Parkinson’s disease-linked D620N VPS35 knockin mice manifest tau neuropathology and dopaminergic neurodegeneration

Generation and Molecular Characterization of D620N VPS35 KI Mice.

We obtained a newly developed conditional D620N VPS35 KI mouse line from The Jackson Laboratory (stock no. 021807) originally developed by the Michael J. Fox Foundation, as recently reported (24). These conditional mice were developed by introducing a loxP-flanked WT minigene (consisting of a splice acceptor, WT VPS35 exons 15–17, and a polyA signal) downstream of intron 14 and replacing endogenous exon 15 (with a D620N mutant version) by homologous recombination in ES cells (Fig. 1A). WT VPS35 is expected to be expressed under normal conditions (via splicing to the minigene), but after Cre-mediated recombination D620N VPS35 is expressed from the endogenous allele (Fig. 1 A and B). Upon intercrossing of heterozygous floxed mice (VPS35^FLOX/WT) we were unable to recover homozygotes (VPS35^FLOX/FLOX) despite the expected frequency of WT and heterozygous mice, suggesting that floxed homozygotes are most likely embryonic-lethal due to compromised VPS35 expression from the conditional allele. Notably, homozygous deletion of VPS35 in mice is embryonic-lethal (23). As such, germline D620N KI mice were developed by crossing VPS35^FLOX/WT mice to CMV-Cre transgenic mice to remove the floxed WT minigene (Fig. 1 A and B). The resulting heterozygous KI mice (VPS35^D620N/WT) were intercrossed to produce cohorts of VPS35^WT/WT, VPS35^D620N/WT, and VPS35^D620N/D620N littermate mice for subsequent age-dependent analysis. KI mice contain a single loxP site and exon 15 harboring the D620N mutation (Fig. 1B). PCR genotyping and genomic sequencing of the targeted exon 15 confirm correct gene targeting and the presence of the D620N mutation in KI mice (Fig. 1 C and D). Notably, heterozygous VPS35^D620N/WT and homozygous VPS35^D620N/D620N mice are viable, fertile, normal in appearance, produce genotypes at the expected frequency (Fig. 1E), and display normal survival up to 24 mo of age (Fig. 1F). Western blot analysis of ventral midbrain extracts from VPS35^D620N/D620N, VPS35^D620N/WT, and WT mice reveals normal levels of endogenous VPS35 as well as the retromer subunits VPS26 and VPS29 (Fig. 1G), confirming normal expression of VPS35 from the mutant allele and no effect of the D620N mutation on retromer stability. Similar experiments on hemibrains from conditional VPS35^FLOX/WT mice indicate a ∼50% reduction of VPS35 protein levels, as well as a corresponding reduction of VPS26 and VPS29, demonstrating that the floxed WT minigene compromises VPS35 expression from this conditional allele (Fig. 1H). Normal VPS35 mRNA expression levels and splicing in brains of VPS35^D620N/D620N mice are also confirmed by Northern blot analysis of total RNA derived from brain (Fig. 1I). In contrast, we observe ∼50% reduced VPS35 mRNA levels in the brains of VPS35^FLOX/WT mice (Fig. 1I) consistent with Western blot data (Fig. 1H). Therefore, the conditional VPS35^FLOX/WT mice are equivalent to heterozygous VPS35 null mice due to disrupted expression of VPS35 from the conditional allele. Collectively, the D620N mutation in VPS35 is well-tolerated in mice and does not influence retromer stability or survival, suggesting that it is unlikely to function through a loss-of-function mechanism.

Modest Motor Deficits in D620N VPS35 KI Mice.

To evaluate the potentially deleterious impact of D620N VPS35 expression on motor behavior, we conducted open-field, rotarod, and gait analyses on VPS35^D620N/WT and VPS35^D620N/D620N mice compared with WT littermate control mice at different time points as the mice aged. VPS35^D620N/WT and VPS35^D620N/D620N mice do not exhibit altered body weight compared with VPS35^WT/WT control mice (SI Appendix, Fig. S1). In the open-field quadrant, VPS35^D620N/WT and VPS35^D620N/D620N mice exhibit novelty-induced locomotor activity similar to that of WT mice at 3 and 13 mo of age, including measures of distance and speed (SI Appendix, Fig. S1). Rotarod analysis using an accelerating paradigm reveals similar motor balance, coordination, and learning between genotypes at 13 mo of age (SI Appendix, Fig. S1). Gait analysis assessed using a digital Catwalk system in D620N KI mice reveals a nonsignificant trend of gait impairment (i.e., average speed, limb swing speed, and cadence) in VPS35^D620N/D620N mice at 6, 9, and 12 mo (SI Appendix, Fig. S1), whereas base of support (distance between front or hind paws) and paw print position (distance between right or left paws) are normal (SI Appendix, Fig. S1). Taken together, the expression of D620N VPS35 has a rather modest impact on motor function at advanced age.

D620N VPS35 Expression Induces Progressive Dopaminergic Neurodegeneration.

To determine whether the expression of D620N VPS35 in mice influences the integrity of the nigrostriatal dopaminergic pathway with age, cohorts of KI mice were aged up to 13 mo. The number of tyrosine hydroxylase (TH)-positive dopaminergic and total Nissl-positive neurons in the substantia nigra pars compacta (SNpc) of VPS35^D620N/WT, VPS35^D620N/D620N, and WT littermate mice were assessed using unbiased stereological counting methodology. The morphology of dopaminergic neuronal soma and their neurite projections appear grossly normal in VPS35^D620N/WT and VPS35^D620N/D620N mice (Fig. 2A). At ∼3 mo of age, VPS35^D620N/D620N mice display a modest nonsignificant reduction of nigral dopaminergic neuronal number compared with their VPS35^D620N/WT or WT littermates (Fig. 2B). Remarkably, at 13 mo of age, VPS35^D620N/D620N mice exhibit a significant loss (38 ± 4.0%) of dopaminergic neurons, whereas VPS35^D620N/WT mice also exhibit significant dopaminergic neuronal loss (31 ± 3.4%) relative to WT mice (Fig. 2C). The loss of TH-positive neurons is due to neuronal degeneration rather than to a loss of TH phenotype, as indicated by a corresponding significant loss of total Nissl-positive SNpc neurons in VPS35^D620N/D620N (31 ± 3.5%) and VPS35^D620N/WT (22 ± 2.6%) mice compared with WT mice (Fig. 2C). Unexpectedly, the density of TH-positive dopaminergic nerve terminals in the striatum of VPS35^D620N/WT and VPS35^D620N/D620N mice at 13 mo is normal compared with their WT littermates (SI Appendix, Fig. S2), suggesting a compensatory resprouting of the remaining dopaminergic axonal processes. To further examine the impact of D620N VPS35 expression on nigrostriatal pathway function, we monitored the levels of biogenic amines in the striatum of VPS35^D620N/WT, VPS35^D620N/D620N, and WT mice at 14 to 15 mo of age by HPLC with electrochemical detection. Consistent with the normal density of striatal dopaminergic nerve terminals, D620N VPS35 KI mice display normal levels of striatal dopamine and its metabolites [3,4-dihydroxyphenylacetic acid (DOPAC) and homovanillic acid (HVA)] as well as other biogenic amines including serotonin and norepinephrine (Fig. 2D). Dopamine and other biogenic amines are also normal in the prefrontal cortex and olfactory bulb (SI Appendix, Fig. S2), which receive projections from dopaminergic neurons in the ventral tegmental area. Dopamine and serotonin turnover in the striatum, prefrontal cortex, and olfactory bulb are not altered in the KI mice (SI Appendix, Fig. S2). Collectively, heterozygous and homozygous D620N KI mice develop robust and progressive substantia nigra dopaminergic neuronal degeneration with age, thereby recapitulating one of the major pathological hallmarks of PD.

Axonal Damage in D620N VPS35 KI Mice.

Since the neuropathological spectrum associated with familial VPS35 mutations in PD is not yet clear, we next evaluated the development of PD-related neuropathology using a number of classical markers of neuronal damage and degeneration. Reactive gliosis consistently accompanies neuronal degeneration in PD and other neurodegenerative diseases. However, evidence for astrogliosis (GFAP) or microgliosis (Iba1) is not observed throughout the brains of VPS35^D620N/WT and VPS35^D620N/D620N mice at 13 mo (SI Appendix, Fig. S3). Axonal pathology was next evaluated using either immunostaining for amyloid precursor protein (APP), a sensitive marker of axonal damage, or staining with Gallyas silver, which specifically labels degenerating neuronal soma and processes. VPS35^D620N/WT and VPS35^D620N/D620N mice at 13 mo display the accumulation of spherical axonal inclusions containing APP as well as increased cytoplasmic APP within neurons of the striatum and substantia nigra (Fig. 3A). VPS35^WT/WT mice do not develop these abnormalities. The accumulation of APP-positive axonal inclusions suggests that the expression of D620N VPS35 may also promote axonal degeneration. Gallyas silver staining of striatum and substantia nigra sections from 13-mo-old VPS35^D620N/WT and VPS35^D620N/D620N mice reveals neurite degeneration, as indicated by the silver/black staining of neurites (Fig. 3B). Quantitation of silver-positive neurites in the striatum confirms a significant increase in neurite degeneration in VPS35^D620N/WT and VPS35^D620N/D620N mice compared with WT mice (Fig. 3C). No evidence of neurite degeneration is observed in D620N KI mice at 3 mo of age (see Fig. 9E). Neurite degeneration is more prominently detected outside of the nigrostriatal pathway, with extensive silver-positive axonal and neuronal degeneration present in the hippocampal CA1-3 region, cerebellum, and brainstem (but not the cortex) of VPS35^D620N/WT and VPS35^D620N/D620N mice at 13 mo (SI Appendix, Fig. S4). VPS35^WT/WT mice display minimal silver-positive neurites or neurons (Fig. 3 B and C and SI Appendix, Fig. S4). Collectively, our data reveal marked and widespread axonal pathology in the brains of D620N VPS35 KI mice.

Accumulation of Abnormal Tau Protein in Brains of D620N VPS35 KI Mice.

To further determine the pathological consequences of VPS35 mutations in vivo, and given the prominent axonal pathology observed in KI mice (Fig. 3), we next assessed alterations in the microtubule-associated protein tau (MAPT). Tau normally localizes to axonal processes via its interaction with microtubules but can dissociate from axons to form intraneuronal neurofibrillary tangles (NFTs) consisting of paired helical filament tau, the hallmark pathology of neurodegenerative tauopathies such as Alzheimer’s disease (25, 26). Notably, some PD brains harboring LRRK2 mutations as well as a number of LRRK2 rodent models can manifest tau-positive neuropathology (27–29), whereas common variation in the MAPT gene is associated with increased PD risk (30). Given that NFTs contain tau species that are abnormally hyperphosphorylated at multiple sites and adopt abnormal conformations (25, 26, 31), we used antibodies that detect specific disease-associated phosphorylation (AT8, PHF1, or CP13) or conformational (MC1) epitopes of tau. Intriguingly, we find evidence for prominent tau-positive neuropathology in the D620N VPS35 KI mice. Tau abnormalities are prominent by 13 mo of age in VPS35^D620N/WT and VPS35^D620N/D620N mice using antibodies to pSer202/pThr205 (AT8) and abnormal conformation-specific (MC1) tau epitopes in the CA1–CA3 subregions of the hippocampus involving the accumulation of tau in the soma and dendritic compartments of neurons (Fig. 4). Additional phospho-specific tau epitopes, pSer202 (CP13) and pSer396/Ser404 (PHF1), also accumulate but to a lesser extent in the hippocampal subregions of VPS35^D620N/WT and VPS35^D620N/D620N mice compared with WT mice (SI Appendix, Fig. S5). Tau abnormalities (AT8 and MC1) are detected most prominently in the cortex, hippocampus, and cerebellum of aged VPS35^D620N/WT and VPS35^D620N/D620N mice, followed by marked tauopathic changes in brainstem and ventral midbrain, and only modest tau pathology in the striatum (SI Appendix, Fig. S6). Somatodendritic tau pathology is generally more prominent in VPS35^D620N/WT mice relative to VPS35^D620N/D620N mice, compared with general background staining observed in VPS35^WT/WT mice (Fig. 4 and SI Appendix, Figs. S5 and S6).

Immunofluorescent colocalization studies in the hippocampus further reveal that abnormal tau accumulation (AT8 and MC1) is largely confined to the dendritic arbor of pyramidal neurons (i.e., stratum oriens, radiatum, and moleculare layers) and to occasional NeuN-positive neuronal soma in the stratum oriens and pyramidale layers of KI mice at 13 mo (Fig. 5A). Quantitation of AT8 and MC1 fluorescence intensity levels in the hippocampal CA1 and CA3 subregions confirms a significant increase of tau pathology in the VPS35^D620N/WT and VPS35^D620N/D620N mice compared with WT mice (Fig. 5B). In the substantia nigra, the significant accumulation of total (Tau5) and abnormal (AT8 and MC1) tau is observed specifically in TH-positive dopaminergic neurons of D620N KI mice at 13 mo of age (Fig. 6) and is detectable as early as 3 mo of age (SI Appendix, Fig. S5). Throughout the brains of D620N KI mice, we do not generally find evidence for tau-positive NFTs, inclusions, or neuritic pathology containing fibrillar forms of tau (Figs. 4–6 and SI Appendix, Figs. S5 and S6), suggesting instead that abnormal tau most likely accumulates rather than aggregates in somatodendritic compartments most consistent with earlier forms of “pre-tangle” pathology (25, 31).

To further explore the aggregation state of tau in D620N KI brains, we conducted biochemical studies to determine the steady-state levels, solubility, or abnormal phosphorylation of tau. At 13 mo of age, 1% Triton-soluble extracts derived from D620N KI mice fail to reveal consistent alterations in the steady-state levels (Tau5), phosphorylation (AT8), or abnormal conformation (MC1) of tau protein in the ventral midbrain and hippocampus of VPS35^D620N/WT and VPS35^D620N/D620N mice (SI Appendix, Fig. S7). A central feature of human tauopathies is the loss of tau solubility and the accumulation of hyperphosphorylated tau that can be isolated from brain tissue with sarkosyl extraction (25). Accordingly, sequential detergent extraction (low-salt-soluble, 1% Triton X-100–soluble, 1% sarkosyl-soluble, and sarkosyl-insoluble fractions) was conducted on brain regions with prominent tau pathology (hippocampus and cerebellum) from D620N KI mice at 13 mo. In general, D620N VPS35 expression does not robustly or consistently alter the detergent solubility of endogenous tau or significantly alter AT8 levels across fractions (SI Appendix, Fig. S7), consistent with the absence of NFTs in the KI mice (Figs. 4–6).

To explore the potential mechanism of tau accumulation in the D620N KI mice, we conducted similar studies in VPS35^FLOX/WT mice with diminished VPS35 expression that serve as heterozygous KO mice (refer to Fig. 1H). Abnormal tau (AT8 and MC1) does not accumulate in hippocampal subregions of VPS35^FLOX/WT mice compared with their VPS35^WT/WT littermate mice at 12 mo of age (SI Appendix, Fig. S8). These data demonstrate that VPS35 deficiency is not sufficient to induce the accumulation of abnormal tau and suggest that tau pathology in D620N KI mice does not manifest via a simple loss-of-function mechanism.

Retromer Subunit Assembly and Levels in Brains of D620N VPS35 KI Mice.

Western blot analyses of ventral midbrain extracts from VPS35^D620N/D620N, VPS35^D620N/WT, and WT mice at 3 mo reveal normal levels of the cargo-selective complex subunits VPS35, VPS26, and VPS29 (Fig. 1G). Therefore, the D620N mutation does not obviously influence overall retromer stability. To evaluate the interaction of VPS35 with retromer complex subunits in the brain of D620N KI mice, we conducted coimmunoprecipitation (co-IP) assays with a VPS35-specific antibody using soluble hemibrain extracts from VPS35^D620N/D620N and WT mice. WT and D620N VPS35 interact robustly and equivalently with VPS26 in brain (Fig. 7A). However, we are unable to detect an interaction of WT or D620N VPS35 with the WASH complex subunits FAM21 and WASH1 (Fig. 7A). We could further confirm the equivalent interaction with retromer subunits VPS26A, VPS26B, and VPS29 by mass spectrometric analysis of VPS35 IPs from WT or D620N KI brains, but we are unable to detect any WASH complex subunits (Fig. 7B). Using similar co-IP assays, we can detect the robust interaction of V5-tagged WT VPS35 with endogenous WASH1 and FAM21 in HEK-293T cells and the marked impairment of both interactions with D620N VPS35 (Fig. 7C), as previously reported (16, 19). Our data suggest that endogenous VPS35 in the mouse brain does not appreciably interact with the WASH complex. Next, we conducted confocal colocalization studies to understand the pathogenic effects of the D620N mutation in dopaminergic neurons of the substantia nigra. Intriguingly, we find a marked significant reduction of VPS35 and WASH1 levels selectively in intact TH-positive dopaminergic neurons of both VPS35^D620N/D620N and VPS35^D620N/WT mice compared with their WT littermates at 3 mo (Fig. 7 D and E). These data suggest a selective retromer deficit (VPS35 and WASH complex) in nigral dopaminergic neurons due to the D620N mutation, whereas retromer levels and assembly are generally normal in the brains of KI mice (Figs. 1G and 7 A and B).

Lack of α-Synuclein Pathology or Functional Interaction in D620N VPS35 KI Mice.

α-Synuclein plays a central role in the pathogenesis of PD, where it comprises a key component of Lewy pathology (30, 32–34). Prior studies have revealed a functional interaction of VPS35 and α-synuclein, whereby VPS35 deletion in yeast and worm models can exacerbate human α-synuclein–induced toxicity (35), whereas α-synuclein accumulates in the brains of mice with VPS35 depletion (21, 22). As such, we evaluated brains from aged D620N VPS35 KI mice for alterations in α-synuclein levels or aggregation. VPS35^D620N/WT and VPS35^D620N/D620N mice do not reveal consistent alterations in the levels, distribution, or morphology of phosphorylated Ser129-α-synuclein above background levels throughout the brain (SI Appendix, Fig. S9), a robust marker of pathological α-synuclein specific to Lewy pathology in PD (36). Biochemical analyses of detergent-soluble (1% Triton X-100) extracts from the ventral midbrain of 13-mo-old D620N KI mice indicate normal steady-state levels and phosphorylation (Ser129) of α-synuclein (SI Appendix, Fig. S9), consistent with immunohistochemical data. Therefore, D620N KI mice lack evidence of α-synuclein accumulation or neuropathology, in contrast to VPS35 KO mice (21, 22).

To further explore a potential pathological interaction of VPS35 with α-synuclein, we crossed the D620N VPS35 KI mice with human A53T-αSyn transgenic mice (line G2-3), a robust and well-characterized model of α-synuclein–induced neurodegeneration (37). VPS35^D620N/D620N mice were first crossed with hemizygous A53T-αSyn mice to generate F1 progeny (i.e., VPS35^D620N/WT and VPS35^D620N/WT/αSyn^Tg) that were subsequently intercrossed to generate F2 cohorts for direct comparison (Fig. 8). A53T-αSyn transgenic mice display shortened lifespan largely owing to the robust degeneration of brainstem and spinal cord neurons leading to limb paralysis, autonomic dysfunction, and premature death (37). Accordingly, we assessed the impact of physiological D620N VPS35 expression on the survival of A53T-αSyn mice (Fig. 8). A53T-αSyn mice exhibit premature lethality occurring over a period of 8 to 18 mo of age. However, the survival of VPS35^D620N/WT/αSyn^Tg and VPS35^D620N/D620N/αSyn^Tg mice does not significantly differ from that of VPS35^WT/WT/αSyn^Tg mice. VPS35^D620N/WT and VPS35^D620N/D620N mice exhibit normal survival relative to VPS35^WT/WT mice over 24 mo (Fig. 8). Our data indicate that the lethal neurodegenerative phenotype of A53T-αSyn transgenic mice is not influenced by endogenous D620N VPS35 expression, in contrast to simpler models with VPS35 depletion (35, 38). We next conducted preliminary experiments by unilaterally delivering mouse α-synuclein preformed fibrils (PFFs) to the striatum of 6-mo-old VPS35^D620N/D620N and VPS35^WT/WT mice to explore the impact of the D620N mutation on the development of α-synuclein pathology (SI Appendix, Fig. S10). At 30 d postinjection, we observe pSer129-α-synuclein–positive pathology localized to dopaminergic neurons of the ipsilateral substantia nigra of VPS35^WT/WT mice; however, qualitatively similar α-synuclein pathology is also observed in VPS35^D620N/D620N mice (SI Appendix, Fig. S10). These data suggest that the D620N mutation does not regulate the initial progression and propagation of α-synuclein pathology in the αSyn-PFF model. Taken together, these data fail to provide robust evidence for a pathological interaction of α-synuclein in the D620N VPS35 KI mouse model.

The D620N Mutation Supports Survival and Is Inconsistent with a Full Loss-of-Function Mechanism.

Our data indicate that homozygous D620N VPS35 KI mice display a normal lifespan (Fig. 1F), in contrast to homozygous KO mice (and VPS35^FLOX/FLOX mice; Fig. 1A) which are embryonic-lethal (23), suggesting that the D620N mutation is not likely to act via a loss-of-function mechanism. To further address this mechanism, we crossed VPS35^D620N/WT mice with VPS35^FLOX/WT mice (containing a single VPS35 null allele) (Fig. 1 A, H, and I) to create compound heterozygous VPS35^D620N/FLOX progeny. These mice harbor a single D620N allele and a VPS35 null allele and exhibit ∼50% VPS35 protein levels in the brain similar to VPS35^FLOX/WT mice, as expected (Fig. 9A). Importantly, we find that the VPS35^D620N/FLOX mice are viable, grossly normal, produced at the expected genotypic frequency (Fig. 9B), and exhibit normal survival over 24 mo (Fig. 9C). At ∼3 mo, the VPS35^D620N/FLOX mice display a normal number of substantia nigra dopaminergic neurons (Fig. 9D) and lack axonal pathology (Fig. 9E) relative to WT littermate mice. These data are in contrast to conditional KO mice with homozygous VPS35 deletion in DAT-positive neurons that develop robust dopaminergic neuronal loss at 2 to 3 mo of age (22). We further evaluated tau pathology in the VPS35^D620N/FLOX mice at 24 mo of age. Both VPS35^D620N/WT and VPS35^D620N/FLOX mice markedly accumulate abnormal tau (AT8 and MC1) in hippocampal subregions that is completely absent from VPS35^FLOX/WT mice (Fig. 9F). The burden of tau pathology in the CA3 subregion is similar between VPS35^D620N/WT and VPS35^D620N/FLOX mice, whereas in the CA1 subregion of VPS35^D620N/FLOX mice the accumulation of abnormal tau shifts from dendrites to the soma of pyramidal neurons (Fig. 9F). As compound heterozygous VPS35^D620N/FLOX mice alone are sufficient for the development of tau pathology, these data suggest that D620N VPS35 can induce pathology irrespective of the presence of WT VPS35. Furthermore, VPS35 depletion in the VPS35^FLOX/WT mice is not sufficient to induce tau pathology (Fig. 9F and SI Appendix, Fig. S8), implying that the D620N mutation impacts tau protein via a gain-of-function mechanism. Since VPS35^−/− and VPS35^FLOX/FLOX mice are embryonic-lethal (23), our data collectively demonstrate that a single copy of the D620N allele (in compound VPS35^D620N/FLOX mice) is sufficient for maintaining normal survival and early dopaminergic neuronal health in mice, indicating that D620N VPS35 is fully functional and highly unlikely to manifest disease via a loss-of-function mechanism (i.e., via a haploinsufficient or dominant-negative effect).

Animals.

Mice were maintained in a pathogen-free barrier facility and provided with food and water for ad libitum consumption and exposed to a 12-h light/dark cycle. Animals were treated in strict accordance with the NIH Guide for the Care and Use of Laboratory Animals (57). All animal experiments were approved by the Van Andel Institute Animal Care and Use Committee. Conditional (“floxed”) D620N VPS35 KI mice (strain 021807) and CMV-Cre transgenic mice (strain 006054) were obtained from The Jackson Laboratory. Human A53T α-synuclein transgenic mice (line G2-3) have been described (37). Mice were identified by genomic PCR (24, 37).

RNA Extraction and Northern Blot Analysis.

Total RNAs were extracted from hemibrains and subjected to Northern blot analysis using a ³²P-labeled DNA probe (nucleotides 163–984 of mVPS35). Signal was detected using a Typhoon phosphor-imaging system (Amersham).

Expression Plasmids and Antibodies.

Human V5-tagged VPS35 plasmids were obtained from Addgene (no. 21691; ref. 58) and previously described (18). Mouse VPS35 cDNA was obtained from Dharmacon. Primary antibodies used were VPS35 (ab57632 or ab154647; Abcam), VPS26 (ab23892; Abcam), VPS29 (ab10160; Abcam), WASH1 (Sigma), Fam21A-D (S-13; Santa Cruz), V5 and V5-HRP (Life Technologies), TH (NB300-109; Novus Biologicals or clone TH-2; Sigma), α-synuclein (clone 42; BD Biosciences), pSer129-α-synuclein (clone EP1536Y; Abcam), APP (clone 22C11; Millipore), tau (clone Tau5; Millipore), phospho-tau (clones AT8, PHF1, CP13, and MC1, provided by Peter Davies, Feinstein Institute for Medical Research, New York), NeuN (ABN78; Millipore), Iba1 (Wako), GFAP (DAKO), and β-tubulin (clone TUB 2.1; Sigma). Secondary HRP-conjugated IgG, light chain-specific (Jackson Immunoresearch), biotinylated IgG (Vector Labs), and AlexaFluor-488 or -594 IgG (Thermo Fisher) were used.

Cell Culture and Transient Transfection.

HEK-293T cells were maintained and transfected using X-tremeGENE HP Reagent (Roche) as described (59).

Co-IP, Western Blotting, and MS.

For co-IP assays, hemibrains from 3-mo-old mice or transfected HEK-293T cells were prepared in IP buffer [20 mM Hepes-KOH, pH 7.2, 50 mM potassium acetate, 200 mM sorbitol, 2 mM EDTA, 0.1% Triton X-100, and 1× Complete Mini protease inhibitor mixture (Roche)] and subjected to IP with antibodies to VPS35 or V5 (5 µg) as described (59). IPs and input lysates were subjected to Western blotting, detection, and quantitation as described (59). For MS, anti-VPS35 IPs from WT or VPS35 KI mouse brain were subjected to analysis by nano liquid chromatography/tandem MS (LC-MS/MS) (ThermoFisher Q Exactive) at MS Bioworks as previously described (59).

Sequential Fractionation of Brain Tissues.

Brain tissues from KI mice were harvested and subjected to sequential fractionation in low-salt, Triton X-100, and sarkosyl buffers as described previously (54). For standard biochemical analysis, brain tissues were homogenized in lysis buffers containing 1% Triton X-100 or 2% SDS (Triton-insoluble), as described (18, 60).

Measurement of Biogenic Amines.

Homogenized tissues were separated on a 150- × 4.6-mm Microsorb MV C18 100-5 column (Agilent Technologies) and analyzed by HPLC-electrochemical detection (ECD) to detect dopamine, HVA, DOPAC, serotonin (5-HT), 5-hydroxyindoleacetic acid (5-HIAA), norepinephrine, and 3-methoxy-4-hydroxyphenylglycol (MHPG), as described previously (61, 62).

Immunohistochemistry and Immunofluorescence.

Mice were perfused with 4% paraformaldehyde (pH 7.4), cryoprotected in 30% sucrose, and dissected into 35-μm-thick coronal sections. Immunostaining by incubation with primary and biotinylated secondary antibodies (Vector Labs), ABC reagent (Vector Labs), and DAB (Vector Labs) has been described (18, 60). Images were captured using an Axio Imager M2 (Zeiss) with color CCD camera (AxioCam; Zeiss). Silver staining was conducted on serial sagittal sections with the FD NeuroSilver kit II (FD Neurotechnologies) according to the manufacturer’s instructions. Images were captured using a ScanScope XT slide scanner (Aperio) at a resolution of 0.5 µm per pixel. Images were quantified using a custom CellProfiler v3.1.5 pipeline to isolate black neurites and particles. The percent of total pixels associated with black neurites was calculated for each image, sampled across three to six images per animal. Fluorescence immunostaining of brain sections was conducted using primary and secondary antibodies conjugated to AlexaFluor-488 or -594 (Life Technologies), imaged using a Nikon A1plus-RSi Laser Scanning Confocal microscope (Nikon Instruments), and subjected to quantitation by corrected total cellular fluorescence (CTCF) as previously described (59). CTCF values were derived by sampling across five sections per animal for hippocampal subregions, or from ≥40 TH-positive neurons per animal for the SNpc.

Stereological Analysis and Optical Densitometry.

Unbiased stereological quantitation of SNpc dopaminergic neurons was performed using the optical fractionator probe of StereoInvestigator software (MicroBrightField Biosciences), as previously described (18, 60). The optical density of TH-immunoreactive nerve terminals in the striatum was determined using the ScanScope XT slide scanner (Aperio) and optical densitometry using NIH ImageJ (v1.3). The mean optical density across six to eight sections per animal was determined.

Stereotactic Delivery of Preformed Fibrils.

Mouse α-synuclein was expressed in Escherichia coli and purified as previously described (63). To generate fibrils, 5 mg/mL of purified α-synuclein in 1× PBS, pH 7.4, was shaken at 1,000 rpm for 7 d, sonicated with 60 pulses at 10% power (30 s total time, 0.5 s on, 0.5 s off), aliquoted, and stored at −80 °C. Sedimentation assays and thioflavin T assays were performed as described (63, 64) to ensure fibril quality. Male adult littermate mice (n = 3 mice per genotype) were subjected to stereotactic surgery for the unilateral delivery of PFF-αSyn into the dorsal striatum. The following coordinates were used relative to bregma: anterior–posterior, +0.2 mm; medial–lateral, +2.0 mm; and dorsal–ventral, −2.6 mm. Mice received PFF-αSyn (5 µg) in 1 µL at a flow rate of 0.1 µL/min. Animals were euthanized at 30 d postinjection.

Behavioral Assessments.

Behavioral testing was performed on male and female mice at ages indicated for each test, as described (60). Open-field testing using an Ethovision XT video tracking system (Noldus Information Technology), accelerating rotarod (5 to 50 rpm over 5 min), and digital Gait analysis using the automated CatWalk XT system (Noldus) were conducted.