Mitochondrial division inhibitor-1 is neuroprotective in the A53T-α-synuclein rat model of Parkinson’s disease

Alpha-synuclein (α-syn) is involved in both familial and sporadic Parkinson’s disease (PD). One of the proposed pathogenic mechanisms of α-syn mutations is mitochondrial dysfunction. However, it is not entirely clear the impact of impaired mitochondrial dynamics induced by α-syn on neurodegeneration and whether targeting this pathway has therapeutic potential. In this study we evaluated whether inhibition of mitochondrial fission is neuroprotective against α-syn overexpression in vivo. To accomplish this goal, we overexpressed human A53T-α- synuclein (hA53T-α-syn) in the rat nigrostriatal pathway, with or without treatment using the small molecule Mitochondrial Division Inhibitor-1 (mdivi-1), a putative inhibitor of the mitochondrial fission Dynamin-Related Protein-1 (Drp1). We show here that mdivi-1 reduced neurodegeneration, α-syn aggregates and normalized motor function. Mechanistically, mdivi-1 reduced mitochondrial fragmentation, mitochondrial dysfunction and oxidative stress. These in vivo results support the negative role of mutant α-syn in mitochondrial function and indicate that mdivi-1 has a high therapeutic potential for PD.


Results
Mdivi-1 prevents motor function deficits and neurodegeneration. We previously demonstrated that blocking mitochondrial fission, using whether rAAV-gene transfer technology or mdivi-1 to block Drp1 function, reduced neurodegeneration and synaptic dysfunction in MPTP-treated and Pink1-null mice 1 . However, due to the inherent limitations of these models, it was not feasible to assess motor function in those animals. In our newly developed and characterized virally induced human A53T-α-synuclein (hA53T-α-syn) rat model, there is progressive motor impairment and neurodegeneration 34 . Combined with the significant role of α-syn in PD, this model is highly suitable for testing the translational potential of mdivi-1. As we previously reported 34 , four weeks after receiving stereotaxic injection of hA53T-α-syn into the substantia nigra pars compacta (SNc, Fig. 1a), rats exhibited progressive and severe motor impairment as compared to the control group that received rAAV-GFP (Fig. 1b). Twice daily intraperitoneal (i.p.) injection of mdivi-1 [20 mg/kg, a dosage regimen that we previously characterized 1 ], however, completely normalized this abnormality -even up to the 8 week-end point of this study (Fig. 1b). At the end of this behavioral study, animals were euthanized and further processed for biochemical and neuropathological alterations.
To correlate the change in striatal dopamine (DA) to motor function, we used High Performance Liquid Chromatography (HPLC) to quantify striatal DA content in these animals ( Fig. 1c-f). Consistent with the observed motor function, severe DA depletion was detected after 8 weeks of expressing hA53T-α-syn (Fig. 1c). Other metabolites of DA were not affected (Fig. 1d,e). hA53T-α-syn also accelerates the turnover rate of DA (Fig. 1f). Mdivi-1 prevented the reduction of DA (Fig. 1c) and its high turnover rate (Fig. 1f) induced by hA53T-α-syn. Of note, mdivi-1 did not affect the total levels of DA and its metabolites in the GFP control animals, suggesting that blocking Drp1 does not have a detectable detrimental effect on healthy neurons.
Next, we performed stereological cell counting for accurate measure of the population of nigral dopaminergic neurons in these animals. Density of striatal dopaminergic fibers was also quantified. Consistent with the alterations in motor function and striatal DA content, hA53T-α-syn induced a dramatic loss of nigral dopaminergic neurons (Fig. 2a,b) and their striatal terminals (Fig. 2c,d). Mdivi-1 treatment conferred a significant protection in this nigrostriatal pathway. Together, these data strongly support the translational value of using mdivi-1 to reduce neurodegeneration and associated motor function deficits in PD.
Mdivi-1 reduces proteinase K-resistant and phosphorylated α-synuclein. Aggregated and phosphorylated α-syn are prominent features in PD. Reducing such aggregation has long been proposed as a therapeutic strategy for this disease. To evaluate if mdivi-1 would reduce such modified α-syn, we performed immunohistochemistry on coronal sections containing substantia nigra (Fig. 3). In animals with hA53T-α-syn transduction, significantly increased levels in phosphorylated serine 129 α-syn and proteinase K-resistant α-syn were detected. However, mdivi-1 significant reduced the levels of these pathogenic α-syn forms, despite the fact that it did not change the overall levels of hA53T-α-syn. Taken together, these results indicate that mdivi-1 is capable of reducing protein aggregation and toxic phosphorylated α-syn.
Mdivi-1 attenuates mitochondrial fragmentation. α-syn has been demonstrated to induce severe mitochondrial fragmentation in a number of cell culture studies [16][17][18][19][20] . In a recent in vivo study, mice with inducible expression of hA53T-α-syn in dopaminergic neurons exhibited severe mitochondrial fragmentation in a time dependent manner prior to axonal damage and progressive loss of these neurons 19 . However, to date, it is not clear whether mdivi-1 would prevent mitochondrial fragmentation in a hA53T-α-syn rodent model. To this end, we performed immunohistochemistry using Heat Shock Protein 60 (HSP60) as a mitochondrial marker and quantified mitochondrial morphology. Consistent with this previous animal study 19 , we also observed significantly fewer numbers of mitochondria with tubular morphology in rats that received rAAV9-hA53T-α-syn (Fig. 4). Mdivi-1 treatment provided protection against such morphological alterations in mitochondria. Although there was a trend of less tubular and more spherical mitochondria in the GFP control group, the difference is not statistically significant. Nevertheless, the absence of elongated mitochondria as we previously observed in mice 1 was unexpected. Based on our previous extensive dose-response study of this molecule and overexpression of proteins that promote mitochondrial fusion (Mfn2), promote fission (Fis1) and blocking fission (Drp1-K38A) 23 , it is possible that the appearance of this spherical morphology was due to enhanced fission inhibition. In our previous study, using rat dopaminergic neuronal cells, we observed that mdivi-1 lost the ability to promote mitochondrial elongation in some cells at increasing concentrations. High levels of expression of Mfn2 and Drp1-K38A produced similar effects 23 . Together, these results suggest that a gene-dose effect of promoting mitochondrial elongation is an inverted U-shape response. Further supporting this idea, studies from other laboratories also observed smaller/fragmented mitochondria in cells with overexpression of Opa1 35 and Mfn2 36 . Of importance in the present study is that mdivi-1 prevented mitochondrial fragmentation induced by α-syn-A53T.
Mdivi-1 improves mitochondrial function in striatal synaptosomes. α-syn has been demonstrated in rodent models to reduce striatal presynaptic DA release 19,[37][38][39][40][41] and impair mitochondrial respiration in the striatum 42 . Together, these observations are consistent with the critical role of mitochondria in synaptic function. To address whether blocking Drp1 improved mitochondrial function in the present study, we isolated synaptosomes from the striatum of rats with hA53T-α-syn or GFP with or without mdivi-1 treatment and measured mitochondrial function using XFe96 Extracellular Flux Analyzer as described 43 . This technology allows measurement of mitochondrial respiration in relatively small quantity and thus facilitates the assessment in specific brain regions. As seen in Fig. 5, one week after gene delivery (an early time point that induces about 50% cell death) 34 , hA53T-α-syn significantly impaired maximal rate of mitochondrial respiration as compared to the control group, resulting in reduced spare respiratory capacity (Fig. 5b). These results are consistent with a previous observation rats (6 weeks old) were assessed for locomotor function using stepping test. Afterwards, these animals were either injected with mdivi-1 (20 mg/kg, i.p) or vehicle control twice daily for the whole duration of the experiment. Three days after the initiation of mdivi-1 treatment, animals were stereotaxically injected with either AAV-hA53T-α-syn or AAV-GFPdegron control. Motor function was assessed every second week. Mdivi-1 treatment prevented motor deficits induced by hA53T-α-syn. HPLC was used to quantify total striatal dopamine (DA) content (c) as well as DA metabolites, (d) 3,4-Dihydroxyphenylacetic acid (DOPAC) and (e) Homovanillic acid (HVA). hA53T-α-syn induced significant DA depletion and increased DA turnover as evidenced by the ratio of DOPAC/DA (f). Mdivi-1 reduced this high turnover rate of DA (e) and restored the normal striatal level of DA (c). Mdivi-1 had no effect when injected in control animals (GFP). (b) Values are means ± SEM for 7-8 rats and were analyzed by 2-way ANOVA. (time: F 4,135 = 224.5 p < 0.001; treatment: F 3,135 = 32.81 p < 0.001; time x treatment: F 12,135 = 5.651 p < 0.001) followed by Tukey multiple comparison test (p < 0.001 versus hA53Tα-syn vehicle injected animals). Data from neurochemistry (c-f) were analysed using two-way ANOVA (AAV injection: F 1,26 = 10.57 p = 0.003; treatment: F 1,26 = 7.916 p = 0.009; AAV injection x treatment: F 1,26 = 1.984 p = 0.1708) followed by Tukey multiple comparison test (*p < 0.05; **p < 0.01; ***p < 0.001 compared to A53T vehicle).
using human induced pluripotent stem cell (hiPSC) with hA53T-α-syn 44 . Spare respiratory capacity is the ability of mitochondria to provide substrate supply and electron transport to response to an increase in energy demand. A reduction in spare respiratory capacity leads to energy crisis when energy demand exceeds the supply ability of mitochondria. Indeed, spare respiratory capacity has been considered as a major factor that defines the survival of the neuron 43 . Mdivi-1 significantly improved spare respiratory capacity in animals with hA53T-α-syn.

Mdivi-1 prevents oxidative stress.
It is well-established that mitochondrial fragmentation and dysfunction generate oxidative stress. We asked whether oxidative stress also occurred in our animal model and if so, whether mdivi-1 would reduce it. To address this question, we performed immunohistochemistry to detect 4-hydroxy-2-nonenal (4-HNE), a major product generated from lipid peroxidation as a result of free radical attack. Eight weeks after AAV-hA53T-α-syn injection in rats, a high level of 4-HNE in nigral dopaminergic neurons was evident as compared to those transduced with AAV-GFP. Consistent with the effect of mdivi-1 on preserving mitochondrial morphology and function, mdivi-1 blocked the production of 4-HNE in the group of animals that received hA53T-α-syn (Fig. 6).

Discussion
As we discussed in our recent reviews 13,14 , misfolded α-syn (whether as a result of mutations, exposure to environmental toxins or infection) induces cellular dysfunction and neurodegeneration through several distinct but non-mutually exclusive mechanisms. For example, because of its strong propensity to bind to membranes, AAV injection x treatment: F 1,27 = 1.504 p = 0.23) followed by Tukey multiple comparison test (*p < 0.05; **p < 0.01; ***p < 0.001 to A53T vehicle, ^p < 0.05 to A53T mdivi-1).
(d) The level of phosphorylated α-syn was increased in A53T animals and, mdivi-1 treatment attenuated this upregulation. (e) Following PK treatment, a strong hA53T-α-syn immunoreactivity persisted in A53T animals. This signal was significantly reduced in A53T rats treated with mdivi-1. The quantification was carried out using an automated threshold in one slice per animal. Values are means ± SEM for 7-8 animals analysed by two-way ANOVA (b) AAV injection: α-syn localizes to organelles such as lysosomes, endoplasmic reticulum and mitochondria. Indeed, it has been well-reported that α-syn impairs the function of these organelles. Combined with its ability to inhibit the ubiquitin-proteosomal and autophagic pathways, protein aggregation is a common observation in α-syn associated toxicity. In addition to these cell-autonomous mechanisms, α-syn also activates microglia and induces   neuroinflammation, resulting in a non-cell autonomous-mediated toxicity. With such multiple pathogenic mechanisms, various therapeutic strategies have been developed over years to combat α-syn associated pathologies 13 . In the present study, we addressed the critical question of whether protecting mitochondrial integrity and function would be sufficient to attenuate α-syn-induced neurotoxicity in vivo.
Using our well-characterized virally induced hA53T-α-syn rat model 34 , we demonstrated that the putative Drp1 inhibitor mdivi-1 is highly effective in reducing dopaminergic neurodegeneration and motor dysfunction. At biochemical and cellular levels, mdivi-1 increased striatal DA content, reduced protein aggregation and oxidative stress, as well as reduced mitochondrial fragmentation and improved mitochondrial function. Altogether these results support the negative impact of α-syn on mitochondria and that blocking mitochondrial fission is protective against α-syn neurotoxicity. The results of our study are consistent with the role of α-syn in causing mitochondrial dysfunction in other in vivo studies. For example, Chesselet et al. demonstrated that mice with global overexpression of human wild type α-syn in the brain using the Thy1 promoter exhibited age dependent accumulation of α-syn in mitochondria in the nigrostriatal dopaminergic pathway, impaired electron transport chain function and enhanced oxidative stress 42 . Recently, Greenamyre's laboratory reported that specific forms of post-translationally modified α-syn bind with high affinity to the mitochondrial receptor TOM20, resulting in mitochondrial dysfunction and production of reactive oxygen species 45 . From the Zhuang's group, inducible hA53T-α-syn mice exhibited severe mitochondrial fragmentation that preceded dopaminergic neurodegeneration and other pathologies 19 . Using double immunogold-transmission electron microscopy, presence of the transgenic α-syn protein in mitochondria was clearly identified in these mutant mice. We also observed mitochondrial fragmentation in our present study using rAAV-hA53T-α-syn rat model. More importantly, for the first time, we demonstrated that blocking in vivo application of mdivi-1 is highly protective against hA53T-α-syn-induced neurodegeneration and other associated pathologies. The observation of mdivi-1 reducing accumulation of proteinase K-resistant and phosphorylated Ser129-α-syn is potentially significant. Additional studies are required to elucidate the mechanism of these effects. However, we hypothesize that the mechanism is mediated, at least in part, by improving mitochondrial function and reducing oxidative stress, because the autophagic and ubiquitin proteasomal pathways are energy dependent and sensitive to oxidative stress. A recent study has similarly reported that mdivi-1 reduced the accumulation of amyloid-beta plaque in double transgenic mice APP/PS1 model of Alzheimer's disease 33 .
Manipulating mitochondrial fission/fusion has been considered as a potential novel mitochondrial therapy in recent years [46][47][48] . Within this context, blocking Drp1 is the most highly pursued strategy, partly because of the availability of a putative inhibitor (mdivi-1). Other investigators have demonstrated that blocking Drp1 function is protective in PD cell culture models of PINK1 23 , LRRK2 49,50 or VPS35 26 mutations, and of rotenone 51 , MPP + 52, 53 or 6-hydroxydopamine 54 neurotoxins, as well as in Pink1 −/− and MPTP-treated mice 1 . The protective role of blocking Drp1 in α-syn cell culture models is not quite definitive. Although some studies demonstrated that blocking Drp1 protected mitochondria from α-syn, others show that mitochondrial fragmentation induced by α-syn is Drp1 independent 16-18, 20, 55 . As often, conflicting data in cell culture models could be difficult to interpret due to different cell systems and experimental conditions. We believe that evaluations of the translational values of blocking Drp1 should extend beyond in vitro mitochondrial morphology and function, and that it should be conducted in mammalian animal models with brain pathologies and motor impairment as seen in PD. In this regard, the present study has provided some insights.
One potential concern of blocking mitochondrial fission as a therapeutic strategy is the possibility of developing side effects, because a balance of fission and fusion is necessary for the maintenance of neuronal function. This is an issue that will need to be considered if and when this treatment is to be conducted in clinical trials. Based on current literature, however, naive wild type mice are viable and no abnormal phenotypes are detectable up to ten weeks after the treatments of Drp1 inhibition, whether this is achieved by systemic injection of mdivi-1 1, 28-31, 33, 56-58 or a peptide (P110-TAT) 59 or by localized gene therapy 1 . Cytotoxicity of this molecule was not detectable at cellular, biochemical and functional levels, as shown in these publications and in the present study. Peripheral injections of mdivi-1 do not affect blood pressure, oxygen saturation, pH and blood cell counts 56 . Germline deletion of Drp1, however, induces embryonic lethality and degeneration of Purkinje neurons in mice [60][61][62] . Mice with conditional knockout of Drp1 63 or Mfn2 64 also indicate that nigrostriatal dopaminergic neurons are vulnerable to complete deletion of fission and fusion proteins. Interesting mice with heterozygous deletion of Drp1 have normal lifespan, phenotype, mitochondrial and synaptic structures 65 . Crossing these Drp1 +/− mice with either the transgenic AβPP mice (Tg2576) or with Tau P301L transgenic mice reduced toxic soluble proteins and improved mitochondrial function in these animal models of Alzheimer's disease 66,67 . Together these studies indicate that there is a gene-dose effect of loss of Drp1 function on its associated negative impact on neuronal function and viability. Furthermore, partial Drp1 loss of function appears to be safe and sufficient to confer neuroprotection.
Mdivi-1 has been demonstrated independently by many laboratories to have striking protective effects in a wide-range of disease models both in vitro and in vivo. The interest in the translational potential of this molecule is therefore understandably high. However, recent studies have raised the question of the specific mechanism of action of mdivi-1. Initially discovered by Nunnari and colleagues, mdivi-1 was shown to be specific and potent inhibitor of GTPase activity of yeast Dnm1, a homolog of mammalian Drp1 22 . Subsequently, this small molecule was also demonstrated to block Drp1 GTPase activity in human recombinant Drp1 68 and in mammalian neuronal cells 28 . GTPase activity of Drp1 is required for Drp1 oligomerization. Largely based on evidence from studies such as these ones, mdivi-1 has been considered in general as a Drp1 inhibitor. However, studies from the laboratory of Reddy showed that mdvi-1 also increased the levels of mitochondrial fusion proteins (Mfn1/2 and Opa1) and reduced the levels of Drp1 in neuronal cells 27,28 . It has also been demonstrated in vivo to reduce the levels of phosphorylated Drp1-S616 induced by kainic acid in the mouse hippocampus 58 . Taken together, these studies indicate that mdivi-1 is capable of blocking mitochondrial fission and promoting mitochondrial fusion at both enzymatic and protein expression levels. It is worth noting that also using human recombinant Drp1, a recent study reported that mdivi-1 is a weak and non-specific inhibitor of Drp1 GTPase 69 . Instead, this molecule was reported to be a reversible inhibitor of complex I and ROS production generated via reverse electron transfer mechanism. More studies are clearly required to reconcile this discrepancy and explain why blocking complex I would confer protection observed in other studies. Rather than using recombinant Drp1, perhaps an intact mammalian cell system should be used to determine whether mdivi-1 would block Drp1 function either directly or indirectly. Taken together, it is clear so far that mdivi-1 confers striking protective effects across multiple disease models, but most likely not exclusively through Drp-1 inhibition.
In summary, the present study reports that mdivi-1 is highly effective in reducing neurodegeneration, motor dysfunction and accumulation of toxic α-syn, mitochondrial damage and oxidative stress in a rat α-syn model of PD. This study further highlights the translational potential of mdivi-1. In addition to the present study, this small molecule has been demonstrated to be beneficial in other rodent models of PD 1 , Huntington's disease 28 , Alzheimer's disease 33 , epilepsy 58 , renal 30 , cardiac 29 , brain ischemic damage 31,70 , neuropathic pain 57 and diabetes 32 . With such a striking neuroprotective property in a wide range of diseases, mdivi-1 and perhaps strategies of blocking Drp1 function hold great promise to novel therapies. . Animals had free access to water and food. Veterinary care includes a full program for prevention of disease, daily observation and surveillance for animal health, appropriate methods of disease control, diagnosis, and treatment, appropriate methods of handling, restraint, anesthesia, analgesia and euthanasia as well as monitoring of surgical programs and post-surgical care. One week before surgery, Sprague Dawley rats (32 animals, 6 weeks old), purchased from Charles River, were tested for stepping test and sorted in order to generate 4 groups with no differences in the mean motor performance (p > 0.95). The 4 groups were then submitted to different treatments: 1) AAV-GFPdegron plus vehicle (GFP + veh), 2) AAV-GFPdegron plus mdivi-1 (GFP + mdivi-1), 3) AAV-hA53T-α-syn plus vehicle (A53T + veh) and 4) AAV-hA53T-α-syn plus mdivi-1 (A53T + mdivi-1).

Antibodies.
Mdivi-1 treatment. As previously described 1 , 3-(2,4-Dichloro-5-methoxyphenyl)-2-sulfanyl-4(3H)-quinazolinone (mdivi-1, Sequoia Research Products, UK) was dissolved in DMSO as a stock solution, which was then diluted with sterile 0.9% saline solution (1% DMSO final concentration), sonicated for 30 seconds and then promptly injected. Three days before stereotaxic injection of AAV-hA53T-α-syn or AAV-GFPdegron, animals were pre-treated with either mdivi-1 (20 mg/kg) or vehicle. This dosage was selected based on our previous study 1 , in which we characterized the in vivo pharmacokinetics and dose-response of mdivi-1. We found that with this dosage regimen, mdivi-1 conferred the most neuroprotective effects without detectable side-effects. The intraperitoneal (i.p) injections were performed twice a day (9 a.m.-18 p.m.) for eight weeks. During this period, every two weeks, motor performance was monitored with the stepping test. At the end of the treatment period animals were euthanized and brain collected for histological and neurochemical analysis. Brains were collected fresh and divided in three parts: bilateral mesencephalon and unilateral striatum was post fixed for 5 days in PFA 4% and then sectioned for histological studies, while another striatum was collected fresh, flash frozen and store at −80 °C for neurochemical analysis. Viral vector production. AAV9-GFPdegron and AAV9-hA53T-α-syn were produced by triple transfection into HEK-293T/17 cell line (ATCC, Teddington, UK) in polyethylenimine solution. After seventy-two hours cell were re-suspended with Tris lysis buffer (NaCl 150 mM, Tris-HCl 50 mM pH 8.5) and lysed using the freeze-thaw cycle procedure (−80 to +37 °C). The supernatant underwent to iodixanol gradient step purification (by centrifugation), and the fraction enriched in viral vector stocked at −80 °C. These procedure were well established and well described in our previous publications 34, 71 . Surgery procedure. Rats were bilaterally injected with AAV-GFPdegron or AAV-hA53T-α-syn (titer normalized to 6.9 × 10 13 gcp/ml) as previously described 34  Motor performance assessment. The stepping test was used to assess the forelimb akinesia as we recently described 34 . Rats were held and dragged sideways on a smooth surface at a constant speed for 0.9 m of distance and the number of adjusting steps counted. The performance was evaluated once every two weeks, three sessions over two consecutive days. The scores over the three sessions were averaged and the left and right backhand steps were pooled together. HPLC analysis. Fresh striatum collected after euthanasia was sonicated in HClO 4 0.1 M and the homogenate centrifuged at 4 °C for 30 min at 13000 rpm. 20 uL of supernatant was used for the HPLC analysis as previously described 72 . Monoamines were measured by coulometric detector (Coulochem II, ESA) coupled to a dual-electrode analytic cell (model 50110) with the potential of electrodes set at +350 and −270 mV. The samples were injected with a mobile phase containing NaH 2 PO 4 70 mM, methanol 7%, sodium octyl sulfate 100 mg/L, triethylamine 100 μL/L, EDTA 0.1 mM, into an HPLC Equisil column (C18, 150 × 4.5 mm, 5 μm) 72 . Retention times for noradrenaline, DOPAC, DA, HVA and serotonin were ~250, 420, 660, 1100, 1970 sec, respectively.
Immunostaining. Tyrosine Hydroxylase. Free floating 50 μm-thick slices were rinsed in PBS and treated for one hour with a blocking solution containing BSA 2% and Triton × 100 0.3% in PBS. After being blocked, the tissue was incubated with the TH primary antibody diluted with a solution containing BSA 1%, Triton × 100 0.3% over night at room temperature. The slices were then incubated with the secondary antibody and finally revealed with peroxidase EnVision TM system (DAKO). For the SNc slices the counter coloration with cresyl violet was performed. The slices were mounted in gelatin-coated slides and the coverslip sealed with Eukitt TM mounting medium.
Stereological counting. The unbiased stereological sampling method was used to quantify dopaminergic neurons in SNc as described in many occasions 34 The cell counting was performed using Leica DM600 motorized microscope equipped with Mercator Pro Software (Explora Nova, La Rochelle, France). After SNc boundaries delimitation, TH positive (TH+) cells in SNc are on-line counted at 40X magnification over five 50 μm-thick sections for each brain, collected every 300 μm, encompassing the whole SNc. The optical fractionator stereological probe was then used to estimate the total number of TH+ neurons for the entire SNc volume. Considering the bilateral nature of our model, the data presented in this paper are the sum of neurons counted for both SNc.
Striatal TH quantification. Images were taken with Nanozoomer 2.0 HT (Hamamatsu, Japan) at 20X magnification and analyzed with Image J (NIH, USA). The striatum boundaries were traced and optical density measured in terms of grey levels for 8-bit images as previously described 34 . α-synuclein immunostaining and quantification. For α-syn immunolabeling we applied the same protocol as described above for the TH staining as previously 73 . The slices treated with proteinase K (PK) were incubated for 10 minutes in a solution with 1 μg/mL of PK prior to any step and the tissue processed for the immunostaining. SNc pictures were captured with NanoZoomer using 20X objective and analyzed with Image J software. The quantification was carried out by measuring the percentage of the SNc area occupied by the stained surface using an automated threshold for all the images.
Mitochondrial network staining in vivo. HSP60 immunostaining was used to characterize the mitochondrial phenotype in dopaminergic neurons of SNc. Free-floating coronal SNc sections of 50 µm thickness were rinsed in PBS, incubated for 10 min with H 2 O 2 3% and 10% methanol, then for 20 min with Triton × 100 2%. BSA 3% for 1 hour was used to saturate the unspecific binding site before the overnight incubation with primary antibody (diluted in a solution containing BSA 1% and Triton × 100 at room temperature. Following incubation, sections were rinsed three times for 10 min in PBS and incubated for 1 hour with the secondary antibody. The same protocol was used to co-stain the TH and phosphorylated α-syn. The triple staining was needed to limit the quantification to dopaminergic neurons (TH positive) expressing hα-syn (indicated by the accumulation of phospho-syn). Single layer pictures of rat SNc were taken with confocal microscope DM6000 TCS SP5 (Leica, Germany) using 63 X (1.4 NA) magnification.

Mitochondrial network quantification in vivo.
Every single neuron picture was analyzed through an Image J macro purposely designed for the quantification of the relative area, shape descriptors and number of mitochondrial particles occupying the cytoplasmic surface of the neuron. Moreover, the circularity index provided by the macro, allowed us to sort the entire mitochondria population fragments into three different groups based on their shape: 0.8 to 1 spheric; 0.5 to 0.79 intermediate; 0 to 0.49 tubular.

4-hydroxynonenal staining and quantification.
TH and 4-HNE were co-stained following the same protocol used for the mitochondrial labeling in vivo. Single layer pictures were captured using confocal imaging at 40X magnification. We used a fixed laser power and AOTF levels for all the images. 16-bit pictures were then analyzed with Image J software measuring the level of grays of 4-HNE staining only in TH positive neurons.
Synaptosomes isolation. Synaptosomes were isolated from Sprague Dawley adult rats based on a previous publication 43 but with some modifications. Briefly, striata were quickly removed, rinsed with ice-cold sucrose buffer (320 mM sucrose, 1 mM EDTA, 0.25 mM dithiothreitol, pH 7.4) and then homogenized (10-12 strokes) in dounce glass homogenizer containing 1-1.5 mL sucrose buffer. The homogenates were then gently layered onto a discontinuous Percoll gradient (2.5 ml of 3%, 10% and 23% in sucrose buffer) in 10 mL centrifuge tubes, and centrifuged at 32,500 g for 10 min at 4 °C in a JA-25.50 fixed angle rotor in a Beckman Avanti J-26 X centrifuge. The band between 10% and 23% from striatum were collected for synaptosomes, pelleted and resuspended into ionic buffer (20 mM HEPES, 10 mM D-glucose, 1.2 mM Na2HPO4, 1 mM MgCl2, 5 mM), protein concentrations were measured using nano drop 2000.