1 The Paradox of Prion Disease Rosalind M. Ridley and Harry F. Baker 1. introduction The study of that group of diseases now collectively known as the prion diseases has always been a source of excitement and argument between scien- tists. These obscure diseases usually have been of extremely rare occurrence and have had little impact on the general public. When an epidemic occurs, however, as in the case of bovme sponglform encephalopathy (BSE) m the United Kmgdom, the bizarreness of the prion diseases and the profound differ- ence between them and any other “mfectlous” condition can (and did) lead to public consternation. Three features of prlon diseases give them an apparently diabolical quality. 7.1. Their Occurrence Is Largely Unpredictable 1. In acquired casest here 1s a silent mcubatlon period that can vary from a few weeks to up to 40 yr, dependmg on the speciesa nd the circumstances 2 During the incubation period there are no symptoms and no detectable signs, e.g., there IS no measurable Immune response that will predict subsequent illness. 3. No mode of transmission could be established for the majority of human cases that appear to occur “out of the blue.” 1.2. They Exhibit Certain Impossible Properties In relation to established principles of microbiology, the prion diseases exhibit certain “impossible” properties. 1 “Infectious” disease seems to arise spontaneously. 2. Inherited cases give rise to a disease that is transmissible but acquired cases do not produce a heritable disease From Methods m Molecular Medicrne Prron Dmeases Edlted by H Baker and R M Rdley Humana Press Inc. Totowa, NJ 1 2 Rciley and Baker 3. Procedures and chemical treatments that specifically destroy nucleic acids (which all replicatmg organisms possess) do not destroy the mfecttous agent 4 The mfecttous agent can persist m the envtronment almost indetimtely 1.3. They Are Invariably Fatal 1 No treatments have been found that ~111 alter the progression of the disease 2 In animal transmtssion studies, the duratton of symptoms from onset to death can be as short as a few hours followmg an mcubation period of many months, during which the animal appears to be completely healthy These three features have combined to stroke fear into the general publtc, who regard priori dtseases( perhaps with some justification) as uncontrollable, mcomprehensible, mdestructable, and incurable. This fear 1s probably only matched by an excessive dread of radioactivity, which IS similarly imbued with great “strangeness.” In sharp contrast, the public steadfastly refuses to worry about “ordmary” factors, such as diet, exercise, and safety, which are known to have an enormous effect on morbidity. 2. Scientific Investigation In recent years the screntific investigation of prron diseases has expanded greatly, leading to strongly held convictions and arguments that have some- times degenerated into acrimony. Scientists are particularly exercised by the fact that some accounts of the pathogenesis of prion disease seem to contradict basic tenets of molecular biology: 1 That proteins are made using information encoded in nucleic acid, 2. That different proteins are encoded by different nucleic acid sequences; 3 That protein synthesis cannot occur without nucleic acid; and 4. That any “agent” that increases m titer must involve rephcatton of nucleic acid. The central argument has been about “viruses vs genes (or proteins).” As in other great scientific debates, e.g., “particles vs waves” or “nature vs nur- ture,” the resolutton of the problem 1su nhkely to be that one vtew is correct and the other incorrect but that a new way of understanding the phenomenon allows the dichotomy to dissolve. For many years the body of data on priori diseases was like a Rubik cube; when one face was intact the others were not. Thus, rt was easy enough on the basis of experimental transmtssion studies to make the case for disease being caused by a virus-except that it was not possible to explain how the majority of cases were “caught.” The transmis- sible agent appeared to behave like a replicating, infectious agent--except that treatments that destroy nucleic acid did not reduce infectivtty. Further- more, some human cases clearly were inherited, whereas other cases clearly were not. The Paradox of Prion Disease 3 2.1. Misinterpretations These difficulties m producing a coherent story for prton diseases were com- pounded by certain mismterpretations of the available evidence. 1 The demonstration of experimental transmissibihty led to the presumption that all cases were acquired by mfectron. 2 The difficulty in demonstrating contact between cases of human prion disease led to exaggerated concern about the mfecttousness of the disease. For example, the apparently higher incidence in cities (probably owmg to sampling and demo- graphic effects) could be interpreted as meaning that casual contact between strangers was all that was necessary to transmit disease 3 The not infrequent occurrence of priori disease m another family member of the proband led to the fear that the disease was transmissible vertically through con- tamination zn utero or through the incorporation of the agent (perhaps a retrovnus) m the genome of an acquired case. Where the disease occurred m a proband and a parent who had been separated for many years the suspicion grew that transmisston of “agent” had occurred m childhood. Where disease occurred m a proband and a distant relative who had met for only the briefest of meetings it was supposed that genetically susceptible people were quite exquisitely vul- nerable to mfection 4 The transmtssibility of the disease and the apparent rnformatron-carrying prop- erty of the agent, as demonstrated by the phenomenon of strain of agent m experimental scrapie, lead to the insistence that failure to find the virus was only a temporary technical failure rather than evidence that there might not be such an independent orgamsm m these diseases The baste phenomenon that the various hypotheses of the etiology of prion diseases have sought to address 1s experimental transmrsstblltty, and advocates for each side draw on a number of facts to support then view. 2.2. Evidence in Favor of the Viral Hypothesis 1. The increase m mfectivity titer in brain during disease progression is taken as evidence of a replicatmg agent 2. No bacteria or other organism can be found so that, by default, the disease is attributed to a vnus. 3. In experimental transmisston the long incubation period followed by the rapid course of disease resembles the occurrence of diseases caused by lentivn-uses. 4 Different isolates of priori disease “agent” appear to behave differently in terms of species specificity, incubation time, and lesion profile on experimental trans- mtssion to other animals leading to the concept of “strain variation ” 5. The existence of different “strains of agent” is regarded as evidence of an infor- mattonal, rephcating molecule that is presumed to be composed of nuclerc acrd 6 The familial occurrence of priori disease m humans is explained as a genetic susceptibility to an infectious agent in the environment. 4 Ridley and Baker 2.3. Evidence in Favor of a Gene (Protein) Hypothesis No vnus particles are associated with infectivity No immune response to mfectron 1s found m an affected host No nucleic acid IS associated specifically with infectivity Infectivity titer is associated with levels of the abnormal isoform of priori protein known as PrPSC PrPSC and the normal form of prton protein (PrPc) do not differ m primary struc- ture, thus negating the need for PrPsc-specific nucleic acid PrPSC has the primary structure dictated by the PrP gene of the host, thus negating the need for agent-specific nucleic acid in the production of PrPSC When priori disease is experimentally transmitted from one species to another the PrPSC m the recipient has the primary structure of the recipient species, not of the donor species or agent Agent replication does not, therefore, need to occur for levels of PrPSC to increase 8 Most (but not all) of the variation m species specificity, mcubation time, and lesion profile can be accounted for by variations m the PrP protein of the host and the homology between that and the PrP protein of the agent (1 e , donor), thus reducing the variation attributable to “stram of agent ” 9 Familial prion disease occurs m pedigrees with a mutation m the PrP gene 10 Transgemc mice carrying mutations m the PrP gene produce a transmissible spongiform encephalopathy despite being kept m controlled conditions where accidental contamination with an exogenous agent can be avoided (I) 11 PrPSC can be reversibly denatured with guamdme-chloride and renatured by dilutmg out the guamdine chloride If radiolabeled PrPC IS added during the rena- turation stage, protemase K digestion results m the formation of radiolabeled proteinase K-resistant fragments of similar size to protemase K-digested PrPSC from scrapie-affected ammals (2) (and see Caughey et al , this volume) This suggests that PrPC can be converted directly mto PrPSC by mteraction with other PrPSC molecules 2.4. Resolution of the Dichotomy A number of arguments are leading to a resolutton of the dichotomy. 2.4.1. Acquired, Familial, and Sporadic Cases It is now recognized that cases of prton disease arise m three different ways-acquired, familial, and sporadic, and that data from one type of case do not have to be applied to other cases. This resolves the argument between “mherited” and “acquired” factors in the etiology of different cases 2 4 1 1. ACQUIRED CASES Acquired cases of prion disease arise because a person or an animal has ingested or absorbed a quantity of the infectious agent. In humans, the majority The Paradox of Prion Disease 5 of these acquired cases occurred in Papua New Guinea m the first half of this century. The diseasew as known as kuru and spread slowly around various related ethnic groups of the mdigenous population. Epidemiological analysis imph- cated funerary practices and the people admitted to cannibalism, especially by the women and the young children m then care (3). It is now supposed that the epidemic began with a sporadic case of prion disease. The epidemic has now almost run its course but the extremely long incubation periods that can occur, especially following low oral exposure to the agent, means that a few cases still occur each year. In the West, prion diseasei n humans has not been associated with any dietary practice, but has occurred following the use of human tissue for medical pur- poses. The largest number of cases (currently approx 100 worldwide) has resulted from the prolonged use of human pituitary-derived growth hormone. A few further cases are associated with the use of human dura mater tissue m reconstructive surgery of the head and a small number of cases have occurred as a result of contammatron during cornea1 grafting or neurosurgery. In the latter cases mstruments were used that had previously been used m patients with priori disease and that probably were still contaminated despite sterihza- non treatments that were standard at that time. The tissues that have been implicated m these iatrogemc cases-pituitary gland, dura mater, and cornea- all come from close to the brain, which is known to have the highest levels of infectivity. The largest epidemic of animal priori disease has affected cows m Britam Bovme spongiform encephalopathy (BSE) was first recogmzed m 1986 and the number of cases has followed a large distribution with the peak (approx 3500 cases/ma) occurrmg m 1993. The total number of cows affected so far is approx 150,000 and the epidemic is expected to run beyond the year 2000. The epidemic is believed to have resulted from changes m the method of rendering carcasses( including ovine and bovine) for the production of animal food pel- lets such that the mfectious agent associatedw ith scrapie was not destroyed and affected the calf population. Subsequently, carcassm aterial from BSE-affected cattle inevitably would have been incorporated m the calf food pellets, adding to the infectivity titer of this food preparation. The number of cases occurring in animals born after the feeding of rummant-derived protein to ruminants was banned m 1988 has dropped dramatically, mdicatmg that, apart from a possible very minor level of residual contamination in the system, the epidemic now is under control. The long incubation period that can occur, however, implies that there will be several years before the disease is eradicated. In addition to BSE, the contamination of animal feed has led to a small number of cases of prion disease m other ammals, notably cats and exotic ungulates, and other food-related outbreaks have occurred in farmed mink and captive deer 6 Ridley and Baker 2.4.1.2 FAMILIAL CASES About 15% of cases of human prion disease occur m families where at least one other family member has been affected and where rt can be established that a large number of cases have occurred in one family, it can be seen that the dtsease runs m an autosomal dommant pattern Linkage of disease to a muta- tion m the PrP gene open reading frame was first estabhshed by Hsrao et al (4). There are now known to be several point mutations and a number of expansions m an octapepttde repeat sequence within the PrP gene open readmg frame (5) that are associated with disease. 2.4 1 3. SPORADIC CASES Sporadic cases of human prton disease occur, by definition, without any known antecedent event through which the disease could have been acquired and without any known family history or mutation m the PrP gene. Some genetic influence ts now being recognized, however, because most sporadic cases are homozygous for a common polymorphism in the PrP gene (6) although, because the disease is exceedingly rare (affecting approx one m a mullion people worldwtde), the vast majortty of people who share this genetic feature are not affected. 2 4.1.4. SCRAPIE: ACQUIRED, FAMILIAL, AND SPORADIC CASES Scrapre in sheep occupres a particularly important position among the prron diseases because it is a relatively common, naturally occurrmg condition that shares features of the acquired, inherited, and sporadic ettologtes. It is thought to be the ortginal source of infection m the BSE epidemic, although this IS difficult to prove. Persistent efforts to find an eptdemtological or case-con- trolled association between scrapie and human priori disease, fortunately, have been unsuccessful. In the 1960s natural scrapie was reported to occur m a pattern consistent with autosomal recessive inheritance m at least some breeds of sheep (7). More recently rt has been shown that m some breeds naturally occurring disease is linked to polymorphisms m the PrP gene that etther are of partial dommance or recessive (8), largely vindicating the original observation despite the many intervening years m whtch the experimental transmisstbility of mouse-adapted scrapre blinded researchers to the possibrlity that the natural sheep disease might wholly be genetic. These PrP polymorphisms also are linked to suscep- tibility to experimental mfection m sheep but it remains unproven whether linkage to naturally occurrmg disease is because of an influence of the poly- morphisms on susceptibility to infection m the field or whether the malortty of naturally occurring cases have a genetic disease, in the absence of any contact with agent m the environment. The more recent demonstration that the latter 1s The Paradox of Prion Disease 7 the case m a proportion of human cases and that transgemc mace carrymg mutations m the PrP gene (1) or additional copies of normal PrP alleles (9) can develop prion disease without contact with an exogenous agent make a purely genetic cause of natural scrapie, in at least some cases,m ore plausible than was believed until recently. The possibtlny that natural scrapie may arise without exogenous infection has, however, been recogmzed by some scientists for many years (10). One of the arguments in favor of the infection hypothesis of the eptdemiology of natural scrapie was the high coincidence of scrapte m dam and progeny that was thought to be indicative of maternal transmission (1 l), but such a pattern clearly is compatible also with a recessive model, especially where the flock shares one ram, such that the variance is determined largely by the status of the dam. Furthermore, although scrapie has been seen m sheep fed with placentas from scrapie-affected sheep (22) this does not prove that oral transmission necessarily determines the natural incidence of disease. Indeed, it is not even clear that the placentas were the source of the disease, since the only sheep that can be guaranteed to be scrapie-free prior to exposure to infected tissue come from flocks that are genetically resistant to experimental infection and that therefore are largely unsuitable as recipients in this type of experiment. The recent demonstration that scrapie develops in genetically sus- ceptible sheep born to unaffected, genetically unsusceptible dams followmg embryo transfer (13) greatly reduces the probable relevance of perinatal events m natural scrapie. In the origmal experiment, however, the embryo donors had been challenged with scrapie, therefore further experiments are required to establish whether this is relevant or whether genetically susceptible animals resultmg from embryo transfer will develop scrapie without exposure of the embryo donor to the agent. The time is clearly right for a reevaluation of theo- ries about the occurrence of natural sheep scrapie in the light of the finding of dtsease-specific polymorphisms in the PrP gene of sheep and other species and the demonstration of pathogenic mechanisms that do not require exposure to the infectious agent. “Naturally occurring scrapie m sheep is not the same as experimental scrapie in any species. It should be studied, m depth, as a unique disease” (14). 2.4.2. Transmissible Disease in Genetic Cases The recent demonstration that genetically manipulated mice spontaneously can develop neurological disease, which then can be experimentally transmit- ted to other rodents, has bridged the gap between the two opposing mecha- msms of “inheritance” and “infection.” It is well established that diseasec an be transmitted experimentally to animals by mtracerebral injection of brain tissue homogenate from affected members of those families in which typical prion disease 1s inherited as an autosomal 8 Ridley and Baker dommant. It is also the case that in each family disease is associated with one of a number of different pathogenic mutations within the open reading frame of the PrP gene. Currently about 17 such mutations have been described, most of which appear to be fully penetrant. That an apparently genetic dtsease can involve a transmissible agent raises the question of whether the gene mutations are causing disease, sui generis, or whether they are predisposing Individuals to develop disease as a result of mfection with an exogenous agent. The argu- ment that the genetic mutations are causing disease receives support from the work on transgemc macee xpressing mutant prion protein. These animals spon- taneously develop neurological disease in adulthood despite bemg protected from possible Infection with exogenous agents. Furthermore, disease 1s then transmissible from these animals by serial passage to other animals (1). These expertments suggest that attempts to classify at least some cases of priori dis- ease as genetic or infectious may be inappropriate. 2.4.3. ConversIon of Prp to Prpc Recent experimental findmgs have suggested that the fundamental feature of pathogenesis in this group of diseasesc onsistso f the conversion of a pre-existmg molecule to a new shape or conformation rather than to replzcation of an agent. 2 4 3.1. THE RELATIONSHIP BETWEEN PRP~ AND PRP~~ The most important aspect of molecular pathogenesis is the relationship between PrPC and PrPSCA. number of important features of this relationship have been established. 1 PrPC and PrPsca re both encodedb y the sameg ene and have the samea mmo acid sequencea nd relatrve mol wt of 33-35 kDa. 2. PrPC and PrPSCa re antrgenically mdtstmguishable. 3 PrPSCIS probably formed by the posttranslational modrficatron of PrPC. 4. PrPC is destroyed completely by proteolytic enzymes but PrPSCIS reduced to a protease-resistantc ore of mol wt 27-30 kDa, termed PrP27-30. 5 PrPC IS soluble m the presence of detergent, whereas PrPSCf orms msoluble aggregates 6 Turnover is rapid for PrPc but very slow for PrPSC 7 PrPC 1sf ound m brain and some other tissueso f all mammals, whereas PrPSCis found only m the brain and, to a lessere xtent, other tissueso f animals or humans with prron disease 8 PrPC levels are constant throughout disease,b ut PrPSCle vels m brain rise during diseasep rogresstoni n caseso f prton disease. 9. PrPSCa ggregatesh ave a fibrrllar structure with the tmctorral properties of amy- loid, whereas PrPC does not aggregate in this way. 10. Levels of PrPSCa re associatedw tth mfectlvrty titer, whereas no such relationshtp is seen for PrPc. The Paradox of Prion Disease 9 2.4.3.2. WHY IS PRP~ CONVERTED TO PRP~~? Since the conversion of PrPC to PrPSCis irreversible, it would seem that thts conversion consists of moving from a higher to lower energy state and it is only the intervening higher activation threshold that prevents the process happening more frequently. Four factors may contribute to the crossing of this barrier: 1 The presenceo f PrP sc PrP seems to belong to a small group of protems that can have one of several conformations, the presence of one form predlsposmg the other molecules to take up that shape (15-17). Such a hypothesis can explain: a Transmission of disease in the absence of agent-specific nuclerc acid, b The exponential accumulation of PrPSC durmg disease, c The observation that, on transmission, the PrPSC in the recipient 1s host coded; d. The posslbihty that there is more than one self-perpetuating conformation allowmg for the observation of “strains of agent,” and e That where the PrP of the donor and host are different, the ability of the donor PrP to interact with the host PrP will be lessened, leading to the observation of a “species barrier.” 2. The amino acid sequence of PrP. Mutations in the PrP gene may produce dlffer- ences m the stability of the secondary structure of the PrP protein such that a spontaneous increase in the P-pleated sheet content 1s more probable. This could lead to the occurrence of genetic cases 3 Availability of PrPC Transgenic mice with multiple copies of the normal PrP gene can develop spontaneous prion disease, those with only one copy show pro- longed mcubatlon times, and mice with no PrP gene cannot be infected Factors such as an increase in general metabolic rate, which may inter &a affect the rate of production and breakdown of PrPC, could occasionally tip the balance toward the production of PrPsc and therefore contrlbute to the occurrence of prlon dls- ease m sporadic cases 4 Time. In a dynamic equlllbrium, the probability that a rare event will have occurred increases with time. The amount of PrPSC required to imtlate the mas- sive production of PrPsc is not known, but it is likely to consist of a threshold quantity. Since PrPSC IS not disposed of, the total amount present m one area will accrue with time. Naturally occurring prlon disease has a characterlstlc age at onset, which usually IS later m sporadic than inherited cases, but which is con- fined to middle age and beyond. Whether this reflects the effect of real time or the effect of age-related changes m the availability of PrPC is not yet clear 2.4.4. Prion Disease and Amyloidosis The recognition that PrPSCh as a propensity to take on the structure described as “amyloid” has allowed prion diseases to be compared to other amyloldoses rather than other mfectlous diseases. PrPc IS a soluble protem containing about 40% a-helical domains and almost no P-pleated structure PrPSCis largely insoluble with reduced a-helix content 10 Ridley and Baker (30%) and increased P-pleated structure (43%). The P-pleated content of PrP27-30 IS even higher (>50%). Pathogemc mutations m the PrP gene are thought to alter the stability of the tertiary structural relations between the alpha hellcal structures m the PrP molecule, thereby altering the probablllty that this structural relationship will break down into a P-pleated sheet conformation (IS). Protems with high P-pleated content have a propensity to form fibrlls that themselves aggregate to form large deposits that have the tmctorlal properties of “amylold.” For example, they exhibit blrefrmgence when stamed with Congo red and are viewed under polarized light The most Important CNS amyloldosis occurs m Alzhelmer disease (AD), which IS characterized by the presence in brain of P-amylold plaques, neu- rofibrlllary tangles (NFTs), and often cerebral amylold angiopathy (CAA), the deposition of P-amyloid m cerebral blood vessels Similarities between AD and prlon diseases include: 1 Approximately 15% of cases of AD and prlon disease occur with an autosomal dominant pattern of mherttance. In famlhal cases, AD may be associated with one of several mutations m the gene that makes the precursor protein (amylold precursor protein, APP) whose modlficatlon produces P-amylold, whereas prlon disease 1s associated with mutations m the PrP gene (AD also may be associated with mutations m other genes that interact with the APP gene ) 2 AD and prlon disease both have an age at onset largely confined to later middle age and beyond with onset being earher In familial than m sporadic cases 3 j3-amylold plaques are found in AD, whereas PrP-amylold plaques are seen in about 20% of cases of prlon disease Even where PrP-amylold plaques are not visible by light microscopy, electron microscopy reveals that the PrPSC exists m free fibrlls m the neuropll, suggesting that prlon disease involves amyloldosls even m cases where plaques are absent (19). 4. In a few cases of prlon disease the degree of CAA and P-amyloid either m sepa- rate plaques or integrated into PrP plaques IS greater than would be expected by chance association, suggesting an interaction m the pathogenesls of both types of amyloldosls. 5 In AD, P-amylold plaques are formed when one of two alternative processing pathways of APP results m the production of a truncated protem sequence that takes on a P-pleated sheet conformation. These molecules then polymerize into fibrlls that aggregate into the characterlstlc form of a plaque In prlon disease, PrP-amyloid may consist of a truncated form of PrPSC (20) It is, however, not yet clear whether the crucial difference between PrPC and PrPSC that associates the latter with mfectlvlty 1s the formatlon of PrPsc into amylold fibrlls or whether PrPSC exists m a pathogemc form prior to Its subsequent breakdown mto PrP-amylold. 6 Although AD IS not a transmissible disease, the laminar distribution of P-plaques in the cortex 1s consistent with the spread of pathology around the brain by a self- sustaining mechanism (21). The mJectlon of P-amylold containing bram into the brains of monkeys has been found to result m the formation of P-amylold plaques