34 Prion-protein-related diseases of animals and man

Scrapie, bovine spongiform encephalopathy (BSE), Creutzfeldt-Jakob disease (CJD), and related diseases of mink (transmissible mink encephalopathy), mule deer and elk (chronic wasting disease) are the founder members of a group of diseases called the transmissible degenerative (or spongiform) encephalopathies (TSE). These diseases can be transmitted by prions from affected to healthy animals by inoculation or by feeding diseased tissues. Prions are cellular proteins that can transfer metabolic and pathological phenotypes vertically from parent to progeny or horizontally between cells and animals. TSEs are characterized by the accumulation of the prion form of the mammalian prion protein (PrP^C) in the central nervous system or peripheral tissues of animals and humans. Mutations of the human PrP gene are linked to rare, familial forms of disease and prion-protein gene polymorphisms in humans and other species are linked to survival time and disease characteristics in affected individuals. Iatrogenic transmission of CJD in man has occurred, and a variant form of CJD (vCJD) is due to cross-species transmission of BSE from cattle to humans. Atypical forms of scrapie and BSE have been identified during large-scale monitoring for TSEs worldwide. This chapter outlines our current understanding of scrapie, BSE, CJD and other TSEs and highlights recent progress in defining the role in disease of the prion protein, PrP.

Prions are cellular proteins that can transfer metabolic and pathological phenotypes vertically from parent to progeny or horizontally between cells or animals. Frequently, the conversion of a normal cellular protein into a prion form involves a conformational change that affects its degree of self-association and ability to interact with other molecules. The consequences of these molecular events can be measured by their effects on a population or on an individual organism, by the gross morphological change or damage to tissues and cells, by the metabolite flux or most directly by monitoring the change in shape and aggregation of the prion protein (Wickner et al. 2009).

Transmissible spongiform encephalopathies (TSEs) are diseases which are characterized by the accumulation of the prion form of the mammalian prion protein (PrP^C) in the central nervous system or peripheral tissues of animals and humans (Prusiner 1997).

Scrapie is the TSE of sheep and goats.

Creutzfeldt-Jakob disease (CJD) is the most common type of human TSE, and a novel variant (vCJD) is believed to be caused by the transmission of the cattle TSE.

Bovine spongiform encephalopathy (BSE) to humans.

Scrapie of sheep and goats has been known in Europe for centuries and has spread to most parts of the world due to trade in livestock. It has several colloquial names—goggles, staggers, traberkrankheit or trotting disease, la tremblante—that reflect a range of clinical conditions—altered behaviour, hypersensitivity to sound or touch, loss of condition, pruritus, and associated fleece loss and skin abrasions, incoordination of the hind limbs. These signs may be diagnostic to the experienced shepherd but usually require confirmation of disease by examination of brain tissue for histopathological signs, notably gliosis and vacuolation of the brainstem particularly in the obex at the level of the dorsal motor nucleus of the vagus (Hadlow 1995). In 1998, Benestad and her colleagues in Norway recognized a neurological disorder in sheep which they termed ‘Nor98’ (Benestad et al. 2003). Subsequent investigations have shown this is an atypical form of scrapie with epidemiological, histopathological, biological and molecular characteristics distinct from those of the type described classically in veterinary textbooks: however, both classical and atypical scrapie are prion-protein-related diseases.

Classical scrapie has been reported in most breeds of sheep and goats and, within a flock or herd, it appears to occur in related animals. The within-flock/herd incidence is usually 1–2 cases/100 sheep or goats/year but there have been several instances of 40–50% of animals of a flock succumbing to the disease within a year. The scrapie status of the dam is a major risk factor for the development of disease in progeny, and introduction of a new sire into a previously clear flock/herd has been noted (anecdotedly) to provoke outbreaks of clinical disease. Onset of the natural clinical disease peaks in flock/herd animals at 3.5 years, with most cases occurring in the age range of 2.5–4.5 years (Hunter et al. 1992). In the incubation period, the infected animal is clinically normal and indistinguishable from its uninfected flockmates. Shedding of prions into the environment via urine or faeces may occur but the lack of an in vivo diagnostic test for the infectious particle means that the true prevalence of infection and of the carrier status of unaffected animals within a flock are also usually unknown. The prognosis on observation of clinical signs is invariably death within a few days or months, hence in veterinary work this usually results in a recommendation to cull the affected animal from the flock/herd, and similarly to slaughter its dam and other maternally related animals.

The introduction of PrP¹ gene typing has greatly facilitated interpretation of field studies on the incidence of natural and experimental disease (Goldmann et al. 1994; Dawson and Del Rio Vilas 2008) and national scrapie control schemes in sheep based on elimination of PrP alleles of high susceptibility to clinical disease (VRQ2) or the expansion of relatively resistant alleles (ARR) have been highly effective in reducing the number of reported cases of classical scrapie (EFSA 2006; Dawson and Del Rio Vilas 2008). Goat scrapie is less common but the increase in intensive goat dairy farming in the last few years has become associated with animal health problems such as scrapie (Gonzalez et al. 2009) and a large outbreak in Cyprus has stimulated studies on the feasibility of breeding for TSE resistance in goats in the EU (EFSA 2009).

In 1998, the molecular and histopathological spectrum of TSEs in sheep was extended by the discovery in Norway of an experimentally-transmissible, PrP-related, neurological disease of sheep that was distinguishable from classical scrapie and was therefore considered to be an ‘atypical’ form of scrapie (Benestad et al. 2003). These Nor98 cases, the prototypes of ‘atypical’ TSE, have little or no vacuolation and sparse staining for abnormal PrP at the obex (dorsal motor nucleus of the vagus), and in many, but not all, cases exhibit an intense cerebellar PrP^Sc deposition/accumulation characterized at a molecular level by a smaller and less stable protease-resistant core of PrP^Sc. Nor98 and other ‘atypical’ cases subsequently identified are more often but not uniquely, found in animals carrying alleles of the PrP gene not usually associated with classical scrapie (ARR, AHQ). For Nor98, this genotype correlation has been further refined to implicate another dimorphic codon in the PrP open reading frame, L141F (Moum et al. 2005; Saunders et al. 2006). Other ‘atypical’ TSE phenotypes, including those that are similar to or the same as Nor98, have now been from France (Buschmann et al. 2004), Germany (Buschmann et al. 2004), Sweden (Gavier-Widen et al. 2004), Ireland (Onnasch et al. 2004), Portugal (Orge et al. 2004), Belgium (De Bosschere et al. 2004) and the UK (Konold et al. 2006).

Atypical scrapie is transmissible by intra-cerebral inoculation to transgenic mice (Le Dur et al. 2005; Griffiths et al. 2010) and by both the intra-cerebral and oral route to sheep (Simmons et al. 2007, 2010). Clinical disease in sheep resembles classical scrapie (Konold et al. 2007) although, in contrast to most forms of classical disease, little abnormal PrP has been detected in peripheral, non-neuronal tissues. Case definition of atypical scrapie (EFSA 2005) has allowed its monitoring in Europe and worldwide and its epidemiology clearly differs from that of classical scrapie (Fediaevsky et al. 2008). Several cases of atypical scrapie have recently been described in goats where it seems to be associated with the H154 codon (Arsac et al. 2007; Seuberlich et al. 2007).

Creutzfeldt-Jakob disease was first described by Creutzfeldt (1920) and Jakob (1921) as a progressive dementia with clinical signs suggesting dysfunction of the cerebellum, basal ganglia, and lower motor neurones. The clinical signs are variable but most commonly the disease is associated with gradual mental deterioration leading to dementia and confusion, and a progressive impairment of motor function, including myoclonus. CJD occurs mainly in the fifth and sixth decades of life. Most patients die within 6 months of onset of clinical signs and there are no verified cases of recovery. Pathologically the lesions of the brain included variable vacuolation of the neuropil, astrocytosis and, in about 10% of CJD cases, kuru and other types of amyloid plaques. By 1968, recognition of its similarity to kuru [(a form of human prion disease first recognized in Papua New Guinea (Hadlow 1995; Gajdusek 2008)] and scrapie had stimulated the transmission of CJD to a chimpanzee by intracerebral inoculation of biopsy tissue and confirmed its classification with these diseases. Gerstmann Straussler syndrome, a progressive dementia with cerebellar amyloid plaques, is a familial variant of CJD with an extended clinical time course (Hainfellner et al. 1995).

Epidemiologically, these human forms of prion disease can be classified as familial, sporadic, and iatrogenic. The incidence of CJD-related disease in man is remarkably constant at 0.5–1 cases per million of population per year throughout the world. This rarity, and the heterogeneity of animal prion diseases, makes it difficult to categorically rule out a link between the two but there is no evidence of an association. The low incidence casts doubt on the role of infection in its propagation within the human population but there have been several cases of human to human transmission via cadaver-derived therapeutics or tissues—as, for example, an unfortunate consequence of corneal transplantation (Kennedy et al. 2001), pituitary growth hormone injection (d’Aignaux et al. 1999) or dura mater grafting (Yamada et al. 2009). Several studies of risk factors from classical CJD have been carried out, but there are no consistently observed associations with occupation or foods (EFSA 2010).

Some 13–14% of cases are familial and linked to mutations in the open reading frame (ORF) of the PrP gene. There have been many clinical and pathological studies on human cases of neurological disease which seem to be associated with these rare mutations of the PrP gene, including Jakob’s original family and the first GSS case (Kovacs and Budka 2009). In some families, there is complete penetrance of the phenotype and so the mutation is regarded as the cause of the disease. Apart from iatrogenic cases induced by transplantation of infected tissues or inoculation of contaminated pharmaceuticals of human origin, there is no epidemiological evidence for horizontal transmission of the disease. A stochastic event involving conversion of the PrP protein to its disease-associated isoform or the chance mutation of a benign, ubiquitous viral-like agent are two mechanisms which have been suggested to explain the incidence of sporadic cases. There is no cure for the clinical condition although genetic counselling, where applicable, may effectively prevent transmission of disease from one generation to the next.

There is considerable clinical and pathological heterogeneity in the human prion diseases, including fatal familial insomnia (Gambetti et al. 1995; Parchi et al. 1995; Tateishi et al. 1995) and although genetic typing and nucleotide sequencing of the PrP ORF has provided some unifying concepts, other genes and epigenetic effects are implicated in this variety of phenotypes (Kovacs et al. 2005).

In April 1996, Will and colleagues (Will et al. 1996) reported a novel variant of CJD (vCJD). Ten cases, all in young adults or teenagers presented with behavioural and psychiatric disturbances and early ataxia. The duration of illness was prolonged (up to 2 years) and typical EEG changes of CJD were absent. There was extensive kuru-type, amyloid plaque formation surrounded by vacuoles. Spongiform changes were most evident in the basal ganglia and thalamus with high-density, abnormal PrP accumulation on immuno-cytochemical analysis, especially in the cerebellum.

The initial, and subsequent, focus of vCJD in Great Britain and its molecular (Collinge et al. 1996) and transmission (Lasmezas et al. 1996; Bruce et al. 1997) similarities to BSE immediately implicated the cattle disease as the source of vCJD infection and beef and cattle by-products were put under restriction to limit the spread of disease. Nevertheless, an estimated three million infected cattle may have entered the human food chain (Ghani et al. 2000) and the impact and cost of preventing a secondary, human-to-human wave of infection are still being felt in the UK in 2010. To date (March 2010), there have been 169 primary cases, and three secondary cases related to transfusion of blood products, in the UK, 25 cases in France, 5 in Spain and 11 in the rest of Europe; other cases have been reported in the USA, Canada, Saudi Arabia and Japan (www.cjd.ed.ac.uk/vcjdworld).

Bovine spongiform encephalopathy (BSE) has devastated the UK cattle industry for the past twenty years. From isolated cases first reported in 1986 and some retrospectively identified in May 1985, a major epidemic was under way by 1988 which has to date claimed over 185,000 cattle within the British Isles, and several thousand more cases within Europe (www.defra.gov.uk/vla/science/sci_tse_stats).The disease produces a progressive degeneration of the central nervous system and was named because of the sponge-like appearance of BSE brain tissue when seen under the light microscope (Wells et al. 1987). Warning signs of the illness include changes in the behaviour and temperament of the cattle. The affected animal becomes increasingly apprehensive and has problems of movement and posture, especially of its hind limbs. The cow (or bull) has increased sensitivity to touch and sound, loss of weight and, as the disease takes hold of its nervous system, a creeping paralysis sets in. This clinical phase of BSE lasts from a fortnight to over 6 months (Konold et al. 2004). Although the majority of animals affected have been dairy cows, this neurological disease can occur in either sex with a modal age of onset of 4–4.5 years (range 1.8–22 years). Most early cases of BSE occurred in cattle between the ages of 3 and 5 years in Great Britain, but as the epidemic has waned and level of exposure declined, the average age of cattle with confirmed signs of disease has increased gradually to over 13 years (EFSA 2009). This has occurred in different European countries at different times and reflects the staggered nature of the BSE epidemics that have spread throughout Europe from the UK. For most of its development time the disease gives no tell-tale sign of its presence (Wilesmith et al. 1988) and the inability to detect the asymptomatic carrier of BSE (or scrapie) limits refinement of the measures which can be taken to prevent infected bovine or ovine tissues from use in feed and pharmaceutical products.

The neurological lesions in BSE-affected cow brains are virtually identical to those found in scrapie-affected sheep and include the spongiform change which gives BSE its name. From its clinical and neuropathological signs, BSE was immediately suspected to belong to the scrapie family of transmissible spongiform encephalopathies. This was confirmed by biochemical studies (Hope et al. 1988) and by experimental transmission of BSE to mice (Fraser et al. 1988), cattle (Dawson et al. 1990), mink (Robinson et al. 1994), marmoset (Baker et al. 1993), cynomolgus macaques (Lasmezas et al. 1996), sheep and goats (Foster et al. 1993) and pigs (Wells et al. 2003).

Epidemiological analyses of BSE-affected herds identified a protein feed supplement to be the most likely source of infection (Wilesmith et al. 1988). During the late 1970s changes in the rendering process which salvages compounds of nutritional and commercial value from abattoir waste are thought to have led to a less efficient system for inactivating prion-infected offal and, in turn, to a contaminated protein supplement. Subsequent recycling of BSE-infected cattle waste in this process may have contributed to the persistence of the disease.

Ruminant feed legislation aimed at removing the source of infection from cattle born after 1988 was introduced in 1989–90 in the UK, and reinforced throughout Europe in October 2000. At its peak, over 1,000 cases were reported each week in Great Britain in 1993 and this has now dwindled to one or two per month in 2010. Although the feed bans have had a dramatic effect on the epidemic curve, cases of BSE continue to be confirmed in cattle born after the reinforced bans of 2000 and feeding of contaminated protein to calves is suspected as the reason for most of these ‘born after the real ban’ (BARB) cases (Wilesmith et al. 2010). Their biological and biochemical characteristics appear similar to those seen earlier in the epidemic and differ from those of atypical BSE cases that have been recognized recently (see below).

In parallel with the BSE epidemic, natural cases of transmissible spongiform encephalopathies have been also been reported for the first time in cattle-related species —greater kudu, eland, nyala and gemsbok, Arabian and scimitar-horned oryx (Kirkwood and Cunningham 1994; Cunningham et al. 2004)and in the cat family—

puma, cheetahs (Kirkwood et al. 1995), and domestic cats (Pearson et al. 1992). Apart from some cases in the greater kudu, contaminated feed is suspected but difficult to prove because of the absence of detailed feeding records.

Experimental oral dosing of cattle with BSE-affected cattle brain homogenates has confirmed that as little as 1 milligram of brain (with ∼ 10–100 mouse ic ID₅₀ units) can induce disease after extended incubation periods of 8–10 years (Wells et al. 2007; Arnold et al. 2009). Larger doses (up to 100g) have been used to study oral pathogenesis of BSE in cattle in the UK (Wells et al. 2007) and Germany (Hoffmann et al. 2007) and confirmed by PrP IHC and bioassay the limited, early distribution of prions to parts of the lower alimentary tract (distil ileum, jejunum) and spread via the autonomic nervous system from the gastrointestinal tract to the central nervous system via either the coeliac and mesenteric ganglion complex, splanchnic nerves and the lumbal/caudal thoracic spinal cord or via the vagal nerve. This experimental tissue distribution of infectivity has been used to refine the list of specified risk materials from various age-cohorts of cattle banned for human consumption and has underpinned several assessments of human and animal exposure risk that have defined UK and European policy for control and management of BSE over the years (EFSA 2011).

BSE surveillance testing of cattle for abnormal prion protein in Europe has allowed the identification of two further, distinct types of cattle TSE, termed H- and L-(or BASE) type BSE (Casalone et al. 2004; Jacobs et al. 2007; Biacabe et al. 2008; Polak et al. 2008). Similar cases were also detected outside Europe (Japan and USA) (Hagiwara et al. 2007; Clawson et al. 2008). About 60 atypical BSE cases have been described worldwide (from testing ∼ 50 million healthy animals and fallen stock) although there is no statutory requirement to distinguish typical and atypical types of BSE in reporting and this figure is derived from research literature.

In France, a retrospective study of all the TSE-positive cattle identified through the compulsory EU surveillance programme between 2001–2007 was recently published (Biacabe et al. 2008). This study indicated that all BSE H and L cases detected by rapid tests were observed in animals over 8 years old in either the ‘at risk’ (9) or ‘healthy slaughtered’ surveillance target group (4). In this study, the reported frequency of H and L type TSE was respectively 1.9 and 1.7 cases per million of over 8 years old tested animals. All EU atypical cases were born before the extended or real feed ban that came into law in January 2001. Hence, as with classical BSE, exposure of these animals to feed contaminated with low titres of TSE cannot be excluded. However, the distribution of H-and L-type cases in France by year of birth differs markedly from that for classical BSE and could be interpreted to indicate that both forms of atypical BSE are sporadic diseases which arise spontaneously.

H- and L-(or BASE) type BSE have been transmitted by intra-cerebral challenge to inbred mice and Tg mice expressing bovine and ovine PrP. L-type BSE has also been transmitted to transgenic mice expressing alleles of the human prion protein (Beringue et al. 2007; Beringue et al. 2008; Buschmann et al. 2006; Capobianco et al. 2007; Kong et al. 2008). Transmission and serial passage in inbred mice and Tg VRQ mice have been interpreted to indicate that, after interspecies passage, BASE could generate classical BSE (Beringue et al. 2007; Capobianco et al. 2007). However, it should be noted that L-BSE—classical BSE phenotypic convergence has not been observed in other Tg mice, including mice expressing the ARQ allele of sheep PrP (Buschmann et al. 2006; Beringue et al. 2007). This phenomenon needs to be confirmed in an independent set of experiments but does raise the issue of a possible classical BSE re-emergence originating from atypical BSE cases.

The sensitivity and specificity of the TSE rapid screening tests are known for classical BSE but not for H- or L-type BSE. These tests use brainstem as the target tissue because this is where pathological lesions and PrPres are first detected in the CNS of cattle (Hope et al. 1988; Wells et al. 1998). Unlike classical BSE, little is known about the pathogenesis of atypical BSE and the brainstem may not be the optimal target site for the detection of H- and L-type BSE (Casalone et al. 2004). Consequently the BSE H- and L-type prevalence of 1–2 per million may be an under-estimation. No data are yet available on distribution of the infectivity in peripheral tissues and body fluids of cattle with H- or L-type BSE.

Foster and colleagues showed cattle BSE could be transmitted to ARQ/ARQ sheep and goats by feeding and intra-cerebral inoculation (Foster et al. 1993) and several subsequent studies have documented that there is wide-spread dissemination of prions in ARQ/ARQ sheep similar to the pathogenesis of natural cases of classical scrapie (Van Keulen et al. 2000). The biological and biochemical characteristics of ‘BSE in small ruminants’ are sufficiently distinct to allow their discrimination in ‘blinded’ tests although there have been concerns ‘mixed’ infections might pass as ‘scrapie’. Historically, small ruminants were known to have been fed the same type of protein supplements implicated as the source of BSE in cattle and fear of a second wave of vCJD due to infection from sheep and goat products stimulated intensive surveillance in the EU of TSEs in sheep and goats and the application of laboratory tests aimed at at a diagnosis of ‘NOT BSE’ or ‘BSE NOT Excluded’; the final confirmation of ‘BSE in small ruminant’ requires the application of bioassay in the same panels of inbred mice used to characterize vCJD and BSE (Bruce et al. 1997). By these stringent criteria, only two cases of BSE in SR, both in goats, have been confirmed (Eloit et al. 2005; Jeffrey et al. 2006) and current estimates of the likely prevalence of BSE in SRs in Europe is very low.

Modern virus classification uses the morphology and biochemistry of virions and their mode of replication as the basis for taxonomy; for example, the nature of the virion nucleic acid—DNA or RNA—its size, symmetry, the presence or absence of a lipid envelope, genome integration, mechanism of cell entry, use of vectors, etc. Prions remain undefined in this sort of detail and so their classification has not been easy, although many structures have been described as specific for TSE-infected fractions: 14 nm particles (Cho and Greig 1975): ‘nemaviruses’ (Narang 1990), scrapie-associated fibrils (Merz et al. 1981; Merz et al. 1983) or prion rods (McKinley et al. 1986) and small, pentangular structures of 10 nm diameter (Ozel et al. 1996) or in tissue sections as spheres and tubes (Baringer et al. 1981) or tubulo-vesicular vesicular structures (TSVs) (Liberski et al. 1988). Only prion rods and scrapie-associated fibrils have been shown unequivocally to be different morphological forms of prion-protein (PrP) aggregates by immuno-gold electron microscopy and the orientation and packing of abnormal PrP within these structures has been exhaustively investigated by modelling, 2D-electron crystallography and X-ray fibre crystallography (Wille et al. 2002; Wille et al. 2009). Natural and synthetic prion structures recovered from brain tissue have a cross-β∼structure characteristic of the protein fibrils of amyloid diseases and, intriguingly, while X-ray fibre diffraction patterns of these two brain PrP amyloids are similar, they differ considerably from the recombinant PrP amyloid used to induce synthetic prions in transgenic mice; the relationship between these structures and infectivity remains elusive (Wille et al. 2009).

Viruses, bacteria, bacteriophages, and all other conventional forms of life show phenotypic or strain variation which is encoded by their nucleic acid genomes. Strain variation is also a common feature of various TSE isolates in mice, hamsters, sheep, and goats, but while selection and mutation of murine TSE strains is documented, a coding molecule has yet to be defined. Prions and viruses share the cell’s trafficking and biosynthetic systems and the perturbations they induce may share common mechanisms and appear very similar. Similarly the concepts of a ‘strain’ and ‘mutation’ have been adopted into prion biology from virology although, by definition, a prion strain lacks a nucleic genome and propagates independently of nucleic acid replication and its errors. For prions, the different effects of infection on cell integrity, host behaviour and tissue pathology are believed to be mediated by the conformation of the prion protein and its epigenetic interactions with the host. Western blotting (WB) and protease hydrolysis have provided a useful, low resolution combination of techniques for monitoring prion protein shape and several, but not all, prion ‘strains’ or phenotypes have been described with unique WB profiles (Somerville et al. 1997). However, defining a prion strain (and, by implication, its disease phenotype) solely by its WB pattern or inferring a ‘mixed infection’ of prion strains from seeing a mixture of protein fragments by WB is not yet possible.

Each scientific discipline—epidemiology, virology, pathology, molecular biology, biochemistry—uses its own techniques and language to categorize prions and their effects and this has led to some confusion when a lack of one or more of these types of data prevents the cross-referencing needed from a comprehensive description of prion disease phenotype. In the past few years, the discovery of apparently-novel cattle and sheep prion-protein related abnormalities by rapid surveillance testing of healthy slaughter animals in the EU has highlighted this confusion.

Historically, the two main criteria used to distinguish strains of mouse-passaged scrapie were (1) the ranking of the incubation periods they produce in mice of the three Sinc genotypesem—s7s7, s7p7; and p7p7; and (2) the severity and location of vacuolar degeneration induced in the brains of terminal cases of disease (Fraser 1976; Bruce et al. 1991). Different alleles of the PrP gene are linked to the susceptibility and disease incubation period of an animal naturally or experimentally exposed to prions (Westaway et al. 1987; Goldmann et al. 1990; Hunter et al. 1996) and their relative effects can change depending on the prion type or strain (Goldmann et al. 1994). The two mouse alleles, Prn-i^a and Prn-i^b, encode prion proteins differing in two amino acids (codons L108V and T183V) and are congruent to the s7 and p7 alleles of Sinc, respectively (Westaway et al. 1987; Moore et al. 1998). Ablation of the gene in transgenic mice is non-lethal and effectively protects mice from infection with prions (Bueler et al. 1992), and survival time, histo-pathology and the molecular characteristics of disease in mice (and other experimental models) infected with prions can be manipulated by changing the allotype, location and level of prion protein expression in exposed animals (Prusiner et al. 1990; Manson et al. 1994; Telling et al. 1996).

In vitro formation of PrP^Sc from PrP^C has been shown in infected cell cultures (see below) and in a cell-free system where the conversion is driven by addition of PrP^Sc template (Kocisko et al. 1994). The cell-free system mimics several aspects of the in vivo disease, including species and strain specificities (Bessen et al. 1995; Kocisko et al. 1995; Raymond et al. 1997). From these test-tube studies, two distinct models for the formation of PrP^Sc have evolved: in both, exogenous PrP^Sc forms catalytic heterodimers with PrP^C which results in the formation of more PrP^Sc; in one, these ‘heterodimers’ are real (Kaneko et al. 1995) while in the other they actually represent the growing face of a PrP^Sc fibril or aggregate (Caughey et al. 1995). This latter, ‘seeded’ polymerization model fits better with the kinetics of in vitro conversion PrP^C to protease-resistant PrP. This mechanism of conversion resembles a crystallization process in that it is rate-limited by nucleus formation and accelerated by seeding (Caughey et al. 1995).

Early studies claimed over-expression of mutant forms of human PrP (P101L-PrP) induces spontaneous brain degeneration, accumulation of protease-sensitive abnormal prion protein and apparent de novo synthesis of infectious particles (Hsiao et al. 1990, 1994); these findings were met with initial scepticism but have been reinforced by similar effects in Tg mice using different constructs (MoPrP170N, 174T (Sigurdson et al. 2009) and the de novo formation of PrP amyloid and infectious particles in vitro (see below, Barria et al. 2009) or re-folding from bacterial, recombinant PrP (Legname et al. 2004, 2005; Bocharova et al. 2006; Makarava et al. 2006, 2009, 2010; Benetti and Legname 2009).

Soto and colleagues have developed a protein mis-folding cyclic amplification technique (PMCA) where brain homogenates from normal and prion-infected animals are incubated together with regular sonication to create new ‘nuclei’ and promote an exponential increase in PrP^Sc; this has drastically improved the efficiency of the cell-free conversion system and animal bioassays have confirmed the serial propagation of prion infectivity and strains in vitro (Saborio et al. 2001; Castilla et al. 2005, 2008; Green et al. 2008), and the de novo production of prions with novel biological properties (Barria et al. 2009).

Although the structures and pathway of conversion between PrP^C and PrP^Sc have yet to be worked out, the atomic coordinates of a soluble, independent folding domain of the protein (residues 121–230) were defined by nuclear magnetic resonance spectroscopy more than 15 years ago (Riek et al. 1996). The prion-protein ‘fold’, a C-terminal, three-helix bundle with short, interacting β-strand segments, is highly conserved across mammalian species (Wuthrich and Riek 2001; Lysek et al. 2005) although the N-terminal half of the molecule (residues 23–120) containing the metal-ion binding histidine-repeat domain has a flexible, more dynamic structure. There are PrP genes and paralogues (Doppel, Dpl, and Shadoo, Sho) in most vertebrates including fish (Watts and Westaway 2007) and evolutionary descent of these mammalian prion proteins from the ZIP family of metal-ion transporters has been inferred by structural comparisons and ‘interactome’ analysis (Schmitt-Ulms et al. 2009; Watts et al. 2009).

Knowledge of the full structure of PrP^C and PrP^Sc may help the design of chemicals engineered to prevent the conversion process and so help predict transmission between species and limit the effects of these diseases. Early comparisons of the primary sequences of PrP from different species failed to aid the prediction of whether or not a particular source or new strain of TSE will transmit from one species to another (Goldmann et al. 1996; Krakauer et al. 1996). The design and construction of multiple ‘synthetic’ strains by re-folding and annealing the same sequence of recombinant prion protein may lead to better insight into the structural determinants of infectivity and cross-species transmission (Colby et al. 2009).

High titres of infectivity are recovered in preparations of membranes purified from TSE-affected brain and other tissues; this infectivity is not significantly reduced by disruption of membranes using deoxycholate (DOC), sarcosinate (Sarkosyl), and other mild detergents, and can be concentrated and pelleted by differential centrifugation. This infectious material is heterogeneous in size and physical properties and this hindered its biophysical characterization. Viewed by electron microscopy, these highly enriched fractions of infectivity contain fibrils of various shapes and sizes as well as ferritin particles and amorphous material. Surprisingly, these fractions are homogeneous biochemically. One isolate of hamster scrapie (263K) survived prolonged treatment with high concentrations of proteinase K and contained little else but a 27–30 000 M  _r protein; this was the prion protein (PrP27-30) (McKinley et al. 1983).

Early investigations reported a stoichiometry of 100,000 molecules of PrP^Sc per infectious particle using rodent models of these diseases (Scott et al. 1991)and the particle-to-infectivity ratio of brain-derived prions has been further refined and estimated at 3,000 (Safar et al. 2005). The fibrils (scrapie-associated fibrils or rods) are aggregates of PrP^Sc and provide a morphological marker of infection/disease; in some models, they can be visualized in tissue by thin-section electron microscopy and shown to be composed of PrP by immunogold staining (Jeffrey et al. 1992; Jeffrey et al. 1994). In retrospect, these sparingly soluble fibrils of a normal cellular protein (PrP^c) had been seen as plaque-like deposits of amyloid in human (kuru) and mouse (experimental scrapie) brain several years before their biochemical characterization (Beck and Daniel 1979).

PrP^c is a phosphoinositol-glycolipid-anchored membrane glycoprotein found in brain and, to a lesser extent, other tissues. The primary structure of the PrP is virtually constant in mammalian species (Schatzl et al. 1995), including man, mouse, and cow and there are homologues in fish, birds and other phyla (Premzl and Gamulin 2007). In mammals, it is a glycoprotein of 33–35 000 Da which is anchored to the cell plasma membrane by a phosphatidyl-inositol glycolipid attached to its carboxy-terminal amino acid (Stahl et al. 1987). The hamster protein (PrP^c and PrP^Sc) has 208 amino acids (PrP^23–231) and the normal isoform is completely degraded by proteases under conditions which leave a 27–30 000 Da, protease-resistant core of the PrP^sc isoform intact (PrP^27–30) (Bolton et al. 1982; McKinley et al. 1983; Hope et al. 1986). PrP^81–230 is equivalent to the proteinase-K-resistant core of mouse PrP^Sc (Hope et al. 1988) and its expression in transgenic mice has been shown to be sufficient to support replication of infectivity and the development of disease (Fischer et al. 1996).

All TSEs are characterized by the accumulation of an abnormal form of the prion protein (PrP^Sc) in brain and peripheral tissues and this has become a biochemical marker for disease and the infectious agent (Hope et al. 1986, 1988). Different isolates from sheep and other species are more susceptible to proteases than the 263K-PrP^Sc protein and so are sometimes harder to detect; classical types of scrapie, CJD and BSE accumulate forms of PrP^Sc which degrade to a protease-resistant core similar to the PrP27-30 of 263K-scrapie but a novel, atypical scrapie in sheep (Benestad et al. 2003; Klingeborn et al. 2006) and human TSE cases (Gambetti et al. 2008) have been characterized by the accumulation of PrP^Sc with a different spectrum of protease-resistant fragments. While the relative protease-resistance of PrP^Sc in classical TSEs facilitated the characterization of its gene and the key association with disease, it is now clear that abnormal PrP with a range of stabilities is found in tissues of both classical and atypical forms of disease and this heterogeneity continues to confound attempts to ‘define’ an infectious structure for the protein. Protease-sensitive synthetic prion models have been developed to investigate this further (Colby et al. 2010).

The physical heterogeneity of prions has been one of the major drawbacks preventing their molecular characterization. During fractionation of tissue homogenates of scrapie-infected brain by rate-zonal density gradient centrifugation, infectivity ranges in size from 40S to > 500S (Prusiner et al. 1978). This behaviour is probably due to its association with the PrP protein. The range in size observed may be due to the interaction of this protein with various membrane fragments and cell debris. The use of filters, ultracentrifugation, gel filtration, and other sizing techniques to define the minimum size of the pathogen have been hindered by this problem. Most recently, light scattering and non-denaturing gel electrophoresis and size fractionation of the prion form of PrP (PrP^Sc) using detergents and proteases to minimize lipid and PrP protein/glycoso-aminoglycan interactions has shown that (with respect to the protease-resistant form of PrP^Sc, PrPres) specific infectivity and ability to convert PrP^C to PrP^Sc  in vitro peaks in association with amorphous, 17-27 nm (300–600 kDa) particles. Much less of these activities were found in large fibrils or small oligomers of less than or equal to five PrP molecules (Silveira et al. 2005).

Irradiation techniques using high-energy ionizing particles are unaffected by the purity of the infectious particles and have consistently given estimates of less than 200 kDa for the target size of the replicating particle (reviewed by Alper 1997), although these estimates are disputed (Rohwer 1991). Ultra-violet light irradiation at 254 nm reduced 1000-fold the titre of purified fractions containing ∼3000 PrP^Sc molecules per ID₅₀ unit and has supported historical evidence in favour of an unconventional structure for the pathogen containing no nucleic acid genome greater than 25 bases (Safar et al. 2005). Detergent lysis, membrane filtration and mouse titration using a mouse-passaged isolate of CJD have been used to define a maximum size of this agent at less than 25 nm (Tateishi et al. 2001). These data have practical implications particularly for the production of pathogen-free, human blood plasma fractions for human therapeutic use; prions have been detected by bioassay in filtrate and retentate after nanofiltration of plasma spiked with prions using 15 nm filters (Yunoki et al. 2010).

Physical and chemical treatments of infectious fractions can alter the pathogenesis of infection in the animals used in the quantal titration of infectivity, and these complex effects are often neglected when fractionation/titre data is analysed. For example, boiling reduces the titre of inocula as measured by an intracerebral route but not by peripheral inoculation (Dickinson and Fraser 1969) and removal of water or lipid prior to or during heating can radically affect the extent and rate of thermal inactivation (Gale 2006; Fernie et al. 2007; Gale 2007; Muller et al. 2007). The cellular and molecular basis of these effects on titre are poorly understood and hence much of our knowledge on inactivation is anecdotal and difficult to interpret in terms of molecular structure. To complicate the situation still further, it appears that different strains of scrapie and other TSEs may have different relative resistance to thermal (and possibly other forms of) inactivation (Somerville et al. 2002). Critically, the stability of prions in one species may change on transmission to another and, ideally, prion inactivation procedures should be validated by bioassay using the most relevant strain/host model (Giles et al. 2008); of course, this is impractical for strains of BSE and CJD and experimentation using tissues from natural cases of these diseases and bioassay in transgenic mice expressing human or bovine PrP transgenes has been used to infer stability in these cases.

In general, prions are only completely inactivated by oxidizing agents (2% available chlorine, bleach), high alkalinity (1–2 M sodium hydroxide), extremes of temperature (138°G for 20 min) and other conditions which destroy proteins and their biological activities (Taylor 1993). These chemicals and treatments are corrosive to surgical instruments and difficult to apply routinely in laboratories, hospitals and on farm. In the wake of the BSE and vCJD epidemics much work has been done to improve their usefulness and to optimize inactivation methods incorporating detergents (Race and Raymond 2004; Peretz et al. 2006), organic solvents (Beekes et al. 2009, 2010), high and low pH (Appel et al. 2006; Lemmer et al. 2008; Murphy et al. 2009), bacterial proteases (Langeveld et al. 2003; Yoshioka et al. 2007; Dickinson et al. 2009), metal dioxides (Paspaltsis et al. 2009; Russo et al. 2009), radio-frequency plasmas (Baxter et al. 2005) and peroxides (Johnson et al. 2009; Lehmann et al. 2009) or combinations of these priono-cides.

The life cycle of an infectious agent can be defined at three levels, its mechanism of transmission within and between populations; at the level of the organism, where one needs to understand route of entry, spread, replication and shedding from an individual; this also includes its pathogenesis; and, thirdly, for intracellular parasites such as viruses it is beneficial to understand how the agent hijacks the cell and uses its biosynthetic machinery to reproduce. These three cycles of a typical TSE-like infection are documented in this section.

One of the main problems of understanding the epidemiology of human TSE disease is to explain its low incidence, an incidence which appears incompatible with a sustainable infection within the population. There are two main views on this dichotomy: that the disease is not infectious but arises de novo in each individual as the result of a somatic or germline mutation in the prion protein gene; the other stresses our lack of knowledge of the prevalence of infection rather than the incidence of clinical disease and proposes that there is a widespread inapparent infection of the population by a benign agent and only in certain genetically susceptible individuals or by its mutation to a pathogenic form will this ubiquitous agent produce disease. In either case, the predominant form of natural transmission is predicted to be vertical in accordance with field observation in man.

In classical scrapie of sheep, there is evidence for both vertical and horizontal transmission of disease. Maternal transmission of the infection from ewe to offspring either in utero or immediately after birth is thought to be the major route of propagation of the disease within a flock, but lateral transmission is also documented (Dickinson et al. 1974). Factors associated with the horizontal spread of infections, such as host susceptibility, source, and route of infection have been investigated for many years in sheep and rodent models of disease. The low incidence of clinical disease in affected flocks has been interpreted to mean the agent is not highly contagious but this historical view takes no account of the variation in genetic susceptibility of flockmates. Medium levels of infectivity in placentae, foetal membranes and amniotic fluids may contaminate pasture or pens and surrounds where lambing occurs and persist for long periods (Pattison et al. 1972; Andreoletti et al. 2000, 2002). Dissemination of prions by ewes or female goats to their offspring by suckling colostrum and milk, or by sharing these secretions within and between flocks and herds is now recognized (Konold et al. 2008; Maddison et al. 2009) while very low levels of prions have been detected in faeces, saliva and urine by PMCA techniques. Of the common routes of entry (or re-entry) of pathogens into the body—ingestion, inhalation, contact, and coitus—the natural route for prions is probably via the mouth or skin abrasions. In general, the timing, scale, speed of movement of prions about the body and shedding to the environment depends on their strain characteristics and the genetics of the host; for example, experimental BSE appears to have the same general dynamics and spread in small ruminants as scrapie, but little is yet know about the pathogenesis of atypical scrapie and some natural forms of the disease (eg., chronic wasting disease of mule deer and elk) appear to be highly contagious and easily shed from an affected animal (Haley et al. 2009; Mathiason et al. 2009; Tamguney et al. 2009).

Infectivity has been detected in the eyes and lungs of natural cases of disease and conjunctival instillation of scrapie in mice can produce disease (Scott et al. 1993) but there have been no accounts of experimental aerosol transmission, and infectivity in lungs may be due to secondary transport and infection; however, this emphasizes the need for adequate protection when handling tissues infected with TSE and recommended safety precautions for laboratory workers include the use of face masks, avoidance of aerosols, and eye protection.

The presence of PrP^c in an organism appears to be a necessary, if not sufficient condition for it to be able to replicate prions. Although the PrP gene is expressed in many embryonic and adult mouse tissues (Manson et al. 1992), its deletion from the mouse genome does not appear to affect normal development, behaviour, and fertility. Post-natal deletion of the PrP gene using the Cre-Lox conditional expression system in transgenic mice produces only minor phenotypic deficits (Mallucci et al. 2002). In adult animals, the protein is found at its highest levels in neurones of the brain and spinal cord, a lower levels in glial cells and cells of the immune system (Moser et al. 1995; Ford et al. 2002). Animals homozygous (0/0) and heterozygous (0/+) for this mutation can be inbred to produce stable lines of mice and appear normal (Bueler et al. 1992), and many experiments show that expression of the PrP gene is a prerequisite for replication and development of disease in the mouse (Bueler et al. 1993; Manson et al. 1994; Sailer et al. 1994). This has stimulated the production of PrP (0/0) domestic livestock such as sheep (Denning et al. 2001), goats (Zhu et al. 2009) and cattle (Richt et al. 2007) as a strategy for the elimination of prions from the feed and food chain.

In vitro replication also appears dependent on the expression in cells of PrP^C protein and, in the absence of other markers for infectivity, the site of conversion of PrP^C to PrP^Sc has been studied to elucidate where in the cell replication is taking place. Most PrP^C is localized on the cell surface in lipid rafts via a C-terminal, glycosyl-phosphatidyl-inositol (GPI) anchor and its synthesis on ER-attached ribosomes, trafficking (and complex N-glycan addition) in the Golgi body and translocation to the cell lumen and surface mirror that of other membrane and secreted proteins. Early time-course studies in infected neuroblastoma cells indicated that the conversion to PrP^Sc takes place following transit of PrP^c to the cell surface, either at the surface or during constitutive endocytosis via clathrin-coated pits and re-cycling in the endosomal-lysosome system (Caughey and Raymond 1991; Westergard et al. 2007); in vivo PrP (either PrP^c or PrP^Sc) can be observed accumulating at the plasma membrane and in the intercellular space well before neuronal loss, vacuolation, and gliosis (Jeffrey et al. 1996) and may have a direct toxic or mitogenic effect on neighbouring cells. PrP^C may play a role in protecting cells from oxidative stress (Crozet et al. 2008), and has a cytoprotective effect in yeast, mammalian cells and mice (Bounhar et al. 2001; Atarashi et al. 2003; Bounhar et al. 2006). Hence, the observed cellular pathogenesis of prion replication may be the product of a toxic gain of function due to the formation of small oligomers of PrP^Sc (Winklhofer et al. 2008; Wang et al. 2009) coupled to loss of normal cyto-protective activity.

Much further work needs to be done to understand these cellular interactions in the brain and peripheral tissues, not least because their understanding should guide the development of therapeutics. In recent years, research to develop an effective treatment for humans affected by vCJD has increased enormously and PrP and its known ligands have been the targets for drug discovery programmes based on in-silico modelling (Ouidja et al. 2007; Larramendy-Gozalo et al. 2007; Ludewigs et al. 2007; Vana et al. 2009), in-vitro binding of small molecules (Schroder and Muller 2002; Gayrard et al. 2005; Klingenstein et al. 2006; Rhie et al. 2003; Sayer et al. 2004) or mAb fragments (Wuertzer et al. 2008), depleting PrP^Sc from infected cell cultures (Nordstrom et al. 2005; Cordes et al. 2007; Cronier et al. 2007; Campana et al. 2009), or prolonging the survival time of animals experimentally-infected with TSEs (Fai Mok et al. 2006; Kawasaki et al. 2007; Doh-ura et al. 2004). However, reducing PrP^Sc levels does not necessarily lead to better survival times (Zuber et al. 2008) and few animal experiments have shown drug efficacy unless the treatment starts before or at the time of infection. Nevertheless, clinical studies in humans are currently in progress (Collinge et al. 2009; Stewart et al. 2008; Tsuboi et al. 2009) and the identification of PrP^C-specific pharmaceuticals may have a broader utility in the treatment of human neurodegenerative diseases; PrP^C has recently been shown to be a mediator of (Alzheimer’s) amyloid-β-oligomer-induced synaptic dysfunction (Lauren et al. 2009).