Human prion diseases: Molecular, cellular and population biology


  • Mark W. Head

    Corresponding author
    • National CJD Research & Surveillance Unit, Centre for Clinical Brain Sciences, School of Clinical Sciences, The University of Edinburgh, Edinburgh, UK
    Search for more papers by this author

Correspondence: Mark W. Head, PhD, National CJD Research & Surveillance Unit, Western General Hospital, Bryan Matthews Building, Edinburgh EH4 2XU, UK. Email:


The past 20 years have witnessed a dramatic resurgence of interest in a hitherto obscure neurodegenerative disease, Creutzfeldt-Jakob disease (CJD). This was driven partly by the novelty of the prion hypothesis, which sought to provide an explanation for the pathogenesis of transmissible spongiform encephalopathies, involving a unique epigenetic mechanism, and partly by events in the UK, where an outbreak of a new prion disease in cattle (bovine spongiform encephalopathy or BSE) potentially exposed a large section of the UK population to prion infectivity through a dietary route. The numbers of cases of the resultant novel disease variant CJD (vCJD), have so far been limited and peaked in the UK in the year 2000 and have subsequently declined. However, the effects of BSE and vCJD have been far-reaching. The estimated prevalence of vCJD infection in the UK is substantially higher than the numbers of clinical cases would suggest, posing a difficult dilemma for those involved in blood transfusion, tissue transplantation and cellular therapies. The clinico-pathological phenotype of human prion diseases has come under close scrutiny and molecular classification systems have been developed to account for the different diseases and their phenotypic spectra. Moreover, enhanced human and animal surveillance and better diagnostic tools have identified new human and animal prion diseases. Lastly, as the prion hypothesis has gained widespread acceptance, the concepts involved have been applied to other areas, including extra-chromosomal inheritance in fungi, long-term potentiation in memory formation and the spread of molecular pathology in diverse conditions, such as Alzheimer's disease, Parkinson's disease and amyotrophic lateral sclerosis. Studies at the molecular and cellular level have helped to provide a better understanding of human prion diseases, aided pathological diagnosis and helped inform public health decision-making.


Prion diseases

Prion diseases are a group of rare fatal neurodegenerative diseases. They affect humans, agricultural, captive and free-ranging animals. Unusually, they have genetic, apparently sporadic and acquired forms, and even the genetic and the sporadic forms are experimentally transmissible. The acquired forms themselves can have extremely lengthy incubation periods, up to 40 years in the case of kuru. Clinically, the diseases are characterized by dementia and ataxia and neuropathologically they are characterized by neuronal loss, gliosis and spongiform change in the brain. Historically, the prion diseases have been known collectively as the transmissible spongiform encephalopathies or TSE (Table 1). These diseases have for some time sat at the border of the infectious disease scientific research community and that of neurosciences and neurodegeneration, viewed by some as a somewhat arcane and hermetically sealed subject, with limited general relevance.

Table 1. The transmissible spongiform encephalopathies or prion diseases
 Acquired (oral and other potential routes of transmission)Genetic (associated with PRNP or prnp mutations)Idiopathic (sporadic in occurrence, possibly spontaneously arising)
Agricultural, domestic and free ranging animals

– Scrapie in sheep and goats

– Transmissible mink encephalopathy (TME)

– Chronic wasting disease in deer and elk (CWD)

– Classical bovine spongiform encephalopathy in cattle (C-BSE)

– Feline spongiform encephalopathy (FSE)

– Atypical scrapie

– H-type bovine spongiform encephalopathy in cattle (H-BSE)

– L-type bovine spongiform encephalopathy in cattle (L-BSE)


– Kuru

– Iatrogenic Creutzfeldt-Jakob disease (iCJD)

– Variant Creutzfeldt-Jakob disease (vCJD)

– Gerstmann-Straussler-Scheinker disease (GSS)

– Familial or genetic Creutzfeldt-Jakob disease (fCJD, gCJD)

– Fatal familial insomnia (FFI)

– PrP-cerebral amyloid angiopathy

– Sporadic Creutzfeldt-Jakob disease and its subtypes (sCJD), including sporadic fatal insomnia (sFI)

– Variably protease-sensitive prionopathy (PSPr or VPSPr)

Molecular basis: an epigenetic agent

There have been two paradigm shifts in our understanding of TSE in the past 30 years. The first being the formulation, promotion and subsequent general acceptance of the prion hypothesis as the best available explanation for TSE.[1, 2] The second (which is currently ongoing) is the extension of the prion paradigm into areas of normal cellular physiology, protein-based inheritance (especially in yeast) and the formulation of a general model for the mechanism involved in a wide variety of neurodegenerative diseases.[3-5]

The prion hypothesis posits an epigenetic agent, composed largely, if not exclusively, of an altered from of the normal host-encoded prion protein (PrPC), refolded and aggregated into the disease-associated form (termed PrPSc). This conversion process is proposed to be autocatalytic, PrPSc being synonymous with the infectious agent, and the production of PrPSc being the key causative event in neurodegeneration. Within this paradigm some of the more unusual features of the TSE become more comprehensible: sporadic forms of the disease resulting from rare (perhaps stochastic) conversion of PrPC to PrPSc, or the failure of quality control mechanisms for PrPSc suppression or degradation. Genetic forms (all known examples of which are associated with mutations of the prion protein gene, PRNP) resulting from an increased likelihood of conversion to the pathogenic form. Lastly, the acquired forms result from iatrogenic or oral exposure to PrPSc.

The question of agent strain

In addition to tissue-based studies of human prion diseases themselves, some of these diseases have been successfully transmitted to rodents (both wild-type mice and humanized PRNP transgenic mice) and to a variety of non-human primate species. This has allowed the clinicopathological heterogeneity of human prion diseases to be viewed through the prism of prion or TSE agent strain,[6-9] a concept originally developed to explain the elaborate series of meta-stable murine scrapie phenotypes that can be derived from field isolates of sheep scrapie. In order for the prion hypothesis to be correct, a biochemical correlate must be found for a strain within the structure of PrPSc. Animal transmission studies indicate different human prion strains may be enciphered in the secondary and higher order structure of PrPSc.[10] More recently cell-free PrP conversion assays have been developed that can be used to model this fundamental aspect of prion biology more rapidly and cheaply and avoiding the ethical concerns associated with animal experimentation.

The genetics of an epigenetic agent

Although the conversion from PrPC to PrPSc occurs at the epigenetic level, PrPC is a gene product of the host. Mutations in PRNP are closely associated with disease, but the human PRNP gene (and its animal orthologues) are polymorphic and these polymorphisms can have quite dramatic effects on prion disease susceptibility and on disease phenotype.[8, 11, 12] In human prion disease genetics the common methionine/valine (M/V) polymorphism at codon 129 of the PRNP gene exerts a particularly powerful effect (Table 2).

Table 2. The effect of the human PRNP codon 129 methionine (M)/valine (V) polymorphisms
Normal Caucasian population37%51%12%
Sporadic CJD and its phenotypic subtypesMM1/MV1 sporadic CJD (70%)VV1 sporadic CJD (1%)

MM2 (cortical) sporadic CJD (2%)

MM2 (thalamic variant or sporadic fatal insomnia) sporadic CJD (2%)

MV2 sporadic CJD (9%)VV2 sporadic CJD (16%)
Variably protease sensitive prionopathyLeast frequently affectedIntermediate frequencyMost frequently affected
Variant CJD (vCJD)

All definite clinical cases of primary vCJD

All known clinical cases of secondary (iatrogenic) vCJD

Single possible clinical case of vCJD

Asymptomatic secondary cases of peripheral infection (n = 2)

Assumed asymptomatic peripheral infection (n = 2)
Iatrogenic CJD (iCJD)MM most frequently affected by dura mater associated iCJD in the UK
VV and MV most frequently affected by growth hormone associated iCJD in the UK
D178N PRNP mutationCoupled to M associated with fatal familial insomnia
Coupled to V associated with familial CJD

Cell biology

The clinical symptoms of human prion diseases most probably derive from selective neuronal dysfunction and cell death, suggesting that neurons are the most significant site of PrP conversion and prion replication. Expression of PrP is a prerequisite for prion replication and pathology.[13] However, neurons are not the only cells of the nervous system implicated in prion disease pathophysiology. A variable degree of astrogliosis and microglial activation accompany neuronal loss. The role of microglia and astrocytes, whether protective or destructive in human prion disease pathogenesis is unresolved (as it is in many neurodegenerative disease), but astrocyte-targeted expression of PrP appears to be sufficient to generate neuronal pathology.[14] Moreover, in the orally acquired prion diseases, neuroinvasion involves the peripheral nervous system, the lymphoreticular system and perhaps cells within the blood. The role of follicular dendritic cells in the germinal centers of secondary lymphoid organs in trapping, concentrating and replicating prions in the periphery has been intensively studied, and it has offered a tool to diagnose and to investigate the epidemiology of one human prion disease in particular, vCJD.[15, 16]

Population biology

Sporadic CJD (sCJD) occurs world-wide with a uniform incidence of around one case in one million per annum. The epidemiology of genetic CJD (gCJD), Gerstmann-Straussler-Scheinker disease (GSS) and fatal familial insomnia (FFI) each reflect founder cases, migration and reproduction. The epidemiology of the acquired forms is arguably more interesting, tractable, and pertinent to their elimination. Kuru for example, is virtually extinct now, despite its very long incubation periods.[17] It had a circumscribed geographical and temporal epidemiology, restricted to ethnic groups in a prescribed region of Papua New Guinea beginning early in the 20th century, presumably originating from a case of sCJD.[17, 18] Cases of iatrogenic CJD (iCJD), as transmitted by dura mater grafting and human pituitary-derived growth hormone are similarly in sharp decline, exposures by these routes having ceased. iCJD in dura mater and growth hormone recipients can probably be viewed as problems that occurred in, and were resolved during, the 20th century.[19]

It might appear that vCJD similarly belongs to the past. The epidemic of bovine spongiform encephalopathy (BSE) in cattle that occurred in the UK peaked in 1986 and the peak of resultant zoonosis (vCJD) occurred in 2000, with 28 patients dying of the disease, and five or fewer patients dying of the diseases per annum in 2005 onwards. There have been no cases of vCJD reported in 2012 in the UK at the time of writing (late 2012).[20] Cases of BSE in cattle have occurred outside the UK, but on a very limited scale by comparison to the UK. The total number of vCJD cases in the UK is 176. The total number of cases in France is 27 and the other 10 affected countries have had five cases or fewer in total.[21]

Why Then Should We Still Be Interested in Human Prion Diseases?


It is important to note that the scale of exposure to BSE in the UK is probably of a different order of magnitude than any previous exposure of a human population to prion infectivity. It is estimated that greater than 400 000 infected cattle entered the human food chain in the UK during the BSE epidemic. A number of post-hoc explanations for the apparent discrepancy in likely exposure and resultant cases have been advanced, including a substantial species barrier between cows and humans, effects of dose, genetic susceptibility related to variations in both PRNP and non-PRNP genes, age-related susceptibility, and the possible necessity for co-factors, such as inflammation. A role for the codon 129 polymorphisms is plausible, but methionine homozygotes constitute 37% of the normal population, so this can only be part of the answer. All definite clinical cases of vCJD that have been tested are MM at codon 129 of the prion protein gene, although a single case of possible vCJD has been reported in a PRNP codon 129 heterozygous patient.[22] However, a retrospective prevalence study carried out in the UK, based on the immunohistochemical detection of abnormal prion protein in appendix and tonsil tissue, indicated a prevalence of infection much higher than the numbers of clinical cases would suggest.[16] The official estimate based on on-going prevalence studies has been revised upwards from 1:4000 to 1:2000 at the time of writing.[23] When positive appendices in these studies have been tested their codon 129 genotype has not been found to be restricted to the MM genotype.[24] Whether individuals of these non-MM genotypes would go on to develop clinical vCJD is unclear; however, it is now clear that blood transfusion can transmit vCJD from asymptomatic donors who subsequently developed vCJD.[25, 26] Interestingly the clinicopathological phenotype of secondary (transfusion-related) vCJD is indistinguishable from that of primary (BSE-related) vCJD, indicating that distinguishing between these two etiologies depends upon epidemiological studies such as the Transfusion Medicine Epidemiology Review.[27, 28] Additionally, an individual of the MV genotype has been found to be susceptible to vCJD infection by blood transfusion as judged by peripheral infection.[29] Evidence of a pre- or sub-clinical state existing in a hemophiliac patient who died of other causes, suggests that plasma products may also be a risk for vCJD transmission.[30] Although modeling exercises indicate that blood-borne vCJD transmission is unlikely to be self-sustaining in the UK population,[31] it may yet be premature to consider BSE and vCJD as things entirely of the past.

Potential future zoonoses

Scrapie is endemic in many countries around the world, yet there is no evidence to suggest that it is pathogenic for humans. The intense investigations of ruminant TSEs that followed the BSE epidemic have resulted in the identification of several distinct animal prion diseases, atypical or Nor98 scrapie in sheep and H-type and L-type BSE in cattle.[32] Moreover, BSE is experimentally transmissible to sheep and there are concerns that if BSE were to have infected the national flock in the UK its presence might be masked by endemic scrapie, but it might retain its pathogenicity for humans.[33, 34] Another concern, particularly for the North American countries, is the spread of chronic wasting disease in farmed and free-ranging deer and elk.[35] There is no known epidemiological link between any of these animal prion diseases and human disease, but there are active efforts to try to quantify strain-related species barriers between the diseases known to be a risk (BSE/vCJD), those thought not to represent a risk (scrapie) and those for which data is lacking (atypical scrapie, H- and L-type BSE and BSE in sheep).[36] In assessing whether or not human prion diseases might have an animal origin, it is important to have a proper understanding of the clinicopathological heterogeneity of the sporadic human prion diseases, because it is against this backdrop that any new acquired forms of the disease will be seen and from which it will need to be distinguished.

Sporadic CJD and variably protease-sensitive prionopathy

Sporadic CJD is the most commonly occurring human prion disease; it occurs world-wide and it has long been known to be clinically and pathologically heterogeneous. The molecular basis for this heterogeneity is currently thought to reside in a combination of the PRNP codon 129 polymorphic status of the patient (MM, MV, or VV) and the type (type 1 or type 2) of the protease-resistant component of PrPSc determined by protease K digestion and Western blotting (termed PrPres).[37, 38] The original sCJD sub-classification system of Parchi et al. that recognized six sCJD subtypes (MM1/MV1, MM2c, MM2t, MV2, VV2 and VV1) has had to be modified to accommodate the growing number of cases recognized to contain both type 1 and type 2 PrPres in different or sometimes the same regions of the brain.[39, 40] Moreover, intensive surveillance and investigation of forms of human prion disease that lack PRNP mutation and known risk factors has identified another sporadic human prion disease, termed protease-sensitive prionopathy (VPSPr).[41] While intensively investigated, the etiology and diversity of the sporadic human prion diseases remain poorly understood.

Fundamental biological mechanism

The prion hypothesis itself is of intrinsic interest. The expectation, implicit in the prion hypothesis, that in prion diseases the infectivity, the neurotoxicity and the strain-like properties of the agent (a prion) depend fundamentally on the structure and production of PrPSc presents a major challenge to molecular biology. However, it is a challenge that is beginning to be met. If one defines a prion as a protein-based inheritance unit conferring a trait on the basis of a post-translational switch in conformation involving the acquisition of β-sheet structures and multimerization, then a group of yeast proteins, Ure2p, Sup35p, Rinq1p and HETs, are prions; associated with a variety of yeast cytoplasmic inheritance-based traits when present in their prion forms, URE3, PSI+, PIN+ and Het-s respectively.[4] These yeast and fungal prions do not cause disease; instead they appear to represent an effective and common epigenetic mechanism for rapid cellular responses to environmental stress.[42, 43] Neither does this prion-like mechanism appear restricted to microbes. The Aplysia cytoplasmic polyadenylation element binding protein (CPEB), which is involved in long-term potentiation, is regulated by a prion-like switch.[3, 44]

Perhaps more controversially within neuropathology circles, the prion paradigm is being invoked as a way of understanding the behavior of proteins such as tau, α-synuclein, superoxide dismutase-1, TAR DNA-binding protein 43, FUS (Fused in Sarcoma) and huntingtin in their neuropathological context.[45-49] The analogy being drawn relates to: (i) a templated or seeded conversion mechanism; (ii) the possible existence of different molecular strain types; or (iii) the ways in which the proteopathy spreads within the nervous system.[50-53] The idea that neurodegenerative change in such diseases is non-cell autonomous, but instead represents the spread of molecular pathology, is of particular interest with respect to sporadic forms of disease. If we equate sporadic forms with spontaneous protein misfolding and self-sustaining amplification, then it follows that the misfolding event that initiates disease must occur at a defined location and at a given time and then spread. One might speculate that different clinicopatholgical features would follow depending on the regional propensity for such events to occur for any given protein, much in the same way that Braak and Braak staging describes typical Alzheimer's disease progression.[54]

There is also a potentially important practical corollary to the idea of prion-like spread, which may affect future stem cell therapies for neurodegenerative diseases. Presumably therapeutic stem cell-derived neurons would be equally susceptible to “infection” (with misfolded protein aggregates) as the patient's own cells, unless steps were taken to prevent this,[55] the most obvious of which would be to prevent expression of the gene product that can be converted to a pathological prion-like isoform. The suggestion that a prion-like mechanism of spread of molecular pathology underlies diseases as diverse as Alzheimer's disease and Parkinson's disease has led some researchers to explore whether the molecular pathology of these diseases is transmissible in an experimental setting[56-58] and to suggest that perhaps some cases of these more common neurodegenerative illnesses might, like CJD, be acquired.[58, 59]

Recent progress in the development of in vitro approaches

The apparent absence of a nucleic acid-based genome and the difficulties associated with culturing prions has meant that much prion disease research (including human prion disease research) continues to be done in experimental animals. However, this is beginning to change. The development and application of techniques that can be used to probe the conformation and/or aggregation state of human prions extracted from human tissue have allowed for “molecular strain typing” as an alternative to biological strain typing by animal transmission.[37, 38, 60] Specific cell lines and strategies that allow for the replication of a widening range of prions in cultured cells are being developed. This has practical application in the form of rapid end-point titration of scrapie prions and the possibility of scrapie prion strain differentiation using a cell panel assay.[61, 62] These technical innovations can be put to basic scientific purpose as demonstrated by the recent finding that, although devoid of nucleic acid, scrapie agent replication in culture displays properties analogous to mutation, competition and selection.[63] Cell-free PrPSc seeded conversion assays, such as protein misfolding cyclic amplification (PMCA) allow prion propagation to be studied in vitro, in a flexible system in which the effects of species, strain and genotype of the seed (containing PrPSc) and substrate (containing PrPC) can be controlled and manipulated.[64, 65] Ancillary molecules involved in PMCA can also be studied and the minimal components required for the formation of infectious prions defined.[66-68] Lastly, prions can be created de novo from refolded recombinant prion protein and their biochemical and biological properties compared.[69] Paradoxically, many of these in vitro approaches to prion disease research have been developed using materials from high-titer rodent models of sheep scrapie. The challenge for human prion disease research is to apply these emerging techniques to the study of human prions in humans.

Molecular Basis of Human Prion Diseases

Molecular strain typing

Molecular strain typing in the form of classifying the mobility and glycoform ratio of protease-resistant prion protein by Western blotting is a remarkably useful adjunct to neuropathological assessment during the post-mortem diagnosis of human prion diseases (Fig. 1). The glycoform ratio difference between vCJD and all forms of sCJD is a remarkably robust phenomenon, although the mechanism underlying it remains obscure. All cases of vCJD examined show type 2B PrPres, irrespective of brain region assayed and the PrPres type is also found in lymphoreticular tissues, albeit with presumably tissue-specific minor modification of mobility and an accentuation of the glycoform ratio. Similarly sCJD cases are characterized by a narrow range of glycoform ratios, distinct from vCJD, and the presence of either type 1 or type 2 PrPres (type 1A and type 2A). The PrPres types found in the brain in iCJD and kuru resemble those found in sCJD (type 1A and type 2A), from which they were presumably derived. Individual cases of gCJD, GSS and FFI usually have type 1 or type 2 PrPres, but with a glycoform ratio in which the non-glycosylated component is under-represented (which we have termed A/B). However, this is not always true and a broad spectrum of glycoform ratios can be found in genetic prion diseases. Moreover, some cases of GSS are characterized by an approximately 8 kDa (N- and C-terminally truncated) PrPres fragment, and some cases of FFI have little detectable PrPres at all.

Figure 1.

Diagrammatic representation of the protease resistant prion protein (PrPres) types found in the human prion disease brain as determined by proteinase K digestion and Western blot analysis. The classification of the banding pattern has two components, one numerical depending on the migration of the bands and the other alphabetical depending on their relative abundance. The pattern is termed type 1 if the non-glycosylated (bottom) band is ∼21 kDa, type 2 if the non-glycosylated band is ∼19 kDa or type 1 + 2 if both bands are found. The pattern is given the suffix A if the middle or bottom (mono- or non-glycosylated) bands predominate, B if the top (diglycosylated) band predominates or A/B if the glycosylated bands (middle and top) predominate at the expense of the non-glycosylated (bottom) band. A pattern dominated by a low molecular mass unglycosylated band is here termed 8 kDa. The faint ladder of bands that sometimes accompanies the 8 kDa band is shown in grey. Types 1A, 2A, 1 + 2A are characteristic of sporadic CJD, iatrogenic CJD and kuru. Type 2B is associated with vCJD. Types 1A/B and 2A/B are often found in genetic CJD, Gerstmann-Straussler-Scheinker disease (GSS) and fatal familial insomnia (FFI). The 8 kDa pattern is characteristic of some cases of GSS and of protease-sensitive prionopathy (VPSPr).

Despite the diagnostic utility, a simple one-to-one correspondence between PrPres type and disease phenotype (and by implication to agent strain) seems unlikely in principle and is complicated by the facts. First, the choice of analyzing only that fraction of PrPSc which survives a particular concentration of protease may seem arbitrary. Second, the interpretation of a molecular population variable, such as glycosylation site occupancy, as conforming to two or three discrete types, could be seen as simplistic. Lastly, protease digestion may be considered to be a somewhat blunt instrument to distinguish secondary and higher-order conformational differences in PrPSc. Even when genotype (mutations and polymorphisms) is taken into account, three major types (1, 2, 8 kDa) and three wild-type genotypes (MM, MV and VV) provide insufficient molecular variation to account for all the phenotypic variations observed. For example, two forms of sCJD share methione homozygosity and type 2A PrPres but one form closely resembles FFI (without the causative mutation) and the other is CJD-like.[8] Two more substantial problems, which may point toward a more subtle and perhaps informative approach to PrPSc analysis, are discussed below.

The co-occurrence of PrPres types 1 and 2

Once controversial, the idea that PrPSc in individual cases might be composed of mixtures (or different types co-occurring) is now well recognized and accepted.[40, 70] There are probably two phenomena at play here. One is the finding of different predominant types in individual samples from different parts of the brain or more rarely approximately equal amounts of type 1A and type 2A in the same sCJD brain samples. The other is the observation made using antibodies that specifically recognize type 1 or type 2 PrPres, that a minority type always accompanies a majority type in sCJD and vCJD, albeit at sub-detectable levels when conventional antibodies are used.[71-75] The former issue is more tractable and a consensus is beginning to emerge that when multiple brain sampling and sensitive co-detection is performed on cohorts of sCJD cases, a plateau is reached at between 30–40% of cases showing co-occurrence. Our own data examining four regions (temporal cortex, parietal cortex, occipital cortex and thalamus) instead of frontal cortex only, shows a rise in detected co-occurrence from 3% to 24% of cases.[76] Interestingly, only very rarely did this re-analysis involve a change in the predominant type found in the brain overall. Parchi et al. have offered a revised version of their 1999 sporadic CJD classification system that adds mixed type to the original “pure” types and have shown that the most common of these 12 sCJD subtypes can be recognized on histological grounds, without reference to biochemical analysis.[39, 40, 77] It will be interesting to see in the fullness of time whether this additional complexity reflects a more refined series of discrete clinicopathological phenotypes or whether it is indicative of a spectrum of phenotypes depending on the spacio-temporal accumulation of PrPSc types set against the patient genotype.[78]

Variably protease-sensitive prionopathy

The phenotypic complexity of the sporadic forms of human prion disease has increased with the report of a new sporadic human prion disease, termed variably protease-sensitive prionopathy (VPSPr) that is distinct from previously recognized sub-types of sCJD.[41, 79] There are no mutations in the open reading frame of PRNP. The patients have no known risk factors for the disease, but the disease is most common in the VV genotype, as opposed to sCJD, which is most common in the MM genotype. The neuropathology involves medium-sized vacuolation and characteristic microplaques. Durations of illness can be very long and this coupled with symptoms that do not conform well to CJD have prompted speculation that the condition may be under-ascertained. The most interesting aspect of the disease from a biochemical perspective is that although PrPSc is abundantly present in the brain, PrPres is difficult to detect because of its sensitivity to proteolysis and because what remains after proteinase K (PK) digestion is both C- and N-terminally cleaved by PK digestion and seen as a faint 8 kDa band on Western blots (Fig. 2). The degree of protease resistance is reported to reflect the codon 129 genotype, with VV being least resistant and MM being most resistant, despite having the same 8 kDa PrPres fragment predominating.[79] We have identified two cases of VPSPr prospectively in the UK[80, 81] and recently completed a retrospective review for such cases confirming many of the original observations by Gambetti and colleagues.[41, 79] Our work has shown that some areas of the VPSPr brain contain PrPres similar in appearance to that found in sCJD and conversely that some cases of sCJD have a very minor PrPres band similar to the 8 kDa PrPres band that typifies VPSPr.[82]

Figure 2.

Western blot analysis of the protease resistant prion protein (PrPres) type found in variably protease sensitive prionopathy (VPSPr). Five microliters of the 10% weight-to-volume brain homogenate shows readily detectable prion protein when analyzed without prior digestion with proteinase K (-[5]). Following digestion with proteinase K at 50 μg/mL, PrPres is undetectable when 0.5 μL of a the same 10% weight-to-volume brain homogenate is loaded (+[0.5]), barely detectable when 5 μL of a the same 10% weight-to-volume brain homogenate is loaded (+[5]), and only becomes readily detectable when 50 μL of the same 10% weight-to-volume brain homogenate is concentrated before loading (+[50]). The protease-sensitive prionopathy (VPSPr) sample is shown run between typical examples of sporadic CJD MM1 subtype (1A) and vCJD (2B) loaded at 5 μL of a 10% weight-to-volume brain homogenate and in which PrPres type 1A and 2B respectively are abundantly present. Markers are shown with their mass given in kDa. The anti-PrP monoclonal antibody used was 3F4. A full description of this case can be found in Head et al.[80]

Protease-sensitive PrPSc

The idea that protease-sensitive forms of PrPSc (senPrPSc) exist is not new, but until recently its significance was uncertain. Additionally, senPrPSc is difficult to detect directly, requiring techniques, such as conformation-dependent immunoassay (CDI), that identify PrPSc on the basis the exposure of specific PrP epitopes by treatment with chaotropic salts. The application of CDI to the post-mortem sCJD brain showed that more than 80% of the PrPSc (as defined by CDI) is sensitive to proteolytic degradation under the conditions generally used for Western blot PrPres typing.[83] We have confirmed these results and extended them to vCJD, which also has more senPrPSc than PrPres present in the brain (Fig. 3).[84]

Figure 3.

Conformation dependent immunoassay analysis (CDI) of vCJD brain samples measuring total PrP, PrPSc, protease-sensitive forms of PrPSc (senPrPSc) and protease-resistant component of PrPSc (PrPres). (A) shows the time resolved fluorescent readings (TRF counts ×103) obtained in the native state (grey bar = PrPC) and denatured state (black bar = PrPC and PrPSc) when the analysis was done after no treatment (0), treatment with 2.5 μg/mL proteinase K (2.5), or treatment with 50 μg/mL proteinase K (50). In (B) the black bars represent the percentage of the denatured signal at 2.5 μg/mL proteinase K that survives treatment with 50 μg/mL proteinase K in three separate samples of vCJD frontal cortex. The black bar is a measure of PrPres, and the white bar therefore represents senPrPSc. CDI used the capture and detection antibody pairing of MAR-1 and Europium-labeled 3F4. Cited from Choi et al.[84]

It is worth pointing out that by definition senPrPSc does not figure in conventional Western blotting analyses and cannot therefore be ascribed a PrPres type. It is therefore possible that phenotypic and strain-related aspects of human prion diseases could be engendered by senPrPSc. The progressive unfolding of PrPSc with increasing chaotrope concentration had previously been shown to produce complex rodent scrapie strain-specific CDI readings or “melt curves”.[60] Direct application of this methodology to human brain specimens is fraught with difficulties; however, we have been able to show that when detergent insoluble PrPSc is analyzed, the stability of vCJD and sCJD PrPSc differs. The stability of PrPSc in the MM1 and VV2 sCJD subtypes is indistinguishable but their PrPSc is more stable than that of vCJD (Fig. 4).[85] Interestingly synthetic prions made by refolding recombinant PrP display a diverse conformational stability, as judged by CDI-like methods[86] and this property has a phenotypic correlate: those strains of synthetic prions with least stability have the shortest incubation periods.[87] Moreover, protease-sensitive synthetic prions can be made and serially passaged in a specific transgenic mouse host.[88] The full implications of these synthetic prion studies are hard to judge at present, but they seem to suggest that our focus has been on the readily detectable, but comparatively inert end product of PrP conversion (PrPres), when a more informative view could be had by developing and refining techniques to look at more the labile and dynamic senPrPSc.

Figure 4.

Comparison of vCJD (vCJD) and sporadic CJD (sCJD) brain PrPSc using the conformation dependent immunoassay (CDI) in (A) and the conformation stability assay (CSA) in (B). In (A) sarkosyl insoluble PrPSc was exposed to increasing concentrations of the chaotrope guanidine hydrochloride (0–7 mol/L GdnHCl), and the CDI values obtained for the denatured fraction were plotted as a percentage of that measured in 7 mol/L GdnHCl. vCJD is shown as white squares, sCJD MM1 subtype as black diamonds, and the sCJD VV2 subtype as black triangles. CDI used the capture and detection antibody pairing of MAR-1 and Europium-labeled 3F4. In (B) vCJD and sCJD (MM1 subtype) brain homogenates were exposed to a range of guanidine hydrochloride concentrations (0–4 mol/L GdnHCl) and then digested with 20 μg/mL proteinase K before being analyzed by Western blotting using the monoclonal antibody 3F4. Cited from Choi et al.[85]

The molecular basis of infectivity and neurotoxicity

Recent studies of the biochemical basis of prion infectivity and neurotoxicity also appear to point away from large stable fibrillar aggregates: As one might expect, the accumulation of oligomeric PrP aggregates precedes the accumulation of PrPres in a rodent models.[89] However, even at end-stage disease, biochemical separations based on molecular size and density implicate non-fibrilar oligomeric species of PrP as the most infectious forms and there appears to be a strain-specific element to the size classes represented.[90-92] Experimental evidence in favor of a role for oligomeric species of PrP in poisoning the proteasomal system in prion diseases has been reported.[93, 94] The differing kinetics of prion infectivity and neurotoxicity in murine scrapie models has been used to argue for the existence of a neurotoxic form of the cellular PrP termed PrPL (for lethal) generated during prion propagation.[95] PrPL may or may not correspond to the toxic monomeric α-helical species TPrP independently identified by a toxicity testing approach.[96]

We have recently examined PrPSc aggregation state in the vCJD brain in an effort to try to understand regional differences in pathology.[97] The approach taken was to combine sucrose density gradient centrifugation with CDI detection of PrPSc in regions of the vCJD brain that differed in their pathological hallmarks. The most marked contrast was between cortical regions (in which vacuolation is intense and PrP plaques and plaque-like structures are common) and the thalamus (which is characterized by intense astrogliosis and neuronal loss, but in which plaques are rare and spongiosis patchy). In cortical samples PrPSc, as defined by CDI, was predominantly in the bottom (heavy or aggregated) fractions whereas the PrPSc found in the thalamus was more polydispersed across the gradient, including a readily detectable fraction with the sedimentation properties of PrPC, that was not observed in cortical regions (Fig. 5).[97] A similar correlation between regional disease severity in sCJD and the presence of PrP oligomers has been previously reported.[98] It is tempting to speculate that these observations might represent the in vivo detection of a form of oligomeric or monomeric PrP directly associated with neurotoxicity.

Figure 5.

Comparison of PrPSc aggregation state (A) and pathology (B–G) of thalamus and frontal cortex of the vCJD (vCJD) brain. In (A) brain homogenates were subjected to sucrose density gradient fractionation, the fractions were digested with 2.5 μg/mL proteinase K to remove PrPC and then analyzed using conformation dependent immunoassay (CDI) in the native (N) and denatured (D) states to detect PrPSc. The values for each fraction of thalamus (black bars) and frontal cortex (grey bars) were expressed as a D/N ratio. CDI used the capture and detection antibody pairing of MAR-1 and Europium-labeled 3F4. The pathology of vCJD thalamus (B,D,F) and vCJD frontal cortex (C,E,G) were compared in sections that had been stained with HE (B,C), immunostained for PrP using the monoclonal antibody KG9 (D,E) and immunostained for GFAP (F,G). Cited from Choi et al.[97]

The biochemical basis of the strain phenomenon

The results of transmission of individual samples from single examples of the six different Parchi et al.[39] sCJD subtypes (MM1/MV1, VV1, MM2c, MV2, VV2) into humanized transgenic mice suggest the existence of four distinct sCJD agents, termed M1, M2, V1 and V2, and a fifth strain corresponding to MM2t or sporadic fatal insomnia.[99, 100] Interestingly, when we performed formally analogous experiments in the cell-free PMCA reaction, similar results were obtained: The PrPres type of the seed was conserved in the PMCA product and the efficiency of conversion appeared to be determined by compatibility at codon 129 of PRNP.[101] The behavior of the seeds from heterozygous patients were particularly interesting, in that MV1 sCJD seeds selectively amplified in MM substrate producing type 1 PrPres and MV2 sCJD seeds selectively amplified in VV substrate producing type 2 PrPres (Fig. 6). These results reinforce the association between methionine at codon 129 and the production of type 1 PrPres and valine at codon 129 and the production of type 2 PrPres.

Figure 6.

Cell-free conversion of humanized transgenic mouse brain PrPC substrate by seeds from sporadic CJD brain using protein misfolding cyclic amplification. Samples from three cases of sporadic CJD MV1 subtype (sCJD MV1) and three cases of sCJD MV2 subtype (sCJD MV2) were used to seed PRNP codon 129-MM (MM) and PRNP codon 129-VV (VV) humanized transgenic mouse brain homogenates. Reactions were either subjected to a single round of 48 cycles of protein misfolding cyclic amplification (PMC) over 24 h (+) or frozen immediately without PMCA (-). The samples were then digested with 50 μg/mL proteinase K and analyzed for PrPres by Western blotting using the monoclonal antibody 3F4. The positions of molecular mass markers are indicated in kDa. Cited from Jones et al.[101]

Emerging Risks

Zoonotic disease

BSE is the only animal prion strain with demonstrated pathogenicity for humans. While it is tempting to suggest that scrapie might represent the animal reservoir that results in some cases of sCJD, there is no epidemiological evidence to support this hypothesis. The pathogenicity of new or newly described animal prion diseases for humans is unclear and this is particularly true for H- and L-type BSE, atypical scrapie and for chronic wasting disease (CWD), all of which are probably consumed. Human susceptibility has been modeled by attempted transmission to (humanized) transgenic mice with sometimes conflicting results, depending on the transgenic model used and depending upon whether central or peripheral tissues are examined.[102-106] We have attempted to establish whether PMCA can model the molecular component of these hypothetical cross-species transmission events.[107] The existing data correspond well with the established facts. First, PrPSc in vCJD brain samples amplifies most efficiently in humanized mouse MM substrate, less efficiently in MV substrate and not at all in VV. Cattle BSE PrPres is less efficient than vCJD, but shows the same substrate genotypic preference. Sheep scrapie fails to amplify detectably in any of the three substrates; however, sheep BSE PrPres does amplify, again with a codon 129 preference for methionine (Fig. 7). We are currently extending this approach to encompass atypical scrapie, H- and L-type BSE and CWD using human rather than humanized PMCA substrates.

Figure 7.

Cell-free conversion of humanized transgenic mouse brain PrPC substrate by seeds from brain samples of animal and human prion diseases. Samples from bovine spongiform encephalopathy in a cow (BoBSE), sheep scrapie (OvSc), experimental bovine spongiform encephalopathy in a sheep (OvBSE) and vCJD from a human patient were used to seed PRNP codon 129-MM (MM), 129-MV (MV) and 129-VV (VV) humanized transgenic mouse brain homogenates. Reactions were either subjected to a single round of 48 cycles of protein misfolding cyclic amplification (PMCA) over 24 h (+) or frozen immediately without PMCA (-). The samples were then digested with proteinase K and analyzed for PrPres by Western blotting using the monoclonal antibody 6H4. The positions of molecular mass markers are indicated in kDa. Cited from Jones et al.[107]

Secondary infection

In the same way that animal reservoirs cannot be completely excluded as causes of individual sCJD cases, neither can other environmental sources, such as medical procedures. The known routes of iatrogenic CJD acquisition are historically growth hormone therapy, dura mater grafting, corneal grafting and certain highly specialized neurosurgical procedures. The secondary transmission of vCJD by blood transfusion and experimental evidence showing the efficiency of the transfusion of viable blood cells between scrapie and BSE-infected and naive sheep have prompted a reappraisal of transfusion-transmitted CJD, including consideration being given to the possibility of prion blood testing or filtration.[25, 26, 108, 109]

Blood transfusion is the original and most extensively used cellular therapy, but we may be on the threshold of a new era of cellular therapies based on embryonic stem cell and induced pluripotent stem cell technologies. Although the potential for stem cell therapy-mediated prion transmission might be judged remote, this was also considered to be the case for transfusion transmission of CJD before 2004.

Future stem cell therapies may depend on a limited number of cell lines currently under development. These lines will have been extensively cultivated and exposed to a wide variety of human and animal-derived biological products, and in some cases exposed to other (feeder) human cells, before being used on a one-donor-to-many-recipients basis. We have begun to investigate the potential for stem cell-mediated prion transmission by examining how self-renewing populations of human stem cells respond to transitory exposure to BSE or vCJD brain homogenates in vitro.[110] Cellular uptake of PrPSc from culture medium is rapid, extensive and does not depend on species or codon 129 compatibility. It is most likely a non-specific uptake mechanism also involving brain components other than PrPSc (Fig. 8). The cells do not appear to become infected as such; instead the majority of cells clear the exogenous PrPSc by as yet undetermined mechanisms. We do not know what the long-term consequences (if any) might be of transitory exposure of stem cells to prion infectivity, nor do we know what effect neuronal differentiation of pluripotent progenitors might have on prion replication in such cells and their derivatives.

Figure 8.

Exposure of human embryonic stem cells (hESC) in vitro to human and cattle brain homogenates and demonstration of brain protein uptake by immunocytochemistry. (A–J) hESC (PRNP codon 129-MM) exposed to bovine spongiform encephalopathy (BSE) (A,F), vCJD (B,G), sporadic CJD VV2 subtype (sCJD) (C,H), Alzheimer's disease (AD) (D,I) 1% brain homogenates or cells grown in control medium without brain homogenate (Unexposed) (E,J) for 48 h. hESC cultures were then fixed, permeabilized, treated with proteinase K and guanidine and immunostained for PrP with either the 6H4 monoclonal antibody (6H4) (A–E) or the 3F4 monoclonal antibody (F–J) (green), and the nuclei counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). (K–M) show the same hESC line exposed to vCJD (K), AD (L) 1% brain homogenate or grown in control medium for 24 h (M). hESCs cultures (K–M) were then fixed, permeabilized, treated with proteinase K and immunostained for GFAP (red), and the nuclei counterstained with DAPI (blue). Scale bars represent 50 μm. Cited from Krejciova et al.[110]

Summary and Perspective

While the prospect of a major epidemic of vCJD in the UK and elsewhere seems to be receding, there remain a series of uncertainties surrounding the eventual numbers of individuals that will suffer from this devastating condition. The issues include the effects of genotype on susceptibility and the possible existence of substantial numbers of asymptomatic infected individuals that may pose risks of onward transmission. sCJD remains the most frequently occurring human prion disease and arguably the least well understood. Other idiopathic forms of human prion disease (such as VPSPr), characterized by protease-sensitive forms of the prion protein, also exist and their true prevalence may be hard to ascertain. The possible risks from newly described animal prion diseases and from emerging cellular therapies are currently poorly quantified. On a more theoretic level the prion hypothesis has provided a unifying conceptual framework for TSE research and provided a paradigm to interrogate the similarities and differences between the diverse neurodegenerative conditions involving prion-like mechanisms of molecular pathology.


I would like to thank Professor Akiyoshi Kakita and Professor Hitoshi Takahashi for their generous invitation to attend the 53rd Annual Meeting of the Japanese Society of Neuropathology at Niigata. I would also like to acknowledge Japanese colleagues with whom it has been a pleasure to collaborate and spend time with over the years, including Akiko Iwaki, Akiyoshi Kakita, Katsumi Doh-ura, Kensuke Sasaki, Mari Tada, Masanori Morita, Masahito Yamada, Tetsuyuki Kitamoto, and last, but by no means least Toru Iwaki. The work discussed here from my own research group is that of Alex Peden, Helen Yull, Marcelo Barria, Michael Jones, Young Pyo Choi and Zuzana Krejciova. I am particularly grateful to my graduate students for all that they have taught me. I am also grateful to Professor James Ironside for his generous support and his encouragement. The monoclonal antibody MAR-1 used in CDI was generously supplied by Dr Albrecht Groener (CSL Behring, Marburg, Germany). The transgenic animal brains used in PMCA experiments were generously provided by Dr Rona Barron and Professor Jean Manson (Roslin Institute, The University of Edinburgh, UK). The analysis of animal prion diseases was carried out in collaboration with Drs Martin Jeffrey and Lorenzo Gonzalez (AHVLA, Lasswade, UK). Animal prion disease specimens were obtained by request from the AHVLA Biological Archive Group (Weybridge, UK). All human brain specimens were obtained by request from the Medical Research Council NCJDRSU Brain and Tissue Bank. Ethical approval for their use is covered by LREC 2000/4/157 (Prof J. W. Ironside). The development of PMCA was funded by the Chief Scientists Office of the Scottish Government (Grant reference CZB/4/357 and CZB/4/688) and through collaboration with the Scottish National Blood Transfusion Service (Prof Marc Turner, Dr Ian MacGregor and Dr Christopher Prowse) and UK Forum funding. The investigation of human stem cell responses to human prion infectivity was also supported by a Chief Scientists Office grant to Dr Paul De Sousa and colleagues (MRC Centre for Regenerative Medicine, University of Edinburgh) (Grant reference CZB/4/588). The work of the NCJDRSU is funded by the Department of Health, UK and by the Scottish Government. This is an independent report commissioned and funded by the Policy Research Program in the Department of Health, UK. The views expressed in the publication are those of the author and not necessarily those of the Department of Health.