Transmissible encephalopathy agents

Abstract

Transmissible spongiform encephalopathies (TSEs) are caused by infectious agents with stable virulence characteristics that are not encoded by the host. Agent-specific features of virulence include disease latency and tissue pathology in a given host, as well as the ability to spread to many species. Such cross-species infections contradict predictions based on the prion hypothesis. Recent transmissions of several human agents to normal mice, and to monotypic neural cells in culture, underscore the existence of unique agent clades that are prevalent in particular geographic regions. Examples include the epidemic UK bovine agent (BSE) and the New Guinea kuru agent. The virus-like biology of unique TSE agents, including epidemic spread, mutation and superinfection, can be used to systematically define features of virulence that distinguish common endemic from newly emerging strains.

Transmissible spongiform encephalopathies (TSEs) were originally classified as neurodegenerative diseases until neural tissue from sheep with scrapie and humans with kuru and Creutzfeldt-Jakob Disease (CJD) were shown to be infectious by inoculating normal animals.¹ In cross-species transmissions, as in classical viral transmissions, TSE agents progressively adapted to their new hosts and displayed decreasing incubation times to disease during serial passages.^2,3 TSEs have now been reclassified as prion diseases, based on the proposal that a host-encoded prion protein (PrP) of 34 kd folds into an infectious form, typically referred to as PrP^sc. Before 1990, this “bad” PrP^sc designation indicated that different infectious agents, such as CJD and scrapie, were ultimately encoded by the species PrP sequence, i.e., were defined as mouse, human or hamster prions. This terminology obscured the existence of a substantial number of TSE agent strains in the environment that displayed species-independent and stable characteristics.^4,5 In fact, the concept that host PrP “enciphers” individual agent strains has not been compatible with a growing body of epidemiological and experimental data (reviewed in ref. 6). The recent transmission of several TSE agents, including human kuru, to normal mice and to monotypic cell cultures,^7,8 has underscored geographic agent clades that encode distinct patterns of virulence.

The recent UK BSE epidemic in cows was caused by a new and more virulent infectious agent that has transmitted to many species including humans. It thus became obvious that local TSE agents could mutate and adapt to a foreign host regardless of PrP sequence homology. Proponents of the prion hypothesis argued that the PrP sequence alone is the sole basis of the species barrier, necessary for “infectious PrP” to convert other PrP molecules through homologous binding. However, recent experiments show that the BSE-linked vCJD agent, after replicating in humans for >5 years, could readily infect mice that express high levels of murine PrP, but not those that express homologous human PrP.⁷ It also has become clear that many TSE agents maintain remarkable species-independent characteristics, even after serial passage and adaptation to a new species. This was first shown in 1965 by re-inoculating sheep with mouse passaged scrapie.⁹ Additionally, several sporadic CJD (sCJD) isolates with characteristically long incubation periods and very limited brain lesions in mice rapidly re-infected hamsters despite serial murine adaptation. These sCJD agents retained their original virulence for hamsters and induced widespread cerebral lesions, similar to those seen in humans with sCJD.¹⁰ Such data implicate a relatively stable agent genome. These data also define the virulence of a particular TSE agent by its propensity to infect and cause fatal progressive disease in only one or in many different species. In these terms the BSE agent is quite virulent. Other relatively avirulent or rarely pathogenic TSE agents in cows were reported in Europe as early as 1883.¹¹ Interestingly, this type of “atypical BSE agent” can also infect primates.¹² A growing cervid epidemic caused by a TSE agent in the US has spread rapidly to deer and elk in a natural setting. Its potential for spread to humans remains unknown.

Experimental models have further solidified the variety of TSE agent strains by comparing their phenotypes in a single host. Normal wild-type (wt) mice inoculated with a variety of sheep and human derived TSE agents display stark differences in incubation time, regional neuropathology and lymphoid tissue involvement.⁸ Florid infection of lymphoreticular tissue along with the demonstration that TSE agents can spread via the bloodstream in experimental animals, as well as humans,^13,14 represents another dimension of cellular invasion. Non-neuronal myeloid cells that lack abnormal PrP¹⁵ are also highly infectious.

Where do agent strains originate? The recovery of distinct TSE infectious agents from restricted geographic areas, such as kuru from New Guinea, and the repeated isolation of a distinct CJD agent strain found in Japan, but not the western hemisphere, accentuates an exogenous and environmental source. Some of these environmental agents can be eradicated through minimizing ritual exposures (as with kuru), and others decline precipitously when the environmental source is removed (as with BSE). These data undermine the idea that host PrP, in any self-inflicted or spontaneously generated form,¹⁶ is the infectious agent. A more parsimonious view is that PrP is a required host receptor for the invading infectious agent, and PrP amyloid pathology is a consequence of this host-agent interaction. Indeed, PrP pathology may be part of a reactive cascade to limit the spread of these agents.¹⁷

To summarize the above, TSE agent virulence can be minimally defined by six criteria. Within one species it can be characterized by: (1) its ability to rapidly induce fatal disease, particularly after few peripheral exposures. These include natural gastrointestinal routes and superficial wounds; (2) its ability to cause widespread lesions in the brain, and (3) its ability to survive for many years. In addition, an agent's virulence should be defined by: (4) its ability to spread to many unrelated species, (5) the severity and rapidity of disease progression in these other species, and (6) the cumulative frequency or incidence of infection. Some of these infections, such as sheep scrapie, are endemic and can affect as many as 10% of particular sheep breeds in the UK, but not in sheep of the same genotype in import-restricting Australia.¹⁸

One can further evaluate the comparative virulence of TSE agents in a single animal by simultaneously infecting it with two distinct agent strains. The more virulent agent should overwhelm infection caused by a less pathogenic strain that replicates and spreads less effectively. This dominance is quite evident when one inoculates a fast-incubation TSE agent together with a slow-incubation agent. As with faster replicating viral strains, an attenuated, slowly replicating agent will be at a disadvantage with each doubling, and progressively lost by selection. This result is clear experimentally. Only the rapid agent is commonly recovered on subsequent passages, and only extensive cloning at limiting dilution can bring out the less virulent member in the mixture.¹⁹

One can also examine the ability of one agent strain to interfere with another that is inoculated at a later time. This is a classic viral vaccination strategy. It depends on the close similarity of the preventive agent with the challenge agent, typically a consequence of the host's production of neutralizing antibodies in response to the first virus. If the two viruses are dissimilar, they can instead lead to a more rapid disease with complementary virulence factors supplied by the challenge virus.

We used this challenge strategy to examine the interaction of geographically distinct CJD agents. When mice infected with a slow sCJD agent were later challenged with a fast CJD agent from Japan, the more virulent fast incubation agent was dramatically suppressed: challenged mice eventually died from the slow agent at 370 days, 120 days later than the mice not protected by the slow agent. The protected mice also showed low levels of pathologic PrP and a restricted neuropathology characteristic of sCJD.²⁰ Interestingly, 2 of these 9 challenged mouse brains displayed slightly greater pathology. To determine if both strains were present in these 2 mice, or if a new chimeric agent with intermediate features was formed, we did a second serial passage. Recipient mice came down with only the slow or only the fast agent. This showed that each of these two agents maintained its own unique identity when replicating together in a single mouse brain for >200 days. According to the prion hypothesis, infectious PrP-PrP interactions from two agent strains would lead to a “chimeric” agent with an intermediate phenotype, but this was not the case.²¹ Moreover, in repeat experiments when the slow sCJD agent was given in very low non-pathogenic doses, it completely suppressed superinfection by the more virulent-fast strain for the entire lifespan of the mouse (>600 days) without any detectable PrP pathology or clinical disease.²² Since there was no evidence of neutralizing antibodies, complex innate immune responses to the initial protective agent appeared to underlie this protection.

On the other hand, a more virulent agent should be able to produce overwhelming superinfections. Experiments designed to determine if virulence could be defined by the criterion of superinfection were performed in neural GT1 cells in culture. Prolonged growth of many different human and scrapie agents has now been successfully achieved, and infections are stable for >1 year in these monotypic cells.^7,8 Because the “PrP^sc” bands in GT1 cells are quite different from those in mouse brain, it was essential to first determine if changing the PrP pattern would alter the fundamentally different slow and fast agent characteristics. It did not, as shown by re-inoculation of mice.²³ This further contradicted the premise that PrP folding encodes agent-specific properties. The GT1 culture model permitted the rapid and repeatable observation that cells could be either protected or overwhelmed by particular challenge agents.²⁴ For example, overwhelming superinfection was found with one sheep scrapie agent but not another. Moreover, successful superinfection was not inhibited by elaboration of high levels of abnormal PrP, and was demonstrable by a unique PrP band linked to one challenge agent. Since in-situ analysis also showed PrP amyloid was present in most cells before challenge, it was apparent that two infectious agents could both replicate in a single cell. Conversely, a lack of abnormal PrP was not necessary for effective protection. An sCJD agent that induced no detectable pathologic PrP clearly prevented superinfection by a fast CJD agent, as well as by two different scrapie agents. These data emphasize profound agent-encoded features as well as innate host cellular responses to TSE infection. Early innate responses of an animal to TSE agents, but not to pathological PrP, also showed the host can recognize more virulent agents, whereas attenuated or less pathogenic members may live as unrecognized “time bombs” in a symbiotic and latent state.²⁵

While the essential molecular component(s) of infectious TSE particles remain indeterminate,⁶ the biological data are fascinating and present a useful roadmap for defining both common and newly evolved TSE agents with enhanced virulence in our environment. The recognition of distinct TSE agent strains is important for human and animal health because these agents can mutate and suddenly cause new epidemic disease. It is likely that agent-specific (non-host) molecules will eventually be identified to aid in their surveillance. Virus-like 25 nm particles that lack PrP are present in infected cells, and comparable particles, as well as nucleic acids and capsid proteins copurify with infectivity in subcellular fractions.⁶ Although not fashionable, this experimentalist suggests a generally stable, yet mutable nucleic acid will be the essential molecule that encodes the variable virulence of TSE agents.

Finally, one of the most intriguing aspects of TSE agent clades is their “precise” incubation time, a phenomenon not adequately explained by any agent model. The clockwork and effective doubling time of each agent²⁶ is a consequence of both agent replication and destruction (or arrested spread). This is most obvious after intracerebral inoculation, or, in a natural setting, after the agent has begun to induce neurological disease. The variable long incubation times to disease (8–30 years) in people inadvertently inoculated peripherally with sCJD-contaminated growth hormone demonstrates that the long latent state is not so precisely defined. Indeed, stress or aging itself may bring these agents out of hiding in lymphoid tissues,²⁷ as is seen with many other types of inapparent infections. This is true not only of viruses, but also for more complex organisms such as bacilli (e.g., tuberculosis bacillus) and parasites (e.g., Chagas trypanosomiasis).

Thus the timing or clockwork of a given TSE agent changes with the host. Many sCJD isolates from humans can induce rapid and widespread brain disease in guinea pigs and hamsters.^2,8,10 However, the same agent, regardless if isolated from human or passaged animal brains, remains excruciatingly slow in mice and provokes very restricted anatomical lesions.⁸ Hence the incubation program of this single sCJD agent types is defined by its interactions with the cells of a particular host. Yet these interactions are much less studied as compared with PrP interactions. The agent's challenge to the host, the host's ability to discard the agent or limit its spread, and the response of the agent to ongoing host restrictions, are all likely to be at the root of the adaptive evolution, clockwork and relative virulence of TSE agents.