein only' hypothesis of infectiousness and indicating that strain variation might be encoded by a combination of PrP conformation and glycosylation.

Sporadic and acquired CJD

Typically, three bands are seen on western blots of protease-resistant PrP from CJD cases. The two larger-molecular-mass bands represent the main glycosylated forms of PrP and the smaller band represents unglycosylated PrP23 (Fig. 1). Two distinct patterns have been described in sporadic CJD: a type-1 pattern seen in the majority of CJD cases, with homozygosity for methionine (MM) at polymorphic residue 129 of PrP, and a type-2 pattern seen in a minority of MM cases and in all methionine/valine heterozygotes (MV) and valine homozygotes (VV)23. We used western blotting to analyse protease-resistant prion proteins for a total of 26 neuropathologically confirmed sporadic CJD cases representing all three PRNP codon- 129 genotypes. Cases that pre-dated and cases that were contemporary to the BSE epidemic were included (Table 1). We confirmed the finding that in sporadic CJD, MM cases had two distinct banding patterns23 (designated types 1 and 2), whereas VV and MV cases all showed the type-2 banding pattern. In our sample of sporadic CJD cases, we found type-2 MM to be more frequent that type-1 MM (Table 1). In addition, differences were previously noted between the proportion of protease-resistant PrP in each of the three bands between the type-1 and type-2 sporadic CJD patients23. In our larger series, a statistically significant difference was seen between type-1 and type-2 cases with respect to the proportion of the low-molecular-mass glycoform, but not with respect to the other two bands .

TABLE 1 PRNP genotypes and banding type of protease-resistant PrP in sporadic, iatrogenic and new variant CJD
Classification Codon-129 genotype Type 1 Type 2 Type 3 Type 4 Total
Sporadic MM 5 13 0 0 18
Sporadic MV 0 4 0 0 4
Sporadic VV 0 4 0 0 4
Iatrogenic (growth hormone) MM 1 0 0 0 1
Iatrogenic (growth hormone) MV 0 0 1 0 1
Iatrogenic (growth hormone) VV 0 0 3 0 3
Iatrogenic (gonadotrophin) VV 0 0 1 0 1
Iatrogenic (dura mater) MM 0 1 0 0 1
New variant MM 0 0 0 10 10
MM, methionine-homozygous genotype at PRNP codon 129; MV and VV, methionine/valine heterozygotes and valine homozygotes, respectively.
We then studied a range of iatrogenic CJD cases, including patients who had received human pituitary-derived growth hormone, with all three PRNP codon-129 genotypes, a human pituitary-derived gonadotrophin patient, and a patient who developed CJD after a dura mater graft. A third distinct banding pattern of protease-resistant PrP was detected in all of the pituitary hormone- related cases which were of PRNP codon-129 genotype MV or VV (Fig. 2). In this type-3 pattern, all three bands are shifted, consistent with a decrease in relative molecular mass (Mr) of roughly 2,000 to 3,000 (2K-3K) of the protease-resistant PrP detected as compared with type-2 sporadic CJD. There were no significant differences in glycoform ratios between type-1 and type-3 or between type-2 and type-3 CJD in this sample (Fig. 4 legend). The single MM growth-hormone case had a banding pattern indistinguishable on western blot analysis from type-1 sporadic CJD cases (Fig. 2). The dura mater-related case had a banding pattern indistinguishable from type-2 sporadic CJD (Fig.2). The possibility that there may be further heterogeneity within these individual PrPSc types is under investigation.

Transmission studies to mice

We are studying the transmission characteristics of CJD in transgenic mice that express human PrP but not murine PrP (designated HuPrP+/+ Prn-p0/0 as they are homozygous for the human PrP transgene array)10. These mice lack a species barrier to human prions, and most inoculated mice succumb to prion disease with consistently short incubation periods10-24, usually in the region of 180-220 days (data not shown). This is markedly different from studies with non-transgenic mice using the same inocula, in which transmissions are infrequent and when they do occur they are associated with prolonged and variable incubation periods.

Transmission of type-2 CJD (of all three codon-129 genotypes) or type-3 CJD (which were all of genotype MV or VV) resulted in production of protease-resistant human PrP in the mice with an identical banding pattern to the primary inoculum (type 2 or 3, respectively) (Fig. 3). The HuPrP+/+ Prn-p0/0 mice used encode valine at residue 129 (ref. 25). However, transmission of type-1 CJD (which are all of genotype MM) resulted consistently in a type-2 banding pattern of human protease-resistant PrP produced in the mice (Fig. 3). The proportion of different glycoforms on western blots of brain homogenates from these mice was indistinguishable from that of the human cases themselves (data not shown).

New variant CJD

Proteinase K treatment of PrPSc from new variant CJD revealed the characteristic band shift seen after digestion to the protease-resistant fragment, confirming complete digestion of the protease-sensitive N-terminal region of PrPSc in the conditions used . All patients studied with new variant CJD were homozygotes for methionine at polymorphic residue 129 (ref. 22). No known or new coding mutations of PRNP were seen in the new variant CJD cases or the sporadic CJD cases that were sequenced. Western blot analysis of these ten new variant CJD cases revealed a consistent and distinct pattern of protease-resistant PrP forms, which could be clearly differentiated by band sizes from type-1 and type- 2 sporadic CJD cases, and from type-3 CJD by a striking and distinctive pattern of band intensities which differed from all three CJD types. Deglycosylation with PNGaseF resulted in a single band, indicating that there might be a consistent proteolytic cleavage site irrespective of glycosylation state (Fig. 1e) and which differs from that seen in sporadic CJD. The high- Mr glycoform was the most abundant, with relatively little unglycosylated PrP when compared with type-1, type-2 and type-3 CJD. These differences in band intensity from types 1- 3 CJD were all highly statistically significant (Fig. 4 legend). A scattergram of the relative proportions of the high-Mr and low-Mr glycoforms reveals two non-overlapping populations of cases: new variant CJD has a distinctive pattern that differs markedly from allpreviously recognized types of sporadic and iatrogenic CJD.

Comparison with BSE

As the band intensities of protease-resistant PrP in new variant CJD differed markedly from sporadic and iatrogenic CJD, we investigated whether this distinctive glycoform pattern was also seen in naturally or experimentally transmitted BSE. First, we compared transmissions of BSE and CJD to the same inbred mice. Comparative transmission data in transgenic mice was not available as BSE has not transmitted, to date, to HuPrP+/+ Prn-p0/0 mice (500 days post-inoculation). The glycoform ratios seen in CJD transmissions to wild-type FVB mice were indistinguishable from those of the three types of CJD (Fig. 5b). However, BSE transmission into both wild-type FVB and C57BL mice resulted in ratios that were closely similar to those of new variant CJD. Similarly, band sizes of protease-resistant PrP seen on transmission of BSE to wild-type mice were shifted to a lower Mr as compared to type 2 CJD transmission (data not shown). We then studied naturally transmitted BSE in the domestic cat (feline spongiform encephalopathy26) and experimental BSE in a macaque27. These cases also closely resembled new variant CJD and experimental murine BSE. BSE itself was not detectable on western blots using the 3F4 monoclonal or R073 polyclonal antibody. However, a PrPSc signal was detected from homogenates of brainstem from naturally infected BSE using a rabbit antibody to a synthetic human PrP peptide (95-108) (provided by B. Ghetti) and the pattern of the glycoforms (51.2% high Mr, 33.9% low Mr, 14.9% unglycosylated) was closely similar to transmitted BSE and new variant CJD. This antibody also detected PrPSc from domestic cat, macaque and humans, producing similar results to R073 and 3F4 antisera (data not shown).

Discussion

Sporadic and iatrogenic CJD seems to be related to the production of three distinct types of human PrPSc which can be differentiated on western blots after proteolytic cleavage by their differing band sizes. Types 1 and 2 are associated with different clinicopathological phenotypes of sporadic CJD23, and type 3 is seen in cases of iatrogenic CJD where exposure to prions has been via a peripheral route (intramuscular injection of human cadaveric pituitary-derived hormones) rather than by a direct central nervous system (CNS) route (dura mater grafting). It is well recognized that such peripherally acquired cases have a distinct phenotype, presenting with cerebellar ataxia and psychiatric disturbance rather than as a dementing illness28. Iatrogenic CJD resulting from CNS exposure typically resembles classical sporadic CJD28. New variant CJD, while it has PrPSc band sizes similar to type-3 CJD, can be clearly distinguished from all three types of CJD by a characteristic pattern of band intensities. This distinctive molecular marker, which clearly differentiates new variant CJD from sporadic CJD, serves to support the proposal, on the basis of comparative clinicopathological studies and epidemiological surveillance21, that new variant CJD is a distinct and new subtype of prion disease, related to a previously unrecognized prion strain. The spontaneous occurrence of a novel human PrPSc type, that is the same in twelve individuals in the United Kingdom, over the last two years, seems extraordinarily unlikely as an explanation for 'new variant' CJD. The alternative conclusion is that these cases have arisen from a common source of exposure to a new prion strain, and the lack of any history of common iatrogenic exposure indicates that this is probably a new animal strain. That the glycoform 'signature' of new variant CJD is seen in BSE itself, in experimental murine BSE (whereas CJD transmission to these types of mice produces the CJD 'signature') and in naturally transmitted BSE in domestic cat and experimental BSE in macaque, is consistent with the hypothesis that new variant CJD results from BSE transmission to humans. Transmission studies of new variant CJD in HuPrP+/+ Prn-p0/0 mice are in progress; it will be of interest to compare incubation periods and patterns of neuropathology with other CJD transmissions (which are also currently under investigation). PrPSc typing will be of immediate application in wider epidemiological studies of CJD. It is possible that BSE could produce other clinicopathogical phenotypes in humans, particularly in different age groups and different ethnic populations, that are not recognized as new variant CJD. Furthermore, this method may also allow typing of various animals to see if BSE has also transmitted naturally to these species. There is particular concern that BSE may have transmitted to, and be surviving in, the sheep population29. Typing of known scrapie strains which pre-dated the BSE epidemic, and recent isolates, will be important in this regard.

This molecular marker can already be used in differential diagnosis of new variant CJD. New variant CJD is atypical both in its clinical features and electroencephalogram, such that diagnosis is dependent on neuropathology, either at autopsy or in some cases on brain biopsy. However, although the brain biopsy may demonstrate spongiform encephalopathy and PrP immunoreactivity adequate for a diagnosis of CJD, the characteristic neuropathological features necessary for a diagnosis of new variant CJD may not be present in the biopsy sample21. As PrP is expressed in the lymphoreticular system, and prion replication occurs in spleen and in other lymphoreticular tissues30, it may be possible to detect this molecular marker of new variant CJD in tonsil31 or lymph-node biopsy and thereby avoid brain biopsy.

The aetiology of sporadic CJD remains unclear but may involve somatic PRNP mutations or spontaneous conversion of PrPC to PrPSc as a rare stochastic event. Sporadic CJD is associated with type-1 or type-2 PrPSc. Type 1 is always associated with genotype MM, type 2 with all genotypes (MM, MV or VV). Type 3 is seen in iatrogenic CJD of genotype MV or VV. Only human PrP M129 seems to form type-1 PrPSc, whereas either human PrP M129 or human PrP V129 can produce type-2 PrPSc. As type-3 PrPSc is only seen in MV or VV individuals, it is possible that only human PrP V129 can form this type. As type-3 PrPSc is seen in peripherally acquired iatrogenic CJD cases, it is possible that this strain is selected in or preferentially produced by the lymphoreticular system, where prion replication first occurs in the experimentally transmitted disease in mice32. If such a PrPSc type were formed preferentially from human PrP V129, this could explain the excess of the PRNP codon-129 VV genotype among pituitary-hormone related CJD cases19,20,33. Further studies will be required to determine whether the type-3 pattern is a consistent marker of peripheral, as opposed to central, prion exposure or sporadic CJD. It is of note, however, that similar band sizes are seen in the type 4 (new variant CJD) pattern (which are of PRNP genotype MM), which seems to have arisen by peripheral (presumably dietary) exposure to bovine prions. The formation of particular PrPSc types seems to be constrained by host PRNP codon-129 genotype. This finding is supported by the observation that type-1 PrPSc converts to type 2 on passage in transgenic mice expressing the human PrP gene encoding valine at residue 129, whereas types 2 and 3 remain unchanged on such passage.

The different types of human PrPSc are seen in association with distinct clinicopathological phenotypes of CJD, and can be maintained on passage in mice, which indicates that these represent distinct human prion strains. The finding that strains appear to involve different post-translational modifications of PrP, which persist or (when PrP genotypes are mismatched) can be predictably converted between discrete strains on passage in mice, is consistent with a protein-only model of prion propagation in which strains are encoded by post-translational modification of PrP itself without the need for a nucleic acid or other cofactor. The bands seen on western analysis of PrP following proteolytic cleavage represent diglycosylated, monoglycosylated (at either of the two N- linked glycosylation sites) and unglycosylated PrP, and two separate features of these bands, shifts in mobility and differences in relative intensities, seem to be associated with strain type. The mobility shifts after cleavage, seen in all three bands, imply different PrP conformations (which may include differing states of assembly). The differences in glycoform ratios could indicate preferential conversion of particular glycoforms into particular conformational states. It has been argued that differing PrP glycosylation in different brain regions could provide a mechanism for the targeting of neuropathology seen with different strain types, as prions may replicate most efficiently in cell populations expressing a similarly glycosylated PrP on the cell surface34. Both PrP conformation and glycosylation may contribute to strain type but further studies will be required to investigate whether these two post-translational modifications of PrP can contribute to strain type independently, or are coupled phenomena.

Selection and molecular genetic analysis of patients.

Twenty-six neuropathologically confirmed sporadic CJD cases including all three codon-129 genotypes, seven neuropathologically confirmed iatrogenic CJD cases and ten neuropathogically confirmed new variant CJD cases referred to the National CJD Surveillance Unit or the Prion Disease Group, from which frozen brain tissue was available, were studied. Brain samples used were from the cerebral cortex, usually frontal cortex. DNA was extracted from blood or brain tissue and analysed for the presence of known or new coding mutations in the prion protein gene (PRNP) and to determine codon-129 genotype. In 8 of 10 new variant CJD and the majority of sporadic and iatrogenic CJD patients, the complete PRNP open reading frame was amplified by polymerase chain reaction (PCR) using oligonucleotide primers chosen so as not to overlay an intron polymorphism 5' to the open reading frame, which can lead to non-amplification of certain alleles, as described previously35. The PCR product was size-fractionated in agarose gels to exclude insertional or deletional mutations, and were then sequenced on both DNA strands using an ABI 373 or 377 automated DNA sequencer (Applied Biosystems).

Western blot analysis.

Between 10 and 20 mg of brain tissue was homogenized in lysis buffer (0.5% NP-40, 0.5% sodium deoxycholate in phosphate buffered saline) by serial passage through needles of decreasing diameter. The homogenate was cleared by centrifugation at 2,000 r.p.m. for 5 min. Proteinase K (BDH) was added to a final concentration of 50 gml- 1 and the samples incubated at 37 ºC for 1 h. The reaction was terminated by the addition of Pefabloc (Boehringer) to 1 mM. The samples were mixed with twice the volume of SDS loading buffer (125 mM Tris-HCl, 4% SDS, 20% glycerol, 0.02% bromophenol blue, pH 6.8) and boiled for 10 min. They were then centrifuged at 14,000 r.p.m. in a microfuge for 5 min before electrophoresis. Between 1 and 10 l of sample was electroporesed on 16% Tris-glycine gels (Novex)36. The gels were then electroblotted onto PVDF membrane (Millipore) using either a tank or semi-dry blotting system. The membranes were blocked in 5% BLOTTO (5% non-fat milk powder in PBS with 0.05% Tween 20) for 1 h at room temperature. No single antibody detects all species studied efficiently. After washing in PBST (PBS with 0.05% Tween 20), the membranes were incubated with anti-PrP monoclonal antibody 3F4 (ref. 38) (diluted 1:5,000 in PBST), rabbit polyclonal antibody R073 (ref. 39) (diluted 1:10,000 in PBST), or rabbit anti-human PrP peptide (95- 108) (diluted 1:20,000) (from B. Ghetti) for between 1 h and overnight. After washing, the membrane was incubated with a horseradish peroxidase-conjugated rabbit anti-mouse antibody (Sigma) or goat anti-rabbit antibody (Sigma) at a dilution of 1:10,000 in PBST for 1 h. The membranes were washed again in PBST and developed using a chemiluminescent substrate (ECL; Amersham) using Biomax MR film (Kodak).

Deglycosylation of prion proteins.

A portion (25 µl) of 10% brain homogenate, pre- treated with proteinase K, was denatured in 0.5% SDS, 1% -mercaptoethanol for 10 min at 100 ºC. NP-40 was added to 1% and the proteins were incubated in 500 units PNGaseF (New England Biolabs), using the proprietary buffer, at 37 ºC for 2 h. After digestion, proteins were precipitated with 4 volumes of methanol and resuspended in SDS loading buffer and western blotted.

Quantitation of PrP glycoform ratios.

Blots were scanned on a Hoefer scanning densitometer (model GS300) and the relative amounts of the different glycoforms were obtained by computerized integration of peaks representing each of the three distinct bands. Scanning was done on exposures within the linear range of the photographic film.

Transmission studies in transgenic and non-transgenic mice.

Strict biosafety methods were followed. Transgenic mice expressing human PrP were bred and maintained in an animal microbiological containment level I facility and moved to a containment level II facility where intracerebral inoculation was performed in a class I microbiological safety cabinet. Preparation of inocula and removal of tissues were done in a microbiological containment level III facility. All mice were examined twice weekly for the development of clinical signs of scrapie. At onset of signs, mice were examined daily. Mice were killed if exhibiting any signs of distress. Criteria for clinical diagnosis of scrapie in mice were as described previously40. Transgenic lines expressing human PrP but not murine PrP were established by breeding mice transgenic for human PrP (designated Tg152, produced as described previously) with mice homozygous for PrP null alleles41. All mice were genotyped to confirm the presence of the HuPrP transgene or PrP null alleles by PCR with genomic DNA obtained by tail biopsy, as described42. Tg152 mice have expression levels of human PrP of 200% of that seen in normal human brain42. Mice were anaesthetised with halothane/O2 and intracerebrally inoculated into the right parietal lobe with 30 µl of a 1% brain homogenate in phosphate- buffered saline.

1. Prusiner, S. B. Science 252, 1515-1522 (1991).
2. Caughey, B. Raymond, G. J. J. Biol. Chem. 266, 18217-18223 (1991).
3. Pan, K.-M. et al. Proc. Natl Acad. Sci. USA 90, 10962- 10966 (1993).
4. Prusiner, S. B. et al. Cell 63, 673-686 (1990).
5. Weissmann, C. Nature 349, 569-571 (1991).
6. Palmer, M. S., Dryden, A. J., Hughes, J. T. Collinge, J. Nature 352, 340-342 (1991).
7. Kocisko, D. A. et al. Nature 370, 471-474 (1994).
8. Wells, G. A. H. Wilesmith, J. W. Brain Pathol. 5, 91-103 (1995).
9. Bruce, M. et al. Phil. Trans. R. Soc. Lond. B 343, 405-411 (1994).
10. Collinge, J. et al. Nature 378, 779-783 (1995).
11. Bruce, M. E., Fraser, H., McBride, P. A., Scott, J. R. Dickinson, A. G. in Prion Diseases in Human and Animals (eds Prusiner, S. B., Collinge, J., Powell, J. Anderton, B.) 497-508 (Ellis Horwood, London, 1992).
12. Marsh, R. F. Kimberlin, R. H. J. Infect. Dis. 131, 104-110 (1975).
13. Bessen, R. A. Marsh, R. F. J. Virol. 66, 2096-2101 (1992).
14. Bessen, R. A. Marsh, R. F. J. Virol. 68, 7859-7868 (1994).
15. Bessen, R. A. et al. Nature 375, 698-700 (1995).
16. Windl, O. et al. Hum. Genet. 98, 259-264 (1996).
17. Weller, R. O. Psychol. Med. 19, 1-4 (1989).
18. Baker, H. F. et al. Lancet 337, 1286 (1991).
19. Collinge, J., Palmer, M. S. Dryden, A. J. Lancet 337, 1441-1442 (1991).
20. Brown, P. et al. Neurology 44, 291-293 (1994).
21. Will, R. G. et al. Lancet 347, 921-925 (1996).
22. Collinge, J., Beck, J., Campbell, T., Estibeiro, K. Will, R. G. Lancet 348, 56 (1996).
23. Parchi, P. et al. Ann. Neurol. 39, 669-680 (1996).
24. Telling, G. C. et al. Cell 83, 79-90 (1995).
25. Telling, G. C. et al. Proc. Natl Acad. Sci. USA 91, 9936- 9940 (1994).
26. Wyatt, J. M. et al. Vet. Rec. 129, 233-236 (1991).
27. Lasmezas, C. I. et al. Nature 381, 743-744 (1996).
28. Brown, P., Preece, M. A. Will, R. G. Lancet 340, 24-27 (1992).
29. Page, G. Nature 382, 381 (1996).
30. Kimberlin, R. H. Walker, C. A. Virus Res. 12, 201-211 (1989).
31. Schreuder, B. E. C., van Keulen, L. J. M., Vromans, M. E. W., Langeveld, J. P. M. Smits, M. A. Nature 381, 563 (1996).
32. Kimberlin, R. H. Walker, C. A. J. Gen. Virol. 69, 2953-2960 (1988).
33. Laplanche, J.-L. et al. Neurology 44, 2347-2351 (1994).
34. Haraguchi, T. et al. Arch. Biochem. Biophys. 274, 1-13 (1989).
35. Palmer, M. S. Hum. Mutat. 7, 280-281 (1996).
36. Laemmli, U. K. Nature 227, 680-685 (1970).
37. Towbin, H., Staehelin, T. Gordon, J. Proc. Natl Acad. Sci. USA 76, 4350-4354 (1979).
38. Kascsak, R. J. et al. J. Virol. 61, 3688-3693 (1987).
39. Serban, D., Taraboulos, A., DeArmond, S. J. Prusiner, S. B. Neurology 40, 110-117 (1990).
40. Carlson, G. A. et al. Cell 46, 503-511 (1986).
41. Bueler, H. et al. Nature 356, 577-582 (1992).
42. Whittington, M. A. et al. Nature Genet. 9, 197-201 (1995).

ACKNOWLEDGEMENTS. We thank all clinicians and neuropathologists who have referred CJD cases to the National CJD Surveillance Unit or to the Prion Disease Group. We also thank M. Desbruslais, J.Beck, I. Gowland, T. OMahony, A. Clarke and J. Sealby for their assistance with these studies; D.Dormont and C. Lasmezas for generously supplying tissue from macaque BSE; G. Pearson for tissue from feline spongiform encephalopathy, and R. Bradley and M. Dawson for tissues from BSE-affected cattle; S. B. Prusiner for providing R073 antibody and Tg152 mice, R. Kascsak for 3F4 antibody and C. Weissmann and S. B. Prusiner for PrP null mice; and J. Wadsworth for statistical advice. This work was funded by the Wellcome Trust, Biotechnology and Biological Sciences Research Council (Biology of Spongiform Encephalopathies Programme) and the David and Frederick Barclay Foundation.

Nature © Macmillan Publishers Ltd. 1996
Registered No. 785998 England.