Mammalian  prions

Abstract

Protein misfolding is central to the pathogenesis of several neurodegenerative disorders. Among these disorders, prion diseases are unique because they are transmissible. The conversion of the host-encoded GPI-anchored PrP protein into a structurally altered form is crucially associated with the infectious and neurotoxic properties of the resulting abnormal PrP. Many lines of evidence indicate that distinct aggregated forms with different size and protease resistance are produced during prion multiplication. The recent isolation of various subsets of abnormal PrP, along with the improved biochemical tools and infectivity detection assays have shed light on the diversity of abnormal PrP protein and may give insights into the features of the more infectious subsets of abnormal PrP.

Unusual Agents for a Unique Disease

Transmissible Spongiform Encephalo-pathies (TSEs), or prion diseases, are transmissible fatal neurodegenerative disorders affecting human and a wide range of mammals.1,2 Although the outbreak of bovine spongiform encephalopathy (BSE) had devastating effects on British cattle flocks, the prototypic animal TSE is scrapie, a naturally occurring disease affecting sheep and goats. Scrapie is a worldwide endemic disease that has been described for over 250 years. However, the implemented small ruminant surveillance programs led to the recent detection of a previously unnoticed prion agent in sheep. The resulting disease, named atypical scrapie, is now recognized as a worldwide sheep disease, with marked prevalence in the EU. Chronic wasting disease (CWD) is a TSE affecting farmed and free-ranging cervids, mainly recognized in North America. In human, prion diseases are traditionally classified into Creutzfeldt-Jakob disease (CJD), its variant form (vCJD) resulting from human infection by BSE prions, Gerstmann-Sträussler-Scheinker disease (GSS) and fatal familial insomnia (FFI). Prion diseases are very unusual as they can arise without any apparent cause (e.g., sporadic CJD) or they can be due to genetic disorders linked to mutations in the endogenous PrP protein. The diseases can also be acquired by infection, through ingestion of contaminated products or through iatrogenic procedures, and are in most cases experimentally transmissible.

The nature of the causative infectious agent has been strongly debated for decades.3,4 Initially thought to be a slow virus because of the very long incubation time before clinical onset, its unconventional resistance to various inactivation procedures led Griffith to hypothesize that it might be a self-replicating protein.5 This somehow heretical proposal was subsequently worked out by Prusiner and co-workers who provided a wealth of compelling experimental evidence that prion infectivity is tightly associated with the misfolding of the GPI-anchored, host-encoded PrP protein.6 Moreover, the possible involvement of a conventional infectious agent is now ruled out because the whole process of prion multiplication has been achieved acellularly in test tubes subjected to successive incubation and sonication steps (namely protein misfolding cyclic amplification or PMCA).7 Thus, the fact that abnormally folded infectious PrP (PrP^sc) recruits and converts host PrP into new PrP^sc is now widely accepted, even if (1) the structure of abnormal PrP is poorly defined, e.g., the molecular arrangement(s) responsible for the infectious properties of the abnormal PrP protein are unknown and (2) the exact molecular composition of the infectious particle is still obscure.

Diversity of Abnormal PrP

Historically, abnormal PrP was recovered after solubilization of infected tissues or cells in non-denaturing detergents followed by digestion with proteases.8 In these conditions, most proteins (including normal PrP) are degraded by proteinase K (PK) treatment while a PK-resistant PrP species is recovered in the pellet after centrifugation of the infected samples. These large insoluble pelleted aggregates, termed PrP^res, are clearly infectious,9 i.e., are able to transmit disease or infection upon experimental inoculation of animals or of susceptible cultured cells. However, whether prion infectivity is entirely congruent with these large PrP^res aggregates is still a matter of debate.10 First, these species, appearing as fibrils (called scrapie-associated fibrils11 or rods12) under the electron microscope, may form as a consequence of detergent extraction.13 Second, in some models of prion disease, significant levels of infectivity were detected in the absence of detectable typical PrP^res,14,15 while in others, infectivity and large PrP^res aggregates have strikingly distinct sedimentation profiles in density gradients.16 Third, as first suggested by Safar and co-workers,17 it is now increasingly clear that infected cells and tissues also accumulate abnormal PrP forms that lack prototypical insolubility and/or resistance to proteolysis. Fourth, different patterns of abnormal PrP deposits may be observed upon immunohistochemical analysis of non-solubilized infected brains. Proteinase K-resistant PrP deposits typically observed as a diffuse punctate staining can coexist with focal, plaque-like PrP immunoreactivity.18 Altogether, these findings support the view that what is collectively referred to as abnormal PrP is in fact a population of diverse entities. Notably, recent evidence suggests that cloned prion strains become heterogeneous upon multiplication in cultured cells,19 raising the possibility that these variants (or quasispecies) may be related to abnormal PrP diversity. In any case, deciphering the respective roles of these different PrP^sc entities in prion pathogenesis is a renewed challenge.

To explore the diversity of infectious PrP^sc, we chose to analyze persistently infected cultured cells, a much simpler model than brain that has been proven to stably produce high titers of prions without degenerative changes.20 Amongst different paradigms, we selected a sheep prion agent (PG127) for its ability to stably multiply in the ovRK13 (also known as Rov) cell line21 as well as in cultured primary neurons.22 In addition, this prion agent can be rapidly quantified through our recently-developed cell-based assay.23 Upon solubilization and PK digestion of infected cells, we recovered large PrP^res aggregates in the pellets of 100,000x g ultracentrifuged samples.24 As expected, this material was infectious. Strikingly, we observed a previously unnoticed PrP^res species that remained in the corresponding supernatants. These supernatants were no longer infectious after selective immunoprecipitation of PrP^res. Ultrafiltration experiments indicated that these PrP^res species (and their associated infectivity) readily pass through 300-kDa filters, in contrast to typical large PrP^res aggregates. However, these small PrP^res aggregates were eventually pelleted at higher ultracentrifugation forces. Finally, preliminary observations with moRK13 cells25 infected with different murine prion strains suggest that recovery of small PrP^res aggregates is not limited to one strain of prion. Although further purification will be required to assess their molecular composition, these findings indicate that small, infectious PrP^res aggregates can be recovered from infected living cells after a standard solubilization step.24 While these data further document the diversity of abnormal PrP, what do they tell us about the features of infectious PrP^sc? The possibility that PrP^res aggregates do not necessarily need to be large to be infectious is not new. Several studies already showed that in vitro treatments (including sonication and denaturation/renaturation) to break up large PrP^res aggregates did not abolish prion infectivity.26–28 Furthermore, physical disruption of large detergent-extracted PrP^res aggregates led to smaller aggregates that proved to be more infectious than the larger ones.29 This later study raised the interesting possibility that small aggregates may be more efficient to initiate infection, presumably by providing more competent seeds. However, whether this possibility is applicable to real life or is limited to harsh in vitro conditions is an open question. This is a particularly relevant issue as abnormal PrP fragmentation in mammalian cells and how this is achieved remain open questions. This could occur through mechanical stress (in vitro, agitation results in fibril fragmentation30) and/or through enzyme-mediated processes. In yeast cells, the Hsp104-chaperone-mediated fibril scission is required for prion maintenance.31 Since yeast prions can also stably multiply in mammalian cells lacking Hsp104 orthologs,32 other chaperone(s) might be involved. However, the cellular activities involved in abnormal PrP fragmentation remain to be identified. Alternatively, small PrP^res aggregates may be metabolic intermediates in the formation of the larger ones. We have previously shown that prions are actively secreted into the extracellular space of cultured infected cells, in association with microvesicles known as exosomes.33 Infectious exosomes are released upon fusion of endosomal multivesicular bodies (MVB) with the plasma membrane34 and may serve as a vehicle for prion dissemination.35 We have preliminary indications that infectious exosomes carry small PrP^res aggregates and are currently testing the possibility that these species are generated along the endocytic pathway. In any case, the recovery of cell-derived small infectious PrP^res aggregates in the absence of harsh in vitro denaturing treatment provides a strong biological basis to study their role in prion multiplication.

Whether these small aggregates are more infectious than the large ones is an important issue that remains to be addressed. Considering that these entities represent about 10% of total PrP^res and of infectivity, their specific infectivity appears roughly similar to that of their larger counterparts. However, we later found that these small species became highly aggregated as a consequence of the methanol precipitation step used before their inoculation. Therefore, other isolation conditions allowing better preservation of their conformational integrity are being developed to accurately test their infectious potential.

Biochemical detection of PrP^res is much less sensitive than assaying infectivity, implying that infectivity can be readily measured in samples with undetectable PrP^res. Nevertheless, there are few examples of high infectivity levels with barely detectable PrP^res.14,15 In addition, infectivity does not necessarily copartition with the bulk of PrP^res. In a recent study,16 solubilized material from brains infected with different sheep scrapie agents was density fractionated. There was no obvious correlation between the sedimentation profiles of infectivity and PrP^res aggregates, supporting the view that the latter are not the only infectious entities. Strikingly, for two strains, including the PG127 used in our study, the vast majority (>99%) of infectivity was found in the light fractions of the gradients. Whether small PrP^res aggregates, such as those detected in our study24 or those generated in vitro,29 are responsible for the high levels of infectivity in these fractions is therefore an interesting possibility.

PK-Sensitive Prions

In the recent years, PK-sensitive infectivity and PK-sensitive abnormal PrP have received an increasing amount of attention. For some,36 but not all37 prion strains, PK digestion results in a marked reduction of the infectious titer. The simpler explanation for the preceding observation is that most of the prion infectivity is destroyed by the protease treatment, suggesting PK-sensitive prion infectivity. However, this does not rule out alternative possibilities (e.g., truncated PrP27–30 generated by the PK treatment might have a lower specific infectivity than full-length PrP^sc). Interestingly, defined conditions of proteolysis with an alternate protease, pronase, followed by precipitation with sodium phosphotungstic acid preserved most of the infectious titer, indicating that pronase did not destroy the putative PK-sensitive infectivity.38 The next challenging step for characterizing PK-sensitive, pronase-resistant infectivity will require its isolation from PK-resistant prions.

From a biochemical point of view, evidence for PK-sensitive abnormal PrP was originally reported by Safar and co-workers in infected hamsters17 and subsequently confirmed in other paradigms.39,40 These forms differ from normal PrP as they are absent from uninfected samples, and they require denaturation to be immunodetected. Yet, in contrast to typical PrP^res, they are destroyed upon PK proteolysis. PK-sensitive PrP^sc, isolated by gel filtration41 or differential ultracentrifugation,42 contains small PrP aggregates. Although they were able to prime the conversion of normal PrP in PMCA reactions,42 there are no published data showing that these PrP species are infectious. Thus, the role of PK-sensitive PrP^sc in the multiplication of prions remains to be clarified. However, these abnormal forms may become important markers for the diagnosis of prion disease43 if indeed they are more abundant that typical PrP^res.

There has been a recent claim regarding the generation of protease-sensitive prions44 in transgenic tg9949 mice overexpressing a truncated form of PrP. Presumably due to massive (16- to 32-fold) PrP overexpression, tg9949 mice spontaneously develop a neurological disease late in life, in the absence of any detectable neuropathological changes. When inoculated with amyloid fibril preparations generated in vitro with recombinant mouse PrP, they develop brain vacuolization and accumulation of an abnormal PK-sensitive PrP that is active in an Amyloid Seeding Assay (ASA). However, because of the spontaneous neurological disease, no specific clinical signs could be ascribed to the inoculated materials. Brain vacuolization and abnormal PrP deposition were observed after serial inoculations into further tg9949 mice, but not in wild-type mice. Because prions are operationally defined as the transmissible agents responsible for a clinical disease in wild-type animals it is presently unclear whether bona fide prions are involved in the observed vacuolization process. An alternate explanation, yet unproven, would be that the inoculated material has accelerated the onset of the spontaneous vacuolization process in these transgenic mice.

Infectious versus Neurotoxic PrP Species

The cellular and molecular mechanisms that eventually lead to neuronal damage are far from being understood45 and the toxic forms of PrP have not been identified. Toxicity and infectivity may not necessarily be mediated by the same abnormal PrP species46,47 because PrP-dependent neurodegeneration may occur in the absence of infectivity48 and high levels of infectivity may show little pathogenic effects. There are indications that topology, trafficking or signaling of normal PrP may be altered in response to prion multiplication and these could in turn affect cell viability.2 On the other hand, a gain of function of PrP^sc is likely associated with neuronal death, either directly or indirectly. Given what is known about abnormal PrP diversity, all PrP^sc subsets may not be equally toxic. The most harmful may be small PrP species (i.e., oligomers) as has been proposed for misfolded proteins involved in other neurodegenerative diseases (e.g., Alzheimer's disease).49 Accordingly, it will be important to determine whether small PrP^res aggregates obtained in vitro29 or recovered from cells or tissues24 are potent neurotoxic species.

The diversity of abnormal PrP, the possible overlap between the infectious entities and those triggering neurotoxicity are fascinating and complex issues that may be important in further proteinopathies as well.

Acknowledgments

We thank Graca Raposo, Vincent Setola and Marie-Christine Vilette for their helpful suggestions on the manuscript.