Establishment of dopaminergic neuron purification system in mice for the Parkinson’s disease study


 Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterized by the degeneration of dopaminergic (DA) neurons. The key neuropathological hallmarks in the brain of patients with PD are Lewy body (LB) inclusions, consisting of misfolded α-synuclein proteins. Despite extensive efforts, the molecular link between LB inclusions and DA neurodegeneration remains elusive because of the lack of a suitable approach. Here, we aimed to establish a novel dopa-decarboxylase (Ddc) fluorescent reporter mouse model that allows the identification and collection of DA neurons using a fluorescence-activated cell sorter. Successful enrichment of Ddc-expressing cells was validated by RNA-sequencing analysis. This approach allowed us to analyze the effect of α-synuclein accumulation on the DA neuron’s transcriptome prior to neurodegeneration occurrence. We found that lipid-related process genes, followed by protein modification and degradation-related process genes, were upregulated in the α-synuclein-injected DA neurons. The activation of fatty acid-binding protein 1 (Fabp1) was particularly evident and confirmed by immunohistochemistry. Thus, our mouse model system and datasets provide a new method and insights into molecular mechanisms in PD.


Introduction
Dopaminergic (DA) neurons play a critical role in diverse functions, such as locomotion, learning, neuroendocrine control, reward, and motivation 1 . Degeneration or dysfunction of DA neurons results in mental, motor, and neurological disorders 1 . Particularly, Parkinson's disease (PD) is a progressive neurodegenerative disorder characterized by the loss of DA neurons in the substantia nigra pars compacta (SNc) of the midbrain [2][3][4][5] . Abnormalities in other types of neurons, such as serotonergic and cholinergic neurons, are also implicated during PD progression, resulting in non-motor symptoms 3,6 .
Despite extensive studies, knowledge about the underlying cause of selective PD neurodegeneration remains lacking. The histopathological marker of PD is the presence of intraneuronal inclusions of Lewy bodies (LB), which mainly consist of accumulated misfolded α-synuclein protein 4,5 . Under physiological conditions, α-synuclein has been suggested to modulate neurotransmitter release and uptake through the assembly of the soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) complex 4,7 . By contrast, the pathological α-synuclein adopts a beta-sheet conformation and aggregates into proto brils 8,9 . It subsequently undergoes maturation by interacting with proteins, lipids, and organelles to form LB inclusions 9 − 12 .
Increasing evidence has shown that the accumulation of α-synuclein impairs the biological function of DA neurons. For example, overexpression of α-synuclein downregulates the activity of VMAT2, a dopamine-transporting vesicular transporter, resulting in an accumulation of cytosolic dopamine metabolites and reactive oxygen species disrupting dopamine homeostasis 13 . Similarly, dopamine biosynthesis is affected by α-synuclein due to decreased tyrosine hydroxylase (Th) activity 14 . Further investigations revealed a presynaptic dopamine de cit increases DA neurons' vulnerability to oxidative stress 5,15,16 . Other mechanisms, such as activation of the neuroin ammation cascade induced by the accumulated cytosolic reactive oxygen species and the LB formation process, have been proposed to cause DA neuron depletion 1,5,17,18 . However, the exact molecular mechanism(s) leading to neurodegeneration in PD is still elusive.
Despite the extensive efforts in establishing a suitable model for revealing the causal relationship of αsynuclein and DA neurodegeneration, neither the in vivo and the in vitro model was capable of reaching a clear conclusion. The available in vivo models, particularly the α-synuclein overexpressing transgenic mice and stereotaxic α-synuclein-injected mice, recapitulate the pathological signatures and symptoms of PD patients 19 − 22 . However, DA neurons are an anatomically heterogeneous cell population, and its puri cation is challenging because DA neurons form extensive and complex networks of synapses, which require harsh conditions for dissociation 23 − 26 . The collection of homogeneous neurons using a uorescence-activated cell sorter (FACS) further induces stress in the neural cells. On the other hand, a homogenous DA neural population can be analyzed by the cell culture-based in vitro system 27,28 . But, DA neurons only survive for a short period of culture time that is not enough to observe the progressive neurodegeneration caused by the LB inclusions 27,28 . Furthermore, the interactions between DA neurons and the surrounding microenvironment are lacking in the in vitro model. This is likely one of the reasons it fails to reproduce cell death due to α-synuclein accumulation 27 .
To address these problems, we established a novel knock-in mouse model bearing a humanized Kusabira-Orange 1 (hKO1) reporter cassette at the dopa-decarboxylase (Ddc, also known as Aadc or Aaad) gene locus (Ddc-hKO1) to facilitate the detection of DA neurons and developed an e cient puri cation method for the collection and downstream analyses. We applied this method in a PD model study and successfully investigated the gene expression changes in DA neurons caused by α-synuclein accumulation.

Results
Establishment of Ddc-hKO1 knock-in reporter mice To visualize and purify DA neurons, we established a uorescent reporter knock-in mouse targeting the Ddc gene. DDC is an enzyme responsible for the biosynthesis of dopamine and serotonin, and it is expressed in several types of neurons, including DA, serotonergic, cholinergic, and adrenergic neurons 29,30 . Using the CRISPR/Cas9-mediated homology-directed repair approach, we inserted a reporter cassette consisting of hKO1 conjugated with P2A self-cleaving peptide (Fig. 1a). After the transfection and clonal expansion of mouse embryonic stem cells (mESCs), successful knock-in was veri ed using PCR (Fig. 1b).
Ddc is highly expressed in DA neurons and embryonic cardiomyocytes 31 . Therefore, we examined whether the Ddc-hKO1 reporter recapitulated the endogenous expression of Ddc using in vitro differentiated cardiac cells from knock-in ESCs. Speci c expression of hKO1 in spontaneous beating cells was con rmed at 20 days post-differentiation (Fig. 1c).
We then established a Ddc-hKO1 knock-in reporter mouse line through chimeric mouse generation using veri ed knock-in ESCs. No apparent abnormalities in embryonic development or growth were observed in the reporter mouse line (Fig. 1d). The adult mice were also healthy and fertile, producing a litter size similar to that of wild-type mice (data not shown).
hKO1 expression recapitulates the endogenous Ddc expression in reporter mouse brains The hKO1 expression was observed in the left atrium and ventricle of the E15.5 embryonic heart, in addition to parts of the brain ( Fig. 1e and Supplementary Fig. 1a). This expression pattern is consistent with the reported Ddc expression pattern and indicates the successful insertion of a reporter cassette to the Ddc locus 31 .
Next, we evaluated the expression pattern of Ddc-hKO1 reporter in three developmental stages: prenatal (E15.5), neonatal (2 days after birth, P2), and adult (3 months old). At the prenatal stage, hKO1-positive cells resided in the midbrain and hindbrain regions ( Fig. 2a and Supplementary Fig. 1a). Similarly, hKO1positive cells were observed in the VTA and SNc in the midbrain and the dorsal raphe nucleus (DR) in the hindbrain of newborn pups and adult mice (Fig. 2b,c and Supplementary Fig. 1b,c). hKO1-positive cells were also detected in the arcuate nucleus (Arc) in the hypothalamus, retrorubral eld (RRF), and locus coeruleus (LC) in the hindbrain of an adult mouse brain ( Fig. 2c and Supplementary Fig. 1c). These expression patterns of hKO1 in the brain were completely in line with the reported in situ hybridization results of Ddc in the Allen Mouse Brain Atlas (www.brain-map.org) 32 and the Allen Developing Mouse Brain Atlas (http//developingmouse.brain-map.org). Additionally, no hKO1-positive cells were found in other brain regions, which is in good agreement with the reported Ddc expression pattern, and this is consistently observed in different generations of the reporter mice. These observations indicate that the reporter expression re ected the endogenous Ddc expression, and off-target insertion of the transgene cassette was less likely.  We used Ddc-hKO1 reporter mice to enrich Ddc-expressing neurons using FACS for subsequent downstream analyses such as RNA sequencing (RNA-seq). The regions containing hKO1-positive cells in adult mouse brains were micro-dissected and subjected to enzymatic digestion to obtain a single-cell suspension ( Supplementary Fig. 2a). Pre-equilibration of media with 95% oxygen and supplementation of D-(+)-trehalose before and during the dissociation process are the two key factors that greatly enhance survivability of neurons 33 . Ddc-hKO1-positive cells were sorted based on the red uorescence signals, and we successfully collected an average of 14,000 viable hKO1-positive neurons (approximately 1.5%) from each brain (Fig. 3a, Supplementary Fig. 2b). We pooled two adult brains as one biological replicate for the following analyses to ensure su cient yield and to minimize variation among samples. Among the selected pro-apoptotic markers, the expression of Bad and Bax was detected in both hKO1positive and -negative populations ( Supplementary Fig. 2c). Given that both Bad and Bax are expressed in the brain in the Allen Mouse Brain Atlas 32 and the BioGPS RNA-seq data (biogps.org) 34 , the detected expression is likely due to basal level transcription. Notably, the expression of other apoptotic markers was either low or undetectable, and several anti-apoptotic marker genes, such as Bcl2l1 and Mcl1, were highly expressed in both hKO1-positive and -negative populations ( Supplementary Fig. 2c). Therefore, the apoptosis pathway is suggested to be not activated during sorting. Taken together, we reasoned that our protocol is feasible for isolating viable Ddc-expressing neurons.
Gene expression pro ling of the DA neurons during the early stage of α-synuclein accumulation To investigate the gene expression changes in DA neurons caused by α-synuclein accumulation, we performed an intranigral injection of G51D α-synuclein into adult Ddc-hKO1 reporter mouse brains 19 . G51D α-synuclein is one of the mutations identi ed in familial PD, and the loss of DA neurons became detectable from 12 weeks onwards after the G51D α-synuclein (hereafter, α-synuclein) injection 19 . To analyze the effect of α-synuclein accumulation at an early stage, we collected DA neurons at 7 and 12 weeks post-injection prior to the onset of neurodegeneration (Fig. 4a) 19 . The assembly of α-synuclein and deposition of Lewy body-like inclusions in the VTA and SNc were con rmed by immunostaining (Fig. 4b).
Consistent with a previous study, a loss of DA neurons was not observed at 7 weeks (Supplementary  We proceeded with the collection of Ddc-hKO1-expressing cells from the VTA and SNc using the protocol established in this study ( Supplementary Fig. 3c). The ratio and yield of hKO1-positive neurons harvested from saline-injected and α-synuclein-injected brains were comparable at both 7 and 12 weeks postinjection (n = 3, 2 brains per sample) (Fig. 4c, Supplementary Table 1, Supplementary Fig. 2d). This result supports a previous study showing that the loss of DA neurons was not evident as early as 12 weeks post-injection 19 .
In total, 20,907 and 21,161 genes were commonly mapped by RNA-seq analysis in saline-and αsynuclein-injected samples at 7 and 12 weeks, respectively. The enrichment of viable DA neurons in the collected hKO1-positive population was con rmed using marker gene analysis ( Supplementary Fig. 4a). DA neuron markers were highly expressed in all replicates with minimal expression of serotonergic neuron markers, suggesting that the major cell population collected from the VTA and SNc was of DA neurons. The harvested neurons from saline-and α-synuclein-injected brains exhibited a similar expression pattern of apoptosis-related genes to untreated sorted hKO1-positive cells ( Supplementary  Fig. 3b,c). The small effects on apoptosis-related gene expression in both 7-and 12-week samples suggest that α-synuclein did not activate the apoptosis-related pathway. This result is also consistent with previous studies showing that neurodegeneration by α-synuclein accumulation is not mediated by apoptosis 2,13 .
Early Biological Response During The α-synuclein Accumulation We de ned a set of genes in which its expression remained unchanged in the saline group at 7 and 12 weeks post-injection as a control to rule out the possible secondary effects (12,166 genes, FC < 1.5, P < 0.05). Differentially expressed genes (DEGs) were then determined by comparing expression changes between the α-synuclein and saline groups. Results showed 133 upregulated and 81 downregulated genes at 7 weeks post-injection, and 142 upregulated and 162 downregulated genes at 12 weeks postinjection (FC > 1.5, P < 0.05) (Fig. 4d). To analyze the functional enrichment of DEGs, we classi ed them into eight groups: upregulated at both 7 and 12 weeks (class 1, 4 genes), upregulated at 7 weeks but no change at 12 weeks (class 2, 60 genes), upregulated at 7 weeks but downregulated at 12 weeks (class 3, 1 gene), no change at 7 weeks but upregulated at 12 weeks (class 4, 117 genes), no change at 7 weeks but downregulated at 12 weeks (class 5, 122 genes), downregulated at both 7 weeks and 12 weeks (class 6, 2 genes), downregulated at 7 weeks but no change at 12 weeks (class 7, 47 genes), and downregulated at 7 weeks but upregulated at 12 weeks (class 8, 1 gene) (FC > 1.5, P < 0.05) (Fig. 4e). Among them, only classes 2 and 4 showed signi cant enrichment in speci c biological terms. In class 2, a total of six biological processes were signi cantly enriched, and three of them were related to lipid metabolism processes, namely, lipid phosphorylation, glycerolipid metabolic process, and steroid metabolic processes, such as Dgkb, Agk, Mvk, and Fabp1 (Fig. 4f, Supplementary Table 2). This result was supported by the GSEA result of gene expression pro les at 7 weeks without a pre-determined subset of genes ( Supplementary Fig. 5). On the other hand, the gene set related to the protein-associated processes involved in the response to unfolded or misfolded proteins, including Xbp1, Hdac8, Cry1, and Fbxo4, was highly enriched in class 4 DEGs (Fig. 4g).
We focused on Fabp1, which is activated in the DA neurons at 7 weeks post-injection, because Fabp3, another fatty acid-binding protein family member, promotes oligomerization and uptake of α-synuclein in DA neurons 35,36 . We hypothesized that Fabp1 may play a similar role as Fabp3 in the early manifestation of PD progression. Immunohistochemical analysis demonstrated intense staining of FABP1 in DA neurons of the α-synuclein-injected brains (Fig. 5a). Furthermore, the number of FABP1-positive cells in the hKO1-positive neurons was also signi cantly increased (Fig. 5b). Therefore, we concluded that FABP1 expression in DA neurons was activated by α-synuclein accumulation. α-synuclein-injected mouse brains at 7 weeks. Each dot represents the actual value obtained from the individual mouse. FACS, uorescence-activated cell sorter; VTA, ventral tegmental area; SNc, substantia nigra pars compacta; Sal, saline-injected; Syn, α-synuclein-injected. *P < 0.05, t-test, Sal: n = 3, Syn: n = 4. </ g> Discussion Elucidation of physiological function and pro ling of DA neurons has remained challenging due to the di culty in the manipulation and puri cation of homogenous neural population 25 . Currently, Th-green uorescent protein (GFP) transgenic reporter mouse model is widely used to visualize living DA neurons 37,38 . This mouse line has a reporter cassette that is randomly integrated into the genome, and GFP expression is driven by a rat Th promoter. Although GFP expression is mainly observed in DA neurons, it is also detectable in non-DA neurons, potentially caused by the effects of the integration site of the transgene or a species-speci c difference in the Th promoter 38 . The construction of the knock-in uorescence reporter mice established in this study enable the visualization of Ddc-expressing cells and offer a great advantage in DA neurons' analysis. We developed an optimized protocol for e cient recovery of viable neurons by integrating two procedures: supplementation of D-(+)-trehalose and incubation with 95% oxygen. Both procedures signi cantly enhance the survivability of neurons during the enzymatic digestion of brain tissue 33 . We harvested more than 10,000 hKO1-positive cells from a single brain, which are su cient for most downstream analyses such as RNA-seq, bisul te sequencing, and ChIP-sequencing. Single-cell RNA-seq (scRNA-seq) has recently been used to obtain transcriptome pro ling of small cell populations, including neurons 23,24,26,39 . Although scRNA-seq provides a great platform for pro ling individual cells, technical issues such as high noise, low capture e ciency, and substantial cost remain. Our study utilizes a bulk RNA-seq approach using a reporter mouse, providing a reliable and robust protocol for the analysis of speci c neuron types.
We utilized a PD model as a representative of nigrostriatal dopamine abnormalities. Although degeneration of DA neurons is the main hallmark of PD, its underlying molecular mechanism remained unclear 2,28 . This is mainly due to the lack of appropriate approaches. The in vivo α-synuclein injection model can recapitulate neurodegeneration, but it is impossible to harvest a homogeneous DA neuron population from wild type mice. On the other hand, in vitro culture experiments can enrich a homogeneous population, while it failed to reproduce cell death caused by α-synuclein 27 . Our reporter mouse model overcame these problems and succeeded in analyzing the transcriptome of α-synuclein accumulated DA neurons before and at the beginning of degeneration. Our results suggest apoptosis is not the cause of neurodegeneration as activation of apoptotic pathway genes was not detected in αsynuclein accumulated neurons. However, we cannot exclude the possibility of apoptosis at later stages because loss of DA neurons was not signi cant at 12 weeks post-injection. Further analysis at later stages using our reporter mice would allow us to unveil the pathways and molecular machinery of cell death in DA neurons.
Our transcriptome results demonstrated that the gene sets related to lipid-related processes, such as glycerolipid and lipid metabolic processes, were activated in DA neurons of the α-synuclein-injected brain at 7 weeks post-injection. Previous studies have shown that α-synuclein brils interact with lipids and other organelles during oligomerization to form Lewy body inclusions 9,10,12,40 . These gene sets include Dgkb and Agk, which are important for phosphorylation of diacylglycerol to phosphatidic acid, which play critical roles in intracellular signaling and phospholipid synthesis 41,42 . Mvk is involved in cholesterol and isoprenoid biosynthesis in the brain 43 . Given that 7 weeks is a short period of time since the α-synuclein injection, a large amount of lipid probably consumed for the on-going brilization of α-synuclein to form Lewy body-like inclusions. Additionally, the propagation of α-synuclein further attenuates the amount of lipids in DA neurons 12 . Thus, the lipid metabolic process may be upregulated to replenish the depleted pool of lipids in the cells. Our results suggest that abnormal lipid metabolism at the early stage of αsynuclein accumulation is one of the triggers of neurodegeneration.
On the other hand, most enriched gene sets of 12-week upregulated genes were involved in posttranslational protein processes (Supplementary Table 2). For example, Xbp1 is a key modulator of the unfolded protein response pathway associated with endoplasmic reticulum stress 44,45 . XBP1 has a neuroprotective function through the activation of endoplasmic reticulum chaperones, and the overexpression of XBP1 suppresses DA neurodegeneration caused by neurotoxin insult in overexpressing α-synuclein cell lines 44,45 . Another example is Fbxo4, an F-box protein; it is a component of the ubiquitinprotein isopeptide ligase SCF 46,47 . αβ-Crystallin, a binding partner of FBXO4, is a small heat shock protein that plays a role in preventing brillization of intracellular α-synuclein 46,47 . Therefore, activation of these genes may represent the proteolytic stress response of DA neurons induced by accumulated α-synuclein. Taken together, we propose that lipid metabolic processes followed by the unfolded protein response pathway are activated during the accumulation of α-synuclein in DA neurons.
We demonstrated that FABP1 was signi cantly upregulated in DA neurons of the α-synucleinaccumulated brain. Fabp1 is a member of the fatty acid-binding protein family. A study using an in vitro system demonstrated that FABP3, another member of the same family, plays a role in promoting the oligomerization of α-synuclein in DA neurons 35,36 . FABP3 is a critical factor in the transportation of arachidonic acid, a polyunsaturated fatty acid 35 . FABP3 is also shown to directly bind to α-synuclein and promote oligomerization 35 . DA neurons lacking Fabp3 were resistant to the neurotoxin agent MPP + insult. Therefore, Fabp1 potentially plays a role similar to Fabp3 in the oligomerization of α-synuclein at an early stage; however, this should be investigated further.
Apart from the PD model study, our reporter mice and neuron isolation protocol also offer other applications. For instance, it can be applied in the analysis of DA neuron subtypes, which show distinct expression patterns in marker genes 23,26,30,39 . Although scRNA-seq and immunohistochemical analyses demonstrate the heterogeneity of DA subpopulations in the brain, their speci c function is still unclear. Different susceptibilities toward toxin treatment were previously observed among the subpopulations of DA neurons 26 . The dual reporter mouse systems, such as Ddc-hKO1/Th-GFP, would reveal the nature and function of such subpopulations. Investigating the physiological function of these subpopulations may offer valuable insights into selective vulnerability in degeneration and drug development.
Additionally, the reporter mice are suitable for studying the effect of drug addiction and allelic expression dynamics of Ddc in the brain. Ddc was reported as an imprinting gene in the brain, where it shows controversial allelic expression patterns in different brain regions 48,49 . The advantage of Ddc-hKO1 reporter mice is the visualization of Ddc allelic expression in neurons. Heterozygous knock-in mice obtained from reciprocal crossing of Ddc-hKO1 and wild-type mice allow the analysis of allele-biased or mono-allelic expression of Ddc in a parent-of-origin manner. However, this study has some limitations. Since Ddc-hKO1 is expressed not only in DA neurons but also in other types of neurons, sophisticated microdissection is necessary to enrich DA neurons (Supplementary Fig. 3a). Also, the allelic expression of DA neurons at a single cell level may be challenging without the dual reporter mouse model that can visualize Ddc expression from both alleles with different uorescence reporters.
In summary, the novel Ddc-hKO1 reporter mice not only offer an e cient puri cation system of neurons with high quality and quantity of yield, but are also a versatile mouse model applicable to investigate physiological changes in DA neurons and neural disorders, such as PD. To the best of our knowledge, this is the rst transcriptomic dataset for DA neurons during the early accumulation of α-synuclein using a PD mouse model that mimics the symptoms of patients. We believe that our ndings provide important insights into the molecular events that contribute to neurodegeneration which accelerates effective drug and therapies development.

Animals
All animal experiments were conducted with approval from the Animal Care and Use Committee of the Graduate School of Frontier Biosciences, Osaka University. Mice were kept under standard laboratory conditions with controlled temperature and 12 h light/dark cycle with ad libitum access to food and water intake. All mice used in this study, including ICR and C57BL/6J, were purchased from Japan SLC (Hamamatsu, Japan). The day the vaginal plug was observed was de ned as embryonic day 0.5 (E0.5).

Cell Culture And Transfection
A G4 mouse embryonic stem cell line was used in this study 50  penicillin/streptomycin (Invitrogen, MA, USA), and 20% FBS in DMEM/F-12 with GlutaMAX) and transferred to an ultra-low adhesion plate (Wako, Osaka, Japan) for 3 days to allow further differentiation under oating culture conditions. Embryoid bodies were plated on a cell culture dish coated with 0.1% gelatin (Sigma-Aldrich, Missouri, USA), and spontaneous beating cells were observed at approximately seven days after attachment.

Brain Dissociation And Cell Sorting Using A Facs
The brain dissociation protocol was performed with modi cation as previously described 53,54 . RNA Puri cation And RNA-seq Analysis hKO1-positive and -negative cells were collected by sorting into TRIzol LS reagent (Thermo Fisher Scienti c, MA, USA). Cells from the two brains were pooled as one biological replicate. RNA puri cation was performed using a Direct-zol RNA microprep kit (Zymo Research, CA, USA) according to the manufacturer's protocol. Library preparation was performed using the SMARTer Ultra-Low RNA Kit (Clontech, CA, USA), and cDNA was ampli ed according to the manufacturer's protocol. Sequencing was conducted using a next-generation sequencer, an Illumina NovaSeq 6000 platform in 101-base single-end mode. Sequence reads were mapped to mouse reference genome sequences (mm10) using TopHat software (v 2.0.13) combined with Bowtie2 (v 2.2.3) and SAM tools (v 0. 1.19). The FPKM values were calculated using Cu inks software (v 2.2.1). Gene ontology enrichment analysis was performed using DAVID functional annotation bioinformatics microarray analysis (https://david.ncifcrf.gov) and ranked GSEA (v 4.1.0) (http://www.gsea-msigdb.org/gsea/index.jsp).
Stereotaxic injection of G51D α-synuclein into the SNc of mice Generation and intranigral injection of pre-formed G51D α-synuclein brils were performed as described previously 19 . Adult Ddc-hKO1 male mice at 7-8 weeks old were used for the injection. Mice were anesthetized, and saline or G51D α-synuclein (20 µg at 5 µg/µL) were injected into the SNc using a Hamilton microsyringe (Hamilton Co, NV, USA) under stereotaxic surgery (1.3 mm lateral, -2.8 mm posterior from the bregma, 4.3 mm below the dural surface). Unilateral injections were performed for immunohistochemistry, and bilateral injections were performed for RNA-seq analysis. Immunohistochemistry At 7 or 12 weeks post-injection, α-synuclein-inoculated mice were deeply anesthetized and perfused transcardially with 4% PFA/PBS. Dissected brains were post-xed overnight in 4% PFA/PBS and then immersed in PBS containing 30% sucrose solution (30% sucrose/PBS) until sinking as reported previously 55 . Immunohistochemistry was performed on 20 µm serial section cut with a cryostat (Leica Microsystems, Wetzlar, Germany). The primary antibodies used were as follows: mouse antiphosphorylated-α-synuclein (Ser-129, 1:10,000; Wako, Osaka, Japan), rabbit anti-TH (1:1,000; Calbiochem, CA, USA), and mouse-anti-fatty acid-binding protein 1 (FABP1) (1:200, Abcam, UK). For double immuno uorescence staining, appropriate uorescent secondary antibodies conjugated to Cy3 and FITC (1:500; Jackson ImmunoResearch) were used. Incubation was performed in PBS for 1 h at room temperature. Sections were washed with PBS three times, counterstained with 4′,6-diamidino-2phenylindole mounting medium (Vectashield, Vector Laboratories), and observed using BZ-9000 (Keyence, Osaka, Japan). For histological analysis, coronal sections were incubated with a biotinylated secondary antibody (1:500; Vector Laboratories), and the reaction products were visualized with avidinbiotin-peroxidase complex (Vector Laboratories) using 3′-diaminobenzidine (Sigma-Aldrich, Missouri, USA) as a chromogen. For the stereological assessment of the total number of TH-positive neurons, serial sections were prepared as reported previously 55 . Every fourth section was stained through the entire extent of the SNc. Cells were counted based on the method described by Furuya et al. 55 . For the assessment of FABP1, three sections from each mouse were analyzed, and the average percentage of FABP1-and Ddc-hKO1-double positive cells was calculated.