Inoculation of α-synuclein preformed fibrils into the mouse gastrointestinal tract induces Lewy body-like aggregates in the brainstem via the vagus nerve

Mouse α-Syn PFFs were generated as described previously with minor modifications [18]. Escherichia coli BL21 (DE3) (BioDynamics Laboratory) were transformed with plasmid pRK172 encoding the mouse α-Syn cDNA sequence and incubated in LB medium to an optical density of 0.3 at 600 nm. α-Syn expression was induced by 0.1 mM isopropyl β-D-1-thiogalactopyranoside for 4 h. The bacteria were pelleted by centrifugation at 4000 g at 4 °C for 5 min and lysed by freeze/thaw and sonication. The lysate was clarified by boiling for 5 min, followed by centrifugation at 20,400 g at 4 °C for 15 min. The supernatant was subjected to ion exchange using Q sepharose fast flow (GE Healthcare), and α-Syn was precipitated with 50% (% saturation) ammonium sulfate. Purified α-Syn was dialyzed against dialysis buffer (150 mM KCl, 50 mM Tris-HCl, pH 7.5) and cleared by ultracentrifugation at 186,000 g at 4 °C for 20 min. The protein concentration was determined with a Pierce BCA Protein Assay kit (Thermo Fisher). Purified α-Syn was diluted in dialysis buffer containing 0.1% (w/v) NaN₃ to 7 mg/ml, followed by incubation at 37 °C in a shaking incubator (AS ONE, SI-300C) at 1000 rpm for 10 days. The α-Syn PFF pellet was obtained by ultracentrifugation at 186,000 g at 20 °C for 20 min and stored at − 80 °C until use. The pellet was dissolved in 8 M guanidine hydrochloride to determine the protein concentration with a Pierce BCA Protein Assay kit (Thermo Fisher). The α-Syn PFFs in phosphate-buffered saline (PBS) (2 μg/μl) were sonicated with an ultrasonic wave disruption system (Cosmo Bio, Bioruptor) for 2 min before inoculation into the mouse gastric wall.

For sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE), sample buffer (1% [w/v] SDS, 12.5% [w/v] glycerol, 0.005% [w/v] bromophenol blue, 2.5% [v/v] 2-mercaptoethanol, 25 mM Tris-HCl, pH 6.8) was added to α-Syn monomer solution or the α-Syn PFF pellet, which was resuspended by vortex. Samples containing 10 μg protein were loaded in each lane and separated on 12% (w/v) gels for SDS-PAGE. For Coomassie Brilliant Blue (CBB) staining, gels were incubated in CBB staining solution (0.1% [w/v] PhastGel Blue R-350 [GE Healthcare], 30% [v/v] methanol, 10% [v/v] acetic acid) at room temperature. For western blot analysis, proteins were transferred to polyvinylidene difluoride membranes with a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad). The membranes were treated with 0.4% (w/v) paraformaldehyde (PFA) in PBS for 30 min at room temperature before blocking with 5% skim milk to prevent detachment of α-Syn from the blotted membranes [19]. The membranes were incubated with an anti-α-Syn primary antibody (BD Transduction, #610787 [Syn-1], 1:2000) at 4 °C overnight, followed by reaction with horseradish peroxidase-conjugated secondary antibodies (Santa Cruz, #sc-2005, 1:5000) for 1 h at room temperature. Immunoreactive bands were detected with Pierce ECL Western Blotting Substrate (Thermo Fisher), and the chemiluminescent signal was detected with an Amersham Imager 600 imager (GE Healthcare).

α-Syn monomers or PFFs, 7 μg each, were incubated in 250 μl of 5 μM Thioflavin T (Sigma-Aldrich, #T3516) solution for 15 min at room temperature. The fluorescence at 535 nm (excitation 450 nm) was measured with a multi-label plate reader (PerkinElmer, 2030 ARVO X). Thioflavin T solution alone was measured as a blank.

α-Syn PFFs, 5 μg, were placed on a 400-mesh carbon-coated copper grid (Nissin EM). The excess solution was removed with filter paper after the sample had stood for 1 min. The PFFs adsorbed on the grid were negatively stained with a 2% (w/v) uranyl acetate solution. Electron micrographs were acquired using a transmission electron microscope (JEOL, JEM-1400 Plus) at 80 kV.

Mice were anesthetized with Avertin (1.875% [w/v] 2,2,2-tribromoethanol, 1.25% [v/v] 3-methyl-1-butanol). A 2-cm incision was made in the abdominal midline, followed by inoculation of α-Syn PFFs (n = 48) or PBS (n = 3) into the gastric wall. Among the mice inoculated with α-Syn PFFs, 8 mice underwent right cervical vagotomy just prior to inoculation of α-Syn PFFs. Each of 8 sites was inoculated with 3 μl of α-Syn PFFs in PBS (2 μg/μl) or with PBS using a 37 gauge needle (Saito Medical Instruments) fitted to 10 μl syringe (Hamilton, #701LT).

Following anesthesia, a 1-cm incision was made at the midline of the mouse neck. The right vagus nerve was identified between the common carotid artery and the jugular vein behind the submandibular gland. After isolation from the carotid sheath, the vagus nerve was cut with a pair of tweezers.

Verification of vagotomy using FluoroGold was performed as described with modifications [20]. Five days prior to sacrifice, unvagotomized or vagotomized mice (3 mice per group) were injected intraperitoneally with 1.2 mg of hydroxystilbamidine (AAT Bioquest) in 1 ml of saline. Frozen sections were obtained as described below and observed with a BZ-× 710 fluorescence microscope (KEYENCE).

Frozen and paraffin sections were used for histologic and immunohistochemical analysis. Following perfusion with 4% (w/v) PFA in PBS, the brains were removed and immersed in 4% (w/v) PFA in PBS. For frozen sections, the brains were replaced in 15% (w/v) sucrose in PBS and subsequently in 30% (w/v) sucrose in PBS at 4 °C each overnight. The brains were embedded in Surgipath FSC 22 (Leica), and 10-μm sections were obtained with a CM1950 cryostat (Leica). For paraffin sections, the mouse brains were dehydrated and embedded in paraffin, and 8-μm paraffin sections were prepared with a HM 325 rotary microtome (MICROM). For immunohistochemical analysis, the following primary antibodies were used: anti–α-Syn (BD Transduction, #610787 [Syn-1], 1:1000), anti–p-α-Syn (Abcam, #ab51253 [EP1536Y], 1:10000), anti–p-α-Syn (Wako, #015-25191 [#64], 1:2000), anti–p-α-Syn (Abcam, #ab184674 [81A], 1:5000), anti-nitrated α-Syn (Santa Cruz, #sc-32,279 [Syn514], 1:200), anti-choline acetyltransferase (ChAT) (Millipore, #AB144P, 1:1000), anti-p62 (MBL, #PM045, 1:1000), anti-ubiquitin (DAKO, #Z0458, 1:500), anti-vasoactive intestinal polypeptide (VIP) (Abcam, #ab8556, 1:50), anti-nitric oxide synthase 1 (NOS1) (Santa Cruz, #A-11, 1:200), and anti-substance P (Millipore, #MAB356, 1:200). The sections were incubated at 4 °C with primary antibodies for 2 d and then processed for visualization. As secondary antibodies, Histofine (Nichirei Bioscience) was used for diaminobenzidine staining, and Alexa Fluor 488 or 594-conjugated antibodies (Molecular Probes) for immunofluorescence. For p-α-Syn and Thioflavin S (ThS, Santa Cruz, #sc-391005) double-labeling staining, after immunolabeled with p-α-Syn antibody, slides were incubated with 0.05% ThS in 50% ethanol followed by differentiation with 80% ethanol. For assessment of p-α-Syn pathology, every 10^th paraffin section was stained with anti–p-α-Syn antibody (EP1536Y). To assess ChAT-positive neurons in the dmX, every 10^th paraffin section was stained with anti-ChAT antibody. The numbers of p-α-Syn–positive and ChAT-positive neurons were manually counted. Sections were examined with a BX43 microscope (Olympus), a BZ-X710 fluorescence microscope (KEYENCE), and an FV-1000 confocal laser scanning microscope (Olympus).

One-way ANOVA with Tukey’s post-hoc test was used. Statistical calculations were performed with GraphPad Prism Software, Version 5.0.

We used mouse rather than human α-Syn PFFs in this study because mouse α-Syn PFFs are more potent for inducing α-Syn pathology and its propagation as compared with human α-Syn PFFs inoculated in wild-type mouse brains [13, 21]. First, we generated mouse α-Syn PFFs and validated their characteristics. SDS-PAGE followed by CBB staining or western blotting showed smear bands of α-Syn PFFs, suggesting α-Syn polymerization (Fig. 1a). High-intensity Thioflavin T fluorescence of α-Syn PFFs indicated that α-Syn PFFs comprised β-sheet–rich structures (Fig. 1b). Transmission electron microscopy revealed fibrillar structures in the α-Syn PFFs (Fig. 1c). We inoculated α-Syn PFFs or PBS into the gastric wall of the mice. To confirm that the PFFs properly inoculated into the gastric wall, we immunostained paraffin sections of the gastric wall with anti-α-Syn antibody. Deposition of inoculated α-Syn PFFs was found in the submucosal layer (Fig. 1d, e) and in the muscular layer (Fig. 1g, h) even 45 days after α-Syn PFFs inoculation. α-Syn immunohistochemistry also showed α-Syn–positive dots around the myenteric neurons, some of which are efferent vagus nerve terminals (Fig. 1f) [20].

Forty-five days after inoculation of α-Syn PFFs into the gastric walls, we examined the brains and stomachs of the mice. Immunostaining for p-α-Syn, a marker of human Lewy pathology, showed p-α-Syn–positive neurons in both rostral and caudal parts of the dmX in mice that had been inoculated with α-Syn PFFs (Fig. 2a–f). All those neurons had p-α-Syn in the cytoplasm, whereas a few also exhibited intranuclear p-α-Syn–positive dots (Fig. 2c). We confirmed that the pathology in the dmX were detectable with other anti–p-α-Syn antibodies and an anti-nitrated α-Syn antibody (Fig. 2g–i) [22]. However, p-α-Syn pathology were rarely observed in the stomachs 45 days after α-Syn PFF inoculation (Additional file 1: Figure S1), nor were they seen in either brains or stomachs of mice inoculated with only PBS (Additional file 2: Figure S2).

Next, we examined the pathological features of the p-α-Syn pathology in the dmX. We confirmed that it was present in ChAT-positive neurons of the dmX (Fig. 3a–d). The p-α-Syn pathology was positive for ThS, indicating that they were composed of β-sheet–rich amyloid fibrils (Fig. 3e–h). Most of the p-α-Syn pathology was also immunopositive for p62 (Fig. 3i–l) and ubiquitin (Fig. 3m–p), both of which are LB markers. All these observations indicated that the p-α-Syn pathology in the dmX shared properties in common with the LBs seen in patients with PD.

Anatomically, it was presumed that α-Syn PFFs inoculated into the gastric wall retrogradely induced p-α-Syn–positive LB-like aggregates in the dmX via the vagus nerve. To verify this, we inoculated α-Syn PFFs into the gastric wall immediately after hemivagotomy. We confirmed that vagotomy was successful using a retrograde neuronal tracer, FluoroGold. Intraperitoneal injection of FluoroGold labeled neurons in the dmX as has been reported in rats (Fig. 4a) [20]. When FluoroGold was injected after hemivagotomy, neurons on the vagotomy side were not labeled, indicating that the vagotomy had succeeded (Fig. 4a). However, it was reported that vagotomy causes neuronal loss over time in the murine dmX on the vagotomy side [23], so we analyzed some vagotomized mice 23 days after α-Syn PFF inoculation in addition to 45 days. The numbers of ChAT-positive neurons on the vagotomy side decreased to 40% and to 20% compared with those of untreated mice 23 days and 45 days after surgery, respectively (Fig. 4b, f). Although a certain number of ChAT-positive neurons survived on the vagotomy side, p-α-Syn–positive neurons were observed only on the unvagotomized side at both 23 and 45 days (Fig. 4b–e, g). This suggests that the vagus nerve indeed was the pathway by which α-Syn PFFs inoculated into the stomach resulted in formation of p-α-Syn aggregates in the dmX.

Then, we analyzed p-α-Syn pathology up to 12 months after α-Syn PFF inoculation. Unexpectedly, the number of p-α-Syn–positive neurons in the dmX bilaterally decreased over time (Fig. 5a). No further propagation of p-α-Syn pathology beyond the dmX was observed during this period. At 12 months postinoculation, there were only a small number of p-α-Syn–positive aggregates in the dmX, which appeared as denser inclusions than those observed at 45 days postinoculation (Fig. 5b). There were also a few p-α-Syn–positive neurons in the myenteric plexus at 12 months postinoculation (Fig. 5c). We also examined the thoracic spinal cord but observed no p-α-Syn pathology up to 12 months postinoculation (Additional file 3: Figure S3).

An autopsy study of patients with PD reported that most LBs in the ENS are present in VIP-positive neurons [24]. According to Braak’s hypothesis, Lewy pathology initially occurs in the ENS, suggesting that a specific type of neurons in the ENS may play a key role in the pathogenesis of PD. Finally, we therefore examined which types of myenteric neurons contained LB-like aggregates 12 months after α-Syn PFF inoculation. Although single markers cannot completely distinguish the various types of neurons, many VIP-positive neurons in the myenteric plexus are inhibitory motor neurons and are also immunopositive for NOS1 (Fig. 6a, b) [25, 26]. Excitatory motor neurons are immunopositive for ChAT and substance P (Fig. 6c, d). We conducted double immunofluorescent analysis using antibodies against these markers along with anti–p-α-Syn antibodies. A substantial portion of p-α-Syn aggregates were present in VIP- or NOS1-positive neurons (83% [10/12] and 67% [6/9], respectively; Fig. 6e–l), but some aggregates were also present in ChAT- or substance P-positive neurons (45% [6/13] and 33% [4/12], respectively; Fig. 6m–t). We therefore did not find cell-type specificity in neurons containing p-α-Syn–positive aggregates in the myenteric plexus in this mouse model.