α-Synuclein Responses: Implications for Early Appearance of Sleep Disorders in Parkinson’s Disease

Sleep Disorders (SDs) precede motor symptoms of Parkinson’s disease (PD), suggesting an early effect of disease processes on sleep control neurons. PD processes involve rises in the protein, α-synuclein, which presents early on in a simple, monomeric form, but later in disease progression, a more complex bril form appears. We hypothesize that monomeric α-synuclein has deleterious cellular actions on sleep control nuclei. We monitored cellular responses to identied monomeric and bril α-synuclein in two sleep controlling nuclei, the laterodorsal tegmentum, and the pedunculopontine tegmentum, as well as the substantia nigra, a motor control nucleus which degenerates as a hallmark PD feature. We monitored differential cell death using a uorescent-based assay following exposure to the simpler form of α-synuclein. In sleep control nuclei, both forms of intrinsic α-synuclein induced excitation, and increased intracellular calcium and the monomeric form heightened putatively excitotoxic, neuronal death, whereas, in the substantia nigra we saw inhibition, decreased intracellular calcium and monomeric α-synuclein was not associated with heightened cell death. These nucleus-specic differential effects suggest previously unappreciated, mechanistic underpinnings of SDs’ prodromal PD appearance in PD, and we hypothesize that in the prodromal phase of PD, the early form of α-synuclein compromises sleep-control neurons. axis the percentage of tested and the axis label, Polarity indicates the polarity (inward or outward) of the current or increase or decrease in calcium elicited in the population of responding cells, respectively. Statistical analyses of data were performed in Prism (version 7.0, GraphPad, USA). Results are presented as mean values ± SEM with the degrees of freedom (df). The gures were prepared using Igor Pro software, Graphpad Prism and Matlab R2018b. The n reported reects numbers of single neurons recorded and represents neurons collected from at least 3 different animals which sourced from different litters. Sample sizes were determined from past experiences with electrophysiological recordings in order to limit unnecessary loss of animal life in adherence to 3R guidelines. Differences in numerical data were tested using a one-way ANOVA with Tukey’s posthoc test, or a two-way, Paired or Unpaired Student’s t-test, and reporting of the f or t statistic is included as per APA style guidelines. Differences in categorical data were examined using the Fisher’s exact test, two-way Chi Square test, or a three-way Chi Square test that in cases of overall signicance, was followed by a post hoc Fisher’s exact test to identify which contingency table cells differed signicantly from expected outcomes. The alpha level for statistical signicance was set at 0.05, and the exact p value is presented.


Background
Elucidating the neuronal mechanism of non-motor symptoms which precede development of motor disability in Parkinson's disease (PD) offers the hope of earlier disease detection and treatment. Appearance of a sleeping disorder is currently the most speci c, and powerful prodromal marker for agerelated, α-synucleinopathies such as PD (Postuma et al., 2015;Barber et al., 2017), as these disorders appear up to a decade earlier than motor symptoms. PD movement symptoms are associated with substantia nigra (SN) degeneration and abnormal aggregation of the protein, α-synuclein (α-syn), which changes structure from simple monomers, to more complex aggregates known as oligomers (Bengoa-Vergniory et al., 2017) and, ultimately, to brils (Fink, 2006). However, SN degeneration is unlikely to be the cause of the appearance of sleeping disorders (SDs) as they precede the movement impairments of PD. A more likely explanation is that as α-syn levels rise (Wang et al., 2012), α-syn-mediated neurodegeneration starts earlier in sleep and arousal control nuclei than in motor control centers.
The most common SDs in PD encompass activity changes within two state-regulating, pontine brain stem nuclei: the Laterodorsal Tegmental Nucleus (LDT), and Pedunculopontine Tegmental Nucleus (PPT) (Fig. 1A), which, interestingly, are profoundly degenerated in post mortem brains of PD patients (Braak et al., 2004;Müller and Bohnen, 2013). Additionally, neurodegeneration in pontine brain stem related to αsyn inclusions has been reported in individuals who were asymptomatic for PD, but diagnosed with REM Sleep Behavior Disorder (Uchiyama et al., 1995;Schenck et al., 1996, Boeve et al., 2007b. REM Sleep Behavior disorder is a SD shown to be prodromal to appearance of PD, and believed to involve dysregulation in the pons (Schenck et al., 1996, Boeve et al., 2007a. When taken together, we hypothesized that α-syn alters neuronal signaling in the LDT and PPT. Given the appearance of SDs prior to motor complaints, we further hypothesized that the LDT and PPT exhibit a greater sensitivity to actions of the simpler forms of α-syn than does the SN. To examine these hypotheses, we investigated whether native monomer, as well as brillated forms of highly puri ed α-syn had effects on LDT and PPT neuronal function, calcium dynamics, and whether exposure to α-syn elicited cell death. Further, we compared these effects to those in SN neurons.

Materials And Methods
Animals: NMRI mice (Harlan Mice Laboratories, Denmark) aged 12 to 30 days were used under approval in accordance with European Communities Council Directive (86/609/EEC). For calcium imaging analysis in the SN and PPT, after determination that there were no signi cant sex-based differences in the uorescence change amplitudes, data from males and females were used (SN: 1 female; PPT: 1 female); whereas, all LDT imaging data sourced from males. The animals were housed with a lactating female under controlled temperature (22-23 °C) and humidity conditions (45-65%) in a 12:12 h light-dark cycle, with water and food available ad libitum.
Tissue preparations for electrophysiology and calcium imaging: Following deep anesthesia with inhalation of iso urane (Baxter A/S, Denmark), a block of the brain containing the LDT, PPT, or SN was excised (Fig. 1A1, 2). The brain block was sectioned into 250 µm slices using a vibratome (Leica VT1200S, Leica Biosystems, Germany) in ice-cold arti cial cerebrospinal uid (ACSF Recombinant α-syn and size exclusion chromatography (SEC): Human α-syn was recombinantly expressed and puri ed as previously described (van Maarschalkerweerd et al., 2014) with few modi cations. In brief, α-syn was cloned into E. Coli BL21DE3 cells using a pET-11a vector construct. Harvested cells were lysed by osmotic shock and non-heat-stable proteins were removed by boiling and centrifugation. α-syn was isolated by ion-exchange chromatography and the monomer was isolated by SEC. Monomer fractions from SEC were pooled and kept in PBS buffer stored at -80 o C.
Fibrillation assay: α-syn in PBS buffer (1.5-2.2 mg/ml) was brillated in a microplate reader (FLUOstar Omega. BMG Labtech) at 37 o C with 3 mm sterile glass beads and orbital shaking (700 rpm for 280 s in each 360 s cycle). Samples were aliquoted (150 µl) in a NUNC 96-well optical polymer-based, clearbottom black plate (Thermo Fisher scienti c 265301) and sealed with clear polyole n tape (Thermo Fisher scienti c 232702). Thio avin T (ThT; 20 µM) was added to three samples, retaining in parallel 3 non-ThT containing samples for the electrophysiology experiments. ThT emission was recorded at 480 ± 10 nm upon excitation at 450 ± 10 nm. Fibril samples were collected after 7 days, to ensure full bril maturation, and thereafter stored at room temperature. ThT uorescence for the 7 day brillation assay are shown in Fig. 1B1 as the mean of a triplicate run, with insert detailing the rst 24 h. Circular Dichroism (CD): The secondary structure of α-syn samples was investigated to con rm monomeric and brillar states. CD was measured on monomer and brillated α-syn samples, respectively, using a JASCO J-1500 CD spectrometer. Samples were diluted to a nal concentration and placed in a cuvette with a light path of 1 mm. Spectra were collected from 260 − 190 nm, with a bandwidth of 1 nm, scanning speed of 100 nm/min, and a temperature of 25 o C. The nal specters were made as accumulations of three. In Fig. 1B2, CD spectra of the secondary structure of monomeric and brillated αsyn (0.15 mg/ml) are shown indicating random coiling of monomeric α-syn, consistent with the native, intrinsically disordered form, and β-sheet structure after brillation. α-syn, lysozyme, AMPA preparation and application: The highly puri ed monomeric form of α-syn was stored in solution at -80 o C in aliquots of 10 µl (150 µM) until use. Aliquots (150 µM) of α-syn brils were kept at room temperature and used within a maximum of 2 weeks. In initial experiments we found that 70 nM of α-syn induced a short-duration, membrane response; whereas, in the same cell, 100 nM induced a longer-lived shift in current amplitude, therefore, we decided to continue with 100 nM, which had been used in other studies (See Supplementary Fig. 1). The nal concentration of α-syn diluted in ACSF was applied for 3-4 min after the establishment of baseline conditions. As lysozyme exhibits a similar molecular weight as α-syn (14 kDa), it aggregates and forms brils in solution, and it has been extensively utilized as a control for non-speci c peptide effects in α-syn studies ( Fig. 1C4). Repeatable effects of α-syn in individual neurons indicated that membrane effects were not due to acute excitotoxicity (n = 3; data not shown). Recovery to baseline and rises in calcium induced by a subsequent application of 1 µM AMPA (αamino-3-hydroxy-5-methylisoxazole-4-propionic acid; 3 ml) con rmed that changes in calcium by α-syn were not due to acute cell death (n = 5; data not shown).
Patch clamp recordings: Borosilicate glass patch pipette electrodes were fabricated using a horizontal puller (P-97, Sutter Instruments, USA), and lled with an intracellular solution containing in mM: 144 Kgluconate; 2 KCl; 10 HEPES; 0.2 EGTA; 5 Mg-ATP and 0.3 Na-GTP, with a resulting pipette resistance of 6-11 MΩ. The brain slice was placed in the recording chamber and 1.2 ml/min of carbogenated ACSF was continuously perfused. Cells were visualized with a 60 x water immersion objective coupled to an upright microscope (BX50WI, Olympus; Japan) with an infrared Dodt gradient contrast system (IR-DGC; Luigs & Neumann, Germany) and a CCD camera (CCD-300ETRC; DAGE-MTI, Michigan City, IN). A patch clamp EPC9 ampli er (HEKA, Germany) and the software Pulse (HEKA; version 13.0) were used in voltage clamp mode and the holding voltage was maintained at -60 mV. Following membrane breakthrough, and at least a 7 min stabilization period, data were collected. Recordings of membrane currents were sampled at a rate of 1 kHz using AxoScope 10.2, and an Axon miniDigi 1B digitizer (Molecular Devices Corporation, USA). Combined patch clamp recordings and calcium imaging were conducted with a different setup in voltage clamp mode with a EPC9 ampli er guided by Pulse (version 9.0); and, membrane current recordings were sampled at 10 kHz using AxoScope 10.2 (Molecular Devices Corporation, USA) and an Axon Digidata 1440A digitizer (Molecular Devices Corporation, USA).
Identi cation of cell Phenotype: For post hoc identi cation of the recorded cells as being present within the perimeter of the cholinergic LDT and PPT, or to phenotypically-identify recorded neurons, in some recordings, Alexa-594 was included in the intracellular solution, and immunohistochemistry for brainderived nitric oxide synthase (bNOS), which is a validated marker of cholinergic neurons in the mouse LDT and PPT, was performed (Veleanu et al., 2016). After the recording, slices were submerged in 4% paraformaldehyde and stored overnight. Slices were then cryoprotected by saturation in a 30% sucrose solution and resectioned. Immunohistochemical labeling for bNOS was conducted following previously reported protocols on the resectioned tissue (Veleanu et al., 2016). Figure 1E  Data analysis and statistics: Calcium imaging data were analyzed in Igor Pro 6 (Wavemetrics, USA). Response amplitude and polarity were quanti ed with changes in uorescence below 4% considered too small to reliably discern from background noise. The ratiometric measures of changes in uorescence are shown in graphs as %DF/F. DF/F is the difference between the average of a series of data points at the maximum change in uorescence following drug application and the average baseline uorescence (F) relative to F, with ascendant de ection indicating intracellular calcium elevation. Amplitudes of membrane currents were measured (the difference between baseline and maximum de ection) by using AxoScope 10.5 (Molecular Devices, USA). In Figs. 1 and 2, the axis label, Response Frequency equals the percentage of tested cells which responded with membrane current or changes in calcium, and the axis label, Evoked Polarity indicates the polarity (inward or outward) of the current or increase or decrease in calcium elicited in the population of responding cells, respectively. Statistical analyses of data were performed in Prism (version 7.0, GraphPad, USA). Results are presented as mean values ± SEM with the degrees of freedom (df). The gures were prepared using Igor Pro software, Graphpad Prism and Matlab R2018b. The n reported re ects numbers of single neurons recorded and represents neurons collected from at least 3 different animals which sourced from different litters. Sample sizes were determined from past experiences with electrophysiological recordings in order to limit unnecessary loss of animal life in adherence to 3R guidelines. Differences in numerical data were tested using a one-way ANOVA with Tukey's posthoc test, or a two-way, Paired or Unpaired Student's t-test, and reporting of the f or t statistic is included as per APA style guidelines. Differences in categorical data were examined using the Fisher's exact test, two-way Chi Square test, or a three-way Chi Square test that in cases of overall signi cance, was followed by a post hoc Fisher's exact test to identify which contingency table cells differed signi cantly from expected outcomes. The alpha level for statistical signi cance was set at 0.05, and the exact p value is presented.
3 Results α-syn brillation and characterization α-syn monomer fractions were frozen in solution directly from size exclusion chromatography to avoid aggregate formation. A Thio avin T (ThT) assay was used to monitor the brillation of α-syn, and displayed a typical sigmoidal curve indicating bril formation (Fig. 1B1). The secondary structure of αsyn was investigated with circular dichroism (CD) showing the expected random coil for monomeric αsyn and β-sheet structure for the bril form (Fig. 1B2).

Membrane Responses
The monomeric and bril forms of α-syn induced membrane currents in the majority of LDT, PPT and SN neurons. Surprisingly, the polarity of the elicited currents was opposite in the sleep-regulating LDT and PPT to those elicited in the motor-controlling SN. Inward currents were induced in the sleep-regulating nuclei, whereas, outward currents were elicited in SN neurons (Fig. 1C).  Fig. 2A).

Neurotoxicity assays
As we hypothesized that continuous exposure to early forms of α-syn could have adverse effects on sleep control but not motor control neurons, we examined neuronal death induced by relatively brief, but continuous exposure to α-syn. Accordingly, LDT and SN slices were incubated in monomeric α-syn (100 nM) or control solution for 7 ½ hours and cell viability fractions were calculated as the ratio of living cells (DAPI-positive) to the total number of cells (DAPI-positive + PI-positive). Relative to control slices, αsyn-exposed LDT slices exhibited a greater proportion of dead cells (Control: n = 4; α-syn: n = 7; Unpaired Student's t-test, t(9) = 2.52, p = 0.03; Fig. 2B) and fewer surviving cholinergic cells (Fig. 2C). However, SN slices showed no indication of heightened cell death over control (Control: n = 8; α-syn: n = 7; Unpaired Student's t-test, t(13) = 0.04, p = 0.96, Fig. 2B).
Single cell recordings of membrane effect, calcium in ux and real-time, cell viability assay As we observed effects on membrane currents, calcium levels, as well as heightened cell death by monomeric α-syn in separate populations of cells in sleep control nuclei, we next examined whether those effects co-occurred in the same cell. Thus, in LDT slices exposed to a high concentration of monomeric αsyn (500 nM), we tracked neuronal viability in real-time using PI uorescence and, within the same neuron, we monitored the membrane current and intracellular calcium response. As control, the same protocols were conducted in the other half of the slice, which was not exposed to α-syn. In all neurons tested, α-syn induced inward currents (-140.5 ± 11.3 pA, n = 4), and an increase in intracellular calcium levels (39.7 ± 10.7% DF/F, n = 4), which were signi cantly different from control conditions (Fig. 3A, B, C).

Conclusions
The (Relevance to SN, see Supplemental Material).
Using multiple, naturally-occurring isoforms of α-syn, which were carefully prepared and validated, our study is the rst to report membrane effects of pure monomeric and con rmed bril forms of α-syn in native, mammalian neurons when extracellularly applied; moreover, our ndings indicate the toxic potential of the monomeric form of this protein. Furthermore, our data document a functional difference in cellular responses to α-syn between LDT, PPT and SN neurons, thusly providing the rst data for a mechanistic explanation regarding why SDs precede motor symptoms in PD. These results should open the path to investigations leading to a deeper understanding of the cellular mechanisms underlying α-syn effects resulting in non-motor symptoms of PD, which could serve as targets for earlier disease intervention.   Fluorescent images of LDT (B1) and SN brain slices (B2) exposed to monomeric α-syn showing dead cells in red (PI-positive, left) and live cells in blue (DAPI-positive, middle) with the two elds overlaid (right). As shown in the bar graphs to the right, there was a signi cantly greater percentage of cell death over control in the population of LDT cells exposed to monomeric α-syn, whereas in the SN, cell death was not signi cantly different between control and α-syn-exposed slices. C) Quanti cation of bNOSpositive cells in bisected slices revealed a smaller number of cholinergic neurons within α-synM-exposed half slices (C2) when compared to the number in the equivalent region in the other half of the slice incubated in control solution (C1) for the same time period. (C1, 2). A count of bNOS-positive cells revealed 14.7% fewer cholinergic cells in the α-synM-incubated, hemi-slices (200 nM; n=2) relative to control, and 47.3% fewer when incubated in 500 nM (n=2), which constituted a signi cant difference (Paired Student's t-test; df: 6; p = 0.04 * indicates p < 0.05). These data are consistent with current hypotheses regarding the role played by pontine cholinergic dysfunction in prodromal, or early-appearing, non-motor symptoms of PD (Müller and Bohnen, 2013).