The expanding realm of prion phenomena in neurodegenerative disease

Abstract

The aggregation of a soluble protein into insoluble, β-sheet rich amyloid fibrils is a defining characteristic of many neurodegenerative diseases, including prion disorders. The prion protein has so far been considered unique because of its infectious nature. Recent investigations, however, suggest that other amyloidforming proteins associated with much more common diseases, such as tau, α-synuclein, amyloid β and polyglutamine proteins, while not infectious in the classical sense, share certain essential properties with prions that may explain phenotypic diversity, and patterns of spread within the nervous system. We suggest a common mechanism of pathogenesis of myriad sporadic and inherited neurodegenerative diseases based on templated conformational change.

Templated Conformation Change and Phenotypic Diversity

The discovery of prions as a cause of sporadic, inherited and infectious neurodegenerative disease has revolutionized our perspectives on mechanisms that underlie neurodegeneration. In each case, the disease relies on the conversion of the normal form of the prion protein, PrP^C, to misfolded form, PrP^Sc through a process termed “templated conformation change,” whereby a normally folded protein is converted to an abnormal conformation via direct contact with the misfolded species.¹ This may involve other factors within the cell, but can occur in vitro with purified protein.² Prion protein misfolding is likely balanced by cellular quality control pathways that degrade or help refold the polypeptide. If the balance tips toward accumulation, the PrP^Sc load increases exponentially, and can cause a rapid neurodegenerative phenotype, although more chronic forms of degeneration are possible.¹ PrP^Sc can assemble into an assortment of structurally distinct species.^3,4 Different PrP^Sc conformers, or “strains,” generate different prion diseases, with characteristic progression rates and regions of pathology. For example, Kuru, iatrogenic CJD and variant CJD are caused by three conformationally distinct prion strains.¹

The phenotypic diversity of prion disorders is mirrored by that of many other neurodegenerative diseases. For example, Parkinson disease (PD), dementia with Lewy bodies, and multiple system atrophy are all associated with aggregation of α-synuclein.⁵ Similarly, tau aggregation is occasionally seen in prionopathies and synucleinopathies⁶ and is the pathological hallmark in over twenty other neurodegenerative diseases, the tauopathies.⁵ Distinct fibrillar protein conformations thus could account for the phenotypic diversity in more common protein aggregation disorders. Conformational diversity and templated conformation change of wild-type protein has been described for three different neurodegenerative disease-associated proteins: tau, amyloid β (Aβ, and α-synuclein). Although not discussed here, evidence also supports this mechanism of pathogenesis for a variety of systemic amyloidoses.⁷

Tau.

The microtubule associated protein tau is the most common misfolded protein associated with human neurodegenerative diseases. Tau filament deposition in Alzheimer disease (AD), frontotemporal dementia and other tauopathies correlates closely with cognitive dysfunction and cell death.⁸ Sporadic tauopathies involve only wild-type tau, and account for about 90% of tauopathy cases. Mutations in the tau gene cause various forms of frontotemporal dementia with Parkinsonism, termed FTDP-17.⁵ Tau mutations often result in increased rates of tau fibrillization and induce structurally distinct aggregates.⁹ Both sporadic and familial forms of tauopathy vary in brain region involvement, disease duration, age of onset, tau isoform expression and fibril morphology.⁵ In vitro studies now indicate that wild-type tau monomer is capable of forming fibrils of various conformations that propagate via templated conformational change, depending on the conformation of the initiating seed.¹⁰ These experiments demonstrate that tau aggregates, like prions, can exist in multiple, self-propagating conformations. Distinct wild-type tau conformers thus might underlie the myriad phenotypes of the tauopathies.

Conformational diversity of prions is reflected in the “species barrier,” which represents the inability of prions derived from different animals to cross-seed one another due to amino acid differences.¹ Species barriers block inter-species transmission in many cases, e.g., sheep to human, but can be overcome by an intermediate, such as cow, in the case of variant CJD.¹ Seeding barriers have recently been observed between different tau mutants. P301L mutant tau fibrils are unable to seed fibrillization of wild-type monomer, unlike fibrils formed from wild-type, R406W,¹¹ P301L/V337M double mutant¹⁰ and ΔK280.¹⁰ This suggests that fundamental characteristics of prions are shared by other amyloid proteins such as tau.

Aβ.

Although it is typically conceived as a single disease, AD has been characterized as a heterogeneous spectrum of phenotypes. Much of this variation can be attributed to mutations in AD-associated proteins. Mutations in presenilin and the amyloid precursor protein (APP) cause early-onset AD, but age of onset and presence or absence of cerebral hemorrhage vary for a given APP mutation.¹² Sporadic disease also exhibits variation in age of onset, rate of cognitive decline, and the location, presence and abundance of senile plaques and neurofibrillary tangles. Aβ accumulation also occurs within myofibers to cause inclusion body myositis, a common age-related muscle disease.¹³

Distinct Aβ conformers might help generate these disease subtypes. Wild-type Aβ fibrils exhibit conformational diversity that is induced by varying growth conditions.¹⁴ Aβ_1–40 fibrils prepared with or without gentle circular agitation form two different conformers, termed “quiescent” and “agitated.” Quiescent fibrils have a periodic twist, while agitated fibrils are straight and frequently paired. Based on solid-state nuclear magnetic resonance imaging, quiescent and agitated fibrils are also structurally distinct, and both fibril types propagate their conformations through serial seeding reactions. Quiescent fibrils are significantly more toxic to primary rat embryonic hippocampal neurons.¹⁴ Thus Aβ fibrils readily form conformationally distinct structures with unique consequences for cells. Further studies will be required to determine whether distinct pathologies in vivo can be linked to unique conformers.

α-synuclein.

Deposition of α-synuclein into Lewy bodies occurs in many neurodegenerative diseases.⁵ Lewy bodies contain a dense core of filamentous and granular material surrounded by radially oriented fibrils, which are composed primarily of hyperphosphorylated α-synuclein.⁵ Missense mutations in α-synuclein induce PD and dementia with Lewy bodies.⁵ Like other amyloid-forming proteins, α-synuclein is natively unstructured when prepared as a monomer in vitro, but forms β-sheet rich aggregates when incubated at a high concentrations.¹⁵ Alternatively, incubating α-synuclein monomer with preformed wild-type or mutant (A30P) seeds induces its aggregation into distinct conformations in a manner similar to tau.^10,16 These distinct wild-type α-synuclein aggregates differ in protease sensitivity, propensity for shedding, and immunoreactivity to different conformation-specific antibodies.¹⁶ As for tau and Aβ, the conformational diversity of α-synuclein may underlie or participate in the distinct clinical phenotypes of synucleinopathies.

Mechanisms of phenotypic diversity.

It is clear that propagation of a fibrillar protein structure via templated conformation change is not unique to prion proteins. Future studies should help determine whether the structural diversity of these proteins underlies phenotypic diversity of the associated diseases in vivo. Detailed studies of the yeast prion, Sup35, have shed light on potential mechanisms by which fibrils of distinct conformations might affect the pathological phenotype. Like mammalian prions, Sup35 aggregates into β-sheet rich fibrils, which provide an epigenetic mode of inheritance.¹⁷ Aggregated Sup35 has reduced efficiency as a translation termination factor, its normal cellular role. Multiple synthetic color phenotypes are produced when yeast with mutations in ADE1 or ADE2 genes are infected with distinct Sup35 conformers. The rate of fibril growth and fragility are key parameters in determining the strength of this phenotype, as the optimal rate of transmission of the Sup35 prion conformation to daughter cells depends on the ability to create a sufficient number of amyloid seeds.¹⁸ Recent work on mouse prion models also indicates that distinct PrP conformers underlie the variation in incubation times, which correlates with sensitivity to urea denaturation.⁴ In other human amyloid diseases, phenotypic variation could derive from expression of specific splice isoforms or post- translational modifications of the target proteins, which would alter the propensity for initial misfolding, the conformation of the resultant fibril, and the efficiency of its propagation.

Propagation of Misfolding

All neurodegenerative diseases begin with dysfunction in a discrete brain region. For example, the first obvious sign of AD is memory loss derived from hippocampal dysfunction, whereas in PD the first prominent sign is a movement disorder derived from degeneration of the substantia nigra. Ultimately, many neurodegenerative diseases spread along paths of neuronal proximity and connectivity to involve much larger areas of the brain. In tauopathies, tau aggregation along neuronal networks was described in the earliest studies.¹⁹ In AD brains, the distribution of neurofibrillary tangles correlates with neural connections, and vulnerability to degeneration correlates with distance from the affected areas.¹⁹ If basic mechanisms of prion pathogenesis apply, then misfolded tau itself might spread neurodegeneration within the brain. A variety of independent studies now support this idea for tau and other amyloid-forming proteins.

Propagation of aggregation requires a misfolded protein in one cell to influence protein folding in an adjacent or connected cell. For example, PrP can travel from cell to cell in exosomes²⁰ and through tunneling nanotubes.²¹ Although prions are infectious, since they can transmit between individuals, propagation of other types of protein misfolding within the brain would not require true “infectivity,” since the aggregates need only move between cells. Emerging data suggests that several types of protein aggregates are capable of entering cells, including Aβ,²² polyglutamine²³ and tau.²⁴ However, detailed mechanisms of cell-cell spread are not yet known.

Tau.

Tau is known principally as an intracellular protein: it binds, stabilizes and promotes the polymerization of neural microtubules.⁸ In disease, tau dissociates from microtubules and forms large, primarily intracellular, β-sheet rich fibrils. In tauopathy patients, extracellular tau is found as “ghost tangles” in the brain,²⁵ and in the cerebrospinal fluid (CSF). Emerging evidence suggests that tau levels in the CSF correlate with the risk of AD onset.²⁶ Extracellular tau aggregates, but not monomer, can enter cultured cells and stimulate fibrillization of intracellular tau, which is capable of further seeding of tau monomer in vitro.²⁴ Moreover, intracellular tau aggregates are capable of trans-cellular spread to co-cultured cells.²⁴ These data are consistent with tau aggregates being an agent of propagation in vivo.

Aβ.

Several lines of evidence suggest that exogenous Aβ seeds can induce pathology in different animal models, consistent with templated conformational change. For example, marmosets express Aβ with high sequence homology to humans. These animals develop Aβ deposits in old age without developing AD. Injection of Aβ-rich brain homogenates from patients with early-onset AD, but not age-matched controls, induces plaque formation within 6–7 years.²⁷ Induction of Aβ pathology from an exogenous seed has also been described in mice expressing APP, the holoprotein from which Aβ is cleaved. In these mice, Aβ deposition and Aβ angiopathy develop with age or can be induced prematurely by injecting cortical extracts from either AD patients²⁸ or aged βAPP mice.²⁹ Injecting these extracts into wild-type mice has no effect and Aβ-immunodepleted extracts fail to induce pathology. Aβ aggregates formed in vitro and injected into mice have not been shown to induce pathology, suggesting that there may be cofactors present in brain extracts that confer transmissibility between animals. Alternatively, there may be a seeding barrier between Aβ aggregates prepared in vitro and Aβ monomer expressed in vivo. While there is no evidence that AD is transmissible between humans, these experiments show clear parallels between prion proteins and Aβ.

α-synuclein.

Like tauopathies, synucleinopathies are temporally and anatomically progressive. Sporadic PD has been divided into six stages reflecting involvement of the entire brain, beginning with the medulla and spreading ultimately into the neocortex.³⁰ Involvement of the substantia nigra occurs midway through this process, and brings patients to medical attention primarily because of the induced motor symptoms. Pathology beginning in the brainstem follows a principally upward course, ultimately reaching the neocortex. This is reminiscent of the spread observed in prion diseases and tauopathies, and raises the possibility that α-synuclein pathology travels between neighboring or synaptically connected neurons.³⁰ Indeed, recent studies of PD patients who received fetal mesencephalic dopaminergic neuron transplants describe the presence of ubiquitin and α-synuclein-positive Lewy bodies in the engrafted neurons. Many of the deposits in the engrafted neurons were indistinguishable from lesions in the diseased host.³¹ This suggests that seeds from the surrounding diseased tissue may have spread into the graft to induce aggregation of native α-synuclein.

Summary

Non-infectious amyloid-forming proteins share many essential properties with prions. The conformational diversity of wild-type aggregates, and the capacity for propagation of these conformations via templated conformation change is shared by a growing number of amyloid-forming proteins. It remains to be determined whether such conformational diversity has pathologic correlates, as has been described for prions. The propagation of neurodegenerative diseases within the brain is also reminiscent of prion pathology. Extracellular aggregates can gain entry into cells, protein aggregates transfer between co-cultured cells, and exogenous aggregates can induce and spread pathology in animals. These data strongly suggest that protein aggregates in a variety of common sporadic and inherited neurodegenerative diseases can propagate misfolding from cell to cell. The emerging similarities between prions and other amyloid forming proteins will require further testing in vivo to determine physiologic relevance. If such commonalities truly underlie pathogenesis, it would suggest new therapeutic strategies to target transfer of aggregates between cells, and might also predict a limited utility of stem cell transplantation.

Abbreviations

PrP

prion protein

CJD

Creutzfeldt-Jakob disease

AD

Alzheimer disease

FTDP-17

frontotemporal dementia with parkinsonism linked to chromosome 17

PD

Parkinson disease

Aβ

amyloid beta

APP

amyloid precursor protein

CSF

cerebrospinal fluid