Accumulating evidence indicates that misfolded protein pathologies can spread throughout the nervous system in a prion-like fashion [2, 17]. It appears that neuronal connections are more important than physical distance for the propagation of a-syn deposits among brain regions. The hypothesis that synaptic transmission of a-syn seeds is responsible for the stepwise expansion of Lewy lesions has been proposed.
In the previous reports, this hypothesis has been discussed from a neuropathological point of view, which stipulated that a-syn accumulation only occurs in the regions where anatomical neural connections exist [4, 17, 19, 25, 28, 29, 33, 40, 41, 49]. Holmqvist et al. have reported that a-syn propagated to the brain stem through the vagus nerve from the intestinal tract [19]. Rey et al. reported that monomers and oligomers of human a-syn injected into the olfactory bulb of mice were transported along the axonal pathways within a few hours [40]. Furthermore, Peelaerts et al. observed that dopaminergic neurons take in all types of a-syn including oligomers, fibrils and ribbons to complete the trans-synaptic transportation [35].In this study, we used callosotomy, a different experimental method from the previous reports, to investigate whether a-syn seeds are specifically transmitted across neural connections, and we experimentally demonstrated the speed of a-syn transmission.
First, in order to confirm the direction of a-syn seeds axonal transport and propagation, we analyzed the accumulation of p-syn in the input and output fibers of the striatum, into which the a-syn seeds were injected. We confirmed that injected a-syn seeds were incorporated into the MSNs of the striatum, to which the a-syn PFFs were administered, and showed the formation of p-syn inclusion bodies. In the neurofilament H-positive axons projecting from the cortex into the striatum, TH-positive axons projecting from neurons in the SNpc, as well as β4- or Nav 1.2-positive fibers projecting from the striatum to the globus pallidus and SNr, thread-like inclusion bodies were detected by anti-p-syn antibody. These results strongly suggest that bidirectional (anterograde and retrograde) transport of seeds occurs in axons from the striatal MSNs, or axons projecting into striatum (Fig. 1a-c).
In our experiments, there were no p-syn inclusions observed inside astrocytes in any of the images. Loria et al. demonstrated that astrocytes degrade a-syn immediately after uptake [23]. Thus it may have been difficult to detect astrocytic p-syn inclusions in the images, obtained at a single time point. In addition, the ability of neurons to degrade a-syn may be lower than that of astrocytes. This presumably indicates that a-syn tends to be accumulated by neurons.
We analyzed the propagation of pathological a-syn over time after the injection of a-syn PFFs into the right striatum. The results showed that accumulation of pathological a-syn tended to increase over time. We observed spreading in the contralateral striatum and in regions connected to the striatum through multiple synapses, such as the contralateral striatum and SN, although the amount of spreading decreased. These data suggest that the transmission of a-syn amounts depended on connectivity. Paumier et al. reported a reduction of the a-syn pathology in SNpc within 180 days after administration of the a-syn fibrils, based on the observation of neuron degeneration [33]. Similarly, Rey et al. observed that the density of the pathology as a whole in the brain tends to decrease over the longer term after administration of the a-syn PFFs [39]. In the present study, brains of mice were analyzed over a period ranging from 0.75 months to 6 months after administration of a-syn PFFs, the p-syn positive deposits in the SN increased within 6 months. The difference could likely be attributed to factors such as the adjustment of synthetic recombinant a-syn. However further longer-term observation is warranted to assess this hypothesis.
Moreover, analysis of the areas and quantities of individual p-syn deposits over time revealed the size of each p-syn deposit to be larger on ipsilateral side as compared to the contralateral side. However, the maximum size of the deposits was equivalent on both sides (Additional file 1: Figure S3). These results suggest that the injected a-syn PFFs initially spread to each area as seeds and formed inclusions over time by recruiting endogenous a-syn. P-syn accumulation in the ipsilateral cortex is the most frequent, and transmission to the striatum and contralateral SN to the site of a-syn PFFs administration tended to be less frequent. This is presumably because the transmission of seeds via multiple synapses is required to reach the striatum and SN on the contralateral side, and the number of seeds declines during transmission.
We further confirmed that a-syn seeds are transmitted through neural circuits using callosotomy (Fig. 3). We designed an experiment using callosotomy to determine whether nerve fiber disconnection inhibits a-syn propagation, in a neural circuit-dependent manner. When callosotomy disconnected the contralateral side from the injected side before the injection of a-syn seeds, the transmission and propagation of pathological a-syn to the contralateral side substantially decreased. This is likely because the delivery of the a-syn seeds to the contralateral side was blocked by severing the axons of the corpus callosum. From this result, we confirmed that the seeds were transported along the path of the nerves. Some propagation of p-syn deposits in the limbic system (EC and Amyg), including routes other than the corpus callosum, such as hippocampal traffic and the anterior commissure, also seem likely to be involved [8], however, the decreased p-syn deposits in these regions after callosotomy suggest that the transmission through the striatum may affect the results through the striatum-Amyg and Amyg-EC connection. In contrast, when the callosotomy was performed 24 h after injection of the a-syn seeds, p-syn accumulated in the contralateral side in a similar fashion as in the control (without callosotomy). This result suggests that exogenous seed migration occurs within a 24-h period.
We also examined the dynamics of the seeds transmission itself. Human a-syn PFFs can be differentiated from mouse a-syn PFFs by the human a-syn-specific antibody LB509 (Fig. 4a). Exogenous human a-syn PFFs injected into the right striatum spread to the contralateral side, in the cortex, striatum, Amyg, and EC, forming visible aggregates (inclusions) 3 weeks (0.75 months) after injection. However, 12 weeks after seed administration, the exogenous human a-syn deposits were no longer detected. Meanwhile, endogenous mouse a-syn inclusions began to appear (Fig. 4b). These results suggest that exogenous human seeds interact with each other and bind other human seeds more rapidly than they convert endogenous mouse a-syn to the misfolded form, due to conformational and species-specific sequence differences (Additional file 1: Figure S10).
The administration of human a-syn PFFs resulted in a slower and reduced formation of a-syn inclusions compared with the administration of mouse a-syn PFFs. This observation has been previously reported [41]. As suggested in previous studies, the species barrier may be the reason for this effect [6]. This observation was also noted by Rey et al. [41]. Further, Luk et al. proposed that the seeding efficiency on a molecular level is determined by the sequence homology of the a-syn PFF seeds and the soluble a-syn monomer, and the result of the present study which showed delayed propagation ability when human a-syn PFFs was administered to mice is also consistent with this proposal [24]. In the present study, the human a-syn deposit itself was undetectable 3 months after seeds administration. We suppose that the majority of exogenous seeds were degraded when the human a-syn deposits disappeared. The later disappearance of human a-syn deposits was also confirmed by the previous studies [2, 3, 24, 29, 39, 41]. Rey et al. stated [39] that this elimination could be due to the degeneration of cells itself, or degradation caused by the autophagy/lysosomal, ubiquitin-proteasome systems [48], or phagocytosis by microglia [7] or astrocytes [23].
Interestingly, the exogenous seeds were not phosphorylated, while the endogenous a-syn aggregates were phosphorylated, suggesting that the exogenous seeds are not phosphorylated inside the cell and can recruit and convert endogenous p-syn more easily than the non-phosphorylated form. Sections of SN from autopsied brains of patients with Parkinson’s disease were double stained with two antibodies, LB509 and phospho S129. Stained Lewy bodies were detected by both LB509 and phospho S129 and were colocalized. Thus, it was inferred that these antibodies did not compete for their epitopes, and this confirmed that the human a-syn PFFs was not phosphorylated. We also examined the spread of the seeds themselves and their speed using callosotomy with the administration of human a-syn PFFs (Fig. 4c, d). When the callosotomy was performed prior to the administration of human a-syn PFFs, transmission to the contralateral side decreased and the emergence of mouse p-syn deposits also decreased. When the callosotomy was conducted 1 day after the injection of human a-syn PFFs, transmission of human a-syn to the contralateral side was observed, and the emergence of mouse p-syn was also observed. From these results, we confirmed that the spread of the exogenous seeds occurs within 24 h. This rapid dissemination/transmission of a-syn phenomenon is very surprising and raises several questions. In clinical cases of PD and dementia with Lewy bodies (DLB)/PD with dementia (PDD), a-syn pathologies spread slowly over a period of 5–10 years [4, 5, 18]. Incidentally, in reports of Lewy body pathology in fetal SN transplants, which provide important evidence for a-syn transmission, it is estimated that Lewy bodies cannot be positively detected until more than 10 years after transplantation [21, 22]. There are several possible explanations for the difference in progression from around 24 h to over 10 years, between experimental transmission systems and the clinic. First, it is plausible that administering artificial protein aggregates to the experimental transmission systems creates a unique situation which does not necessarily reflect the clinical disease. Alternatively, it could be that the seeds themselves, which form the initial explosive trigger for the transmission and aggregation of the pathological protein, can spread within 24 h, while it takes more time for the exogenous seeds to recruit endogenous proteins by altering their conformation. This could represent a novel pathological concept in the disease progression of synucleinopathies. Indeed, since the transmission experiments involve injecting a large dose of seed proteins directly into the brain, they could be viewed as a way to capture the clinical pathological condition in a shorter period of time. Meanwhile, prion disease has been shown to spread extremely quickly after onset, both experimentally and clinically [27]. This difference could arise from the fact that prion disease causes changes in the cell membrane [36] which lead to a rapid worsening of disease, while synuclein causes changes within the cytoplasm [20, 45, 47] which may lead to slower progression.
In recent years, the propagation of pathological a-syn to anatomically adjacent neurons or cells has been observed in other studies as well, further supporting the trans-neuronal and trans-synaptic transportation of a-syn seeds [4, 9, 19, 28, 33, 35, 40, 41]. However, a study using cultured primary neurons reported that the transmission of pathological a-syn may not necessarily occur via synapses. Indeed, the transmission between axons and soma has been observed during the early culture phase (1 to 4 days), prior to the formation of synapses [11]. There are additional studies that report mechanisms other than synaptic transmission, such as transport via nanotubes [1], exocytosis or exosomes, or by receptor-mediated endocytosis [10, 17, 38]. However, our results strongly suggest that a synaptic mechanism underlies the trans-neuronal transmission of a-syn seeds in-vivo. Understanding the pathways involved in the pathological propagation of proteins is an important step for developing therapeutic interventions. The results of the present study using callosotomy suggest that the rapid dissemination of seeds through synapses does indeed occur. Finally, we used BoNT/B to investigate whether the transmission of seeds was inhibited by the cessation of synaptic release (Fig. 5). BoNT blocks synaptic vesicle fusion in the presynaptic terminal by breaking down and inactivating SNARE proteins in a highly specific manner (VAMP-2 by BoNT/B). BoNT has been shown to block vesicle exocytosis, but not endocytosis [31]. The role of VAMP-2 in the presynaptic SNARE complex that mediates vesicle fusion is widely accepted [46]. In our present study, BoNT/B injected into the striatum inhibited the exocytosis of a-syn seeds from the synapses of axons entering the striatum from the cortex. Transmission and propagation of pathological a-syn in the contralateral striatum and Amyg were inhibited. As a result, we observed that BoNT/B treatment reduced accumulation of p-syn in an area in which the seeds transmission should occur, confirming the presence of trans-synaptic transmission using an in vivo model, for the first time. Previously, it was reported that BoNT inhibited the spread of huntingtin protein in primary cultures, but the actual substances used were β-bungarotoxin [34], at least as far as the product numbers suggest. Thus, ours is the first study to definitively demonstrate trans-synaptic transmission of a pathological protein using BoNT in vivo. In recent years, the association of inflammation with PD or other diseases has been discussed. A study reported that inflammation alters neuronal functions, leading to an increase in cell death [43] and another study reported that inflammatory environments enhance a-syn spreading [13]. The possibility that inflammation was caused by a surgical procedure such as callosotomy and injection of BoNT cannot be denied. Thus in this study, we performed callosotomy or injection of BoNT before or after administration of a-syn PFFs. When we performed callosotomy or injection of BoNT before a-syn PFFs administration, the spreading of a-syn inclusions decreased. However, when the administration of a-syn PFFs was performed after callosotomy or injection of BoNT, spreading of a-syn inclusions occurred equally as that with a-syn PFFs administration alone. Therefore, we concluded that inflammation had little effect in this study. Combined with the results of the experiments using callosotomy, we confirmed the rapid transmission and dissemination of a-syn seeds through circuit and synapses. Our findings provide empirical support for a dissemination mode (Fig. 6b) of a-syn propagation, where a-syn seeds are disseminated and then induced to form aggregates, and not a step-by-step mode (Fig. 6a) where each neuron forms inclusions and releases seeds. Typical PD shows only limited affected regions and usually has a slowly progressive course. However, the distribution and rapid dissemination of a-syn seeds observed in this study seems to represent a different point in the recognition of PD, beyond its usual disease course. The dissemination mode may be a mechanism underlying the pathology of dementia with Lewy body disease, in which larger regions are affected in the cortex, including in the rapidly progressive phenotype [12]. Further studies will be necessary to elucidate this regulatory mechanism and to develop therapeutic strategies for synucleinopathies.