Temporal evolution of microglia and α-synuclein accumulation following foetal grafting in Parkinson’s disease

Abstract

Introduction

Previously, we described robust survival of healthy appearing dopamine neurons 18 months following transplantation of foetal mesencephalic tissue into the striatum of patients with advanced Parkinson’s disease (Kordower et al., 1995, 1996; Olanow et al., 2003). In these studies, embryonic dopamine neurons derived from either one or four donors per side, were implanted into the post-commissural putamen bilaterally using eight needle tracts per side and four deposits per tract. Cell counting demonstrated 80 000–120 000 healthy appearing tyrosine hydroxylase (TH)-positive neurons in each putamen, with extensive organotypic innervation of the host striatum. However, two doubleblind studies failed to meet their primary end-points (Freed et al., 2001; Olanow et al., 2003), and the procedure was complicated by the development of often disabling graft-induced dyskinesias that persisted despite stopping levodopa (Olanow et al., 2009a). Based on these results and other considerations, we discontinued our transplant programme (Olanow et al., 2009b).

Approximately 14 years following the initial transplant procedure, our group and others co-discovered Lewy pathology in cell bodies and axons of the embryonic dopamine neurons that had been implanted into the striatum of these advanced Parkinson’s disease patients (Kordower et al., 2008a, b; Li et al., 2008). The inclusions stained positively for α-synuclein, ubiquitin, and thioflavin S, and were proteinase K resistant, indicating protein misfolding with beta-sheet formation and aggregation. The inclusions in the implanted dopamine neurons were morphologically identical to Lewy bodies seen in neurons of the host substantia nigra pars compacta (SNc). Importantly, the implanted cells were derived from multiple unrelated donors, and inclusions were seen in all deposit sites in each patient, and at a time when Lewy pathology is virtually never seen under normal circumstances. We speculated that Lewy pathology developed in the healthy implanted dopamine neurons as a result of their being placed into a Parkinson’s disease milieu, and represented transmission of misfolded α-synuclein from host Parkinson’s disease-affected neurons in a prion-like manner (Olanow and Prusiner, 2009). Indeed, similar results have been seen in experimental models (Kordower et al., 2011; Angot et al., 2012). In subsequent years, a body of experimental evidence has been developed supporting the prion hypothesis in Parkinson’s disease (Olanow and Brundin, 2013). Fundamental to the prion theory is the concept of permissive templating, whereby misfolded proteins promote misfolding of the native unfolded proteins, thereby sustaining an infectious chain reaction.

The mechanism whereby Lewy pathology develops in implanted dopamine neurons remains unknown. Recent experimental studies suggest that immune activation of microglia with cytokine release might play a critical role in promoting α-synuclein misfolding, and in the spread of misfolded α-synuclein (Bassil et al., 2016; Sampson et al., 2016; Wang et al., 2016). In this regard, it is noteworthy that in our initial studies we observed robust CD45 cell staining surrounding graft deposits indicative of an inflammatory reaction, with some cells displaying a morphology suggestive of activated microglia (Kordower et al., 1997; Olanow et al., 2003). However, our initial studies were performed prior to the discovery of α-synuclein mutations in Parkinson’s disease (Polymeropoulos et al., 1997), and the appreciation that this protein is a major component of Lewy pathology in sporadic cases (Spillantini et al., 1997). In the present study, we examined the evolution of α-synuclein accumulation and aggregation in graft deposits, and its temporal relationship with inflammation and microglial activation. We were able to examine brains derived from six transplanted Parkinson’s disease patients who died at 18 months (n = 1 patient), 4 years (n = 2 patients), 14 years (n = 2 patients), and 16 years (n = 1 patient) following the procedure. This is the largest post-mortem sample of transplanted brains reported to date and is of particular interest in view of the increasing evidence suggesting that activated microglia might play a role in promoting the misfolding, seeding, aggregation, and spread of α-synuclein in Parkinson’s disease as part of a prion-like process.

Materials and methods

General immunohistochemistry procedures

Endogenous peroxidase-containing elements were eliminated by 20 min incubation in 0.1 M sodium periodate, and background staining was blocked by 1 h incubation in a solution containing either 2% bovine serum albumin or 5% horse serum. Series of sections through the rostrocaudal striatum and SNc were immunostained with primary antibodies (see below) overnight at room temperature. After six washes, sections were sequentially incubated for 1 h in biotinylated secondary antibody (1:200; Vector) followed by Elite® avidin–biotin complex (1:500; Vector) for 75 min. The immunohistochemical reaction was completed with 0.05% 3, 3′-diaminobenzidine (DAB) and 0.005% H₂O₂. Sections were mounted on gelatin-coated slides, dehydrated through graded alcohol, cleared in xylene, and coverslipped with Cytoseal™ (Richard-Allan Scientific™).

TH, dopamine transporter and ubiquitin

The immunoperoxidase labelling method was used to perform TH, dopamine transporter (DAT) and ubiquitin immunohistochemistry as described previously (Chu and Kordower, 2007; Kordower et al., 2008a, b).

α-Synuclein

Separate series of sections were incubated with antibodies directed against total α-synuclein (LB509) and phosphoserine 129 α-synuclein (ab51253, Abcam, 1:1000) overnight at room temperature, as previously described (Chu and Kordower, 2007). We used proteinase K digestion to determine whether the α-synuclein was soluble (non-aggregated) or insoluble (aggregated). Sections were mounted onto gelatin-coated slides and dried for at least 8 h at 55°C. After wetting with TBS-T (10 mM Tris–HCl pH 7.8, 100 mM NaCl, 0.05% Tween-20), the sections were digested with 50 μg/ml proteinase K (Invitrogen) in TBS-T for 3 h at 55°C. The sections were fixed with 4% paraformaldehyde for 10 min. After several washes, the sections were then processed for α-synuclein immunostaining as described above.

CD45

CD45 staining was carried out to assess inflammation and inflammatory cells, including macrophages and microglia (Cosenza-Nashat et al., 2006; Bennett et al., 2016). Sections were incubated with primary antibody, mouse anti-CD45 (1:500, BD Biosciences) overnight at room temperature (Kordower et al., 1997) and processed as described above.

TMEM119

Transmembrane protein 119 (TMEM119) staining was used to more specifically assess for the presence of brain resident microglia (Bennett et al., 2016; Li and Barres, 2018; Lopez-Atalaya et al., 2018). Tissue sections from all cases were immunostained with TMEM119 antibody (1:500, Abcam) and processed immunohistochemically as described above.

Stereological quantification

TH neurons

CD45 and TMEM119 immunoreactivities

CD45 (leucocyte common antigen) is an essential regulator of antigen receptor signalling and is present on haematopoietic cells, such as leucocytes and macrophages, as well as brain microglia. TMEM119 is a transmembrane protein that is specific for brain resident microglia (Bennett et al., 2016; Lopez-Atalaya et al., 2018). Semi-quantitative optical density analyses were carried out to assess CD45 and TMEM119 immunoreactivities in foetal tissue graft deposit sites. All assessments were made in a blinded fashion using three to five sections per case. Optical density was measured from an area within the transplant region and from an area with a similar volume outside of the graft region in transplanted patients; similar volumes within the caudate nucleus/putamen were also analysed in ungrafted Parkinson’s disease controls and normal controls. Scion Image software was used for the measurement and mean optical density was calculated for each case. Statistical analysis to compare the overall CD45 and TMEM119 optical densities in graft regions of transplanted patients and areas outside of the graft was performed using GraphPad Prism 7.02 software (La Jolla, CA, USA). Paired Student’s t-test was used.

Data availability

All data will be made available upon request.

Results

Neuronal TH expression in graft deposits

The number of surviving grafted dopamine neurons per putamen was estimated in cases where sufficient tissue was available (Table 1). Neuronal survival was relatively robust in each case, albeit with substantial variability. Interestingly, while receiving an identical number of donors, our shortest surviving case (18 months post-grafting) had ∼100 000 surviving dopamine neurons per side (Kordower et al., 1995), while our case with the longest survival (16 years post-transplantion) had ∼300 000 surviving dopamine cells on each side (Kordower et al., 2017). In each case, TH-positive grafted neurons reinnervated the striatum in an organotypic manner with a normal-appearing patch-matrix pattern and synaptic connectivity. Individual TH-positive cells displayed a morphological appearance typical of mesencephalic dopamine neurons. The breadth and density of TH innervation was relatively robust and similar in cases from 18 months to 14 years post-grafting. However, the 16-year post-graft case had significantly greater graft survival with TH reinnervation of the entire putamen at a level approximating that seen in normal age-matched controls, even though grafts were only placed into the post-commissural putamen. In addition, in all cases, TH immunoreactive fibres were seen traversing the internal capsule to reach the caudate nucleus. Nonetheless, in both the 14- and 16-year post-grafted cases we observed functional impairment of TH immunoreactive neurons as evidenced by a marked reduction in DAT (Kordower et al., 2008a, b). These changes were not noted in the 18-month and 4-year post-transplantation cases suggesting a time-dependent loss of expression of dopamine markers.

Pattern of α-synuclein accumulation and distribution

To assess the presence and distribution of α-synuclein pathology within foetal tissue graft deposits, brain sections from each of our cases were immunostained with antibodies directed against total α-synuclein (LB509) and phosphorylated α-synuclein (Ser129-phosphorylated α-synuclein) (Fig. 1). In the 18-month post-surgical case, no α-synuclein immunoreactivity was detected in the grafted cells with either antibody. In both of the 4-year cases, diffuse staining of monomeric, proteinase K digestible, α-synuclein was seen throughout the cytosol of a few implanted cells close to the graft margins. There was no immunostaining for Ser129 phosphorylated α-synuclein or thioflavin S. In the older grafts (14 and 16 years post-transplantation), there was abundant staining for Ser129 phosphorylated α-synuclein inclusions in cell bodies and neurites that were proteinase K resistant. Ser129 α-synuclein positive inclusions were noted in ∼5–30% of TH-positive grafted cells, with greater numbers in the 16-year than in the two 14-year cases. The 16-year post-graft case also had the most extensive Lewy neuritic pathology, involving fibres within the graft as well as within the host striatum. In both the 14- and 16-year post-graft cases, α-synuclein aggregates also stained positively for ubiquitin and thioflavin S indicating the presence of misfolded protein with beta-pleated sheet formation. In these older (14- and 16-year) grafts, less monomeric soluble α-synuclein staining was visible as aggregate accumulation increased. These data suggest that within grafted embryonic neurons, there is a gradual accumulation of soluble α-synuclein that develops over 4 years post-transplantation and misfolded/aggregated α-synuclein that evolves over the course of approximately the next decade.

Figure 1

Time-dependent progression of α-synuclein accumulation and aggregation in foetal dopamine nigral grafts. Sections through the graft were stained for Ser129 phospho-synuclein. Scale bar = 100 μm (A–D); 20 μm (a–d).

Open in new tabDownload slide

Pattern of CD45-labelled cells

To analyse inflammation in graft regions, brain sections from transplanted cases as well as from two non-transplanted Parkinson’s disease patients and two normal age-matched controls were stained with CD45 (leucocyte common antigen) antibody to identify inflammation in the region of graft deposits. In normal brain, expression of CD45 is low, but when microglia are activated, CD45 expression is markedly upregulated and cells assume characteristic morphologies (Cosenza et al., 2002; David and Kroner, 2011). However, this antibody is relatively non-specific and recognizes infiltrating lymphocytes and macrophages as well as activate microglia. In the four transplanted cases where tissue was available for analysis (18 months, 4 years, 14 years, and 16 years post-transplantation), there was a significant increase in CD45-positive cells in each of the grafted areas in comparison to non-grafted regions of the putamen, and in comparison to the putamen of non-transplanted Parkinson’s disease patients and age-matched controls (Fig. 2). At 18 months, CD45 stained cells were round and ramified in morphology. The round cells displayed dark staining without processes suggesting that these cells were likely to be lymphocytes or infiltrating macrophages. The ramified cells exhibited light staining with more processes as compared to what was observed in the non-grafted putamen. In all of the older grafts (Fig. 2), dark stained CD45-positive cells with abundant short and large processes continued to be observed, but round cells suggestive of lymphocytes were no longer detected. The CD45-positive cells in non-grafted regions of the putamen were markedly fewer in number, and displayed lighter staining and longer and finer processes. In the grafts that were 18 months and 4 years post-transplantation, CD45-positive cells were somewhat heterogeneous with some areas relatively devoid of cells, whereas CD45-positive cells were more evenly distributed throughout the implant sites in the 14- and 16-year grafts. In the 18-month-old graft, CD45-positive cells were also observed around blood vessels near the grafted sites suggesting the possibility of either ongoing infiltration into the brain of CD45-positive inflammatory cells or greater activation and accumulation of immune activated cells at these sites (Fig. 3). In the older grafts, the cell morphology and increased CD45 expression was more consistent and restricted to the graft regions. Quantitative optical density measurements are shown in Fig. 2, and demonstrate significant increases in grafted regions of each transplanted patient in comparison to non-grafted regions in these same patients, and in comparison to comparable regions of the putamen in Parkinson’s disease and control subjects who had not undergone a transplant procedure.

Figure 2

Qualitative and quantitative evidence for increased CD45 immunoreactivity was observed in grafted regions of the striatum at all time points examined. Representative low and high power examples are provided from the cases at 18 months (A and B), 4 years (C and D), 14 years (E and F) and 16 years (G and H) after transplantation and (I) an age-matched control. (J) Mean optical density (OD) measures of CD45 immunostaining was calculated for graft regions and comparable regions of the putamen of Parkinson’s disease and controls. Data are presented as mean optical density in grafted versus non-grafted controls (P < 0.0002). Scale bar = 200 µm (A, C, E, G and I); 30 µm (B, D, F, H and inset I). NC = normal control; PD = Parkinson’s disease.

Open in new tabDownload slide

Figure 3

Low (A) and high power (B) photomicrographs from grafted region of 18 month graft illustrating CD45 immunoreactivity. Intense CD45-labelled inflammatory cells were observed within graft (arrowheads; A). Boxed area in A is enlarged in B showing two types of CD45-positive cells in graft: dark stained round cells with absent processes (arrows; B) associated with blood vessels within the graft, and light stained small cells with present processes (arrowheads; B) distributed either in grafted region or non-grafted putamen. Scale bars = 400 µm (A); 40 µm (B).

Open in new tabDownload slide

Pattern of TMEM119-labelled cells

To characterize the inflammatory cells that accumulate in grafted regions, brain tissues were immunostained with a TMEM119 antibody, which stains for a cell surface protein specific to brain resident microglia (Bennett et al., 2016). TMEM119-positive cells displayed ramified morphological features in all brain areas (Fig. 4). TMEM119 staining in grafted patients was much more pronounced than in non-grafted Parkinson’s disease patients or control subjects. Robust dark and dense TMEM119 staining was observed in all grafted regions at all time points examined, with immunoreactivity mainly seen in processes (Fig. 4). TMEM119-labelled cells had enlarged cell bodies with both short and large processes that displayed a bushy-like (non-amoeboid) morphology. In contrast, in non-grafted regions of the putamen, TMEM119-positive cells displayed small cell bodies with long ramified processes (Fig. 5), similar to those seen in patients with Parkinson’s disease who had not undergone a transplant procedure, and greater than in healthy controls. The density of TMEM119-positive cells in 18-month-old grafts was somewhat less than in the 4–16-year-old grafts. Further, in the older grafts, the TMEM119-positive cells accumulated around the periphery of the graft with only a few TMEM119-positive cells in the centre area, similar to the distribution pattern observed for grafted dopaminergic neurons.

Figure 4

Photomicrographs showing TMEM119 immunoreactivity. Photomicrographs from grafted regions (A, B, D, F, G, H, J and K), putamenal area remote from graft deposits (C, F, I and L), Parkinson’s disease control putamen (M), and control (N), illustrating TMEM119 immunoreactivity. The grafted samples were collected from 18 month (A and B), 4 year (D and E), 14 year (G and H), and 16 year (J and K) time points after transplantation. Note that intense TMEM119 labelled microglia were observed within grafts at all time points examined as compared with putamenal area remote from graft deposits and Parkinson’s disease patients who did not undergo a transplant procedure. Scale bar = 100 μm in N (applies to B, C, E, F, H, I, K, L and M); 500 μm in A, D, G and J. Mean optical density (OD) of TMEM119 immunostaining was calculated for graft regions and comparable regions of the putamen of Parkinson’s disease subjects and controls (O). Data are presented as mean optical density versus non-grafted controls (P < 0.05 to P < 0.001). NC = normal control; PD = Parkinson’s disease.

Open in new tabDownload slide

Figure 5

Photomicrographs showing morphological pattern of TMEM119-labelled microglia. High power photomicrographs from grafted regions (A, C, E and G), putamenal area remote from graft deposits (B, D, F and H), Parkinson’s disease control (I), and normal control (J) illustrating morphological pattern of TMEM119-labelled microglia. The TMEM119-positive microglia inside grafts at all time points examined displayed dense swollen and short processes and enlarged cell bodies with bushy morphology (A, C, E and G). However, in putamen remote from grafts (B, D, F and H) and Parkinson’s disease control (I), TMEM119 labelled small cell bodies and long ramified processes. In age-matched control (J), TMEM-positive microglia exhibited light stained cell bodies and fine ramified processes. Scale bar = 30 μm in J (applies to all). PD = Parkinson’s disease.

Open in new tabDownload slide

Quantitative optical density measures (Fig. 4E) confirmed that there was a significant increase in TMEM119 immunoreactivity in grafted regions of each transplanted patient in comparison to non-grafted regions of the putamen, and in comparison to comparable regions of the putamen of Parkinson’s disease subjects who had not undergone a transplant procedure and normal age-matched controls.

Discussion

We demonstrate that pathological α-synuclein aggregation in embryonic dopamine neurons that have been implanted into the striatum of patients with advanced Parkinson’s disease develops gradually over the course of ∼14–16 years. No α-synuclein staining was detected in any of the grafted neurons in any of the 64 deposit sites in the single case examined at 18 months post-transplantation. In the two cases at 4 years, there was a diffuse homogeneous pattern of staining for α-synuclein in transplanted dopamine neurons that was proteinase K digestible and was not associated with staining for thioflavin S. These findings are consistent with the accumulation of soluble α-synuclein, but without misfolding or aggregate formation. In the two cases at 14 years and the one case at 16 years post-transplantation, there was prominent α-synuclein aggregate formation in dopamine nerve cells and terminals in all deposit sites, affecting ∼5–30% of the implanted dopaminergic cells. At these time points, the aggregates were proteinase K resistant, stained positively for ubiquitin and thioflavin S, and were indistinguishable from Lewy bodies in the SNc neurons of these same patients as we have previously reported (Kordower et al., 2008a, b, 2017). These data suggest that in grafted, dopaminergic neurons placed into the Parkinson’s disease brain, there is a slow, gradual, decade-long accumulation of misfolded α-synuclein. In this regard it should be noted that α-synuclein aggregates are virtually never seen in dopamine neurons of this age.

In contrast, staining for inflammation (CD45) and specifically for activated microglia (TMEM119) was markedly increased in all graft sites in comparison to non-grafted regions of the putamen in transplanted patients, and in comparison to all regions of the putamen in non-grafted Parkinson’s disease patients and healthy controls (Fig. 2). This staining pattern was observed at all time points examined (18 months to 16 years). CD45-positive staining and the appearance of the cells is indicative of inflammation, but could represent infiltrating peripheral macrophages particularly at 18 months. However, positive staining with TMEM19, which is a specific marker for brain resident microglia, indicates that the inflammation in graft regions is principally associated with activated microglia even at the 18-month time point. These findings contrast with α-synuclein staining, which was not detected at 18 months, and showed only diffuse soluble, non-aggregated α-synuclein accumulation at 4 years post-transplantation. Ser129 phosphorylated α-synuclein immune-positive, proteinase K-resistant aggregates with thioflavin S-positive staining indicative of misfolded α-synuclein were not observed until the cases examined 14 and 16 years after transplantation. Thus, there is evidence of abundant and persistent inflammation and microglial activation in all graft deposit sites years before the accumulation of α-synuclein aggregates in implanted dopamine neurons.

Our data demonstrating that following dopamine cell transplantation, microglial activation is present many years prior to α-synuclein deposition and aggregation, raise the possibility that microglial activation might contribute to the development of α-synuclein pathology in the implanted dopamine neurons, and support the concept that activated microglia play an integral role in the propagation and spread of α-synuclein pathology (David and Kroner, 2011; Tomé et al., 2013; Dzamko et al., 2017; Valdinocci et al., 2017). Recent studies have begun to shed light on how this might occur. Activated microglia secrete toxic cytokines such as caspase-1 and calpains, which metabolize and cleave α-synuclein at the C-terminus causing the protein to misfold and aggregate (Dufty et al., 2007; Levesque et al., 2010). In cell culture studies, menadione, which promotes caspase-1 activity, induces dose-dependent truncation and aggregation of α-synuclein (Angot et al., 2012) that can be blocked by inhibiting caspase-1 with shRNAs or the caspase-1 inhibitor VX765. Importantly, studies in transgenic mice that overexpress α-synuclein show that VX765 prevents α-synuclein misfolding and aggregation, as well as degeneration of SNc dopamine neurons (Bassil et al., 2016). Further, it has been demonstrated that activated microglia associated with the Parkinson’s disease microbiome are critical for the spread of α-synuclein pathology to the CNS and the development of motor dysfunction in α-synuclein transgenic mice (Wang et al., 2016). In these transgenic animals, antibiotic treatment, which suppresses the microbiome and the associated microglial activation, prevents α-synuclein transfer to the brain and the development of motor impairment. Subsequent gastric infusion of extracts derived from the microbiome of Parkinson’s disease patients, but not from normal individuals, causes recurrence of microglial activation with the transfer of α-synuclein to the CNS and the development of motor deficits.

In our transplanted cases, microglial activation could have been initiated by an immune reaction due to the implantation of foreign tissue or by the surgical procedure itself. Indeed, we observed a prominent increase in CD45 cell staining surrounding blood vessels in the early grafted cases, suggesting the possibility of infiltrating macrophages from the periphery consequent to a surgically-induced disruption of the blood–brain barrier (Dudvarski et al., 2016). It is also interesting to speculate that activated microglia could release toxic cytokines causing host α-synuclein to misfold and aggregate, and that pathological α-synuclein released from cells could be taken up by microglia causing them to become, or continue to be, activated, thus accounting for the continued presence of inflammation and the progressive α-synuclein pathology (Duda et al., 2002; Hoenen et al., 2016; Zhang et al., 2017) (Fig. 6).

Figure 6

Schematic of the viscous cycle between inflammation and Lewy pathology. Schematic illustration demonstrating how a variety of factors including the implantation of foreign tissue could cause microglia to become activated and initiate a vicious cycle in which toxic cytokines cause native α-synuclein to misfold, and the uptake of misfolded α-synuclein causes microglia to become activated. In this scenario, microglia become activated and release toxic cytokines, which can metabolize and cleave native α-synuclein causing it to misfold, aggregate, and form pathological inclusions (Lewy pathology). Misfolded α-synuclein can be released from cells and taken up by microglia causing them to become activated with the release of toxic cytokines, thereby perpetuating the vicious cycle. The release of different toxic cytokines from microglia may cause α-synuclein to form and fold differently to cause the known diversity of pathological species of α-synuclein. In addition, cytokine release might affect α-synuclein in nearby cells in a similar manner, thus facilitating the selective spread and propagation of α-synuclein pathology in vulnerable cell populations. This vicious cycle could explain why α-synuclein pathology developed in implanted dopamine neurons in Parkinson’s disease patients, how templating occurs in Parkinson’s disease, and how toxic cytokines may be involved in the spread of α-synuclein pathology.

Open in new tabDownload slide

Despite evidence of robust inflammation and microglial activation in graft deposit sites in our transplanted patients from the earliest time points evaluated, α-synuclein pathology was only observed many years afterwards. This suggests that initially, cellular clearance mechanisms may have been sufficient to clear misfolded proteins and prevent aggregation. Alternatively, cytokines released from microglia may initially have been present at too low a concentration to cause the accumulation and aggregation of misfolded proteins. However, over time, clearance capacities may be overwhelmed causing the accumulation and aggregation of the α-synuclein protein. Further, the accumulation of aggregated α-synuclein could accelerate the process, as it has been shown to interfere with both proteasomal and lysosomal clearance mechanisms (Snyder et al., 2003; Martinez-Vicente et al., 2008; Choi et al., 2015) and to promote the microglial inflammatory cascade (Hoenen et al., 2016).

It is also interesting to speculate on whether a similar sequence of events, but with a different initiating factor, could account for a ‘templating’ phenomenon in Parkinson’s disease and explain how aggregated α-synuclein could promote misfolding of the wild-type protein. Based on the observations in our present study, we propose that the following sequence of events might occur in Parkinson’s disease (Fig. 6): (i) α-synuclein misfolds and accumulates (possibly due to genetic, toxic, inflammatory, or stochastic causes); (ii) misfolded α-synuclein is released from cells; (iii) extracellular α-synuclein is taken up by microglia; (iv) microglia become activated; (v) cytokines are released from activated microglia; (vi) cytokines cleave/truncate host α-synuclein; (vii) truncated α-synuclein misfolds and aggregates; and (viii) misfolded α-synuclein continues to accumulate.

Thus, a vicious cycle might occur in Parkinson’s disease in which misfolded α-synuclein promotes misfolding of native α-synuclein ultimately leading to the development of Lewy pathology and neurodegeneration (Fig. 4). Indeed, there is evidence of inflammation in sporadic Parkinson’s disease as well as in asymptomatic patients with incidental Lewy bodies (McGeer et al., 1988; Münch et al., 2000; Doorn et al., 2014). Toxic cytokines generated from activated microglia could also extend to affect other vulnerable nerve cells and thus promote the spread of α-synuclein pathology. That native α-synuclein is critical for this templating phenomenon to occur is illustrated by the observation that α-synuclein fibrils do not induce Lewy pathology in α-synuclein null animals (Luk et al., 2012).

The number of transplanted brains we have been able to study is limited, but they represent the largest number of brains that have been examined from patients who have undergone a transplant procedure to date. Unfortunately, the examinations we were able to perform were restricted by tissue availability, so we were not able to assess cytokines or other inflammatory markers. However, these cases are rare, and the material available provides a unique opportunity to examine the temporal relationship between inflammation/microglial activation and α-synuclein pathology in Parkinson’s disease patients. Our findings are also of potential clinical importance as they suggest that this mechanism could account for the templating phenomenon that underlies the prion phenomenon, and that interventions that prevent microglial activation or inhibit cytokines such as caspase-1 or calpains might prevent the continued production of misfolded protein, and act as disease-modifying agents in Parkinson’s disease (Games et al., 2014; Olanow and Kordower, 2017).

Funding

The experiments reported were funded from a grant to J.H.K. from the Parkinson’s Foundation.

Competing interests

C.W.O. has no direct conflict of interest with the present study. He owns stock in Clintrex, a pharmaceutical advisory company. J.H.K., M.S., G.H. and Y.C. have no direct conflict of interest with the present study.

References