Volume 2: Science
2. The Spongiform Encephalopathies - knowledge existing in 1986
Nature of the TSE agent and mode of replication
Virus hypothesis
Nemavirus hypothesis
Virino and replication site hypothesis
Prion hypothesis
Clarification of the pre-1986 theories by post-1986 studies
Summary

2.49 Understanding the nature of the infectious agent and its mode of replication was fundamental to the question of how TSE diseases spread, and how they could be prevented or treated. Most of the work directed at the investigation of the agent had examined scrapie, and had provided contrasting results, suggesting both viral and non-viral properties. The following section describes the major theories that had been proposed by 1986 for the nature and mode of replication of the agent, as well as the evidence for and against them.

Virus hypothesis

2.50 In 1954 Sigurdsson demonstrated that scrapie could be transmitted to sheep from scrapie-brain material that had been passed through filters designed to remove bacteria. This observation led him to speculate that the causative agent of the disease was a filter-passing virus. ¹ Given the prolonged latent period and progressive clinical course of the disease, Sigurdsson suggested that scrapie belonged to the family of 'slow unconventional viruses'. These slow unconventional viruses were later shown to share many properties typical of conventional viruses: ²

1. they were found in filtration studies to be of a comparable size to conventional viruses;
2. they replicated to high titres (concentrations) in the brain following initial replication in the spleen and elsewhere in the body; and
3. they showed a specificity of host range, ie, only affected certain species. Adaptation to a new host was marked by a shortened incubation period.

2.51 However, there was also a significant amount of contrary evidence suggesting that the transmissible agents were non-viral, and that replication of the agent was not mediated by nucleic acid, the only known mechanism for replication. Electron microscopy studies had failed to identify virus-derived structures in brain sections; the agent had been shown to be resistant to treatments known to inactivate viruses and degrade nucleic acid; and infection was not accompanied by an inflammatory response.

Inactivation studies

2.52 The first evidence of the resistance of the scrapie agent to inactivation was provided by Gordon in 1946, when he observed the development of scrapie in over 1,000 sheep that had been vaccinated against louping-ill disease. ³ The vaccine had been prepared from ovine brain, spinal cord and spleen tissue, and treated with formalin, at a concentration of 0.35 per cent, to destroy viruses. On the basis of the virus hypothesis, the resistance of the scrapie agent to formalin was unexpected.

2.53 The scrapie agent was shown by various workers in the late 1950s and early 1960s to be resistant to heating to 100°C, treatment with chloroform, phenol and acetylethyleneimine, and extraction with ether. ⁴ These observations led Pattison in 1965 to conclude that if the transmissible agent of scrapie was a living virus, it was a virus of a kind that was thus far unrecognised, since it could withstand treatments commonly used to disinfect virus-contaminated materials. ⁵

2.54 In 1966 Alper and co-workers investigated the effect of ionising radiation and ultraviolet light on the scrapie agent. ⁶ They concluded that the agent was likely to be of an unusual nature, and if associated nucleic acid were present, it was implausibly small for a viral genome. Further work by Alper and co-workers in 1978 showed that the biological activity of the agent was reduced when exposed to ionising radiation in the presence of oxygen. ⁷ Had the agent contained nucleic acid, loss of activity would not have been observed because it was known that under the conditions used, oxygen protects nucleic acid from degradation. These studies thus suggested that the agent was composed of lipids, polysaccharides and proteins.

2.55 However, in 1984 Rohwer presented data that countered the findings described above and apparently upheld a viral nature for the transmissible encephalopathy agents. Almost complete inactivation of the scrapie agent was demonstrated following treatment of scrapie-affected hamster brain extract with strong disinfectants (ie, 0.525 per cent hypochlorite or bleach or 0.01 per cent molar sodium metaperiodate). ⁸ Further work showed that the agent was sensitive to temperatures of 100ºC or more, in a manner thought to be consistent with the behaviour of conventional viruses. ⁹ Rohwer explained the apparent stability of the agent to heat, as observed by Stamp and co-workers in 1959, by reference to the existence of a small resistant subpopulation of the agent remaining active after treatment. It was concluded that the resistance of the scrapie agent to sterilisation was limited to a small subpopulation of total infectivity, the majority of the population being highly sensitive to inactivation. In Rohwer's view, the results added further evidence that the scrapie agent was a small virus with conventional sensitivities to heat and numerous chemicals.

Lack of immune response

2.56 Further evidence to suggest that the transmissible agents were not viruses had been provided by the absence of an inflammatory (ie, immune) response in the host animal. The inability to demonstrate an immune response to scrapie infection was recognised in 1959; this feature was found to be characteristic of other infections such as kuru, CJD and TME. ¹⁰ In contrast, other neurological disorders, for example experimental allergic encephalomyelitis (EAE), which is characterised by the demyelination of nerve fibres (ie, the destruction of myelin protein from nerve fibres) and chronic inflammation, did provoke an immune response. ¹¹

2.57 Reasonable explanations for the failure to elicit an immune response in scrapie infection had been proposed by various workers since 1959. ¹² For example, tests used to measure antibodies were not sensitive enough or were inappropriate; antibodies were measured at the incorrect time during the course of infection; and the animals used in the studies were not suitable. A study in 1973 by Porter and co-workers which used a very sensitive test (indirect immunofluorescence) similarly failed to detect an immune response, again suggesting that such a response was not generated following scrapie infection.

2.58 In 1982 Kasper and co-workers, who had failed in their attempt to raise antibodies to the scrapie agent in rabbits, added further weight to these findings. ¹³ They considered the possibility that the scrapie agent was capable of producing an immune response, but that the antibodies produced could not be detected by the tests used, for example through cross-reaction of the antibody with normal cellular constituents. It was, however, also possible that the scrapie agent was sufficiently similar to a normal cellular component to make the host tolerant to its presence.

Nemavirus hypothesis

2.59 Although this body of evidence suggests an absence of nucleic acid in the infective agent of scrapie, we should mention a theory advanced after 1986 asserting that nucleic acid is involved. Using electron microscopy, Narang identified structures containing single-stranded (ss-) DNA in scrapie-infected brain. ¹⁴ These studies were performed in 1990, but are relevant to this discussion since they demonstrate the persistence of the theory that nucleic acid is intrinsic to TSE infectivity. The fact that scrapie infectivity was resistant to inactivation by enzymes that specifically degrade nucleic acid was consistent with Narang's view that the scrapie-associated DNA was protected by a protein coat.

2.60 In 1992 Narang coined the term 'nemavirus' to describe these structures. ¹⁵ The nemavirus particle was postulated to be an unusual three-layer structure in which an outer layer of protein surrounds the central layer of ss-DNA, which in turn is coiled around a core of prion protein/scrapie-associated fibrils (SAFs). ¹⁶ It was proposed that the DNA sequence might code for an 'accessory' protein, perhaps acting as an enzyme which facilitates the conversion of the normal, cellular prion protein to the abnormal, disease-inducing form (see paragraphs 2.66-2.70).

2.61 Later work by Narang in 1998 suggested that experimental animals inoculated with ss-DNA derived from scrapie-affected animals developed vacuolation in their brains when also injected with chemicals designed to facilitate the uptake of DNA into cells. Narang concluded therefore that the ss-DNA associated with the nemavirus particle was the genome, or informational molecule, of the scrapie agent. ¹⁷ Independent attempts to identify the ss-DNA have so far been unsuccessful. ¹⁸

Virino and replication site hypothesis

2.62 In 1971 Dickinson and Meikle provided a hypothesis for the replication of the scrapie agent based on the discovery of a single autosomal gene controlling the scrapie incubation period in mice - the gene sinc - and on observations about strains of scrapie agent. ¹⁹ As discussed later (paragraphs 2.84-2.89), the sinc gene was shown to possess two alleles, ²⁰ s7 and p7, associated with short and prolonged incubation periods. ²¹

2.63 This hypothesis proposed that the gene products of each sinc allele contributed to a multimeric protein structure, which then formed a 'replication site' for the scrapie agent. The replication of the agent would depend on how the particular strain interacted with the replication site and what the site was composed of: ie, on whether it was homomeric (composed only of the product from one allele) or heteromeric (composed of the products from two different alleles).

2.64 The fact that different strains of scrapie were known had suggested that the agent was similar to conventional viruses in that it carried a genome composed of nucleic acid. Thus variants could arise during incubation, giving rise to new strains. No host-encoded properties were found to determine scrapie agent strain differences. It was thought that this showed that the genome of the agent could vary independently and, although replicated by normal host mechanisms, was not coded by the host. ²²

2.65 The term 'virino' was coined to reflect the small size, immunological neutrality and virus-like nature of the infectious particles. Thus, in the virino model proposed by Dickinson and Outram in 1979, ²³ the life cycle of the scrapie agent included a stage where the genome was bound to host protein, probably a multimeric protein complex, derived from the sinc gene. In the virino model, the host protein protects the scrapie agent nucleic acid from degradation and prevents the host raising an immune response, since the protein/nucleic acid complex is seen as 'self'. However, as recounted in paragraph 2.54 above, scrapie-associated nucleic acid has not been identified and physical or chemical evidence for its presence is lacking.

Prion hypothesis

2.66 As described above, many experiments had suggested that the scrapie infective agent did not contain nucleic acid and, in 1967, Griffith proposed that the causative agent could be a self-replicating protein. ²⁴ Several mechanisms for agent replication were suggested that challenged the established view that nucleic acid was essential. The basis of Griffith's proposal was that the scrapie agent was a protein which the host animal was genetically equipped to make but which it either did not normally make or did not make in an infectious form.

2.67 The premise that Griffith put forward in 1967 was built upon by Prusiner who published his prion hypothesis in 1982. ²⁵ The term prion was introduced to denote small proteinaceous infectious particles, in response to an accumulation of data showing that the scrapie agent contained protein but was resistant to inactivation by most procedures that modify nucleic acid.

2.68 By 1983 Prusiner had isolated the prion protein from infected hamster brains which had been enriched for scrapie infectivity. ²⁶ Under the electron microscope it was found to aggregate into abnormal fibrillary structures, apparently identical to the 'scrapie-associated fibrils' (SAFs) which had been identified in 1981 byMerz. ²⁷ Merz had shown that SAFs were present in treated homogenates of scrapie-affected brains; they resembled amyloid and were thought to be responsible for the amyloid plaques found in the brains of affected animals. Thus Prusiner proposed that deposition of prion protein in the brain caused the damage characteristic of scrapie.

2.69 Later work by Prusiner in 1985 showed that prion protein was a component of the normal cell and was encoded by the host genome. ²⁸ This was revealed by identifying the sequence of amino acids in fragments (peptides) of the hamster prion protein. The DNA code of these fragments was predicted by the amino acid sequence, which enabled the identification and isolation of the hamster gene. This in turn allowed the identification and sequencing of the mouse and human genes. The amino acid sequence of the scrapie-associated prion protein (PrP^Sc) was found to be the same as the normal prion protein (PrP^C). Prusiner showed that PrP^Sc was partially resistant to digestion by protease enzymes, while PrP^C showed no resistance to enzyme digestion. He postulated that the protease resistance was due to a modification of the three-dimensional structure of the prion protein, and that this conformational change had the ability to convert normal PrP^C into PrP^Sc by a form of chain reaction. The abbreviation PrP^sc came to be used for the abnormal isoform of the prion protein in all TSEs, not just scrapie.

2.70 The conversion provided a mechanism for disease transmission and accounted satisfactorily for the long incubation time and the natural history of the disease. The hypothesis explained why infection with PrP^Sc did not evoke an immune reaction, as the infective agent was not foreign to the host but the product of the host's own prion gene.

Clarification of the pre-1986 theories by post-1986 studies

2.71 Since these findings were made, much more has been learned about prion protein, and it has become widely accepted that the infective agent in TSEs is the protease resistant isoform of prion protein. Although prion protein is present in cell membranes and other cellular components, its function remains unclear (see paragraph 2.77). Computer modelling suggests that the structure of normal prion protein is characterised by alpha helix motifs (Figure 2.1A), whereas the disease isoform is characterised by beta sheet formation (Figure 2.1B). This is the change in conformation that is associated with protease resistance and the accumulation of insoluble aggregates of prion protein. The aggregates comprise the amyloid deposits in the diseased brain and give rise to the SAFs found in treated brain homogenates.

Figure 2.1: Proposed three-dimensional structure (a) PrP^C and (b) PrP^Sc

2.72 The discovery that mutations of the prion protein gene were the cause of GSS and familial CJD provided very strong evidence that PrP^Sc was the infective agent in TSEs (see paragraph 2.176). Further confirmation came in 1994, when the disease was reproduced experimentally in transgenic mice in which the normal prion gene was absent and in which copies of a synthetic gene containing the same mutation that caused familial GSS (P102L) had been inserted. ²⁹ It was already known that the same transgenic mice in which the normal PrP gene was absent, but not replaced by the mutant gene, did not develop a TSE and were not susceptible to experimental TSE. ³⁰ The susceptibility to infection of these mice was restored by replacing the normal mouse PrP gene. Recent evidence obtained in 1999 suggests that the PrP mutation (101L) introduced into mice did not produce any genetic disease in mice but significantly altered the incubation time of TSE infection. ³¹

2.73 The effect of removing the PrP gene has proved to be variable in the hands of different workers. Two independently generated strains of mice have had a normal lifespan and no apparent ill-effects from deletion of the gene, apart from a slight alteration in their circadian rhythms (biological processes or activities set by 'biological clocks') or electrophysiological abnormalities. However, two other gene-deleted (knockout) strains have shown neurodegeneration and fatal ataxia, suggesting that PrP may be involved in long-term survival of certain nerve cell types. The reasons for the discrepancies in the mice strains are unclear, though may be linked to the recent identification of a novel PrP-like gene, doppel, located very near to the PrP gene. ³² The doppel gene has been found to encode a protein that is 25 per cent homologous to all known prion proteins, ³³ and is expressed in higher amounts in the ataxic knockout mice, but not in the knockout mice which do not develop ataxia. This suggests that doppel may provoke neurodegeneration in PrP-deficient mice. It seems that the interaction between the two genes may be important in the development of TSEs, though the functions of both PrP and doppel remain to be determined.

2.74 It is evident that PrP is necessary for disease, though whether it is sufficient on its own for disease is unclear. The mechanisms by which prions cause disease are equally unclear. The 'protein-only' hypothesis states that prion replication occurs by PrP^Sc recruiting PrP^C and converting it into further PrP^Sc. This results in a chain reaction of events whereby PrP^Sc accumulates increasingly quickly. Indeed, it has recently been shown that interaction between PrP^C and PrP^Sc is necessary for the conversion of PrP^C, though it is not sufficient on its own. ³⁴ Moreover, this interaction takes place at a localised site, near to or including the C-terminus of the protein (ie, the right-hand end of the protein when considering it as a chain of amino acids) and has been found to be highly specific. The findings support the idea that PrP^Sc may serve as a critical ligand ³⁵ and/or receptor for PrP^C in the course of PrP^Sc propagation and pathogenesis in living animals. ³⁶

2.75 However, how the conversion of PrP^C to PrP^Sc actually occurs is currently the subject of much debate. One mechanism, a 'template-directed refolding' model, has been proposed by Prusiner. It postulates that a physical interaction between PrP^C and PrP^Sc is prerequisite to the structural changes which underlie the conversion of PrP^C. Genetic evidence has suggested the participation of a further protein (dubbed protein X), which acts as a molecular chaperone to facilitate misfolding of PrP^C. ³⁷ Indeed, the identification of the prion-like gene, doppel, has raised the possibility that protein X and the protein generated by doppel are one and the same thing.

2.76 An alternative mechanism has been proposed whereby PrP^C and PrP^Sc exist in thermodynamic equilibrium in the cell, either in the cytoplasm or in specialised cell compartments where PrP^C can be found. ³⁸ According to this 'seeding' hypothesis, monomeric PrP^Sc is a normal constituent of the cell, the infectious agent being a multimeric, highly ordered aggregate of PrP^Sc. Only if several monomeric PrP^Sc molecules are assembled into a highly ordered seed can further monomeric PrP^Sc be recruited from the surroundings and aggregate into amyloid. The likelihood of spontaneous formation of a seed is a function of the local PrP^Sc concentration. Infectivity would increase when the PrP^Sc aggregate becomes so large that it splits into smaller nuclei, each one being capable of recruiting PrP^Sc and thereby acting as an infectious unit. However, it has thus far not been possible to distinguish which of these models, if either, is correct.

2.77 While there is considerable speculation about the conversion of PrP^C to PrP^Sc, there is also debate as to the normal function of PrP^C in cells. Recent work has focused on the finding that PrP^C is able to bind copper ions in vitro, and that PrP^C could exist in a copper-bound state in vivo. ³⁹ Findings suggest that these copper binding properties of PrP^C may be important in the protection of cells from 'oxidative stress', that is, the destructive effect of highly charged, toxic oxygen free radicals, which are produced in the body in a number of normal biochemical reactions. It appears that PrP^C may influence the activity of a copper/zinc superoxide dismutase enzyme, which mops up and deactivates harmful free oxygen radicals. Indeed, recent evidence suggests that PrP^C itself has superoxide dismutase activity. ⁴⁰ Thus it is possible that the toxicity of PrP^Sc may be tightly associated with the loss of activity of PrP^C in the cell.

2.78 Soto and colleagues have recently reported a most compelling experimental result which provides strong evidence that PrP^Sc is the infective agent in scrapie. ⁴¹ They found that treatment of protease-resistant PrP^Sc with synthetic beta-sheet breaker peptides (molecules which disrupt the -sheet structure of proteins) could reverse the infectivity of PrP^Sc by 90 to 95 per cent in mouse bioassays. A significant decrease in protease resistance of PrP^Sc to a state similar to PrP^C accompanied this reversal of infectivity. This is the first time that reversal of infectivity has been achieved, and the significance of this observation for developing a treatment for TSEs is self-evident. (This is discussed further in paragraph 5.48. Its relevance here is that it confirms almost incontrovertibly the validity of the prion hypothesis.)

Summary

2.79 In the late 1960s there was powerful evidence that the agent responsible for scrapie was resistant to treatments which inactivated all known micro-organisms including viruses. Heat levels short of causing protein degradation, exposure to ultraviolet light and ionising radiation, treatment with formaldehyde and with enzymes which degraded nucleic acids, all failed to prevent successful transmission of infection to experimental animals. Although there was some conflicting evidence, filters which prevented the passage of the smallest viruses failed to retain the scrapie agent. Scientists were forced to consider other hypotheses, and several ingenious ideas were postulated, such as the 'virino' hypothesis, in which the infective agent's nucleic acid core was protected from inactivation by being wrapped inside an immunologically neutral protein coat that somehow did not evoke an immune reaction in the host. Griffith's earlier concept of a self-replicating protein as the infective agent was reconsidered, and from this Prusiner developed his prion hypothesis. In 1986 the prion hypothesis was still regarded as highly controversial. The turning-point was to come in 1989 with the demonstration of a mutation of the prion gene in patients with GSS (paragraph 2.176). Experimental mutation of the prion gene in transgenic mice provided confirmation. Since then there has been mounting experimental evidence in favour of the hypothesis, although certain aspects remain unexplained, including the mechanism by which the infective agent (PrP^Sc) converts the host's normal prion protein into the protease-resistant, disease-producing conformation. However, the development of beta-sheet breaker peptide, and the demonstration that inoculating the scrapie agent with this peptide could reverse scrapie infectivity, provide convincing evidence in favour of the prion hypothesis.