Striatal Dopamine Transporter Function Is Facilitated by Converging Biology of α-Synuclein and Cholesterol

Striatal dopamine transporters (DAT) powerfully regulate dopamine signaling, and can contribute risk to degeneration in Parkinson’s disease (PD). DATs can interact with the neuronal protein α-synuclein, which is associated with the etiology and molecular pathology of idiopathic and familial PD. Here, we tested whether DAT function in governing dopamine (DA) uptake and release is modified in a human-α-synuclein-overexpressing (SNCA-OVX) transgenic mouse model of early PD. Using fast-scan cyclic voltammetry (FCV) in ex vivo acute striatal slices to detect DA release, and biochemical assays, we show that several aspects of DAT function are promoted in SNCA-OVX mice. Compared to background control α-synuclein-null mice (Snca-null), the SNCA-OVX mice have elevated DA uptake rates, and more pronounced effects of DAT inhibitors on evoked extracellular DA concentrations ([DA]o) and on short-term plasticity (STP) in DA release, indicating DATs play a greater role in limiting DA release and in driving STP. We found that DAT membrane levels and radioligand binding sites correlated with α-synuclein level. Furthermore, DAT function in Snca-null and SNCA-OVX mice could also be promoted by applying cholesterol, and using Tof-SIMS we found genotype-differences in striatal lipids, with lower striatal cholesterol in SNCA-OVX mice. An inhibitor of cholesterol efflux transporter ABCA1 or a cholesterol chelator in SNCA-OVX mice reduced the effects of DAT-inhibitors on evoked [DA]o. Together these data indicate that human α-synuclein in a mouse model of PD promotes striatal DAT function, in a manner supported by extracellular cholesterol, suggesting converging biology of α-synuclein and cholesterol that regulates DAT function and could impact DA function and PD pathophysiology.


INTRODUCTION
Striatal dopamine (DA) release is regulated by mechanisms that drive activity in midbrain DA neurons in conjunction with mechanisms in striatum that act on and within DA axons (Sulzer et al., 2016). DA uptake transporters (DATs) on DA axons clear striatal DA from the extracellular space to spatially and temporally limit DA signaling . In addition, it has become evident that striatal DATs also regulate the underlying process of DA release. Inhibition of DATs with pharmacological inhibitors not only limits the uptake but also promotes the amount of DA released (John and Jones, 2007) through a synapsin-dependent pathway (Venton et al., 2006;Kile et al., 2010) suggesting that the DAT limits vesicular mobilization for release. In addition, DATs regulate the short-term dynamic plasticity of DA release, promoting subsequent release at short inter-pulse intervals (IPIs) corresponding to high frequency firing, and limiting subsequent release at longer IPIs corresponding to low frequency firing (Condon et al., 2019).
Besides their role in regulating DA release and uptake, DATs are implicated in the etiology of Parkinson's disease (PD). DATs offer a route of entry to DA neurons-of environmental toxins e.g., MPTP, pesticides (Lehmensiek et al., 2006;Ritz et al., 2009), and higher DAT expression and DAT: VMAT ratios promote oxidative stress via dopamine oxidation and consequent dopamine neuron loss (Miller et al., 1999;Masoud et al., 2015). Variations in the DAT/SLC6A3 gene that are associated with PD susceptibility increase the expression of striatal DAT (van de Giessen et al., 2009;Richter et al., 2017). Furthermore, the DAT is electrogenic, offering depolarizing currents, which can be uncoupled from translocation of DA (Sonders et al., 1997), and so contribute to a metabolic burden, requiring ATP to re-establish ion gradients across the axonal membrane (Pissadaki and Bolam, 2013).
DAT function is regulated by post-translational modifications and lipid-and protein-binding partners (Vaughan and Foster, 2013), including α-synuclein (Lee et al., 2001) and cholesterol (Jones et al., 2012). DATs co-localize with α-synuclein in human post-mortem tissue (Bellucci et al., 2011) and many aspects of the DAT are reported to be affected by α-synuclein, including surface expression, internalization, DA translocation, and sensitivity to ligands. However, there is a lack of consensus about the impact of α-synuclein on DAT function, with conflicting evidence showing that DAT function can be enhanced (Lee et al., 2001;Chadchankar et al., 2011), decreased (Wersinger and Sidhu, 2003;Swant et al., 2011;Lundblad et al., 2012), or remains unaffected (Abeliovich et al., 2000). This lack of consensus might reflect the differing experimental preparations, assays and aspect of DAT function explored, and also might arise from the different structural forms (monomeric, oligomeric or fibrillary, soluble or aggregated) of α-synuclein that might be differently present (Alegre-Abarrategui et al., 2019) and have different pathophysiological outcomes. Determining the structural form and binding partners of α-synuclein is non-trivial and has been reviewed elsewhere (Alegre-Abarrategui et al., 2019;Alza et al., 2019). In addition, α-synuclein is in an equilibrium between cytosolic and membrane-bound states and cholesterol has been shown to affect the position of this equilibrium (Man et al., 2020). Cholesterol is a key component of highly curved membranes including synaptic vesicles, which α-synuclein can associate with (Galvagnion, 2017).
Here, we assessed whether DAT function is modified in the SNCA-OVX mouse model of early PD relative to Snca-null background control mice. SNCA-OVX mice are devoid of mouse α-synuclein but overexpress human wild-type α-synuclein at disease-relevant levels modeling SNCA locus multiplication seen in PD (Singleton et al., 2003). They show early deficits in DA release prior to DA cell loss, behavioral deficits in old age (Janezic et al., 2013), and progressive α-synuclein oligomerization with age compared to Snca-null mice (Bengoa-Vergniory et al., 2020) but in the absence of aggregation (Janezic et al., 2013). The control Snca-null mice do not differ from wild-type mice in DA release levels using our stimulation protocols (Senior et al., 2008). We assessed whether the SNCA-OVX PD model has altered DAT function in DA release or uptake by the DAT, and explored a potential point of convergence with cholesterol biology, using fast-scan cyclic voltammetry, molecular biology and ToF-SIMS.

Mice
All procedures were conducted in accordance with the United Kingdom Animals (Scientific Procedures) Act of 1986 and approved by the local ethical review panel at the Department of Physiology, Anatomy and Genetics, University of Oxford. SNCA-OVX mice (B6.Cg-Tg(SNCA)OVX37Rwm Snca tm1Rosl /J; Jackson Laboratories stock no. 023837) are BAC-transgenic mice that overexpress human α-synuclein from the SNCA genomic locus at Parkinson's disease-relevant levels, and are back-crossed onto a mouse α-synuclein-null (Snca-null) background. SNCA-OVX and their litter-mate Snca-null background control mice were produced as described previously (Janezic et al., 2013). Mice were age-(3-20 months) and sex-matched throughout. Data from male and female mice were combined given there were no apparent differences between the effects of cocaine on DA release between the sexes in both genotypes (Supplementary Figure 5). DAT-Cre mice (DAT IRESCRE Jax stock number 006660) were homozygote 4-month old males, generated by crossing heterozygous mice.

Proximity Ligation Assay (PLA)
DAT:α-synuclein PLA experiments were carried out using Duolink in situ kits supplied by Sigma Aldrich according to the manufacturer's instructions, and an α-synuclein antibody (syn4D6 ab1903, Abcam), and a DAT antibody (ab5990, Abcam). Briefly, the conjugates were prepared using the DuoLink Probemaker kit by incubating 20 µl of each antibody (1 mg/ml) with the Probemaker activated oligonucleotide (Plus and Minus respectively) and conjugation buffer and leaving it overnight at room temperature. Conjugates were incubated with Probemaker stop solution for 30 min at room temperature and then suspended in Probemaker storage buffer. Paraffin-embedded tissue was prepared for fluorescent PLA by dewaxing in xylene and histoclear, rehydrating via graded alcohols, blocking endogenous peroxidases with 10% H 2 O 2 for 15 min at room temperature, followed by antigen retrieval using microwave heat (10 min total) and citrate buffer pH 6 (ab93678, Abcam). All samples were incubated in Duolink block solution for 1 h at 37 • C, followed by conjugates diluted in Duolink PLA diluent (both DAT and α-synuclein 1:100) overnight at 4 • C. After washing in TBS + 0.05% Tween 20, samples were incubated with Duolink ligation solutions and ligase for 1 h at 37 • C, before washing and incubation with Duolink amplification reagents and polymerase for 2.5 h at 37 • C. Samples were then washed with Duolink wash buffer B and mounted/coverslipped with Fluorsave (Calbiochem). Images were blindly acquired at three distinct sites in dorso-medial, dorso-mid and dorso-lateral striatum from a single 5 µm coronal section per mouse, and puncta were automatically counted with ImageJ. Three mice per genotype were assessed.

DAT Immunofluorescence
Paraffin-embedded mouse brains from SNCA-OVX and Sncanull mice were sectioned to 5 µm. Paraffin-embedded tissue was prepared for immunofluorescence by dewaxing in xylene and histoclear, rehydrating via graded alcohols, blocking endogenous peroxidases with 10% H 2 O 2 for 15 min at room temperature, followed by antigen retrieval using microwave heat (10 min total) and citrate buffer pH 6 (ab93678, Abcam). The tissue was blocked for 1 h at room temp (TBS containing 1 M glycine, 10% normal goat serum and 0.1% Triton) prior to incubating in DAT primary antibody (rat anti-DAT; ab5990, Abcam) overnight at 4 • C. Following washes in TBS + 0.1% Triton, the tissue was incubated in secondary antibody (Alexa Fluor 488 Goat anti-rat, Life Technologies) for 1 h at room temperature. Sections were then washed and coverslipped with Fluorsave TM (Calbiochem). Images were taken blindly in dorsomid, dorso-medial and dorso-lateral striatum from two coronal sections, three mice per genotype were quantified automatically with ImageJ.

3H-Cocaine Autoradiography
Mice (three per group; 3-4 months old) were sacrificed by cervical dislocation and the brains immediately collected and snap-frozen in isopentane. Coronal sections (20 µm) were sectioned on a cryostat at the level of the striatum (plates 18-24; Franklin and Paxinos, 2004), thaw-mounted onto gelatinized slides, and stored at −80 • C until use. Slides were defrosted and air-dried at room temperature before being pre-incubated for 10 min in an ice-cold PBS buffer (150 mM NaCl, 10 mM NaHPO 4 , pH7.4). Slides were then incubated at 4 • C for 30 min in a PBS buffer containing [ 3 H]-cocaine (1 µM; Perkin Elmer). Non-specific binding was determined in presence of GBR 12935 (1 µM). After incubation, slides were washed twice in ice-cold PBS and left to dry. Slides were then put in contact with film (Kodak MR) for 6 weeks. Densitometric analysis of the autoradiograms was performed against 3H-microscales (Amersham, UK) using a computerized imaging system (MCID Core, version 7.0, Interfocus Imaging Limited, UK) and subsequently normalized to Snca-null.
Carbon-fiber microelectrodes were manufactured in-house using borosilicate glass (GC200F-10, Harvard Apparatus) and epoxy-free carbon fiber (7 µm diameter; Goodfellow). Voltage waveform (−0.7 V to +1.3 V) was scanned at 8 Hz at 800 V/s across the recording CFM and switched out of circuit between scans using a Millar Voltammeter as described previously (Threlfell et al., 2010(Threlfell et al., , 2012. Single recording sites in either dorsomedial or dorsolateral caudate putamen (CPu) were selected and multiple stimulation paradigms were delivered in a pseudo-random order with 2.5 min intervals between stimulations. Following a period of recording to establish a stable baseline in control or drug conditions, at least three repeats of each stimuli were made within a given condition-e.g., aCSF control or drug. These repeats were then averaged within an experiment and normalized and averaged across different animals to obtain a mean effect of a drug. At least three animals per experiment were used (N).
To assess dopamine reuptake kinetics in SNCA-OVX and Snca-null we used two different approaches. We used exponential decay curve fits and Michaelis-Menten models to fit the falling phases of evoked DA transients using GraphPad Prism 6.0. Exponential curve fits were applied to individual transients concentration-matched between genotypes, and fitted over 1 s to extract rate constants, k. Conditions designed to maximize [DA] o and thereby approach levels that might approach V max were train stimuli of 20 pulses at 100 Hz delivered in the presence of K v +channel blocker 4-AP (100 µM), nicotinic acetylcholine receptor blocker DHβE (1 µM) and D2 receptor antagonist L-741,626 (1 µM).
In experiments applying cholesterol, slices were pre-incubated for 1 h in water soluble (ws)-cholesterol (50 µg/ml) or vehicle (methyl-β-cyclodextrin; 1 mM) prior to transferring to the recording chamber. For probucol treated experiments, slices were incubated in probucol (5 µM) or DMSO vehicle control for 30 min then transferred to recording chamber with probucol (5 µM) present in the superfusate throughout recording. For nystatin treated experiments slices were incubated in nystatin (25 or 100 µg/ml) or vehicle control (DMSO) for 30 min and then transferred to recording chamber due to potential interactions with the CFM.

Secondary Ion Mass Spectrometry (Tof-SIMS)
Frozen brain samples were sectioned at −20 • C to 6 µm thickness and the slices containing both CPu and NAc were mounted on ITO glass slides. The brain tissue samples were stored in capped containers at −80 • C until SIMS analysis. The sections were dehydrated in a vacuum desiccator for 30 min. ToF-SIMS analysis was performed using a J105 3D mass spectrometry imaging instrument (Ionoptika, Limited, U.K.). The instrument was equipped with a continuous 40 keV primary gas cluster ion beam (GCIB) to analyze the samples and clusters of CO 2 gas (AGA, Sweden) with a nominal size of 6,000 [(CO 2 ) 6,000 + ] were utilized as the primary analysis ion beam. The ion dose density was kept below the static limit (1 × 10 13 ions/cm 2 ), in order to minimize surface damage. Spectra were obtained over a mass range of 100-1,000 Da and the mass resolution was approximately 5,000-8,000 in the range of intact phospholipids. In order to perform statistical analysis and comparison between SNCA-OVX and Snca-null mice, data were collected from three animals per genotype and from four slices per mouse. Data were normalized to the total ion counts obtained in the measured mass range.

Statistical Analysis
All data are expressed as means ± SEM (unless otherwise stated) and the sample size, n = technical repeats and N = number of animals. The number of animals in each data set was ≥3. Parametric tests were used for data sets with equal variances and passed normality tests (D'Agostino and Pearson normality test). Subsequently, comparisons for differences in means were assessed using unpaired t-tests, or Mann-Whitney tests or oneor two-way ANOVAs with post hoc Sidak's multiple comparison tests using GraphPad Prism 6.0. Michaelis-Menten fits used GraphPad Prism.

DAT Function Is Potentiated by α-Synuclein
In mice expressing human α-synuclein on a mouse αsynuclein-null background (SNCA-OVX; Janezic et al., 2013), we used fast-scan cyclic voltammetry to test whether α-synuclein expression affected properties of DA release in ex vivo slices that are regulated by DAT. Initially, we ensured that we could replicate previously published finding that striatal DA release is impaired in SNCA-OVX compared to Snca-null mice by 3 months- (Janezic et al., 2013). In control conditions, [DA] o release evoked by a single local electrical pulse in the dorsal striatum was 34% lower in SNCA-OVX compared to Snca-null (Figures 1A,B, 1.69 ± 0.11 µM vs. 2.57 ± 0.20 µM), in line with previous findings. Local electrical stimuli evoke DA release by direct depolarization of DA axons and indirectly by activating nAChRs on DA axons following evoked ACh release from cholinergic interneurons (Cachope et al., 2012;Threlfell et al., 2012;Wang et al., 2014). Therefore, we also tested whether the DA deficit in SNCA-OVX mice is present when the ACh-driven portion of DA release is prevented. In the presence of nAChR antagonism (DHβE 1 µM), levels of [DA] o evoked by a single electrical pulse were lower than in drug-free media as previously published , and, remained 34% lower in SNCA-OVX vs. Snca-null mice (Figures 1A,B, 0.37 ± 0.03 µM vs. 0.69 ± 0.05 µM).
Next, we explored DA uptake kinetics using several approaches. We first extracted decay constants, k, from singlephase exponential decay curves fitted to concentration-matched [DA] o transients evoked by either single pulse or stimulation protocols that drive elevated [DA] o (20p 100 Hz in DHβE, 4AP, L741-626). k was significantly greater in SNCA-OVX than Snca-null mice indicating faster uptake kinetics, for both stimulus protocols when analyzed either separately (Figure 1C) or in combination (Figures 1D,E). Second, we constructed a Michaelis-Menten-like plot, of the maximum decay rate (V) found on falling transients plotted vs.
[DA] o and found that both V max and K m of the best-fit curves were greater in SNCA-OVX than Snca-null mice (Figure 1F, V max : 11.8 vs. 7.2 µM/s; K m : 4.0 vs. 1.8 µM). We then tested whether the altered DAT function in SNCA-OVX was reflected by differential effects of DAT inhibition on [DA] o . Indeed, inhibition of DAT with either cocaine (5 µM), GBR 12935 (10 µM) or nomifensine (10 µM) resulted in a greater enhancement of evoked [DA] o in SNCA-OVX vs. Snca-null mice (Figures 1G,H). We explored whether the changes to DAT function might be an adaptation to, or conversely arise independently from, a deficit in evoked DA release levels in SNCA-OVX mice (Janezic et al., 2013) as chronic DA levels can impact DAT function (Calipari et al., 2013;Siciliano et al., 2018;Brodnik et al., 2020). In particular we tested whether the effect of cocaine was greater in SNCA-OVX vs. Snca-null mice in the NAc where, in contrast to dorsal striatum, there is no deficit in evoked [DA] o (Janezic et al., 2013), and found that the effect of cocaine on 1p-evoked [DA] o was similarly enhanced in NAc of SNCA-OVX mice (Supplementary Figure 1) indicating that altered DAT function in SNCA-OVX mice occurs is not due to a DA release deficit.
In addition, we have recently shown that the DAT is a powerful regulator of short-term plasticity in DA release, whereby the DAT promotes a strong inverse relationship between [DA] o released at a subsequently paired pulse (pairedpulse ratio) and interpulse interval (IPI; Condon et al., 2019). We tested whether elevated DAT function in SNCA-OVX resulted in this relationship being promoted. Indeed, we found a slightly steeper inverse relationship between paired-pulse ratio and IPI in SNCA-OVX vs. Snca-null (Figure 1I; in the presence of nicotinic receptor antagonism to prevent the confounding effects of ACh . In turn, for short bursts of 4 pulses, there was a steeper relationship between frequency and evoked [DA] o in SNCA-OVX vs. Snca-null ( Figure 1J). These differences between genotypes were ameliorated by cocaine (Figures 1I,J), which weakened the relationship of [DA] o to IPI and frequency in both genotypes, as seen previously in wild-type mice (Condon et al., 2019).

DAT Availability Is Increased by α-Synuclein
To understand the enhanced DAT function in SNCA-OVX striatum we explored how DATs and α-synuclein interact. Using a proximal ligation assay (PLA) we identified that DATs form a close spatial arrangement with α-synuclein (<40 nm) in SNCA-OVX mice (Figures 2A,B), supporting the hypothesis that αsynuclein and DAT associate. We tested whether α-synuclein expression increased DAT levels using western blot. However, under reducing conditions we found no difference in total striatal DAT levels between wild-type mice (Snca-wt), Sncanull, or SNCA-OVX mice (Figures 2C,D). To validate these data, and in particular, as a positive control for the ability of western blotting to detect differences in DAT levels, we confirmed that we could detect the lower DAT levels seen in DAT-Cre +/+ mice compared to controls (Bäckman et al., 2006). In parallel with these low levels of DAT we confirmed in DAT-Cre mice that uptake rates for released DA were correspondingly low and that there were only modest effects of cocaine on peak [DA] o and paired-pulse ratios detected with FCV (Supplementary Figures S2A,B). Consequently, the elevated DAT function in SNCA-OVX mice seems not due to different level of total DAT, suggesting that DAT function might be affected at the level of the cell surface. We tested whether αsynuclein could be revealed to modify the levels of DAT levels under more native conditions using two different assays. We immunolabeled DAT in striatum and found greater levels of DAT-immunoreactivity in SNCA-OVX vs. Snca-null striatum (Figures 2E,F). We also used autoradiography with [ 3 H]-cocaine to determine the relative density of striatal DAT binding sites, and found levels to be about 2.5 times higher in SNCA-OVX mice compared to Snca-null mice, with intermediate levels seen in wild-type mice Snca-wt; Figures 2G,H). The relative level of transporter binding, SNCA-OVX > Snca-wt > Snca-null tallies with the order of relative effect size of cocaine on increasing peak 1p-evoked [DA] o seen with cocaine in these genotypes in our hands, which is respectively 2.5-fold in SNCA-OVX > 2-fold in wild-types (Condon et al., 2019) > and 1.8-fold in Snca-null (Figure 1). These differences between SNCA-OVX and Sncawt mice also correlate with the relative expression levels of pan-species α-synuclein, whereby SNCA-OVX expresses human α-synuclein at 2-fold the level of mouse α-synuclein seen in Sncawt mice (Figure 2I), as previously published (Janezic et al., 2013). Together these data illustrate that DAT and α-synuclein form a close spatial interaction and that α-synuclein promotes DAT availability to bind its ligand.

Cholesterol Promotes DAT Function
There is a complex reciprocal relationship between α-synuclein and cholesterol, whereby cholesterol can influence α-synuclein levels and structural forms, and cholesterol levels have been reported to depend on α-synuclein expression ( Barceló-Coblijn et al., 2006;Don et al., 2014;Ronzitti et al., 2014;Mazzulli et al., 2016;Galvagnion, 2017;Hsiao et al., 2017;Alecu and Bennett, 2019) and furthermore, somewhat analogously to our findings with α-synuclein and DAT function, cholesterol has been shown to promote ligand binding to DAT (Hong and Amara, 2010;Morissette et al., 2018) and enhance DA transport (Jones et al., 2012;Morissette et al., 2018). We therefore explored whether there is converging biology between α-synuclein and cholesterol and DAT function in SNCA-OVX mice. Firstly, we tested whether applied cholesterol could phenocopy the effect of α-synuclein level on DAT function. Application of water-soluble cholesterol (50 µg/ml) decreased [DA] o evoked by a single pulse in Snca-null mice (Figure 3A), and increased the rate constant k for exponential curve-fits to the falling phase of concentrationmatched transients, indicating that applied cholesterol increases DA clearance (Figure 3B, k, 2.54 vs. 3.14 s −1 , P < 0.0001). Furthermore, applied cholesterol enhanced the effect of cocaine on peak [DA] o evoked by a single pulse in Snca-null mice ( Figure 3C) and also in SNCA-OVX mice ( Figure 3D). There was no difference in the effect of cocaine between vehicle and drug free conditions (Supplementary Figure 3). To explore whether cholesterol levels might be modified in SNCA-OVX mice, we measured cholesterol content of striatal tissue membranes using ToF-SIMS. We found that levels of cholesterol in SNCA-OVX mice were significantly lower than those in Snca-null mice, whereas levels of galactosylceremide (32:1) were higher (Figures 3E,F), which was not due to any difference in glucocerebrosidase (GBA) activity, unlike previous findings (Galvagnion, 2017 ; Supplementary Figure 4).
The finding that cholesterol content is lower in SNCA-OVX mice might at first seem counterintuitive, given that DAT function is enhanced in SNCA-OVX mice and by application of ws-cholesterol. However, these findings would be consistent   (E) Secondary ion mass spectrometry (SIMS) analyses of Snca -null (top) and SNCA-OVX (bottom) striatum groups obtained from the striatum region of brain tissue in positive ion mode. Cholesterol (green arrow) and galactosylceramide (galcer; purple arrow) signal intensity (normalized to the number of selected pixels for the spectrum). (F) Statistical comparison of striatal lipid species cholesterol and galcer (32:1). In SNCA−/− (unfilled bars) and SCNA-OVX (filled bars) Mean ± SEM (normalized to the total ion count). 2-sample T-test, n = 12, * * * P < 0.001.
with α-synuclein (soluble monomeric and oligomeric forms) being able to potentiate cholesterol efflux, via the ATP-binding cassette (ABC) transporter ABCA1 (Hsiao et al., 2017). Elevated cholesterol efflux could reduce cellular cholesterol level, consistent with the low cellular cholesterol content seen in SNCA-OVX measured by ToF-SIMS, and cholesterol efflux could promote extracellular cholesterol, consistent with elevated DAT function seen in SNCA-OVX mice. To test whether elevated DAT function in SNCA-OVX mice is due to enhanced extracellular cholesterol we tested whether an inhibitor of ABCA1 activity (probucol, 5 µM) could decrease the effect of cocaine in peak [DA] o . After incubation of striatal slices from SNCA- OVX mice in probucol, we found that the effect of cocaine was less than in vehicle-treated slices (Figures 4A-C). In addition, we tested whether nystatin (25-100 µg/ml), which chelates extracellular/sequesters lipid raft cholesterol, could also decrease the effect of cocaine. Indeed, nystatin, decreased the effect of cocaine on peak [DA] o in a concentration-dependent manner (Figures 4D-F), and both nystatin and probucol led to a significant slowing of the falling phases of DA transients ( Figure 4G). Together these data support the hypothesis that α-synuclein potentiates DAT function via its ability to potentiate cholesterol efflux via the ABCA1 transporter, leading to an elevation in extracellular cholesterol that in turn potentiates DAT function.

DISCUSSION
Here, we report that striatal DAT function is enhanced in the SNCA-OVX mouse model of PD, compared to Snca-null background controls. Furthermore, we show changes to cholesterol and other lipid levels, that can directly impact DAT function. These findings demonstrate that the intersecting biologies of α-synuclein and cholesterol can powerfully regulate DAT function.
In SNCA-OVX mice, we found close spatial localization of α-synuclein and DAT, and increased DAT function in several aspects relative to Snca-null controls. DAT function in wild-type mice represented an intermediate level between the extremes of SNCA-OVX and Snca-null mice. Extracellular DA clearance kinetics were enhanced, the effects of DAT inhibitors on peak [DA] o were promoted, and there was enhanced facilitation of DA release at short IPIs in SNCA-OVX mice. Correspondingly, we found evidence of elevated membrane DAT levels, from in situ autoradiography and DAT-immunocytochemistry, but not greater total DAT expression. We found that both the V max and K m for DA uptake were greater than in Snca-null control mice. These data therefore indicate that α-synuclein is promoting DAT availability, and also changing DA binding. These outcomes could also potentially be driven by an increase in forward trafficking of DAT to the membrane, a decrease in reverse trafficking and/or a change in conformation, including the relative outward-vs. inward-facing DAT. Our findings that K m appears increased in SNCA-OVX mice in addition to V max , appears to indicate that α-synuclein not only increases the number of DA-translocation sites, but also affects how DA binds to DAT. One candidate explanation might be an underlying change to the stoichiometry of DATs e.g., an increase in monomeric/oligomeric (dimer or tetramer) ratio, which was not assessed here but is known to affect the number of actively translocating DAT units and their availability to bind substrate (Gur et al., 2017;Zhen and Reith, 2018). Interestingly, cholesterol has been shown to affect how DA binds to DAT by promoting an outward facing confirmation (Hong and Amara, 2010;Jones et al., 2012). Intriguingly, our lipid analyses identified significantly lower levels of cellular cholesterol in SNCA-OVX mice, which is consistent with a report that soluble monomeric α-synuclein promotes cholesterol efflux, via the ABCA1 transporter (Hsiao et al., 2017). Potentially, the level of ceramide and cholesterol are related in a way that an increasing level of ceramide at the cell membrane promotes the release of cholesterol to the extracellular space, resulting in a lower level of cholesterol at the cell membrane (Mohammadi et al., 2018). Cholesterol efflux via ABCA1 transporter can occur from neurons or astrocytes (Lee et al., 2020). Correspondingly, we found that probucol, which decreases ABCA1 transporter function, or a cholesterol chelator, decreased the effect of cocaine on elevating peak evoked [DA] o . These findings together suggest that the elevated DAT function in SNCA-OVX mice is dependent on extracellular cholesterol, reflected in reduced membrane cholesterol levels. This hypothesis should be explored in future work and in other PD models.
These findings support a converging biology of α-synuclein, cholesterol and DAT, but the precise mechanism is still to be characterized. DATs are subject to extensive post-translation modifications (phosphorylation, palmitoylation, glycosylation, ubiquitination), express many regulatory binding domains (CRAC, SH3) and associate with other molecular regulators of DA release (D2Rs, RIM, synt-1; Vaughan and Foster, 2013). Both cholesterol and α-synuclein have scope to affect many of these parameters both directly and indirectly e.g., αsynuclein and cholesterol affecting VGCC function (Ronzitti et al., 2014), which in turn interacts with DAT function (Kile et al., 2010;Cameron et al., 2015). Furthermore, although DAT expression is restricted to DA axons, cholesterol and α-synuclein are more ubiquitous. Therefore, our observed effects of α-synuclein induced cholesterol efflux might involve other interacting striatal cellular networks, neuronal and non-neuronal.
The overall up-regulation of DAT function in SNCA-OVX mice also offers insights into another previous observation in this mouse model. DATs can apparently limit the recruitment of vesicles for release (Venton et al., 2006;Kile et al., 2010), and we have previously shown that vesicles within DA axons in SNCA-OVX mice are less dispersed and more tightly clustered, and that DA releasability is restricted (Janezic et al., 2013). Elevated DAT function could be a mechanism that contributes to this redistribution of DA vesicles and limited DA release.
We note that the increases to DAT function in SNCA-OVX mice, compared to background control mice, also similar parallel effects of knocking down a different protein in DA neurons, the Ca 2+ binding protein calbindin-D28k (calb1; Brimblecombe et al., 2019). Calb1 expression in midbrain DA neurons is negatively correlated with vulnerability to degeneration in PD, being lower in SN neurons that are more vulnerable. A multiple hit hypothesis has been suggested for driving DA neuron degeneration in PD, arising from hits that include α-synuclein burden, Ca 2+ burden and oxidative stress due to DA load after uptake (Sulzer, 2007;Mosharov et al., 2009;Post et al., 2018;Surmeier, 2018). We speculate that the enhanced DA uptake rates caused by α-synuclein could have important implications for pathology in Parkinson's. For example, increased DA uptake would decrease the extracellular availability of DA for acting on post-synaptic targets, exacerbating the effects of a deficit in DA release. Furthermore, elevated DAT function might increase intracellular levels of DA and provide an additional source of oxidative stress due to production of toxic metabolites, especially if coupled with low VMAT levels (Miller et al., 1999;Masoud et al., 2015). Our findings that both cholesterol and α-synuclein function are linked to DA uptake biology, indicate that the pathological gain of each individual hit will not act independently, but co-operatively through biological interactions.
In conclusion, we show that human α-synuclein promotes multiple facets of the ways in which DATs regulate DA release and extracellular levels, through mechanisms that depend on extracellular cholesterol. The interacting biology of α-synuclein, cholesterol and DATs leads to changes to DAT function and DA signaling in this mouse model of PD that will diminish DA output, and could potentiate the burden of cytosolic DA, potentially promoting vulnerability to degeneration.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by University of Oxford DPAG/EP AWERB.

AUTHOR CONTRIBUTIONS
KB, ST, and SC conceived the study. ST, AM, BR, NC-R, NP, RA, FS, TS, NB-V, and KB acquired and analyzed data. RW-M, AE, and TS contributed resources and commented on the manuscript. KB and SC co-wrote the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
The work was funded by grant support from Parkinson's UK (J-1403; G-1803). RA was supported by an MRC-iCASE award. NP was supported by an MRC-DTA Studentship.