How strong is the evidence that Parkinson’s disease is a prion disorder?

Abstract

Purpose of review

We describe evidence supporting the hypothesis that α-synuclein has a prion-like role in Parkinson’s disease and related α-synucleinopathies, and discuss how this novel thinking impacts the development of diagnostics and disease-modifying therapies.

Recent findings

Observations that immature dopamine neurons grafted to Parkinson’s disease patients can develop Lewy bodies triggered a surge of interest in the putative prion-like properties of α-synuclein. We recount results from experiments which confirm that misfolded α-synuclein can exhibit disease-propagating properties, and describe how they relate to the spreading of α-synuclein aggregates in α-synucleinopathies. We share insights into the underlying molecular mechanisms and their relevance to novel therapeutic targets. Finally, we discuss what the initial triggers of α-synuclein misfolding might be, where in the body the misfolding events might take place and how this can instruct development of novel diagnostic tools. We speculate that differences in anatomical trigger sites and variability in α-synuclein fibril structure can contribute to clinical differences between α-synucleinopathies.

Summary

The realization that α-synuclein pathology can propagate between brain regions in neurodegenerative diseases has deepened and expanded our understanding of potential pathogenic processes which can lead to the development of novel diagnostic tools as well as the identification of new therapeutic targets.

Introduction

The word prion was established in 1982 by Stanley B. Prusiner. It is derived from protein and infection and is short for "proteinaceous infectious particle". Prions can self transform their shape and propagate. Thus they can be transmitted from cell to cell, between individuals, and even between animals of different species. For his discovery of prions, Prusiner was awarded the Nobel Prize in 1997.

The neuropathology of Parkinson’s disease (PD) is characterized, in part, by severe loss of dopaminergic neurons in the substantia nigra and the development of intraneuronal α-synuclein aggregates called Lewy bodies and Lewy neurites (collectively Lewy pathology) in widespread brain regions (1). Patients with Dementia with Lewy bodies (DLB) also develop intraneuronal α-synuclein aggregates, while in Multiple System Atrophy (MSA) the majority of the α-synuclein aggregates are located in oligodendrocytes. Due to the shared pathological feature of α-synuclein aggregates, PD, DLB and MSA are known as α-synucleinopathies (1). Recently, several proteins that are prone to misfold, form aggregates in the brain and that each define specific neurodegenerative disorders have been shown to transmit from one cell to another in in experimental disease models. This applies to α-synuclein in models of PD, DLB and MSA, as well as Tau in models of Alzheimer’s disease; mutant huntingtin in Huntington’s disease models, and mutant SOD in amyotrophic lateral sclerosis (ALS) models (2,3). The behavior of these disease-related proteins has been called “prion-like”. In this minireview, we briefly describe 1) evidence that α-synucleinopathies are prion-like disorders; 2) mechanisms that underlie intercellular transfer of misfolded α-synuclein; and 3) the relevance to future therapies and diagnostics. We also discuss where the first misfolding events might occur, and controversies regarding the possibility that α-synucleinopathies might be communicable.

What triggered the idea that α-synuclein might be a prion-like protein?

In 2005 John Hardy speculated that “permissive templating” could play a key role in the pathogenesis α-synucleinopathies (4). This was an exciting concept, however, at the time there was no real clinical or experimental supporting evidence. Eventually, support for the ‘prion hypothesis’ came from a surprising place. We published two papers simultaneously showing that Lewy bodies were present, at autopsy, in 2–5% of dopamine neurons in transplants of fetal midbrain tissue implanted into the striatum of PD patients more than a decade before they had died (5,6). Interestingly, based on these first 3 cases and observations in 8 subsequent cases (operated in 5 different centers worldwide) that we are aware of, the Lewy bodies in the grafted neurons first develop after a lag period of about one decade and gradually increase in frequency up to the longest surviving case at 24 years after surgery (5–13). The aggregates stain for phosphorylated serine129-α-synuclein, ubiquitin and thioflavin S, akin to Lewy bodies found in PD brain. In 2008, when the two initial papers were published (5,6), we suggested that misfolded α-synuclein had spread from neurons in the diseased PD brain into the young dopamine neurons inside the transplant and had seeded misfolding of endogenous α-synuclein. We also proposed that these observations could explain neuropathology findings in PD that Braak and colleagues had reported 5 years earlier (14). Braak and his team had mapped α-synuclein aggregates in the nervous system of a large series of post-mortem cases and suggested that Lewy pathology in PD can be can be divided into 6 distinct stages and that, over time, the pathology is propagated from one brain region to another via long unmyelinated pathways (15,16). While Braak and colleagues suggested that the transmitting agent might be a unknown pathogen, e.g. a neurotropic virus (17–19), following our observations in the transplants we instead proposed that it was a misfolded species of α-synuclein (14). At the time, α-synuclein was primarily viewed as an intraneuronal protein with a possible role in vesicular trafficking and synaptic transmission (20). A few studies, notably by El-Agnaf and Lee and coworkers, had already independently shown that α-synuclein could be secreted by neurons into the extracellular space (even be measured in human CSF and plasma) (21–23) and affect nearby cells, e.g. microglia (22). However, the idea that this α-synuclein might be taken up by nearby neurons and act as a template for misfolding was not generally discussed. Thus, the suggestion in 2008 that cell-to-cell transfer of α-synuclein was what caused the development of Lewy bodies in the transplants in PD patients received a mixed reception. In the research community there was both a great deal of excitement, as well as significant skepticism, with suggestions that e.g. neuroinflammation or accelerated cellular ageing inside the transplant had triggered the formation of the Lewy bodies (10,14). During the years immediately following the proposal that misfolded α-synuclein might behave in a prion-like fashion, a large number of experimental studies addressing this topic were published and we summarize them briefly in the next section.

Experiments support the idea that misfolded α-synuclein behaves in a prion-like fashion

Numerous in vitro studies have convincingly shown that cultured neurons can both secrete α-synuclein and take it up from the extracellular space (22–26). In other experiments, recombinant α-synuclein protein added to the culture medium was found to be taken up by cultured neurons (27–29). Once inside a “new” neuron, “exogenous” α-synuclein (both when cell-derived and when added as recombinant protein) can oligomerize with endogenous α-synuclein and seed the formation of aggregates (24,27–29). Similarly, animal experiments have shown that α-synuclein can be released and taken up by neurons. For example, in different paradigms involving transplantation of rodent-derived neurons or neural stem cells into animals overexpressing human α-synuclein (either transgenically or via an injected viral vector), the grafted rodent neurons incorporate the human α-synuclein (24,30–32). Moreover, using strategies involving the intracerebral, intramuscular or intraperitoneal injections of oligomeric or fibrillar α-synuclein derived either from recombinant protein of from post-mortem brain samples obtained at autopsies of patients with α-synucleinopathies (e.g. PD, MSA and DLB), the injected material triggers accumulation of α-synucleinin the host rodent or non-human primate brain (33–43). This occurs within a few weeks of the injection both in rodent brains overexpressing α-synuclein and in normal wild-type brains, indicating that the seeding of aggregation is a powerful mechanism. It is clear, however, that the precise methods used (e.g. preparation of the α-synuclein fibrils and amount injected) strongly influence the outcome, which has been reviewed in detail recently (41). What is also clear is that different forms of α-synuclein fibrils exist, with varying three-dimensional structures. Akin to “prion strains”, different α-synuclein fibril conformers exhibit varying degrees of cellular toxicity (44) and can cause different distributions (cellular and anatomic) of neuropathology when injected into animals (40,45).

While the demonstration of induced α-synucleinopathy in experimental animals is undeniably fascinating, critics have argued that the models are not relevant to the human diseases and comments such as “PD patients do not develop the disease due to an injection of α-synuclein fibrils into the striatum” are frequently heard. In the next section we discuss how animal models that more closely mimic human α-synucleinopathies have recently entered the stage.

Braak neuropathological staging supports the concept of propagation of α-synuclein aggregates

Based on the proposed six stages of Lewy pathology in PD, Braak and colleagues proposed two starting points for the disease process, namely the olfactory system and the enteric nervous system (18,19). The olfactory bulb and enteric nerves (including the dorsal motor nucleus of the vagal nerve) both exhibit a heavy load of Lewy pathology already in early disease (46–48). Therefore they formulated the “dual-hit hypothesis”, mentioned above, stating that an unknown toxic agent or pathogen was inhaled, affecting the olfactory bulb and the gut wall, after being swallowed (18,19). This hypothesis is also consistent with clinical observations that hyposmia and constipation are prevalent at the time of PD diagnosis and often precede the development of overt motor symptoms by several years (49–52). Braak’s team suggested that the offending agent spread along axonal pathways, not unlike a suggestion made 20 years earlier (long before the discovery of α-synucleinin Lewy bodies) by Saper et al. (53), who proposed that “the processes of axonal and transneuronal transport may subserve the transmission from neuron to neuron of a toxic or infectious agent”.

Mouse studies have shown that injections of recombinant α-synuclein fibrils into the olfactory bulb lead to rapid transport of the protein throughout the olfactory system and interconnected brain regions (54). When the injected material is a mixture of different sized α-synuclein fibrils, the Lewy body-like pathology spreads gradually to several additional brain areas over 6 months (41), validating the idea that α-synuclein aggregates can spread from the olfactory bulb along axonal pathways and transcend synaptic connections (46). The vagus nerve is a second potential entry point for Lewy pathology into the central nervous system. Interestingly, a epidemiological study on a large Danish cohort reported a controversial finding describing a significantly reduced future PD risk in patients who underwent truncal vagotomy as a treatment for peptic ulcer in 1977–1995 (55). Recently, it was speculated that changes in gut microbiome (56,57), e.g. as a consequence of smoking or coffee (known modifiers of PD risk), could reduce the likelihood of inflammatory events that might trigger α-synuclein misfolding in the gut wall (58,59). In rats, injections of monomeric, oligomeric or fibrillar α-synuclein or PD patient-derived brain lysates into the gut wall led to transport of α-synuclein retrogradely along the vagus nerve to the brainstem (60). Providing further support for the gut being of interest in PD, injection of a viral vector encoding human α-synuclein into the vagal nerve lead to α-synuclein spreading beyond the cranial nerve nucleus to more rostral brain areas (61). When the brainstem neurons initially involved were damaged, the α-synuclein aggregates spread less rostrally, suggesting that neural connections are essential (62). These initial observations of α-synuclein spreading from the olfactory bulb and gut (via the vagal nerve) in mice are very exciting and await replication in larger animals. It is unclear if α-synuclein pathology can spread to nigral dopamine neurons, or if this takes so long that it cannot be studied in rodents with a lifespan of 2–3 years.

Cellular and molecular mechanisms of α-synuclein intercellular transfer and interregional transport

Several studies, mostly on cultured neurons, have defined molecular mechanisms of α-synuclein release, uptake and axonal transport. Recent reviews describe these areas in detail (3,63,64) and we only provide a short overview. It is important to emphasize that underlying mechanisms differ depending on whether the α-synuclein is present as a monomer, oligomer or fibril. A recent in vivo study indicated that monomeric α-synuclein is released from nerve terminals in the striatum in a process that is potassium-dependent and regulated by presynaptic GABA receptors (65). Monomers of α-synuclein appear to readily translocate across cellular membranes in a temperature-insensitive manner (22). For both α-synuclein oligomers and fibrils, the cellular release and uptake is likely regulated differently because they are not presumed to have normal physiological roles. In cell cultures, α-synuclein is released via exocytosis and the process is enhanced by cellular stress such as inhibition of lysosomes (66–70). A minor portion (frequently estimated at around 5%) of the extracellular α-synuclein is associated with exosomes and it can be present in oligomeric forms, both on the inside and surface of the microvesicles (68,69,71,72). It is unclear whether all of the α-synuclein found in the exosomal fraction is secreted with the exosomes or whether they associate with each other outside the cell. Exosomal lipids can accelerate the process of α-synuclein fibril formation and thereby they might stimulate the disease process (73).

The uptake of extracellular α-synuclein, whether it is oligomeric or fibrillar, is believed to be mediated via endocytosis. Several pharmacologic and genetic experiments have demonstrated that endocytosis is major mechanism for uptake of extracellular α-synuclein (24,30,74–79). How α-synuclein binds the outer cellular membrane is not fully understood, but heparan sulphate proteoglycans appear to play a role because uptake of α-synuclein is reduced when they are blocked (75). What also remains unclear is how fibrillar α-synuclein contained in an endosome can gain access to the cytoplasm so that it can seed the aggregation of the endogenous protein, and it has been suggested that it has the capacity to penetrate lysosomal membrane akin to what viral peptides are capable of (80). In vitro experiments have shown that α-synuclein that has been taken up from the extracellular space undergo bidirectional fast and slow axonal transport, i.e. to and from the cell body (60,81,82). This can explain how α-synuclein pathology spreads in PD between brain regions over considerable distances.

While we have primarily considered transfer of from neuron-to-neuron, astrocytes, microglia and oligodendrocytes can also all take up extracellular α-synuclein (79,83–86). It is beyond the scope of this brief review to describe this in detail but it warrants a few comments. Uptake of α-synuclein has been suggested to activate astrocytes, which can explain the astrocytosis sometimes observed in α-synucleinopathies (87). Uptake of α-synuclein by microglia is enhanced by immunoglobulins binding to α-synuclein, which is relevant to ongoing immunotherapy trials in PD and MSA (88–93). Further, uptake of α-synuclein by oligodendrocytes (79,83) might be explain why these cells exhibit α-synuclein aggregates in MSA (1), although they also can synthesize α-synuclein endogenously (94,95).

New treatment strategies and diagnostics based on α-synuclein acting like a prion

If the cell-to-cell propagation of α-synuclein aggregates is linked to clinical progression of α-synucleinopathies, then preventing its spread may be a powerful clinical target. Four principally different strategies are evident: inhibition of α-synuclein release or its uptake by neurons (for PD and DLB) or oligodendroglia (for MSA); prevention of α-synuclein seeding and promotion of α-synuclein clearance from the extracellular space. It is conceivable that α-synuclein can be reduced by promoting its intracellular clearance by enhancing the lysosomal-autophagy system through inhibition of the mammalian target of rapamycin (mTOR), using, e.g., drugs such as rapamycin or metoformin (96,97). As mentioned previously, endocytic uptake of α-synuclein can be reduced by in cell cultures by pharmacologically interfering with heparan sulphate proteoglycans (75). Further, oligomerization, the initial crucial steps in the seeding process, can be inhibited by certain small molecules such as Anle138b (98,99). Finally, as mentioned above, immunotherapy is currently being pursued both in the laboratory and clinically in attempts to clear α-synuclein from the extracellular space (88–93).

The therapeutic potential of any of these approaches will be maximized once diagnostic and progression biomarkers have been established. Such biomarkers would potentially allow for the identification of PD prior to the onset of the motor symptoms (52). Related to the prion hypothesis for α-synucleinopathies, it is notable that extensive efforts are being made to identify pathological expression of α-synuclein in peripheral tissues, e.g. the submaxillary gland (100,101) or enteric nervous system biopsies (102,103), although the field remains controversial with contradicting results (104).

No evidence that α-synucleinopathy is communicable except in experimental paradigms

The experimental transmission of α-synuclein leads to the question: Is PD (and/or related α-synucleinopathies) a communicable disease? The epidemiology studies conducted so far do not support that PD is a communicable disease. Thus, a retrospective analysis of patients who received cadaveric human growth hormone (c-hGH), the same cohort that hosted an outbreak of more than 200 cases of iatrogenic Creutzfeldt-Jakob disease (iCJD), revealed no cases of PD (105). A caveat of the study is that the follow up period might not have been long enough. The incubation time for iCJD can be as long as 42 years (106) and it might take several decades for misfolded α-synuclein to cause a widespread aggregation. Another caveat is the reliance on death certificates, but not neuropathological analyses, to ascertain if the subjects had PD or not. Finally, we do not know if the protocol used to extract growth hormone affects the stability of the putatively pathogenic α-synuclein conformers. While currently available evidence does not support that PD is communicable disease, the limited observation period and lack of neuropathology data in this key study means that we cannot be certain.

Conclusions

Experimental studies demonstrating cell-to-cell transfer of α-synuclein pathology firmly establish that, in addition to its normal conformation, α-synuclein can exist in altered, self-propagating conformation(s). Once inside animals, the altered α-synuclein conformers are able to propagate to host α-synuclein, resulting in α-synuclein aggregation and, e.g., degeneration of nigral dopaminergic neurons, both of which are pathological hallmarks of PD. These properties of α-synuclein are exactly as those of prions, i.e. 1) able to stably exist in at least two conformational states (one being the self-perpetuating conformer); and 2) able to replicate and cause neurodegeneration when it is inside the body.

Despite several similarities at the cellular level, current evidence does not support that PD other α-synucleinopathies are communicable disease like several other prion diseases. As mentioned above, it might still be too early to exclude the possibility that we will eventually identify cases of contracted PD, but that the need for a ‘high dose’ of pathogenic α-synuclein protein make them rare and the putatively long incubation times will make them difficult to detect. That said, the lack of communicability is not an exclusion criterion per se for a prion disease. Prion diseases are a group of neurodegenerative disorders and only some animal prion diseases are classic communicable diseases, e.g. scrapie in sheep and chronic wasting disease in deer and elk. Further, some human prion diseases are communicable via man-made transmission routes, such as Kuru and iCJD, but only in rare situations (cannibalism and transplantation procedures). On the other hand, certain prion diseases are not communicable, e.g. sporadic and inherited CJD. Thus, even if PD and other α-synucleinopathies are non-communicable diseases, this should not exempt them from membership in the family of prion diseases. The final take home message, however, should be that the most important issue is not classification or not of α-synucleinopathies as prion diseases, but the realization that prion properties of α-synuclein contribute to the pathogenic process and further understanding of mechanistic underpinnings can lead to new therapeutic strategies.

Key bullet points that summarize the article.

• Intraneuronal α-synuclein aggregates are a key feature of Parkinson’s disease neuropathology

• Emerging evidence suggests that α-synuclein pathology can propagate from cell-to-cell in a prion-like manner progressively engaging additional brain regions, and that this spreading might contribute to the symptomatic progression.

• The molecular mechanisms underpinning cell-to-cell spreading of α-synuclein pathology are currently being elucidated and are leading to the identification of new therapeutic targets

• Animal models of Parkinson’s disease that are based on α-synuclein acting in a prion-like fashion offer a powerful test bed for novel therapeutic interventions