ARTICLE https://doi.org/10.1038/s41467-019-08858-y OPEN Brain tyrosinase overexpression implicates age-dependent neuromelanin production in Parkinson’s disease pathogenesis Iria Carballo-Carbajal 1, Ariadna Laguna 1, Jordi Romero-Giménez 1, Thais Cuadros1, Jordi Bové1, Marta Martinez-Vicente 1, Annabelle Parent 1, Marta Gonzalez-Sepulveda1, Núria Peñuelas 1, Albert Torra 1, Beatriz Rodríguez-Galván1, Andrea Ballabio 2, Takafumi Hasegawa3, Analía Bortolozzi 4, 1234567890():,; Ellen Gelpi 5,6 & Miquel Vila 1,7,8 In Parkinson’s disease (PD) there is a selective degeneration of neuromelanin-containing neurons, especially substantia nigra dopaminergic neurons. In humans, neuromelanin accumulates with age, the latter being the main risk factor for PD. The contribution of neuromelanin to PD pathogenesis remains unknown because, unlike humans, common laboratory animals lack neuromelanin. Synthesis of peripheral melanins is mediated by tyrosinase, an enzyme also present at low levels in the brain. Here we report that over- expression of human tyrosinase in rat substantia nigra results in age-dependent production of human-like neuromelanin within nigral dopaminergic neurons, up to levels reached in elderly humans. In these animals, intracellular neuromelanin accumulation above a specific threshold is associated to an age-dependent PD phenotype, including hypokinesia, Lewy body-like formation and nigrostriatal neurodegeneration. Enhancing lysosomal proteostasis reduces intracellular neuromelanin and prevents neurodegeneration in tyrosinase- overexpressing animals. Our results suggest that intracellular neuromelanin levels may set the threshold for the initiation of PD. 1 Neurodegenerative Diseases Research Group, Vall d’Hebron Research Institute (VHIR)-Network Center for Biomedical Research in Neurodegenerative Diseases (CIBERNED), 08035 Barcelona, Spain. 2 Telethon Institute of Genetics and Medicine (TIGEM), 80078 Naples, Italy. 3 Department of Neurology, Tohoku University School of Medicine, Miyagi 980-8574, Japan. 4 Department of Neurochemistry and Neuropharmacology, IIBB–CSIC, August Pi i Sunyer Biomedical Research Institute (IDIBAPS)-Center for Networked Biomedical Research on Mental Health (CIBERSAM), 08036 Barcelona, Spain. 5 Neurological Tissue Bank, Biobanc Hospital Clínic-IDIBAPS, 08036 Barcelona, Spain. 6 Institute of Neurology, Medical University of Vienna, 1090 Vienna, Austria. 7 Department of Biochemistry and Molecular Biology, Autonomous University of Barcelona, 08193 Barcelona, Spain. 8 Catalan Institution for Research and Advanced Studies (ICREA), 08010 Barcelona, Spain. These authors contributed equally: Iria Carballo-Carbajal, Ariadna Laguna. Correspondence and requests for materials should be addressed to M.V. (email: miquel.vila@vhir.org) NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y I n Parkinson’s disease (PD), neurons that contain the dark- Results brown cytoplasmic pigment neuromelanin (NM) are particu- Age-dependent NM formation in tyrosinase-overexpressing larly susceptible to neurodegeneration1. Indeed, while PD rats. The mechanism of synthesis of NM is poorly understood, patients exhibit an abnormal accumulation of alpha-synuclein but it is generally accepted that NM is formed as an inert cellular (aSyn) protein in different brain regions, and even in peripheral by-product of DA synthesis produced by simple non-enzymatic tissue2, neurodegeneration preferentially occurs within NM- DA autoxidation17. In contrast, it is well-established that the containing regions in these patients3–5. In contrast, neuronal synthesis of peripheral melanins (e.g. skin and hair), which occurs loss in non-melanized brain regions is either inconsistent, not within specialized cells (i.e. melanocytes), results from an enzy- specific to PD, or secondary to the loss of interconnected NM- matically driven biosynthetic pathway initiated with the hydro- containing neurons3–5. Such highly vulnerable NM-containing xylation of L-tyrosine to L-DOPA, followed by the oxidation brain regions include the substantia nigra pars compacta (SNpc), of L-DOPA to the melanin precursor DOPAquinone, in which where the loss of dopaminergic (DA) neurons therein leads to the tyrosinase is the key, rate-limiting enzyme18,19. Remarkably, typical motor symptoms of the disease and constitutes the car- tyrosinase expression and activity are not apparently restricted dinal pathologic diagnostic criterion for PD. In the human SNpc, to melanocytes but have also been observed, although at low which is the primary source of NM in the human brain, NM levels, in the brain, including human SNpc20–22 (Supplementary levels are actually so high that this structure can be seen mac- Figure 1). In addition, tyrosinase is also able to oxidize the roscopically as a darkened area (hence the origin of the name catechol ring of DA, which is an essential event required for NM given to this brain region)6. NM is restricted to catecholamine- synthesis19. However, whether brain tyrosinase may actually producing regions and forms only in neurons. It first becomes contribute to NM synthesis is currently unknown. Here we observable in the human SNpc at ~3 years of age and progres- assessed what are the consequences of overexpressing tyrosinase sively accumulates over time within the cells in which it has been in the SNpc of rodents. produced, as neurons apparently lack the mechanisms for Adult rats received a single unilateral stereotaxic injection of an degrading or eliminating this pigment. As a consequence, intra- adeno-associated viral (AAV) vector expressing human tyrosinase cellular NM builds up with age until occupying most of the (hTyr) above the right SNpc (Fig. 1a). By 2–4 weeks (w) post- neuronal cytoplasm7. Importantly, aging is the main risk factor AAV-hTyr injection, up to ~80–90% of ipsilateral SNpc DA for developing PD8. neurons were transduced and expressed the hTyr protein (Fig. 1b). DA-producing cell groups of the normal human midbrain By 2 months (m) post-AAV-hTyr injection, the ipsilateral differ markedly from each other in terms of the percentage of SNpc from these animals could be visualized macroscopically, NM-pigmented neurons they contain1,9,10. In PD, the estimated in the absence of any staining, as a darkened brown area, similar cell loss in these cell groups directly correlates with the percentage to human melanized SNpc tissue (Fig. 1c). In contrast, the of NM-pigmented neurons normally present in them1,9,10. Like- contralateral SNpc from the same animals could not be detected wise, within each cell group in PD brains, there is greater relative macroscopically, as rodents lack NM (Fig. 1c). Similar to humans, sparing of weakly pigmented than of strongly melanized ipsilateral SNpc from AAV-hTyr-injected animals could also neurons1,9,10. Also, classical Lewy bodies (LB), i.e. aSyn- be detected macroscopically as a hyperintense area by NM- containing intracytoplasmic inclusion bodies that represent the sensitive high-resolution T1-weighted magnetic resonance ima- pathological hallmark of the disease, as well as their presumed ging (Fig. 1d). Optical microscopy visualization confirmed that precursor structures, pale bodies (PB), typically appear within the the darkened SNpc area in AAV-hTyr-injected rats corresponded intracellular areas of the cytoplasm in which NM accumulates to the presence of an intracellular dark-brown fine granular and form in close physical association with this pigment11. Along pigment analogous to human NM within ipsilateral SNpc this line, studies in human brains have shown that aSyn redis- neurons (Fig. 1e). Virtually all (~97%) hTyr-expressing SNpc tributes to the lipid component of NM at early PD stages12 and DA neurons from AAV-hTyr-injected animals produced NM that aSyn becomes entrapped within NM granules extracted from (Supplementary Figure 2). As in humans, NM from AAV-hTyr- PD, but not control, brains13. Further linking PD neuropathology injected rats stained prominently with the melanin marker with NM, PD-linked neuroinflammatory changes are highly Masson-Fontana, which reflects the ability of NM to chelate localized within NM-containing areas and are barely observed in metals (Fig. 1f). Ultrastructural examination by electron micro- non-melanized regions, such as the cortex, despite the latter scopy identified NM granules from AAV-hTyr-injected rats as exhibiting PD-related aSyn depositions14. membrane-delimited autophagic structures with an electron- According to the above observations, PD pathogenesis appears dense matrix of irregular shape and size, associated with inextricably linked to the presence of NM. However, despite the characteristic lipid droplets, equivalent to what has been close and long-established association between NM and PD, the described in humans23 (Fig. 1g). Within the SNpc, NM pigment physiological significance of NM and its potential contribution to in AAV-hTyr-injected rats was restricted to TH−immunopositive PD pathogenesis remain unknown. The current lack of knowl- DA neurons (Supplementary Figure 2), thereby mimicking the edge about the role of NM both in healthy subjects and in PD pattern of human NM distribution within catecholamine- patients lies in the fact that, in contrast to humans, laboratory producing neurons6. animal species commonly used in experimental research, such as In human SNpc DA neurons, NM is continuously produced rodents, lack NM15. In fact, the great abundance of NM in the throughout life and progressively accumulates with age until it brainstem is unique to humans, as macroscopic dark pigmenta- occupies a major portion of the neuronal cytoplasm24. Similarly, tion of this brain area is not observed in other animal species16. quantification of intracellular NM optical density in SNpc Consequently, a factor so intimately linked to PD such as NM has neurons of AAV-hTyr-injected rats, which reflects the actual been surprisingly neglected so far in experimental in vivo para- concentration of NM7,10, revealed that intracellular NM levels digms of the disease. To address this major limitation, in the steadily increase in the SNpc of these animals over time, from 0.5 present study we generate the first rodent model of age- up to 24 m post-AAV-hTyr injection, until occupying most of the dependent human-like NM production in SNpc DA neurons at neuronal cytoplasm (Fig. 2a, b). In ~1–2 m, intracellular NM in levels up to those reached in elderly humans. Using this unique AAV-hTyr-injected rats reached levels equivalent to those in model, we show that progressive intracellular NM accumulation post-mortem SNpc tissue from elderly (~80 years old) human above a specific threshold is associated with a PD phenotype. control subjects (Fig. 2b, c). In ~2–4 m, AAV-hTyr-injected rats 2 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE a c Hu d Hu AAV-hTyr b TH hTyr e f Hu g Hu VTA SNpc VTA SNpc * * Fig. 1 Human-like NM production in AAV-hTyr-injected rats. a Schematic representation of the site of AAV-hTyr unilateral stereotaxic injection above the SNpc of the rat brain. b Representative photomicrographs of a 30-μm-thick ipsilateral SNpc section from an AAV-hTyr-injected rat (1 m post-AAV injection) immunostained for TH (red) and hTyr (green). Scale bar, 500 μm. c Left, representative unstained AAV-hTyr-injected rat brain (2 m post-AAV injection) mounted in a cryostat in which ipsilateral SNpc can be detected macroscopically as a brown, darkened area (dashed outline). A hole was made in the contralateral hemisphere as anatomical reference. Right, representative unstained midbrain from a 62-year-old human control subject (Hu) in which the SNpc can be detected macroscopically (bilateral dashed outlines). d Representative NM-sensitive high-resolution T1-weighted magnetic resonance imaging of an AAV-hTyr-injected rat brain at 2 m post-AAV injection (left, ex-vivo) and of a 59-year-old human control brain (Hu, right, in vivo). SNpc can be detected as a unilateral (AAV-hTyr-injected rodent) or bilateral (human) hyperintense area (dashed outlines). e Representative photomicrograph of an unstained 30-μm-thick ipsilateral SNpc section from an AAV-hTyr-injected rat (2 m post-AAV injection) in which NM is shown in brown. Inset, high magnification of a melanized neuron. Scale bars, 100 μm and 12.5 μm (inset). f Representative photomicrographs of Masson-Fontana melanin staining (NM in dark brown) in 5-μm-thick SNpc sections from an AAV-hTyr-injected rat at 2 m post-AAV injection (left) and an 80-year-old human control subject (Hu, right). Scale bars, 25 μm (left) and 12.5 μm (right). g Representative electron micrograph of NM granules in the ipsilateral SNpc of an AAV-hTyr-injected rat at 4 m post-AAV injection (left) and a 75-year-old human control subject (Hu, right). NM pigment is detected as an electron dense matrix. Characteristic associated lipid droplets are indicated with asterisks. Scale bars, 500 nm b c Intracellular NM density (OD) Intracellular NM density (OD) 40,000 a § * # 25,000 * * 30,000 *# 20,000 0.5 m 1m 2m *# Control *# 15,000 20,000 * 10,000 10,000 5000 0 0 0.5 1 2 4 12 24 Control ILBD iPD Time post-AAV-hTyr injection (months) Intracellular NM density (OD) Intracellular NM density (OD) 4m 12 m 24 m 150,000 # 100,000 * * iPD *# § * 100,000 *# *# 50,000 50,000 * 0 0 –50,000 –50,000 0.5 1 2 4 12 24 Control ILBD iPD Time post-AAV-hTyr injection (months) Fig. 2 Age-dependent NM accumulation in AAV-hTyr-injected rats. a Hematoxylin-eosin (H&E)-stained brain sections showing progressive intracellular NM accumulation (brown) within ipsilateral SNpc DA neurons from AAV-hTyr-injected rats. Scale bar, 12.5 μm. b Quantification of intracellular NM optical density in ipsilateral SNpc DA neurons of AAV-hTyr-injected rats. *p < 0.05, compared to 0.5 m; #p < 0.05, compared to 1 m; §p < 0.05, compared to 2&4 m. c Left, quantification of intracellular NM optical density in post-mortem SNpc sections from elderly human (average 80 years old) control subjects, age- matched ILBD subjects and age-matched idiopathic PD patients. *p < 0.05, compared to control subjects. Right, H&E-stained human brain sections. Scale bar, 12.5 μm. In b, c values are mean ± SEM. In b, n = 163 neurons from n = 7 rats (0.5 m), n = 136 neurons from n = 5 rats (1 m), n = 174 neurons from n = 5 rats (2 m), n = 162 neurons from n = 7 rats (4 m), n = 172 neurons from n = 6 rats (12 m), n = 179 neurons from n = 6 rats (24 m). In c, n = 1436 neurons from n = 6 control subjects, n = 640 neurons from n = 3 ILBD subjects, n = 644 neurons from n = 10 PD subjects. See Supplementary Table 1 for additional information on the human subjects used for the analyses in c. Statistical analyses: ANOVA on ranks; Dunn’s post-hoc test. Photomicrographs correspond to 5-μm-thick sections exhibited intracellular NM levels equivalent to those in post- steadily increase over time, reaching first levels equivalent to mortem SNpc tissue from (i) subjects with incidental LB disease those of elderly human brains and further increasing up to levels (ILBD), i.e. clinically healthy individuals exhibiting LB pathology attained in pre-PD subjects and PD patients. at autopsy who are considered to represent early, presymptomatic stages of PD, and (ii) established PD patients (Fig. 2b, c), in both of which intracellular NM levels appeared significantly higher Progressive PD-like neurodegeneration in NM-producing rats. than those from age-matched control subjects (Fig. 2c; see also Using our newly generated NM-producing rats, we next assessed Supplementary Table 1). whether age-dependent intracellular NM accumulation, up to Overall, these results show that hTyr expression in rat SNpc levels reached in elderly subjects and ILBD/PD cases, may be leads to the production of human-like NM within SNpc DA associated to compromised cell viability. Stereological cell counts neurons. As in humans, intracellular NM levels in these animals of SNpc TH-positive neurons in AAV-hTyr-injected rats revealed NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y a Tyrosine hydroxylase Masson-Fontana b c 0.5 m 0.5 m Contralateral Striatal TH (OD; % contralateral side) 25,000 120 Ipsilateral *# SNpc TH-positive neurons 20,000 100 * *# *# 80 *# 1m 1m 15,000 60 10,000 *# § 40 5000 20 2m 2m 0 0 Naive 0.5 1 2 4 12 24 Naive 0.5 1 2 4 12 24 Time post-AAV-hTyr injection (months) Time post-AAV-hTyr injection (months) 4m 4m d e Contralateral 100 *# * * #§ Ipsilateral NM+TH– 25,000 NM+TH+ *# % of NM-laden neurons SNpc total DA neurons 80 20,000 12 m 12 m *# 60 15,000 40 10,000 *#§ 24 m 24 m 20 5000 0 0 0.5 1 2 4 12 24 Naive 0.5 1 2 4 12 24 Time post-AAV-hTyr injection (months) Time post-AAV-hTyr injection (months) Fig. 3 Progressive nigrostriatal degeneration in AAV-hTyr-injected rats. a Left, ipsilateral SNpc and striatum (inset) sections immunostained for TH (blue). Right, ipsilateral SNpc sections stained with Masson-Fontana (NM in dark brown; inset, higher magnification of melanized neurons). Scale bars, 300 μm (left), 2 mm (left inset), 150 μm (right), and 12.5 μm (right inset). b Stereological cell counts of SNpc TH-positive neurons in AAV-hTyr-injected rats. *p < 0.05, compared to respective contralateral (non-injected) side; #p ≤ 0.05, compared to ipsilateral naive, 0.5, 1, and 2 m; §p < 0.05, compared to ipsilateral 4 and 12 m (two-way ANOVA; Student–Newman–Keuls post-hoc test). c Optical densitometry of striatal TH-positive fibers in AAV-hTyr- injected rats. *p < 0.05, compared to 0.5 m; #p < 0.05, compared to naive (ANOVA on ranks; Dunn’s post-hoc test). d Stereological cell counts of SNpc NM + /TH + (white arrowhead) and NM + /TH− (black arrowhead) neurons vs total NM + neurons in AAV-hTyr-injected rats. *p < 0.05, compared to 0.5 m; #p < 0.05, compared to 1 and 2 m; §p < 0.05, compared to 12 m (one-way ANOVA; Student–Newman–Keuls post-hoc test). e Stereological cell counts of total SNpc DA neurons (including TH-immunopositive and TH-immunonegative melanized neurons) in AAV-hTyr-injected rats. *p < 0.05, compared to respective contralateral side; #p < 0.05, compared to ipsilateral naive, 0.5 and 1 m; §p < 0.05, compared to ipsilateral 2, 4, and 12 m (two-way ANOVA; Student–Newman–Keuls post-hoc test). In all panels, values are mean ± SEM. In a–e, n = 6 (Naïve), n = 8 (0.5 m), n = 7 (1 m), n = 5 (2 m), n = 8 (4 m), n = 7 (12 m), and n = 6 (24 m). Photomicrographs correspond to 5-μm-thick sections a progressive, age-dependent loss of TH-positive SNpc neurons, was observed in AAV-hTyr-injected rats, as indicated by the starting by 4 m post-AAV-hTyr injection (Fig. 3a, b). Progressive preservation of the ipsilateral gamma-aminobutyric acidergic loss of SNpc TH-positive neurons in AAV-hTyr-injected rats was nigrotectal pathway in these animals (Supplementary Figure 4), accompanied by a reduction of striatal DA TH-positive fibers, as despite this region being systematically transduced with hTyr in measured by optical densitometry (Fig. 3c). Loss of nigrostriatal AAV-hTyr-injected animals, due to its proximity to the site of the DA fibers in AAV-hTyr-injected rats was accompanied by AAV-hTyr injection (Supplementary Figure 2). Relevant to decreased striatal DA levels, as measured by high-performance humans, these neurons do not degenerate in PD patients3 and liquid chromatography (HPLC) (Supplementary Table 2), and do not contain NM in humans, as they do not use a abnormally enlarged nigrostriatal TH-positive nerve terminals catecholaminergic NM precursor as neurotransmitter. In addi- (Supplementary Fig. 3). tion, no nigrostriatal DA degeneration was observed in aging Similar to PD brains, NM-laden neurons from AAV-hTyr- non-melanized contralateral SNpc from AAV-hTyr-injected rats injected rats also exhibited an early phenotypic loss of TH (Fig. 3b, e) nor in aging non-melanized ipsilateral SNpc from expression, as indicated by an increased percentage of TH- AAV-empty vector (EV)-injected or vehicle-injected control rats immunonegative neurons within the total population of NM- (Supplementary Figure 4). containing neurons (Fig. 3d). TH-immunonegative pigmented We next determined whether the neurodegenerative changes neurons are believed to represent dysfunctional/dying SNpc DA observed in AAV-hTyr-injected rats were accompanied by neurons1. To distinguish between reduced TH expression and functional alterations. At different time-points post-AAV-hTyr actual cell death, we performed stereological cell counts of total injections, from 0.5 up to 24 m, rats were subjected to behavioral SNpc DA neurons (including TH-immunopositive and TH- analyses with the cylinder test, which allows asymmetrical immunonegative melanized SNpc neurons), which confirmed a alterations in nigrostriatal DA function to be detected25. progressive loss of SNpc DA neurons in AAV-hTyr-injected rats Concomitant with the ipsilateral degeneration of NM-laden SNpc (Fig. 3e). In parallel to the loss of melanized SNpc neurons, there neurons reported above, AAV-hTyr-injected rats exhibited was a progressive depigmentation of the SNpc in AAV-hTyr- contralateral forepaw hypokinesia, as indicated by reduced injected rats, as it occurs in PD patients (Fig. 3a, right panel). contralateral forepaw use, compared to AAV-EV-injected animals Correlation analyses in AAV-hTyr-injected rats confirmed a (Fig. 4a). These motor alterations started by 2 m post-AAV-hTyr strong positive and significant correlation between intracellular injection, thus coinciding with the first signs of striatal axonal NM levels and SNpc DA neurodegenerative changes (Supple- damage (i.e. axonal swelling, Supplementary Figure 3) and mentary Figure 3). In contrast, no degeneration of non-DA, non- preceding nigrostriatal neurodegeneration in these animals melanized neighboring gamma-aminobutyric acid nigral neurons (Fig. 3a–e). As motor impairment preceded overt neuron cell 4 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE a AAV-EV-injected AAV-hTyr-injected b Veratridine 50 µM c Amphetamine S1 S2 210 2000 4000 500 1 µM 10 µM 100 µM Extracellular DA (% of baseline) Extracellular DA (% of baseline) Extracellular DA (% of baseline) * Use of contralateral forepaw 175 *# * (% of AAV-EV-injected) 1500 3000 140 350 * * * 105 1000 2000 * 70 200 500 1000 * 35 0 0 0 50 0.5 1 2 3 6 12 24 –1 0 1 2 3 4 5 –2 –1 0 1 2 3 0 1 2 Time post-AAV-hTyr injection (months) Time (h) Time (h) Time (h) Vehicle AAV-hTyr-injected Vehicle AAV-hTyr-injected Fig. 4 Early dopaminergic dysfunction in AAV-hTyr-injected rats. a Contralateral forepaw use in AAV-hTyr- and AAV-EV-injected rats, as assessed with the cylinder test. *p ≤ 0.05, compared to AAV-EV-injected animals at the same time-point; #p < 0.05, compared to 0.5 m AAV-hTyr animals (two-way ANOVA; Student–Newman–Keuls post-hoc test). b Striatal DA release in AAV-hTyr- and vehicle-injected rats measured by microdialysis in the ipsilateral striatum following local amphetamine (left) or veratridine (right) administration by reverse-dialysis. *p < 0.05, compared to vehicle-injected animals (ANOVA for repeated measures of the DA values during the specified time periods; Tukey’s post-hoc test). Baseline DA concentration (fmol/fraction- 20 min): amphetamine experiments, 82.83 ± 19.16 (vehicle) vs 94.38 ± 21.82 (AAV-hTyr); veratridine experiments, 25.52 ± 4.33 (vehicle) vs 29.93 ± 6.46 (AAV-hTyr). c Striatal DA release in AAV-hTyr- and vehicle-injected rats measured by microdialysis in the ipsilateral striatum following electrical stimulation of the MFB for 10-min periods under S1 and S2 conditions (S1, 2.0 Hz, 0.1 mA, 0.2 ms; S2, 10 Hz, 0.1 mA, 1 ms). *p < 0.05, compared to stimulated AAV-hTyr-injected rats (ANOVA for repeated measures; Tukey’s post-hoc test). Baseline DA concentration (fmol/fraction-10 min): 16.8 ± 2.38 (vehicle) vs 23.02 ± 6.37 (AAV-hTyr). In all panels, values are mean ± SEM. In a, AAV-EV-injected rats n = 20 (0.5 m), n = 20 (1 m), n = 9 (2 m), n = 9 (3 m), n = 8 (6 m), n = 12 (12 m), n = 4 (24 m); and AAV-hTyr-injected rats n = 29 (0.5 m), n = 8 (1 m), n = 21 (2 m), n = 14 (3 m), n = 7 (6 m), n = 14 (12 m), n = 6 (24 m). In b, c, experiments were performed at 1-2 m post-AAV injection. In b, n = 6 vehicle-injected, n = 6 AAV-hTyr-injected rats + Amphetamine and n = 8 AAV-hTyr-injected rats + Veratridine. In c, n = 4 animals per group death in AAV-hTyr-injected rats, we next evaluated whether hTyr injections and reached a 50% reduction. These results striatal DA function was already impaired in these animals at concur with equivalent neuroimaging in vivo analyses in PD such early stages. Striatal DA release following chemical or patients, in which neuroimaging reductions in DAT striatal electrical stimulation was measured by microdialysis in the density (i.e. by [11C]d-threo-methylphenidate PET or 123I-2β- ipsilateral striatum of vehicle- and AAV-hTyr-injected rats. At carbomethoxy-3β-(4-iodophenyl)-N-(3-fluoropropyl)-nortro- baseline, no differences in extracellular striatal DA concentration, pane SPECT) can be detected at early stages of the disease, measured by HPLC, were observed between control- and AAV- including in pre-motor idiopathic REM sleep behavior hTyr-injected freely moving rats, indicating that hTyr expression disorder patients28–30 and pre-symptomatic LRRK2 non- per se does not alter basal DA neurotransmission (Fig. 4b). manifesting carriers31–34. In contrast, neuroimaging reductions Stimulation of striatal DA release by local amphetamine in VMAT2 striatal density (i.e. by [11C]( ± )dihydrotetrabenazine administration induced concentration-dependent increases of PET) only become consistent once the disease is fully estab- extracellular striatal DA in both control- and AAV-hTyr-injected lished35–38, probably as a compensatory mechanism to maintain rats, but this effect was much less pronounced in the latter group extracellular levels of DA at early disease stages and thus delay the of animals (Fig. 4b). Similarly, when stimulated with the onset of parkinsonian symptoms37. depolarizing agent veratridine, striatal DA release was also less Concomitant with SNpc DA neurodegeneration, abundant prominent in AAV-hTyr-injected rats compared to control- extracellular NM aggregates, released from dying neurons, were injected animals (Fig. 4b). We next assessed striatal DA release detected in the SNpc of AAV-hTyr-injected rats, as it is usually induced by electrical stimulation of the ipsilateral medial observed in aged and PD brains39 (Fig. 5a, c). Extracellular NM forebrain bundle (MFB) via an electrode implanted in these was associated with sustained microglial/macrophage activation animals. Similar to freely moving rats, no differences in baseline (Fig. 5b, d), with NM fragments as well as entire NM-filled extracellular DA levels in the striatum were detected between residual neurons being surrounded by, or contained within, control- and AAV-hTyr-injected anaesthetized animals (Fig. 4c). activated microglia/macrophages (Fig. 5b–e). This phenomenon, However, following electrical stimulation of the MFB, AAV- known as neuronophagia, is commonly seen in post-mortem aged hTyr-injected animals exhibited a markedly reduced release of and PD brains39 (Fig. 5c) and is indicative of an active, ongoing striatal DA compared to control-injected rats (Fig. 4c). These neurodegenerative process40. Double immunofluorescence ana- results indicate that synaptic DA release by NM-laden nigros- lyses revealed that neuronophagia was directed not only at triatal neurons is impaired before any overt SNpc cell loss has extracellular NM debris but also, in some instances, at occurred, as is believed to be the case in PD patients26,27. dysfunctional/dying NM-laden neurons that have lost their TH To further corroborate the occurrence of a synaptic failure in phenotype (Fig. 5e). In contrast, no neuronophagia was detected these animals, we also measured the striatal densities of dopamine surrounding morphologically intact, apparently healthy, TH- transporter (DAT) and vesicular monoamine transporter 2 positive NM-laden cells (Fig. 5e). Ultimately, and analogous to (VMAT2) immunopositive fibers in AAV-hTyr-injected rats at PD brains, NM-filled microglia from AAV-hTyr-injected rats different times post-AAV-hTyr injections, from 0.5 up to 24 m appeared to migrate to blood vessels to exit the brain along with (Supplementary Figure 3). Striatal DAT-positive fibers in AAV- the NM pigment, as revealed by abundant perivascular NM in hTyr-injected rats exhibited an earlier and more extensive loss these animals (Fig. 5a, c, f). No neuroinflammatory changes were than striatal TH-positive fibers, starting by 2 m post-AAV-hTyr observed in the SNpc of AAV-EV-injected control rats nor in the injections (i.e. at the time of neuronal dysfunction but before the contralateral (non-melanized) SNpc of AAV-hTyr-injected rats occurrence of overt neurodegeneration in these animals) and (Supplementary Figure 4). reaching up to a 90% striatal depletion. In contrast, striatal DA oxidation to quinones is an essential event required for VMAT2 levels were significantly decreased by 12 m post-AAV- NM synthesis41 (Supplementary Figure 5). Because DA-derived NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y a H&E c PD patient AAV-hTyr rat # * 8000 10,000 * Extracellular NM * Extracellular NM 8000 6000 Extracellular NM 6000 4000 4000 2 2000 R = 0.7159 2000 20,000 15,000 10,000 5000 0 0 0.5 1 2 4 12 24 SNpc total DA neurons Neuronophagia Time post-AAV-hTyr injection (months) Pearson r = –0.8461; p = 0.0337 b H&E 8 20 R 2 = 0.8622 15 Neuronophagia 6 Neuronophagia 10 4 * Perivascular NM 5 2 0 0 2000 4000 6000 8000 0.5 1 2 4 12 24 Extracellular NM Time post-AAV-hTyr injection (months) Pearson r = 0.9285; p = 0.0075 d Iba1 e f SNpc TH Iba1 BF Hoechst Lectin Iba1 GFAP Hoechst BF CD68 1.5 R 2 = 0.8884 1.0 Perivascular NM SNpc 0.5 0.0 5 10 15 20 –0.5 Neuronophagia TH Iba1 BF Hoechst Pearson r = 0.9425; p = 0.0049 Fig. 5 Extracellular NM and neuronophagia in AAV-hTyr-injected rats. a Left, extracellular NM aggregates in ipsilateral SNpc from AAV-hTyr-injected rats. *p < 0.05, compared to 0.5&1 m; #p < 0.05, compared to 2 m (one-way ANOVA; Student–Newman–Keuls post-hoc test). Middle, hematoxylin-eosin (H&E)-stained ipsilateral SNpc sections from AAV-hTyr-injected rats (4 m). Scale bar, 25 μm. Right, Pearson correlation analysis between total SNpc DA neurons and extracellular NM. b Left, neuronophagia in ipsilateral SNpc from AAV-hTyr-injected rats. *p < 0.05, compared to 0.5 m (ANOVA on ranks; Dunn’s post-hoc test). Middle, H&E-stained ipsilateral SNpc sections from AAV-hTyr-injected rats (4 and 12 m). Scale bar, 62.5 μm. Right, Pearson correlation analysis between extracellular NM and neuronophagia. c H&E-stained midbrain sections from PD post-mortem brains and AAV-hTyr-injected rats. Scale bars, 12.5 μm (top & bottom) and 25 μm (middle). d Ipsilateral SNpc sections from AAV-hTyr-injected rats (2–12 m) immunostained for the microglial marker Iba1 (blue, top) or the macrophage marker CD68 (blue, bottom). Scale bars, 375 μm (Iba1, low magnification), 12.5 μm (Iba1, high magnification), 150 μm (CD68, low magnification), and 25 μm (CD68, high magnification). e Ipsilateral SNpc sections from AAV-hTyr-injected rats (4 m) immunostained for TH (red) and Iba1 (green) and co-stained with the nuclear marker Hoechst (blue). Arrowheads, intact TH-positive NM-laden neurons; single arrow, neuronophagia directed at extracellular NM debris; double arrows, neuronophagia directed at a TH-immunonegative NM-laden neuron. Scale bar, 25 μm. f Top, Ipsilateral SNpc section from an AAV-hTyr-injected rat (4 m) immunostained with Iba1 (green) and astrocytic marker GFAP (purple), co- stained with blood vessel marker tomato lectin (red) and Hoechst (blue). Arrowheads, microglial cells carrying NM towards blood vessels. Arrow, perivascular NM. Scale bar, 25 μm. Bottom, Pearson correlation analysis between neuronophagia and perivascular NM. In a, b histograms, values are mean ± SEM. In a, n = 8 (0.5 m), n = 7 (1 m), n = 5 (2 m), n = 8 (4 m), n = 7 (12 m), n = 6 (24 m). In b, n = 6 (0.5 m), n = 5 (1 m), n = 5 (2 m), n = 7 (4 m), n = 6 (12 m), n = 6 (24 m). For Pearson correlation analyses each point represents the average value for the corresponding parameter at any given time post-AAV-hTyr injection (0.5, 1, 2, 4, 12, and 24 m). BF bright-field. In all panels, unstained NM appears as brown. Photomicrographs correspond to 5-μm- thick sections quinones can be potentially toxic to DA neurons42–45, we next injection: 1 m (i.e. before the occurrence of NM-linked neuronal determined whether the neurodegenerative process observed in dysfunction in these animals), 2 m (i.e. at the onset of neuronal our NM-producing rats could be attributed to hTyr-mediated dysfunction but before the occurrence of neurodegeneration) and production of potentially toxic oxidized DA species. To address 4 m (i.e. once neurodegeneration is established). No significant this issue, we developed a method to directly measure DA changes were detected between ipsilateral and contralateral oxidation in AAV-hTyr-injected rats by ultra-performance liquid ventral midbrain levels of oxidized DA at any of the times chromatography-tandem mass spectrometry (UPLC-MS/MS). analyzed (Supplementary Figure 5), therefore arguing against a Levels of oxidized DA were measured in both ipsilateral (i.e. major significant contribution of oxidized DA species in the melanized) and contralateral (i.e. non-melanized) ventral mid- pathological changes observed in AAV-hTyr-injected rats. brains of AAV-hTyr-injected rats, at different times post-AAV Further supporting this concept, we did not detect significant 6 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE alterations in DA metabolism in these animals, either preceding MB are spherical eosinophilic nuclear inclusions found in or at the onset of neurodegeneration, as reflected by unchanged pigmented neurons of the human SNpc, the incidence of levels of DA and other DA metabolites in their ventral midbrains which increases with advancing age47. Similar to humans47, (Supplementary Figure 5). one or several MB were detected in the nucleus of pigmented Our results reveal that age-dependent NM production within SNpc neurons from AAV-hTyr-injected rats (Fig. 6c, d, f). In SNpc DA neurons in AAV-hTyr-injected rats is associated with these animals, MB were intensely immunopositive for p62, a neuronal dysfunction and progressive nigrostriatal neurodegen- common component of neuropathological inclusions, and eration, equivalent to that occurring in PD patients46, once a ubiquitin (Fig. 6f), as it has been previously reported in certain threshold of intracellular NM accumulation is reached. humans47. No MB were found in non-pigmented neurons Remarkably, the intracellular NM levels above which AAV-hTyr- from AAV-hTyr-injected rats. Concomitant with the occur- injected rats started exhibiting functional alterations (by 2 m) and rence of MB, NM-laden neurons from AAV-hTyr-injected rats overt neurodegeneration (by 4 m) were equivalent to those also exhibited cytoplasmic PB and LB-like inclusions, often observed in ILBD subjects and PD patients (Fig. 2b, c), thus within the same cell, as is commonly seen in PD brains47 allowing a specific pathogenic threshold for intracellular NM (Fig. 6c–i). No cytoplasmic inclusions were found in non- accumulation to be defined (see dashed red lines in Fig. 2b, c). pigmented neurons from these animals. PB are identified as These results are consistent with intracellular NM levels setting a weakly or non-eosinophilic cytoplasmic areas of varied threshold for the initiation of PD-related neurodegeneration. morphologies that displace NM, lacking the spherical and compact appearance of LB-like structures (Fig. 6d). As in PD, PD-like inclusion formation in NM-producing rats. In addition PB and LB-like inclusions from AAV-hTyr-injected rats were to neurodegeneration, ipsilateral NM-laden SNpc DA neurons from immunopositive for p62 and, in some instances, for aSyn AAV-hTyr-injected rats also exhibited intracellular inclusion bodies (both full length and hyperphosphorylated) and ubiquitin typical of human post-mortem aged and PD brains, including (Fig. 6e, g–i and Supplementary Figure 6). This is consistent nuclear Marinesco bodies (MB) and cytoplasmic PB and LB-like with previous studies in human PD brains reporting a inclusions (Fig. 6). Inclusion body formation in these animals: (i) selective, early, and consistent incorporation of p62 into both was restricted to NM-containing neurons, (ii) peaked at 2 m post- PB and LB, suggesting an important role for p62 in their AAV-hTyr injection, thus coinciding with functional alterations formation, in contrast to a less consistent and less selective and preceding neurodegeneration in these animals, and (iii) was accumulation of ubiquitin within these cytoplasmic inclu- substantially reduced by 4 m and onwards, once neurodegeneration sions47. Cytoplasmic inclusions in AAV-hTyr-injected rats was already established (Fig. 6a, b). emerged within NM-filled areas, either overlapping with or H&E H&E aSyn a b c d e PD patient 15 8 inclusions (average per section) 8 bodies (average per section) * Neurons with cytoplasmic * Neurons with Marinesco Neurons with inclusions PB * (average per section) # 6 6 LB-Like 10 4 4 AAV-hTyr rat 5 2 2 0 0 0 0.5 1 2 4 12 24 0.5 1 2 4 12 24 0.5 1 2 4 12 24 Time post-AAV-hTyr injection (months) Time post-AAV-hTyr injection (months) Time post-AAV-hTyr injection (months) f g h i Ub p62 p62 p62 p62 aSyn p62 TH * * Hoechst BF BF BF BF Ub BF Hoechst BF * * Fig. 6 PD-type inclusion formation in AAV-hTyr-injected rats. a Quantification of NM-laden neurons with p62-positive intranuclear Marinesco bodies (MB) in ipsilateral SNpc from AAV-hTyr-injected rats. *p < 0.05, compared to 0.5 m (ANOVA on ranks; Dunn’s post-hoc test). b Quantification of NM-laden neurons with p62-positive total intracytoplasmic inclusion bodies (left) and pale bodies (PB) and/or Lewy body (LB)-like p62-positive inclusions (right) in ipsilateral SNpc from AAV-hTyr-injected rats. Left, *p < 0.05, compared to 0.5, 1, 4, 12, and 24 m (one-way ANOVA; Student–Newman–Keuls post-hoc test). Right, *p < 0.05, compared to PB at 0.5, 4, 12, and 24 m; #p < 0.05, compared to LB-like inclusions at 0.5, 1, 4, and 24 m (two-way ANOVA; Holm- Sidak post-hoc test). c–e Midbrain sections from PD post-mortem brains (top) and AAV-hTyr-injected rats (bottom) exhibiting MB (c, white arrowheads), PB (d, blue arrowheads) and LB-type aSyn-positive inclusions (e, arrows). In c, d, Hematoxylin-eosin (H&E) staining. In e, aSyn immunostaining (in blue). Scale bars, 12.5 μm. f Ipsilateral SNpc sections from AAV-hTyr-injected rats exhibiting NM-laden neurons with MB (white arrowhead) detected by immunofluorescence with p62 (red) and ubiquitin (Ub, green). Nuclei are stained with Hoechst (blue). Scale bar, 12.5 μm. g Ipsilateral SNpc sections from AAV-hTyr-injected rats exhibiting NM-laden neurons with p62-positive multipunctate cytoplasmic inclusions (yellow arrowhead), PB (blue arrowhead) or LB-like inclusions (arrows). Scale bar, 12.5 μm. h Ipsilateral SNpc section from an AAV-hTyr-injected rat exhibiting a NM-laden neuron with an intracytoplasmic LB-like inclusion immunopositive for p62 (red), aSyn (purple) and ubiquitin (Ub, green). Scale bar, 12.5 μm. i Ipsilateral SNpc section from an AAV-hTyr-injected rat immunostained for p62 (red) and TH (purple) in which a NM-laden neuron containing both a PB (blue arrowhead) and a MB (white arrowhead) exhibits decreased TH immunostaining (asterisk). Nuclei are stained with Hoechst (blue). Scale bar, 25 μm. In a, b, values are mean ± SEM. n = 8 (0.5 m), n = 5 (1 m), n = 6 (2 m), n = 5 (4 m), n = 6 (12 m), n = 5 (24 m). Ub ubiquitin, BF bright-field. In all panels, unstained NM appears as brown. Photomicrographs correspond to 5-μm-thick sections NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y displacing NM granules, and often appeared as early multiple rats and the potential pathogenic role attributed to aSyn aggre- punctate small aggregates immunopositive for p62 that gation in PD, we next assessed whether NM-linked PD-like seemed to progressively coalesce into more compacted LB- pathology in AAV-hTyr-injected animals was dependent on aSyn. like spherical structures (Fig. 6g), as similarly reported in This question was addressed by inducing NM production with post-mortem PD brains47. AAV-hTyr in the SNpc of aSyn-deficient mice (Fig. 7a). Because Taken together, our results show that intracellular NM aSyn has been reported to be able to interact with tyrosinase49 accumulation is associated with the formation of PD-like neuronal and to modulate ultra-violet radiation-induced melanin synthesis inclusions in AAV-hTyr-injected rats. While the mechanism and in melanoma cells50, we first determined whether the lack of aSyn significance of inclusion body formation in PD remain unknown, in aSyn-deficient mice could interfere with AAV-hTyr-induced the temporal pattern of inclusion formation in AAV-hTyr-injected production of NM in these animals. Ruling out this possibility, rats (i.e. preceding neurodegeneration and decreasing in parallel intracellular NM levels within AAV-hTyr-injected SNpc neurons with neuronal death) suggests that inclusion-containing neurons are were comparable between wild-type (WT) and aSyn-deficient those that preferentially degenerate in these animals. In agreement mice at 6 m post-AAV-hTyr injection (Fig. 7b). By that time, WT with this, the number of neuronal inclusions in PD brains at and aSyn-deficient mice exhibited intracellular NM levels similar advanced stages of the disease is much lower than that observed in to the pathogenic threshold of intracellular NM accumulation early PD cases14, as it occurs in AAV-hTyr-injected rats. In defined above for rats and humans (Fig. 7b). We next assessed addition, in both humans48 and AAV-hTyr-injected rats (Fig. 6i), whether the absence of aSyn in aSyn-deficient mice may inclusion-containing neurons often exhibit decreased intracellular prevent or attenuate the formation of PD-like cytoplasmic TH levels, which reflects neuronal dysfunction at early stages of inclusions within NM-laden neurons in AAV-hTyr-injected mice. neurodegeneration1. Consistent with this, inclusion formation in At 2 m post-AAV-hTyr injections (i.e. at the peak of inclusion AAV-hTyr-injected rats peaked at the onset of functional formation defined in AAV-hTyr-injected rats), NM-laden SNpc alterations (i.e. impaired DA release and hypokinesia) in these neurons from WT mice exhibited p62-immunopositive PB and animals. Our results indicate that PD-type inclusion body LB-like inclusions at levels equivalent to those observed in AAV- formation reflects a pathologic process related or contributing to hTyr-injected rats (Fig. 7c and Supplementary Figure 6). Unex- PD-linked neuronal dysfunction and degeneration. pectedly, NM-laden SNpc neurons from AAV-hTyr-injected aSyn-deficient mice exhibited similar p62-immunopositive aSyn is dispensable for NM-linked PD pathology. Given the inclusions at levels comparable to those of WT animals presence of PD-like aSyn-positive inclusions in NM-producing (Fig. 7c), indicating that the lack aSyn did not prevent the a c WT aSyn KO d AAV-hTyr p62 Contralateral 150 AAV-hTyr-injected (OD, % contralateral side) * 100 Striatal TH aSyn aSyn KO / WT 50 b H&E WT aSyn KO BF 0 WT aSyn KO Contralateral AAV-hTyr-injected Intracellular NM density (OD) 75,000 10 PB 150 SNpc TH-positive neurons Neurons with inclusions LB-like (average per section) (% contralateral side) 60,000 8 100 * 45,000 6 * 30,000 4 50 15,000 2 0 0 0 WT aSyn KO WT aSyn KO WT aSyn KO AAV-hTyr AAV-hTyr Fig. 7 Dispensability of aSyn for PD-like inclusion formation and neurodegeneration linked to NM accumulation. a Schematic representation of the site of AAV-hTyr unilateral stereotaxic injection above the SNpc of aSyn knockout (KO) and wild-type (WT) mice. b Top, representative photomicrographs of NM-laden neurons in hematoxylin-eosin (H&E)-stained ipsilateral SNpc brain sections from AAV-hTyr-injected aSyn KO and WT mice at 6 m post-AAV injection. Scale bar, 12.5 μm. Bottom, quantification of intracellular NM optical density in ipsilateral SNpc DA neurons of AAV-hTyr-injected aSyn KO and WT mice. p = 0.879 (two-tailed t-test). c Top, ipsilateral SNpc sections from AAV-hTyr-injected aSyn KO and WT mice exhibiting NM-laden neurons with p62-positive (red) PB (arrowhead) and LB-like inclusions (arrow) at 2 m post-AAV injection. aSyn immunofluorescence is shown in green. Scale bar, 12.5 μm. Bottom, quantification of NM-laden neurons with p62-positive PB or LB-like inclusions in AAV-hTyr-injected aSyn KO and WT mice at 2 m post- AAV injection. p = 0.866 (two-way ANOVA). d Top, optical densitometry of striatal TH-positive fibers in AAV-hTyr-injected aSyn KO and WT mice at 6 m post-AAV injection. *p < 0.05, compared to respective contralateral (non-injected) side (two-way ANOVA; Student–Newman–Keuls post-hoc test). Bottom, stereological cell counts of SNpc TH-positive neurons in AAV-hTyr-injected aSyn KO and WT mice at 6 m post-AAV injection. *p < 0.05, compared to respective contralateral (non-injected) side (two-way ANOVA; Student–Newman–Keuls post-hoc test). In all panels, values are mean ± SEM. In b, n = 76 neurons from n = 4 mice (WT), n = 112 neurons from n = 4 mice (aSyn KO). In c, n = 3 (WT), n = 4 (aSyn KO) mice. In d (top), n = 9 (WT), n = 5 (aSyn KO) mice. In d (bottom), n = 8 (WT), n = 6 (aSyn KO) mice. BF bright-field. In all panels, unstained NM appears as brown. Photomicrographs correspond to 5-μm-thick sections 8 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE formation of cytoplasmic inclusions. While in WT mice these which cell function and survival become compromised56. For inclusions contained aSyn, p62, and ubiquitin, in aSyn-deficient instance, mitochondrial quality control is tightly linked to the mice they were immunopositive for p62 and ubiquitin but lacked proteolytic cytosolic systems and if the latter are blocked, aSyn (Fig. 7c and Supplementary Figure 6). Finally, to determine damaged mitochondria gradually accumulate57. Supporting this whether aSyn actually contributed to NM-linked cell death in concept, NM-laden cells exhibited impaired mitochondrial NM-producing animals, we compared nigrostriatal degeneration respiration, as determined by respirometry with a Seahorse flux between AAV-hTyr-injected WT and aSyn-deficient mice. At 6 m analyzer system (Fig. 8h), and increased production of reactive post-AAV-hTyr injections, WT and aSyn-deficient mice exhib- oxygen species (ROS), as assessed by flow cytometry with the ited comparable levels of striatal DA denervation and SNpc DA green fluorescent dye CellROX (Fig. 8i). Concomitantly, these neurodegeneration (Fig. 7d), indicating that aSyn is not con- cells exhibited reduced metabolic activity, as revealed by a tributing to NM-linked cell death. decreased capacity to reduce resazurin to red-fluorescent Overall, our results indicate that whereas NM accumulation resorufin (Fig. 8j). These functional changes ultimately compro- above a pathogenic threshold is associated with the formation of mised cell viability, as NM-producing cells started to die by 6d PD-like aSyn-positive inclusions, aSyn is not actually required for post-hTyr induction (Fig. 8k). By that time, some NM-filled neither inclusion formation nor nigrostriatal neurodegeneration neurons exhibited a nucleus that was apparently displaced and in AAV-hTyr-injected animals. extruded by NM while leaking their NM contents outside the cell as they degenerate (Fig. 8l). By 8d, most NM-laden cells have disappeared, with only some NM-filled cellular remains and General proteostasis failure in NM-producing cells. The accu- extracellular residual NM persisting in the medium (Fig. 8l). mulation of p62-immunopositive inclusion bodies in NM- These results indicate that the continuous buildup of NM producing rats and the dispensability of aSyn for their forma- within autophagic compartments ultimately exhausts the autop- tion suggests a general failure of cellular proteostasis, beyond hagic capacity of the cell, leading to a general failure of aSyn, in NM-laden neurons. Indeed, p62-positive aggregates are proteostasis and subsequent dysfunction and degeneration of typically formed following inhibition of autophagy or NM-laden cells. proteasome-mediated intracellular degradation51. Both autop- hagy and ubiquitin-proteasome (UPS) systems are defective within NM-laden, but not non-melanized, regions from post- Modulation of NM levels and pathology in NM-producing rats. mortem PD brains52,53. In neural cells, suppression of basal To corroborate whether proteostasis failure was indeed respon- autophagy is sufficient to cause neurodegeneration, even in the sible for PD-like pathology linked to NM accumulation, we next absence of any disease-associated mutant proteins54. assessed the effects of enhancing proteostasis with transcription To determine whether NM accumulation may indeed affect factor EB (TFEB) in NM-producing rats. TFEB is a master reg- cellular proteostasis, additional experiments were performed in ulator of autophagy that upon activation translocates into the differentiated human catecholaminergic neuroblastoma SH-SY5Y nucleus and induces the biogenesis of lysosomes and autopha- cells inducible for hTyr expression. As in AAV-hTyr-injected gosomes, boosts autophagic cellular clearance, promotes lysoso- rats, induction of hTyr expression in these cells resulted in a mal exocytosis, and modulates general proteostasis58–62. By doing progressive intracellular production and accumulation of NM, so, TFEB overexpression has been shown to protect against until most of the cellular cytoplasm was occupied by 6 days (d) pathological substrate accumulation and cell death in several post-induction (Fig. 8a and Supplementary Figure 7). In humans, rodent models of lysosomal storage disorders and neurodegen- intracellular NM is known to continuously accumulate over a erative diseases63. To assess the effects of TFEB on NM- lifetime into autophagic structures23. In agreement with this, NM producing rats, both AAV-TFEB and AAV-hTyr were co- from both NM-producing rats and hTyr-expressing cells injected into the rat SNpc, with additional control groups of appeared enclosed within single-membrane-delimited mature animals receiving equivalent amounts of either vehicle, AAV- lysosomes/autophagolysosomes (Supplementary Figure 8) or hTyr or AAV-TFEB, separately. Co-injection of AAV-TFEB and inside double-membrane-delimited autophagosomes (Fig. 8b), AAV-hTyr resulted in a marked nuclear expression of TFEB as assessed by electron microscopy examination. Consistent with within SNpc neurons, without interfering with hTyr expression or the autophagic nature of NM granules, levels of the lysosomal NM production in these cells at 5 m post-AAV injection (Fig. 9a). structural marker Lamp1 and the autophagosome marker LC3-II/ In agreement with the well-established role of TFEB at inducing I increased in parallel with NM production in hTyr-expressing lysosomal biogenesis, SNpc DA neurons from AAV-TFEB- cells (Fig. 8c and Supplementary Figure 7) and Lamp1 was shown injected animals exhibited increased levels of Lamp1, up to 1 to co-localize with NM granules in AAV-hTyr-injected rodents year post-AAV injection (Fig. 9b). In hTyr-expressing cells, TFEB (Supplementary Figure 8). The continuous buildup of NM within overexpression was also confirmed to induce lysosomal exocytosis autophagic compartments was associated to a parallel decrease in (Supplementary Figure 8), as part of its known effects at pro- lysosomal-mediated proteolysis in hTyr-expressing cells, as moting cellular clearance60,62. In NM-producing rats, TFEB measured by intracellular protein degradation assay (Fig. 8d). overexpression markedly reduced inclusion body formation, both In addition to impaired lysosomal proteolysis, UPS activity was in the nucleus (i.e. MB) and the cytosol (i.e. PB and LB-like) of also markedly reduced in NM-laden cells, as measured by NM-laden neurons, compared to AAV-hTyr (only)-injected quantification of chymotrypsin-like activity of the 20 S protea- animals (Fig. 9c). Remarkably, intracellular NM density was also some (Fig. 8e). Consistent with a failure in autophagy- and UPS- markedly reduced by TFEB within neurons displaying overtly mediated proteolysis, NM-laden cells exhibited a marked activated (i.e. nuclear) TFEB expression (Fig. 9d). TFEB-induced accumulation of p62 protein by immunoblot (Fig. 8f and reductions in intracellular NM and inclusion body formation Supplementary Figure 7). In addition, NM-laden cells also were associated, in turn, with a marked attenuation of NM-linked exhibited extensive deposition of aSyn oligomers, as directly neurodegenerative changes in hTyr/TFEB-expressing rats, visualized by aSyn proximity ligation assay (PLA) (Fig. 8g), a including attenuated loss of SNpc TH-positive cells, decreased method that has been previously used to detect aSyn oligomers in nigrostriatal denervation, reduced TH downregulation, and post-mortem PD brains55. Impairment of both autophagy and diminished total SNpc DA neuronal death, all of which was UPS is indicative of a general, late-stage proteostasis failure, in associated to a marked improvement of contralateral forepaw NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y a Days post-hTyr induction b c Days post-hTyr induction Days post-hTyr induction d OFF 1 3 6 - 100 kDa OFF 3 6 OFF 3 6 hTyr - - 75 kDa - 150 kDa - 25 kDa - 50 kDa * Lamp1 - - 100 kDa LC3-I - LC3-II - - 20 kDa - 50 kDa - 50 kDa - 50 kDa β-Actin - β-Actin - β-Actin - 50 - 37 kDa Lysosomal proteolysis (%) - 37 kDa - 37 kDa 40 2.4 6 * 3000 * * Intracellular NM intensity 2.0 30 Lamp1 levels LC3-II/I ratio 1.6 4 2000 (OD %OFF) 20 1.2 * 1000 0.8 2 10 * 0.4 0 0 0 0 OFF 3 6 OFF 3 6 OFF 3 6 Days post-hTyr induction Days post-hTyr induction Days post-hTyr induction OFF 1 3 6 Days post-hTyr induction f g BF OFF 3d 6d e Days post-hTyr induction OFF 3 6 p62 - - 75 kDa - 50 kDa 1.2 150,000 β-Actin - - 50 kDa * *# aSyn oligomers (OD) 125,000 - 37 kDa aSyn Oligomers 100,000 0.8 10 UPS activity 75,000 * 8 50,000 * * p62 levels 6 25,000 0.4 * 4 0 Hoechst –25,000 2 OFF 3 6 0 0 Days post-hTyr induction OFF 3 6 OFF 3 6 Days post-hTyr induction Days post-hTyr induction h 15 i j k 120 l BF Hoechst OFF ON 500 *# 1500 OCR (pmol/min x gr prot) Number of cells (% OFF) Metabolic activity (MFI) * ROS production (MFI) 100 400 * 10 1000 * 80 300 * * 60 5 * 200 500 40 100 20 1d 3d 6d 8d 0 0 0 0 OFF 3 6 OFF 3 6 OFF 3 6 OFF 3 6 Days post-hTyr induction Days post-hTyr induction Days post-hTyr induction Days post-hTyr induction Fig. 8 General proteostasis failure in NM-producing cells. a Top, Induction of hTyr expression in differentiated neuroblastoma SH-SY5Y cells before (OFF) and after (ON) treatment with doxycycline. Bottom, quantification of intracellular NM optical density post-hTyr induction. *p < 0.05, compared to OFF and ON 1d (ANOVA on ranks; Dunn’s post-hoc test). Scale bar, 12.5 μm. b Electron micrographs of NM granules in hTyr-expressing cells (6d post-hTyr induction). Arrowhead, double-membrane; asterisk, lipid droplet. Scale bar, 0.5 μm. c Immunoblot levels of Lamp1 (left) and LC3 (right) post-hTyr induction. d Lysosomal proteolysis post-hTyr induction. e Ubiquitin-proteasome (UPS) activity post-hTyr induction. f Immunoblot levels of p62 post-hTyr induction. g Representative images (left) and quantification (right) of aSyn oligomers (red) in hTyr-expressing cells as detected by aSyn-proximity ligation assay. Hoechst, nuclei (blue). BF bright-field. Scale bar, 25 μm. h Mitochondrial oxygen consumption rate (OCR) post-hTyr induction. i ROS production post-hTyr induction. MFI mean fluorescence intensity. j Cellular metabolic activity post-hTyr induction. k Number of surviving cells post-hTyr induction. l Representative bright-field photomicrographs with superimposed nuclear Hoechst fluorescent staining (blue) of hTyr-expressing cells before (OFF) and after (ON, 1–8d) hTyr induction. For clarity purposes, cell contour in weakly or no melanized cells are highlighted in white. Scale bar, 12.5 μm. In all panels, values are mean ± SEM. In a, d, e and i, n = 3 independent experiments. In c, h, j and k, n = 4 independent experiments. In a (bottom), n = 499 (OFF), n = 155 (1d), n = 111 (3d), and n = 48 (6d) cells. In g, n = 153 (OFF), n = 201 (3d), and n = 135 (6d) cells. Immunoblot densitometry was normalized to β-actin expression levels. BF bright-field, KDa kilodaltons. Unstained NM appears as brown. In c–f, h, j, *p < 0.05, compared to OFF (one-way ANOVA; Student–Newman–Keuls post-hoc test). In g, *p < 0.05, compared to OFF; #p < 0.05, compared to ON 3d (ANOVA on ranks; Dunn’s post-hoc test). In i, *p < 0.05, compared to OFF; #p < 0.05, compared to ON 3d (one-way ANOVA; Student–Newman–Keuls post-hoc test). In k, *p < 0.05, compared to OFF (ANOVA on ranks; Dunn’s post-hoc test) hypokinesia in these animals (Fig. 9e–i). Our results indicate that peripheral melanins18, the striking resemblance between human NM-linked impaired proteostasis plays a pivotal role in the NM and NM from AAV-hTyr-injected rats raises the question of initiation of PD-like pathology in AAV-hTyr-injected rats. a potential enzymatic contribution to NM synthesis, either by tyrosinase itself, which is present at low levels in human SNpc, or by alternative tyrosinase-related enzymes. Further supporting a Discussion potential enzymatic involvement in NM synthesis, it has been In the present work, we have generated the first experimental reported that increased levels of DA and oxidized DA in mice and in vivo model of age-dependent production and intracellular rats, either with chronic L-DOPA treatment44,64 or TH over- accumulation of human-like NM within PD-vulnerable SNpc DA expression45, is apparently not sufficient by itself to produce NM neurons, up to levels reached in elderly humans. While the aim of in these animals, as might be expected if NM represents a mere generating such a model was to assess the consequences of pro- process of autoxidized DA. In contrast, in the present study, gressive intracellular NM accumulation, our results may also shed conspicuous NM production was obtained solely by over- light on the actual mechanisms of NM synthesis in humans, expressing hTyr in the rodent brain. In addition, our observation which are currently poorly understood. While it is generally of tyrosinase expression being normally present at low levels in accepted that NM is formed by non-enzymatic DA autoxida- the human brain further supports the possibility of a potential tion41, in contrast to the well-known enzymatic synthesis of enzymatic contribution to NM synthesis. The latter results concur 10 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE a b 100,000 Lamp1 Non-injected * AAV-hTyr protein levels (OD) Intracelluar Lamp1 AAV-TFEB/flag Flag hTyr 50,000 AAV-TFEB-injected 0 Non-inj AAV-TFEB c d Flag Flag TH 8 6 PB 500 Neurons with Marinesco bodies * LB-like Neurons with inclusions 400 Intracelluar NM densiry (average per section) (average per section) 6 (OD, % of Flag– cells) 4 300 4 200 100 Hoechst BF 2 2 0 * * * –100 0 0 –200 hTyr hTyr+TFEB hTyr hTyr+TFEB Nuclear flag– Nuclear flag+ SNpc TH-positive neurons e 20,000 Contralateral f 0.08 Contralateral Ipsilateral Ipsilateral SNpc TH-positive neurons 15,000 0.06 § Striatal TH (OD) § 10,000 0.04 *# *# 5000 0.02 0 0 Vehicle TFEB hTyr hTyr+TFEB Vehicle TFEB hTyr hTyr+TFEB g h Contralateral i 100 20,000 75 Ipsilateral * (% of total NM-laden cells) Use of contralateral paw 80 (% of total forepaw use) SNpc total DA neurons 15,000 TH-negative neurons *# 50 60 * 10,000 40 25 5000 20 0 0 0 hTyr hTyr+TFEB Vehicle TFEB hTyr hTyr+TFEB hTyr hTyr+TFEB with previous reports having detected the presence of brain tyr- in the human brain66,67. The apparently conflicting results osinase in humans at the level of mRNA, protein, and enzymatic regarding the presence or absence of tyrosinase in the human activity20–22, as well as with results from three different micro- brain may be attributed to the low levels at which tyrosinase may array experiments in human brain tissue (Entrez_id: 7299; Allen be actually expressed in the brain and thus the requirement of Brain Institute), non-human primate brain tissue (Entrez_id: adequate methodological approaches for its detection. For 705792; NIH Blueprint NHP Atlas), and developing human instance, brain tyrosinase has not been detected in humans prenatal brain tissue (Entrez_id: 7299, BrainSpan Atlas of the when using immunohistochemistry or western blot (WB) from developing human brain) consistently detecting the presence of total protein homogenates66,67, two techniques unable to detect tyrosinase transcripts at low levels. In rodents, tyrosinase such low levels of protein expression at which tyrosinase may expression has been previously detected in mouse SN by in situ be present in the brain. In contrast, tyrosinase expression hybridization (Allen Brain Atlas data portal) and qPCR65. has been systematically detected in the human brain when However, other studies have failed to detect tyrosinase expression using either RT-PCR/qPCR, in situ hybridization or WB with NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y Fig. 9 Therapeutic enhancement of lysosomal proteostasis in NM-producing rats. a Ipsilateral SNpc sections from rats co-injected with AAV-hTyr/flagged- AAV-TFEB immunostained with Flag (blue, left) or hTyr (blue, right) at 5 m post-AAV. NM, brown. Arrowheads, nuclear Flag/TFEB. Scale bar, 25 μm. b Lamp1 immunolabeling (blue) in ipsilateral SNpc of AAV-TFEB-injected rats (12 m post-AAV). *p < 0.05, compared to contralateral (non-injected) SNpc (Mann–Whitney rank sum test). Scale bar, 12.5 μm. c NM-laden neurons with p62-positive MB (left) and PB/LB-like inclusions (right) in ipsilateral SNpc from AAV-hTyr- and AAV-hTyr/TFEB-injected rats (2 m post-AAV). *p < 0.05, compared to AAV-hTyr-injected animals (left, Mann–Whitney rank sum test; right, two-way ANOVA, Student–Newman–Keuls post-hoc test). d Intracellular NM density in ipsilateral SNpc TH-positive neurons from rats co- injected with AAV-TFEB/AAV-hTyr (12 m post-AAV). White arrowhead, Flag/TFEB-positive nucleus; black arrowhead, Flag/TFEB-negative nucleus. Scale bar, 12.5 μm. *p < 0.05, compared to Flag/TFEB-negative neurons (Mann–Whitney rank sum test). e SNpc TH-positive neurons at 12 m post-AAV/ vehicle injections. f Striatal TH-positive fibers at 12 m post-AAV/vehicle injections. In e, f, *p < 0.05, compared to respective contralateral side; #p < 0.05, compared to ipsilateral vehicle- and AAV-TFEB-injected animals; §p < 0.05, compared to ipsilateral AAV-hTyr-injected animals (two-way ANOVA; Student–Newman–Keuls post-hoc test). g TH downregulation within NM-laden neurons (12 m post-AAV). *p < 0.05, compared to AAV-hTyr-injected animals (two-tailed t-test). h Total SNpc DA neurons at 12 m post-AAV/vehicle injections. *p < 0.05, compared to respective contralateral side; #p < 0.05, compared to ipsilateral vehicle- and AAV-TFEB-injected animals (two-way ANOVA; Student–Newman–Keuls post-hoc test). i Contralateral forepaw use in AAV-hTyr- and AAV-hTyr/TFEB-injected rats (12 m post-AAV). *p < 0.05, compared to AAV-hTyr (only)-injected animals (two-tailed t-test). Dashed line indicates average contralateral forepaw use in control-injected rats. In all panels, values are mean ± SEM. In b, n = 138 neurons from n = 4 non-injected rats and n = 203 neurons from n = 5 AAV-TFEB-injected rats. In c, n = 6 (hTyr), n = 8 (hTyr + TFEB) rats. In d, n = 132 nuclear flag- neurons and n = 173 nuclear flag + neurons from n = 4 AAV-hTyr + TFEB-injected rats. In e and h, n = 7 (vehicle), n = 4 (TFEB), n = 5 (hTyr), n = 4 (hTyr + TFEB) rats. In f, n = 7 (vehicle), n = 6 (TFEB), n = 4 (hTyr), n = 3 (hTyr + TFEB) rats. In g, n = 4 animals per group. In i, n = 12 (control), n = 3 (hTyr), n = 4 (hTyr + TFEB) rats. BF bright-field. Photomicrographs correspond to 5-μm-thick sections tyrosinase-immunoprecipitated protein samples20–22 (see also AAV-hTyr-injected WT rodents; (ii) is dependent on aSyn44,45, Supplementary Figure 1). Further studies are required to deter- in contrast to the dispensability of aSyn that we reported for mine in a definite manner a potential contribution of tyrosinase AAV-hTyr-mediated toxicity. In this context, the continuous or other enzymes in NM synthesis. accumulation of NM over a lifetime into undegraded or partly The functional significance of NM production remains spec- degraded autophagic structures, either in an attempt by the cell to ulative. Until recently, NM was considered merely as a waste degrade this pigment or to isolate it from the rest of the cell to product of cellular metabolism that was devoid of any physiolo- prevent its potential pathological interaction with cytosolic cel- gical function. However, increasing evidence indicates that NM lular components, may ultimately exhaust the vesicular storage can interact with the cellular environment and influence cell capacity of the cell, interfere with lysosomal proteases and other function and survival. For instance, similar to peripheral mela- degradative pathways, impair intracellular vesicular trafficking nins, NM possesses antioxidant and free radical scavenging and alter endocytic/secretory tasks17,23. Consistent with such a properties, which might be important for catecholaminergic scenario, here we found that NM accumulation was associated to neurons, as these cells endure a particularly high oxidative bur- a general impairment of cellular proteostasis accompanied by the den68. In this context, it has been proposed that the synthesis of formation of PD-like intracellular inclusions, neuronal dysfunc- NM may represent an antioxidant mechanism to remove exces- tion, and neurodegeneration. Supporting a pivotal detrimental sive cytosolic DA-derived quinones and semiquinones, thereby role for impaired proteostasis in NM-laden cells, enhancement of blocking their potential toxicity68. Supporting this concept, hTyr- lysosomal-mediated proteolysis by TFEB overexpression in NM- induced oxidized DA species did not significantly increase in producing rats reduced intracellular NM density, attenuated PD- hTyr-injected rats, suggesting that the continuous conversion of like inclusion formation, prevented nigrostriatal neurodegenera- these species into NM may be preventing their potentially toxic tion, and reversed motor impairment. accumulation in these animals. Moreover, NM is able to bind and The results presented here could explain the major pathologic sequester potentially toxic metals, in particular redox-active ferric features of PD, including: (i) the selective vulnerability of specific ions, as well as environmental parkinsonian toxins, such as 1- neuronal groups (i.e. those containing NM); (ii) the correlation of methyl-4-phenylpyridinium (MPP+, the active metabolite of PD with age (as NM accumulates with age); (iii) the reported MPTP) or the pesticide paraquat69. Despite a putative beneficial cellular alterations linked to PD, such as mitochondrial dys- role of NM synthesis, our results suggest that the progressive function, increased ROS production, autophagy/UPS defects, buildup of intracellular NM may ultimately interfere with normal inclusion body formation, and neuroinflammation, all of which cell function and trigger PD-like neuronal dysfunction/degen- have been proposed as potential primary pathogenic events in PD eration when reaching a pathogenic threshold of accumulation, but, in light of our results, may in fact occur concomitantly regardless of whether the initial synthesis of NM is a protective or secondarily to intracellular MN accumulation; (iv) the uniqueness inert process. of PD to humans, since only humans accumulate such high levels Neuronal dysfunction/degeneration in NM-producing AAV- of intracellular NM16. However, while NM is present in the brain hTyr-injected rats could potentially result from DA-mediated of all humans, only some individuals actually develop PD (1 to up neurotoxicity by increased hTyr-mediated production of poten- ~5% of the population over 60, increasing with age)70. Based on tially toxic oxidized DA species. However, as discussed above, the our findings, PD will appear in those subjects that have reached continuous formation of NM in AAV-hTyr-injected rats seems to earlier the pathogenic threshold of intracellular NM accumula- prevent the significant accumulation of oxidized DA species in tion. Supporting this concept, we found that intracellular NM these animals. Further arguing against a major pathogenic role for levels appear indeed above this pathogenic threshold in the brains oxidized DA species in AAV-hTyr-injected rodents, it has been of PD patients and pre-symptomatic PD subjects (i.e. ILBD) while recently reported by two independent groups44,45 that DA- in the brains of elderly, aged-matched healthy individuals intra- mediated toxicity in rodents: (i) is only observed in animals cellular NM levels were maintained below this pathogenic displaying additional PD-related alterations, such as DJ-1 defi- threshold. In PD patients, the intracellular buildup of NM might ciency44 or overexpression of PD-linked A53T mutant aSyn45, be accelerated, thereby reaching earlier the pathogenic threshold but not in regular WT animals44,45, in contrast to our results in of intracellular NM accumulation. This could occur, for instance: 12 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE (i) by an abnormal upregulation/increased activity of hTyr or PicoGreen® as previously described74, and the following vector concentrations other putative NM-producing enzymes, such as tyrosinase-related were obtained: AAV-hTyr: 1.7 × 1013 gc/mL (batch a), 2.43 × 1012 gc/mL (batch b); AAV-EV: 2.48 × 1013 gc/mL. protein 2 (i.e. dopachrome tautomerase), the expression of which AAV serotype 2/9 containing the murine Tfeb cDNA fused to 3 Flag epitopes we found significantly increased in induced pluripotent stem cell under control of the cytomegalovirus (CMV) promoter (AAV-TFEB; (iPSC)-derived DA neurons from PD patients;71 or (ii) by concentration: 3 × 1012 gc/mL) was provided by TIGEM AAV Vector Core Facility increased cytosolic DA levels, which may convert to NM, as a (Italy)59. consequence of defects in VMAT2-dependent DA encapsulation within synaptic vesicles, as it has been shown in early PD cases26. Surgical procedures. All surgical procedures on rodents were performed with the Based on our results, strategies to maintain or decrease intra- animals placed in a stereotaxic frame under general anesthesia using isofluorane (5% for the induction phase and 2% for the maintenance phase) (Baxter). Vector cellular NM levels below their pathogenic threshold may provide solutions were injected using a 10 μL Hamilton syringe fitted with a glass capillary unprecedented therapeutic opportunities to prevent, halt, or delay (Hamilton model Cat#701). Animals received 2 μL (rats) or 1 μl (mice) of either neuronal dysfunction and degeneration linked to PD. Given the AAV-hTyr, AAV-EV, vehicle or, in the experiments involving TFEB, a 1:1 mixture putative protective role of NM synthesis (see above), therapeutic of AAV-hTyr/AAV-TFEB, AAV-hTyr/vehicle, or AAV-TFEB/vehicle. In both rats and mice, infusion was performed at a rate of 0.4 μL/min and the needle was strategies aimed at maintaining or decreasing intracellular NM left in place for an additional 4 min period before it was slowly retracted. Injection levels below their pathogenic threshold should be preferentially was carried out unilaterally on the right side of the brain at the following coor- targeted at preventing NM accumulation once NM has already dinates (flat skull position), right above the substantia nigra pars compacta (SNpc): been produced, for instance by promoting its elimination as antero-posterior: −5.2 mm (rats), −2.9 mm (mice); medio-lateral: −2 mm (rats), reported here with TFEB. In contrast, the inhibition of NM −1.3 mm (mice); dorso-ventral: −7.6 mm (rats), −4.2 mm (mice) below dural surface, calculated relative to bregma according to the stereotaxic atlas of Paxinos synthesis itself may prove instead detrimental. For instance, a rare and Watson75. loss-of-function mutation in hTyr linked to albinism has been recently associated to an increased risk for PD72, which has been Transduction efficiency. AAV-hTyr transduction efficiency was determined at attributed to the possible inability of this hTyr variant to syn- 2–4 weeks post-AAV injection (n = 6–8 rats) either by: (i) assessing hTyr thesize NM and the subsequent accumulation of potentially toxic expression within SNpc tyrosine hydroxylase (TH)-positive cells by double DA-derived species that cannot be detoxified by its conversion immunofluorescence with antibodies against TH and hTyr in six coronal midbrain sections through the entire SNpc, using the Cell Counter plugin on ImageJ software into NM. In this context, it would be very important for any (Rasband, W.S., ImageJ, U. S. National Institutes of Health, Bethesda, Maryland, therapeutic strategy aimed at modulating NM levels to tightly USA, 1997-2016); (ii) by counting the number of NM-positive neurons versus the regulate NM levels above its protective threshold and below its total number of SNpc TH-positive neurons, as described below. pathogenic threshold revealed here. Overall, with the introduction into experimental in vivo Brain processing for histological analyses. Animals were deeply anesthetized research of a factor so intimately linked to PD such as NM, which with sodium pentobarbital (50 mg/kg, i.p.) and then perfused through the left has been so-far neglected in PD animal modeling, the results ventricle with saline [0.9% (wt/vol)] at room temperature (RT), followed by ice-cold formaldehyde solution 4% phosphate buffered for histology (Panreac). The presented here may open new research avenues and lead to a brains were removed and post-fixed for 24 h in the same fixative and subsequently paradigm shift in the field of PD and, in a broader sense, brain processed for paraffin embedding following standard procedures or cryoprotected aging. for 24–48 h in 30% sucrose at 4 °C and frozen. Sectioning was performed with a sliding microtome (Leica, Germany) at 5-µm-thickness for paraffin samples or in a cryostat at 20- or 30-µm-thickness for frozen samples (Leica, Germany). Methods Animals. Adult male Sprague–Dawley rats (Charles River), 225–250 g at the time of surgery, were housed two to three per cage with ad libitum access to food and Immunohistochemistry. Deparaffinized rodent brain sections were quenched for water during a 12 h (h) light/dark cycle. All the experimental and surgical proce- 10 min in 3% H2O2-10% (vol/vol) methanol. Antigen retrieval in paraffin sections dures were conducted in strict accordance with the European (Directive 2010/63/ was performed with a 10-mM citric acid solution at pH 6.0 in a microwave for UE) and Spanish laws and regulations (Real Decreto 53/2013; Generalitat de 20 min. Sections were rinsed 3 times in 0.1 M Tris buffered saline (TBS) between Catalunya Decret 214/97) on the protection of animals used for experimental and each incubation period. Blocking for 1 h with 5% (vol/vol) normal goat serum other scientific purposes, and approved by the Vall d’Hebron Research Institute (NGS, Vector Laboratories) was followed by incubation with the appropriate pri- (VHIR) Ethical Experimentation Committee. Rats were randomly distributed mary antibody at 4 °C for 48 h in 2% (vol/vol) serum and with the corresponding into the different experimental groups and control and experimental groups were biotinylated antibody (Vector Laboratories). Sections were visualized by incubation processed at once to minimize bias. When appropriate, additional sporadic with avidin–biotin–peroxidase complex (Ultrasensitive and Immunopure ABC experiments were performed in aSyn-deficient73 and wild-type C57BL/6 mice, Peroxidase staining kits for the striatum and for the SNpc, respectively; Thermo as indicated. Fisher Scientific), using the VectorSG Peroxidase Substrate Kit (Vector Labora- tories) as a chromogen, and then mounted and coverslipped with DPX mounting medium (Sigma-Aldrich). Bright-field section images were examined using: (i) Viral vector production. Recombinant AAV vector serotype 2/1 expressing the Zeiss AX10 LabA1 microscope coupled to an AxioCam ERc5s camera; (ii) Zeiss human tyrosinase cDNA driven by the CMV promoter (AAV-hTyr) and the Imager.D1 microscope coupled to an AxioCam MRc camera; (iii) Zeiss Stemi 2000- corresponding control empty vector (AAV-EV) were produced at the Viral Vector C magnifying glass coupled to an AxioCam ERc5s camera. Images were processed Production Unit of the Autonomous University of Barcelona (UPV-UAB, Spain). with ZEN 2011 software (Zeiss, Germany). For immunofluorescence, a similar Briefly, rAAV vectors were produced by triple transfection of 2 × 108 HEK293 cells protocol was used without the quenching step. Preincubation was performed with with 250 µg of pAAV, 250 µg of pRepCap, and 500 µg of pXX6 plasmid mixed with 5% (vol/vol) NGS and 0.1% (vol/vol) Triton X-100 (Sigma-Aldrich) in phosphate polyethylenimine (PEI; branched, MW 25,000; Sigma-Aldrich). The UPV-UAB buffered saline (PBS) solution. Corresponding primary antibodies were incubated generated a pAAV plasmid containing the ITRs of the AAV2 genome, a multi- together overnight at 4 °C in 2% (vol/vol) serum and adequate Alexa 488, 594, cloning site to facilitate cloning of expression cassettes, and ampicillin resistance and 647-conjugated secondary antibodies (1:1000, Thermo Fisher Scientific) were gene for selection. 48 h after transfection, cells were harvested by centrifugation incubated simultaneously for 1 h at RT in 2% (vol/vol) serum. Nuclei were stained (200 × g, 10 min); resuspended in 30 ml of 20 mM NaCl, 2 mM MgCl2, and 50 mM with Hoechst 33342 (1:2000, Thermo Fisher Scientific) in 1× PBS for 10 min. Tris-HCl (pH 8.5), and lysed by three freeze–thaw cycles. Cell lysate was clarified Sections were coverslipped using the DakoCytomation Fluorescent Mounting by centrifugation (2000 × g, 10 min) and the rAAV particles were purified from the Medium (Dako). Immunofluorescence section images were examined using: (i) an supernatant by iodixanol gradient as previously described (Zolotukhin et al., 1999 Olympus FSX100 microscope with a DP72 incorporated camera and processed Gene Therapy). The clarified lysate was treated with 50 U/ml of Benzonase with the FSX-BSW software (Olympus, Germany), and in an Olympus BX61 (Novagen; 1 h, 37 °C) and centrifuged (3000 × g, 20 min). The vector-containing microscope with a DP72 camera and processed with the Cell Sens Entry supernatant was collected and adjusted to 200 mM NaCl using a 5-M stock v1.8 software (Olympus, Germany). solution. To precipitate the virus from the clarified cell lysate, polyethylene glycol Primary antibodies and dilution factors were as follows: mouse anti-DAT (PEG8000; Sigma-Aldrich) was added to a final concentration of 8% and the (1:500, Novus Biologicals, Cat#NBP2-22164); guinea pig anti-VMAT2 (1:5000, mixture was incubated (3 h, 4 °C) and centrifuged (8000 × g, 15 min). The AAV- Progen, Cat#16085); rabbit anti-TH (1:40000 for SNpc and 1:3500 for striatum containing pellets were resuspended in 20 mM NaCl, 2 mM MgCl2, and 50 mM immunohistochemistry in paraffin sections; 1:2000 for SNpc and 1:5000 for Tris-HCl (pH 8.5) and incubated for 48 h at 4 °C. The rAAV titration method used striatum immunohistochemistry in free-floating sections; 1:1000 for was based on the quantitation of encapsidated DNA with the fluorescent dye immunofluorescence, Calbiochem, Cat#657012); mouse anti-TH (1:1000, Merck NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 13 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y Millipore, Cat#MAB5280); mouse anti-tyrosinase (1:500, Thermo Fisher Scientific, The counting frames were placed randomly by the software at the intersections of Cat#MS-800-P1); mouse anti-FLAG (1:1000, Sigma-Aldrich, Cat#F3165); rabbit the grid within the outlined structure of interest. The cells in one brain side, contra- anti-GAD65 + GAD67 (1:750, Abcam, Cat#ab49832); mouse anti-GFAP (1:1000, or ipsilateral to the injection site, were counted following the unbiased sampling Sigma-Aldrich, Cat#G3893); rabbit anti-Iba1 (1:1000, Wako, Cat#019-19741); rule using a ×100 lens and included in the measurement when they came into focus mouse anti-CD68 (1:100, Serotec, Cat#MCA341R); DyLight 594 Lectin (1:500, within the dissector. A coefficient of error of <0.10 was accepted. Data for the total Vector Laboratories, Cat#DL-1177); guinea pig anti-p62 (1:1000, Progen, numbers of TH-positive neurons and NM-containing neurons in the SNpc are Cat#GP62-C); rabbit anti-Ubiquitin (1:500, Dako, Cat#Z0458); mouse anti-aSyn expressed as the absolute numbers in the non-injected contralateral side and in (1:1000, BD Biosciences, Cat#610786); mouse anti phosphorylated(Ser129)-aSyn the AAV-injected ipsilateral side. The total number of SNpc DA neurons was (1:1000, Wako, Cat#015-25191); rabbit anti-Lamp1 (1:250, GeneTex, calculated by considering all TH+NM+, TH−NM+, and TH+NM− neurons. The Cat#GTX19294); rat anti-Lamp1 (1:4000, Santa Cruz, Cat#sc-19992). percentage of TH down-regulation was calculated by considering the total number of TH+NM+ and the total number of TH−NM+ with respect to the total number of neurons containing NM in the different experimental groups. AAV-hTyr- Masson-Fontana and hematoxylin-eosin stainings. Staining of NM granules in injected rats were analyzed at different time-points after injection: 0.5 m (n = 8), 5-µm-thick paraffin-embedded rat brain sections was performed using the Masson- 1 m (n = 7), 2 m (n = 5), 4 m (n = 8), 12 m (n = 7), and 24 m (n = 6). AAV-hTyr Fontana Staining Kit (DiaPath); this procedure is based on the ability of NM to (n = 5), AAV-TFEB (n = 4), AAV-hTyr + TFEB (n = 4), and vehicle-injected rats chelate metals by reducing silver nitrate to a visible metallic state. Briefly, paraffin (n = 7) were analyzed at 12 m post-injection. AAV-EV-injected rats were analyzed tissue sections were dewaxed and rehydrated by heating at 60 °C for 10 min, at 4 m (n = 7), 12 m (n = 4), and 24 m (n = 3) post-injection. Vehicle-injected followed by xylene (5 min, 3 times) and ethanol serial washes (100-95-70%-H2O, rats were analyzed at 4 and 12 m post-surgery (n = 8 animals/group). AAV-hTyr- 5 min each). Staining was performed by incubating the sections with ammoniac injected mice were analyzed at 6 m post-AAV injections: WT (n = 8), aSyn KO solution for 40 min at 56 °C, followed by sodium thiosulphate for 2 min and a final (n = 6). All quantifications were performed by an investigator blinded to the counterstain with Kernechtrot for 7 min. Between each step, samples were rinsed in experimental groups. distilled water. Standard hematoxylin-eosin (H&E) staining was performed in 5- µm-thick paraffin-embedded SNpc section for each animal. In these sections, SNpc DA neurons were identified by the visualization of unstained NM pigment. Quantification of neuropathological parameters. The absolute number of extracellular NM aggregates was estimated within the same sections in which SNpc Magnetic resonance imaging. 1H-Magnetic resonance imaging (MRI) studies in TH-positive stereological cell counts were performed (i.e. serial 5-µm-thick rats were performed in a 7T Bruker BioSpec 70/30 USR (Bruker BioSpin GmbH, paraffin-embedded sections covering the entire SNpc, taking every 17th section Karlsruhe, Germany) equipped with a mini-imaging gradient set (400 mT/m), a in rats and every 6th section in mice, for a total of 6–8 sections analyzed/animal). linearly polarized transmit volume coil (72 mm inner diameter) and a dedicated The number of neuronophagic events were assessed in serial H&E-stained 5-µm- brain surface coil as receiver. All MRI data were acquired and processed on a Linux thick paraffin-embedded sections covering the entire SNpc. The number of p62- computer using Paravision 5.1 software (Bruker BioSpin GmbH, Karlsruhe, Ger- immunopositive Marinesco bodies, PB, and Lewy body-like inclusions was counted many). Fixed rat brains were placed in a 15-ml Falcon tube and embedded in 2% from SNpc sections fluorescently immunostained with guinea pig anti-p62 (1:1000, agarose to diminish susceptibility artefacts. Low-resolution T2-weighted fast spin- Progen), rabbit anti-ubiquitin (1:500, Dako), and mouse anti-aSyn (1:1000, BD echo images were initially obtained in axial, sagittal, and coronal planes to be used Biosciences). The total number of p62-positive inclusions falling into each category as reference scout images. Imaging parameters for these images were: effective echo was counted from images covering the whole SNpc region in each section. time (TEeff) = 36 ms; repetition time (TR) = 3 s; echo train length (ETL) = 8; field Quantifications were performed in AAV-hTyr-injected at different time-points of view (FOV) = 6 × 6 cm2; matrix size (MTX) = 128 × 128; slice thickness (ST) = post-AAV-hTyr injection: 0.5 m (n = 8), 1 m (n = 5), 2 m (n = 6), 4 m (n = 5), 2 mm; gap between slices (gap) = 0.5 mm; number of slices (NS) = 25 -axial, 12 m (n = 6), and 24 m (n = 5). AAV-hTyr-injected mice were analyzed at 2 m 10-sagittal, 11 -coronal; number of averages (NA) = 1. High-resolution T1- post-AAV injections: WT (n = 3), aSyn KO (n = 4). AAV-hTyr (n = 5), AAV- weighted spin-echo images were acquired afterwards in the axial and coronal TFEB (n = 4), AAV-hTyr + TFEB (n = 4), and vehicle-injected rats (n = 7) were planes containing the region of interest with the following parameters: TE = 9 ms; analyzed at 12 m post-injection. All quantifications were performed by an inves- TR = 500 ms; NA = 128; NS = 20 (axial) and 7 (coronal); FOV = 1.92 × 1.92 cm2; tigator blinded to the experimental groups. MTX = 128 × 128 (axial) and 192 × 128 (coronal); ST = 0.25 mm (axial) and 0.35 (coronal); thus resulting in a spatial resolution of 150 × 150 × 250 µm3 (axial) and 100 × 150 × 350 µm3 (coronal). The acquisition time was 2 h and 16 min for Optical densitometry analyses. The density of TH-positive fibers in the striatum each high-resolution image. was measured by densitometry in serial coronal sections covering the whole region (10 sections/animal). TH-immunostained 5-µm-thick paraffin-embedded sections Intracellular NM quantifications. An average of 165 neurons (94 in mice) were were scanned with an Epson Perfection v750 Pro scanner and the resulting images analyzed at each time-point post-AAV-hTyr injection, i.e. 0.5, 1, 2, 4, 12, and were quantified using Sigma Scan Pro 5 software (Systat Software Inc, USA). 24 months (m), corresponding to 5–7 different animals per time-point. For the Striatal densitometry values were corrected for non-specific background staining by experiments related to TFEB, an average of 108 neurons were analyzed at 12 m subtracting densitometric values obtained from the cortex. Data are expressed as post-AAV injections, corresponding to 3–4 different animals per experimental the percentage of the densitometric value of the equivalent anatomical area from group. In all cases, a representative 5-µm-thick paraffin-embedded H&E-stained the non-injected contralateral side of the same animal. For the AAV-hTyr groups, rat midbrain section was selected for each animal. In each section, all NM-positive animals were analyzed at different time-points after AAV-hTyr injection: 0.5 m neurons were examined at an objective magnification of ×63. Images covering all (n = 8), 1 m (n = 7), 2 m (n = 5), 4 m (n = 8), 12 m (n = 7), and 24 m (n = 6). For the SNpc region were taken with a Zeiss Imager. D1 microscope was coupled to an the AAV-hTyr/TFEB-related experiments, animals were analyzed at 12 m post- Axiocam MRc camera. Quantification of the intracellular density of NM pigment injection: AAV-hTyr (n = 4), AAV-TFEB (n = 6), AAV-hTyr + TFEB (n = 3), and was achieved by means of optical densitometry using ImageJ software (NIH, USA). vehicle-injected rats (n = 7). For control experiments, animals were analyzed at The pixel brightness values for all individual NM-positive cells (excluding the different time-points after injection: AAV-EV-injected rats at 4 m (n = 7), 12 m nucleus) in all acquired images were measured and corrected for non-specific (n = 4), and 24 m (n = 3); vehicle-injected rats at 4 and 12 m (n = 8 animals background staining by subtracting values obtained from the neuropil in the same per group). AAV-hTyr-injected mice were analyzed at 6 m post-AAV injections: images. In animals co-injected with AAV-hTyr and AAV-TFEB, additional WT (n = 9), aSyn KO (n = 5). All quantifications were performed by an investi- quantifications of intracellular NM density were performed as above in bright-field gator blinded to the experimental groups. photomicrographs at ×63 magnification, taking into account whether NM-laden The density of GAD65 + GAD67-positive fibers in the superior colliculus was cells exhibited flagged TFEB-positive or TFEB-negative nuclei. All quantifications measured by densitometry in selected coronal sections around Bregma −5.64 mm. were performed by an investigator blinded to the experimental groups. For NM GAD65 + GAD67-immunostained 5-µm-thick paraffin-embedded sections were quantifications in post-mortem human samples and cultured cells, see details scanned as previously described and densitometry values were corrected for non- below. Of note, because intracellular NM density in both humans and rodents was specific background staining by subtracting densitometric values obtained from the measured as the intensity of the NM signal normalized to the respective neuronal adjacent ventral parenquima. Data are expressed as the percentage of the surface (excluding the nucleus), these values can be compared between species. densitometric value of the equivalent anatomical area from the non-injected contralateral side of the same animal. Animals were analyzed at different time- points: (i) naive rats: 4 m (n = 4) and 12 m (n = 5); (ii) vehicle-injected rats: 4 m Stereological cell counting. Assessment of the total number of SNpc TH-positive (n = 6) and 12 m (n = 5); (iii) AAV-EV-injected rats: 12–24 m (n = 4); (iv) AAV- neurons, the number of SNpc NM-laden neurons and the total number of DA hTyr-injected rats: 4 m (n = 3) and 12–24 m (n = 7). neurons in the SNpc was performed according to the fractionator principle, using Lamp1 protein expression levels in SNpc neurons from AAV-TFEB-injected the MBF Bioscience StereoInvestigator 11 (64 bits) Software (Micro Brightfield). rats (n = 5 animals) were measured by intracellular optical densitometry using Serial 5-µm-thick paraffin-embedded sections covering the entire SNpc were ImageJ software. Five-µm-thick paraffin-embedded SNpc sections were included in the counting procedure (every 17th section in rats and every sixth immunostained for Lamp1 and 200 cells were randomly selected for analysis in the section in mice, for a total of 6–8 sections analyzed/animal). The following sam- AAV-TFEB-injected ipsilateral side and the non-injected contralateral side of the pling parameters were used: (i) a fixed counting frame with a width and length of same animal. All quantifications were performed by an investigator blinded to 50 μm; (ii) a sampling grid size of 100 × 75 μm in rats and of 125 × 100 μm in mice. the experimental groups. 14 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE Cylinder behavioral test. Rats were tested for left and right forepaw use with the UPLC-MS/MS analysis. AAV-hTyr-injected rats (n = 6) were killed at 1, 2, or 4 m cylinder test 1 week before surgery (to establish the basal conditions for each after AAV-hTyr injections and the brains quickly removed and placed over a cold animal) and at different times post-AAV-hTyr injection thereafter (0.5, 1, 2, 3, 6, plate. Contralateral non-injected and ipsilateral AAV-hTyr-injected ventral mid- 12, and 24 m). Rats that presented an asymmetric usage of the right-left forepaws brain from each animal were dissected out, frozen on dry ice, and stored at −80 °C at basal examination were excluded from the analysis. At the beginning of the until analyzed. On the day of analysis, samples were homogenized with 300 µl/ experiment, the number of animals for each condition was: n = 29 (AAV-hTyr- hemisphere of 250 mM formic acid (FA) and protein concentrations determined by injected rats) and n = 20 (AAV-EV-injected rats). Because a randomly selected bicinchoninic acid (BCA) Assay Kit (Pierce). Dopamine-1,1,2,2-d4 hydrochloride group of animals was euthanized for histological analyses at each time-point fol- (DA-d4) was added as internal standard (IS) to a final concentration of 500 nM and lowing the cylinder test, at the end of the behavioral analysis (24 m): n = 6 (AAV- samples were then centrifuged at 20,000 × g for 10 min at 4 °C, and the supernatant hTyr-injected rats) and n = 4 (AAV-EEV-injected rats) animals remained. For the was filtered using an Ostro™ protein precipitation and phospholipid removal plate experiments with TFEB, vehicle (n = 7), AAV-TFEB (n = 5), AAV-hTyr (n = 3), (Waters, USA) and injected in the UPLC-MS/MS system twice to analyze MIX1 and AAV-hTyr + TFEB (n = 4) injected rats were analyzed 1 week before surgery (dopamine, 3-methoxytyramine and oxidized dopamine) and MIX2 (3,4-dihy- (to establish the basal conditions for each animal) and then every 4 weeks until rats droxyphenylacetic acid). A Waters Acquity™ UPLC system was coupled with a were killed 12 m post-injection. For the performance of the cylinder test, rats were Xevo TQ-S triple quadrupole mass spectrometer with electrospray ionization first allowed to habituate to the experimental room for at least 1 h before each test. interface (Waters). Instrument control, data acquisition, and analysis were per- Then, rats were put in a glass cylinder and the total number of left and right formed using MassLynx V4.1 (Waters). The chromatographic separation was forepaw touches performed within 5 min was counted. Data are presented as the performed on a Waters Acquity™ HSS T3 (1.8 μm; 2.1 × 100 mm) column coupled percentage of the contralateral paw usage with respect to either AAV-EV-injected to a Acquity™ HSS T3 VanGuard (100 Å, 1.8 µm, 2.1 mm × 5 mm) pre-column and or control (vehicle and AAV-TFEB-injected) rats, as appropriate. Behavioral a Acquity™ UPLC in-line filter (Waters). Column temperature was set at 40 °C equipment was cleaned with 70% ethanol after each test session to avoid olfactory and samples were maintained at 6 °C in the thermostatic autosampler. The mobile cues. All behavioral tests were performed during the light cycle by an investigator phase consisted of solvent A (methanol 100%) and solvent B (25 mM FA in water) blinded to the experimental groups. at a flow of 0.4 ml/min with the following gradient profile: 0.5% B maintained for 0.5 min, 8% B at 2.6 min, 55% B at 2.9 min, 60% B at 3.3 min, 80% B at 4.3 min, 90% B at 4.4 min and maintained for 0.5 min, and 0.5% B at 5 min followed by In vivo microdialysis. To assess local effects of D-amphetamine sulfate and 1 min of equilibration. The mass spectrometer detector operated under the veratridine on striatal DA release in microdialysis experiments, drugs were following parameters: source temperature 150 °C, desolvation temperature 450 °C, dissolved in artificial cerebrospinal fluid (aCSF: 125 mM NaCl, 2.5 mM KCl, cone gas flow 50 L/hr, desolvation gas flow 1100 L/h and collision gas flow 0.15 ml/ 1.26 mM CaCl2, and 1.18 mM MgCl2) and administered by reverse dialysis at min. Argon was used as the collision gas. The capillary voltage was set at 0.5 kV for the stated concentrations (uncorrected for membrane recovery). Stock solutions MIX1 and at 2 kV for MIX2 detection. The electrospray ionization source was of veratridine were made in dimethyl sulfoxide and were diluted to appropriate operated in both positive and negative modes, depending on the analyte. Multiple concentrations in aCSF to reach 1% dimethyl sulfoxide. Concentrated solutions Reaction Monitoring (MRM) acquisition settings for the targeted metabolites are (1 mM; pH adjusted to 6.5–7 with NaHCO3 when necessary) were stored at summarized in Supplementary Table 3. Oxidized dopamine was detected using the −80 °C and working solutions were prepared daily by dilution in aCSF. following parameters: MRM transition (m/z) 149.61 > 121.91; cone voltage 25 V; Microdialysis procedures in vehicle- (n = 6) and AAV-hTyr-injected rats (n = 6 collision energy 25 eV; capillary voltage kV 0.5 kV. Two independent weighted for the Amphetamine experiment and n = 8 for the Veratridine experiment) (1/x) linear regression curves excluding the origin were constructed for oxidized were conducted essentially as described previously76. Rats were anaesthetized dopamine with 5 calibration points spiked to ventral midbrain samples of control with sodium pentobarbital (60 mg/kg, i.p.) and implanted with 3-mm concentric rats. Linearity was considered acceptable if R2 ≥ 0.99 and more than 75% of the dialysis probes (Cuprophan membrane, 6000 Da molecular weight cut-off) in residuals had a variability smaller than 20%. With respect to sensitivity, the limit the ipsilateral striatum at the following coordinates (in mm): AP, + 0.5; L, −3.0 of detection (LOD) was defined as the concentration level with the signal-to-noise and DV, 6.6 (ref. 75). Microdialysis experiments were performed in freely moving ratio at 3, and the limit of quantification (LOQ) was defined as the concentration rats 24–48 h after surgery, except in experiments involving the electrical level with the signal-to-noise ratio at 10 (see parameters in Supplementary Table 4). stimulation of the medial forebrain bundle (MFB) (see below). Probes were They were calculated as 3 or 10 times the ratio between the standard deviation perfused with aCSF at 1.5 μL/min. Following an initial 100-min stabilization of the Y-intercept value and the slope of each calibration curve, respectively. period, 5 or 7 baseline samples were collected (20 min each) before local drug Samples with a concentration between LOD and LOQ or bigger than LOQ were application by reverse dialysis and then successive dialysate samples were collected. considered acceptable; samples with a concentration lower than LOD were con- In the experiments examining the effects on striatal DA release of electrical sidered as 0 nM. Oxidized dopamine standard (0.5 mM) was freshly prepared as stimulation of the MFB, rats (n = 4 animals per group) were anaesthetized with previously described44,79. In brief, 100 µl of 1 mM dopamine dissolved in water was chloral hydrate and 10 min fractions were collected (flow rate 3.0 μL/min). Bipolar mixed with 100 µl of 2 mM KIO4 dissolved in 100 µM aqueous ammonium acetate stimulating electrodes consisted of two stainless steel enamel-coated wires buffer (pH 5.8) at RT with vigorous shaking for 1 min. Successive dilutions were (California Fine Wire, Grover Beach, CA, USA) with a diameter of 150 μm, a tip made in 25 mM FA. separation of ~200 μm and in vitro impedances of 10–30 KΏ. A stimulating electrode was stereotaxically implanted in the MFB (AP, −4.8; L, −1.0; and DV, −8.2 mm) and secured to the skull with glue and dental cement. Constant Human post-mortem brain tissue. Paraffin-embedded midbrain sections (5 µm) current electrical stimuli were generated with a Grass stimulation unit S-48 from idiopathic Parkinson’s disease patients (PD; n = 10), incidental Lewy body connected to a Grass SIU 5 stimulus isolation unit. Two stimulating conditions disease cases (ILBD; n = 3) and age-matched control individuals (n = 6) were were used: S1 (2.0 Hz, 0.1 mA, and 0.2 ms) and S2 (10 Hz, 0.1 mA, and 1 ms)77. provided by the Neurological Tissue BioBank at IDIBAPS-Hospital Clinic (Bar- The concentration of DA in dialysate samples was determined by HPLC with celona). Detailed information about these subjects is provided in Supplementary electrochemical detection (Hewlett Packard 1049, Palo Alto, CA, USA). Striatal Table 1. Informed written consent was obtained from all human subjects. All dialysates were collected into microvials containing 5 μL of 10 mM perchloric acid procedures were conducted in accordance with guidelines established by the BPC and were rapidly injected into an HPLC. DA was amperometrically detected at (CPMP/ICH/135/95) and the Spanish regulation (223/2004) and approved by the 5–7.5 min with a LOD of 3 fmol/sample using an oxidation potential of +0.75 V. Vall d’Hebron Research Institute (VHIR) Ethical Clinical Investigation Committee Microdialysis results are expressed as femtomoles per fraction (uncorrected for [PR(AG)370/2014]. For intracellular NM quantification in human brain samples, recovery) and are shown in figures as percentages of basal values (individual means three different clinical diagnosis groups were evaluated: control subjects (1436 cells of 5–7 pre-drug fractions). analyzed in a total of 6 cases), ILBD cases (640 cells analyzed in a total of 3 cases) and PD patients (644 cells analyzed in a total of 10 cases). Standard H&E staining was performed on 5-µm-thick paraffin-embedded SNpc sections for each subject. Tissular levels of DA and metabolites. DA, DOPAC (3,4-dihydroxyphenylacetic Identification of DA neurons was ascertained by the visualization of unstained NM acid), and HVA (homovanillic acid) contents were determined by HPLC with pigment. For each section, all NM-positive neurons were examined at an objective electrochemical detection (+0.7 V) as previously described78. AAV-hTyr-injected magnification of ×20. Quantifications of intracellular NM optical density were rats (n = 8) were killed at 4 m post-AAV injection and their brains quickly performed as above by an investigator blinded to the experimental groups. removed and placed over a cold plate. Contralateral non-injected and ipsilateral AAV-hTyr-injected striatum and substantia nigra from each animal were carefully dissected out using a Rat Brain Matrix (TedPella, Spain), frozen on dry ice, and Endogenous brain tyrosinase gene expression. Total RNA from human post- kept at −80 °C until assayed. Tissues were homogenized in 300 or 500 μl (for mortem substantia nigra brain tissue (n = 8 control individuals; mean age at death substantia nigra or striatum, respectively) of buffer containing 0.4 M perchloric 77 ± 3.7 years; detailed information for each case can be found in Supplementary acid, 0.1% sodium metabisulphite, 0.01% EDTA, and 0.1% cysteine and centrifuged Table 1) was extracted using miRNeasy Micro Kit (Qiagen). Total RNA from at 12,000 × g for 30 min. Aliquots of supernatants were then filtered through mouse (n = 5) and rat (n = 8) dissected ventral midbrain tissue was extracted using 0.45 μm filters (Millex, Spain) and analyzed by HPLC as described. The mobile the mirVana PARIS RNA and Native Protein Purification Kit (Thermo Fisher phase consisted of 0.1 M KH2PO4, 1 mM octyl sodium sulfate, 0.1 mM EDTA Scientific). RNA concentration was determined using a NanoDrop ND-1000 (pH 2.65), and 18% methanol. DA and their metabolites were separated on a Spectrophotometer and RNA integrity was assessed by running the samples on Mediterranea Sea column (C18, 3 μm, 10 cm × 6.4 mm) (Teknokroma, ref an Agilent RNA 6000 Nano chip on an Agilent 2100 BioAnalyzer (Agilent TR010042, Spain). Technologies). For human samples, 10 ng of total RNA were retrotranscribed NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 15 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y and amplified using the Ovation Pico WTA System V2 (Nugen Technologies, Cat#DUO92009) or MINUS (Sigma-Aldrich, Cat#DUO92010) oligonucleotides. Inc.). For mouse and rat, 0.5 µg of total RNA were retrotranscribed using oligo(dT) PLA assay was done using the Duolink in Situ Detection Reagents Red assay 12–18 primers (Thermo Fisher Scientific) and SuperScript™ III Reverse Tran- protocol (Sigma-Aldrich, Cat#DUO92008) following manufacturer’s instructions. scriptase kit (Thermo Fisher Scientific). Quantitative real-time PCR (qPCR) Briefly, fixed cells were incubated with 3 drops of tempered blocking solution for was performed with 45–50 ng of cDNA per well in technical duplicates mixed 30 min at RT, shaking at low speed. Slides were then washed with PBS once and with Taqman Gene Expression Master Mix (Applied Biosystems) and Taqman each slide incubated overnight at 4 °C with 20 µl of a mixture of both PLUS gene expression assays (human tyrosinase (Tyr) Hs00165976_m1; rat Tyr and MINUS antibody-conjugated oligonucleotide PLA probes in a humidity Rn01511409_m1; mouse Tyr Mm00495818_m1; human tyrosine hydroxylase (TH) chamber. Ligation was done for 30 min at 37 °C, followed by amplification for Hs00165941_m1; rat TH Rn00562500_m1; mouse TH Mm00447557_m1; Applied 90 min at 37 °C, both in a humidity chamber as well. Nuclei counterstaining was Biosystems) using standard procedures in a 7900HT Fast Real Time Instrument done with Hoechst 33342 (1:10,000, Thermo Fisher) for 2 min. Slides were then (Applied Biosystems). Thresholds cycles (Cts) for each target gene were normalized to washed three times with Washing Buffer B for 10 min each, mounted using an endogenous reference gene (Gapdh Mm99999915_g1 / Rn01775763_g1, Rpl19 fluorescent mounting medium (Dako) and visualized using an Olympus FSX100 Mm02601633_g1 / Rn00821265_g1 or Rplp0 Mm00725448_s1 / Rn03302271_gH for microscope with a DP72 incorporated camera and FSXBSW visualization software rodent samples, and HPRT1 Hs02800695_m1, GUSB Hs00939627_m1, or RPLP0 (Olympus, Germany). Quantification of aSyn-PLA signal was performed by mea- Hs04189669_g1 for human samples). Water was included in the reaction as a non- suring intracellular pixel intensity values for individual cells (OFF, n = 153; 3d, template (negative) control. The relative expression was calculated with the n = 201; 6d, n = 135) in all acquired images using ImageJ. ΔCt-method. Serial dilutions of a DNA fragment corresponding to the full-length hTyr transcript, obtained by enzymatic digestion from a pCDNA4-hTyr plasmid (kindly provided by T. Hasegawa), were used as template for a qPCR reaction of Mitochondrial respiration. Oxygen consumption rate (OCR) was measured in 3 replicates for each dilution with the assay Hs00165976_m1 to specifically TR5TY6 cells using an XF24 Extracellular Flux Analyzer (Seahorse Biosciences, detect human Tyr expression and correlate the number of Tyr molecules and Agilent Technologies, USA). Briefly, 105 cells/well were plated in XF24 microplates, the Ct values obtained. differentiated, and induced for hTyr expression for 0, 3, and 6 days as described above. 24 h prior to the experiment, the instrument was calibrated following the manufacturer’s instructions. Cells were incubated for 1 h in 675 µl respiration TR5TY6 neuroblastoma cell line. A stable inducible SH-SY5Y cell line expressing medium (DMEM base, Sigma-Aldrich, supplemented with 5.5 mM glucose, 31.6 human tyrosinase (TR5TY6) under the transcriptional control of the T-Rex TM mM NaCl, Phenol Red, and 2 mM GlutaMax, Thermo Fisher Scientific) at 37 °C in Tet-On system (Invitrogen) was provided by Dr. T. Hasegawa (Department of a CO2-free incubator. Following overnight calibration of the sensor cartridge, plates Neurology, Tohoku University, Sendai, Japan). Cell line was confirmed negative for were run on the Seahorse XF24 flux analyzer to calculate OCR. Four initial mea- mycoplasma contamination by routine PCR analysis. Cells were maintained in low- surements, corresponding to basal oxygen consumption (routine state, without glucose (1 g/l) Dulbecco’s modified Eagle’s (DMEM) (Gibco) medium with peni- respiratory chain inhibitors), were performed per experiment, with each mea- cillin/streptomycin, and the appropriate selection of antibiotics (7 µg/mL blas- surement cycle consisting of 4 min agitation, 2 min stand, and 1.5 min measure- ticidin and 300 µg/mL Zeocin, both from Life Technologies). Medium was ment. After each run, cells were lysed in RIPA buffer (150 mM NaCl, 10 mM Tris, supplemented with tetracycline-free fetal bovine serum (FBS) (Clontech) to avoid pH 7.2, 0.1% SDS, 1% Triton X-100, 1% deoxycholate, 5 mM EDTA, 1 mM PMSF) unwanted expression of the transgene. TR5TY6 cells were seeded for aSyn-PLA, and OCRs normalized to protein concentration, with measurements made flow cytometry and immunoblot analysis at 2.5 × 105 or 106 cells/plate in 24-, 12-, using BCA. Four independent experiments were performed, each in triplicates. or 6-well plates, respectively. For PLA, cells were grown onto 12-mm slides coated with 50 µg/ml poly-D lysine (Sigma-Aldrich). Unless stated otherwise, 24 h after seeding cells were differentiated with 10 µM retinoic acid (RA) (Sigma-Aldrich) for Intracellular protein degradation assay. Total protein degradation in differ- 3 days, followed by 80 nM 12-O-tetradecanoylphorbol-13-acetate (TPA) (Sigma- entiated TR5TY6 cells was measured by pulse-chase experiments as previously Aldrich) for 3 extra days prior to hTyr induction with 2 µg/ml doxycycline (Sigma- described, with some modification80. Briefly, confluent TR5TY6 cells were induced Aldrich) for up to 6 days. Intracellular NM in differentiated hTyr-induced TR5TY6 for hTyr expression for 0, 3, and 6 days and labeled with [3H]-valine (2 μCi/ml; cells was quantified at the indicated times as follows: (i) optical densitometry in Hartmann Analytic) at 37 °C for 24 h. Cells were then extensively washed with fixed cells, examined under transmitted light at an objective magnification of ×40; medium and returned in complete growth medium containing an excess of unla- each slide was divided in a grid, images were taken from 25 fields (50% coverage) beled valine for a chase period of 5 h. Cells were cultured with or without lysosomal and analyzed as described above for rat sections; (ii) spectrophotometry, by inhibitors (100 µM leupeptin, 10 µM pepstatin A, and 10 µM E64d) to assess the diluting 2 µl of cell lysate in a total volume of 200 µl PBS and measuring the contribution of the lysosomal system to protein degradation. Triplicated aliquots of absorbance at 405 nm with a ELx800 Absorbance Microplate Reader (BioTek the medium at the indicated time were collected and precipitated with 20% tri- Instruments, USA). chloroacetic acid. Acid-soluble fractions were collected after vacuum filtration using a Millipore manifolder and were quantified using a LS6500 liquid scintilla- tion counter (Beckman Coulter). Cells were then lysed with 0.1% NaOH and 0.1% Immunoblot. Differentiated TR5TY6 cells induced for 0, 1, 3, and 6 days were Triton X-100, and the total protein concentration measured for normalization homogenized in RIPA buffer supplemented with protease inhibitors (Roche) and purposes and to measure total surviving cells at the end of the induction times. cell extracts clarified by centrifugation at 10,000 × g for 10 min at 4 °C. Protein Proteolysis was measured using a LS6500 liquid scintillation counter (Beckman concentrations were quantified using the BCA method and subjected to SDS- Coulter) and expressed as the percentage of the initial acid insoluble radioactivity PAGE. Proteins were resolved in 10 or 15% polyacrylamide gels and transferred (total protein) transformed to acid-soluble radioactivity (amino acids and small onto 0.45 µm nitrocellulose membranes (Amersham). Blocking with 5% milk peptides) at the end of the incubation time. Lysosomal proteolysis was calculated powder in PBS was followed by overnight incubation at 4 °C with the primary as the proteolysis sensitive to lysosomal inhibitors. Three independent pulse and antibodies mouse anti-tyrosinase (1:2000, Thermo Fisher Scientific, Cat#MS-800- chase experiments were performed. P1), rabbit anti-Lamp1 (1:250, GeneTex, Cat#GTX19294), rabbit anti-LC3 (1:1000, Novus Biologicals, Cat#NB100-2220), guinea pig anti-p62 (1:1000, Progen, Cat#GP62-C) or mouse anti-β-actin (1:5000, Sigma-Aldrich, Cat#A5441). Incu- Ubiquitin-proteasome system activity. Differentiated TR5TY6 cells induced for bation with the secondary antibodies donkey anti-rabbit, sheep anti-mouse (both hTyr expression for 0, 3, and 6 days were lysed in 10 mM tris pH7.8, 1 mM EDTA, 1:5000, from Amersham), and goat anti-guinea pig (1:1000, Santa Cruz Bio- 5 mM Mg2Cl, 0.1% Triton X-100. Chymotrypsin-like activity was determined technology) was performed for 1 h at RT. Band densitometry, normalized to using 10 µg of total cell extracts in a total volume of proteasome activity buffer (10 β-actin expression, were measured using ImageJ image analysis software. mM tris pH7.8, 1 mM EDTA, 0.5 mM DTT, 5 mM Mg2Cl, 2 mM ATP), incubating for 60 min at 37 °C with 0.5 mM of the fluorogenic substrate Z-Leu-Leu-Glu-AMC (Z-LLE-AMC) (Enzo, Cat# BML-ZW9345). As control, all reactions were also Optical microscopy for NM. Cells fixed with 4% paraformaldehyde for 30 min at performed in parallel in the presence of 50 µM MG132 inhibitor (Enzo BML- 4 °C were blocked with 3% NGS (Atom) and 0.1% (vol/vol) Triton X-100 (Sigma- PI102). Fluorescence intensity was quantified by using the Flx800 multi-detection Aldrich) in PBS solution. Hoechst 33342 (1:10,000, Thermo Fisher Scientific) microplate reader (BioTek) with excitation and emission wavelengths at 360 and was used for nuclei counterstain for 1 h at RT in 2% (vol/vol) NGS. Cells were 460 nm, respectively. UPS activity at different time points is presented as percen- coverslipped using Dako Cytomation Fluorescent Mounting Medium (Dako) tage of the control (OFF cells), and represent the mean of three independent and images acquired using standard filter sets either with an Olympus FSX100 experiments each performed in triplicates. microscope with a DP72 incorporated camera and FSX-BSW visualization software (Olympus, Germany) or with an Olympus FluoView™ FV1000 confocal microscope and FV10-ASW 4.2 visualization software. TFEB-induced lysosomal exocytosis (dot blot). TR5TY6 cells were transiently transfected using Lipofectamine ™3000 Transfection reagent (Thermo Fisher Sci- entific) according to the manufacturer’s instructions. Briefly, 5 × 104 cells were aSyn-proximity ligation assay. TR5TY6 cells were induced for hTyr expression seeded into 6-well plates the day before transfection with 2.5 µg of the construct for 0, 3, and 6 days and fixed in 4% paraformaldehyde for 30 min at 4 °C. Custom pCIP-Flag-TfebAA (a gift from Reuben Shaw, Addgene plasmid Cat#79014)) and PLA probes were prepared by conjugating 20 µl of the Ab-2 mouse monoclonal 5 µl of P3000 reagent. Transfected cells were harvested with medium, centrifuged at aSyn antibody (1 mg/ml) (Thermo Scientific, Cat#MS-1572) with one vial of 800 × g at RT and 150 µl of culture medium were spotted onto a nitrocellulose Duolink In Situ Probemaker containing either PLUS (Sigma-Aldrich, membrane (GE Healthcare) using the Mini-Fold®-1 Dot-Blot system (Whatman, 16 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE Inc.). The membrane was blocked in 5% non-fat milk powder in PBS for 1 h at RT Reporting summary. Further information on experimental design is available in and incubated overnight at 4 °C with mouse anti-Lamp2 (H4B4) antibody (1:1000, the Nature Research Reporting Summary linked to this article. Santa Cruz Biotechnology, Inc. Cat#sc-18822). Incubation with anti-mouse sec- ondary antibody coupled to horseradish peroxidase (1:1000 in 5% non-fat milk powder/PBS; GE Healthcare, #NXA931V and #NA934V) was performed for 1 h at Data availability RT, followed by repeated washings with PBS. Immunoreactive spots were visua- Data relevant to the study is available from the authors on reasonable request. lized using SuperSignal Femto Chemiluminescent Substrate (Pierce) according to the manufacturer’s instructions and quantified on an ImageQuant RT ECL imaging system (GE Healthcare). Lysosomal exocytosis at different time points is presented Received: 26 November 2018 Accepted: 28 January 2019 as percentage of the non-TFEB-transfected cells, and represent the mean of three independent experiments run in duplicates. Cell metabolic activity and ROS production. Metabolic activity was measured in differentiated TR5TY6 induced for hTyr expression for 0, 3, and 6 days. Briefly, References cells were centrifuged at 800 × g for 5 min at 4 °C, resuspended in PBS and incu- 1. Hirsch, E., Graybiel, A. M. & Agid, Y. A. Melanized dopaminergic neurons are bated at 37 °C and 5% CO2 for 15 min with 10 µM C12-resazurin (Thermo Fisher differentially susceptible to degeneration in Parkinson’s disease. Nature 334, Scientific). C12-Resazurin is readily reduced in metabolically active cells to red- 345–348 (1988). fluorescent C12-resofurin (λexcitation = 563 nm; λemission = 587 nm). To quantify 2. Dickson, D. W. et al. Neuropathological assessment of Parkinson’s disease: the production of ROS, TR5TY6 cells were loaded with freshly prepared 25 µM refining the diagnostic criteria. Lancet Neurol. 8, 1150–1157 (2009). CellROX Green Reagent (λexcitation = 485 nm; λemission = 520 nm) (Thermo 3. Sulzer, D. & Surmeier, D. J. Neuronal vulnerability, pathogenesis, and Fisher Scientific), a fluorescent dye that measures ROS production independently Parkinson’s disease. Mov. Disord. 28, 715–724 (2013). of its potential origin or location, for 30 min at 37 °C and 5% CO2. Cells were 4. Bensaid, M., Michel, P. P., Clark, S. D., Hirsch, E. C. & Francois, C. Role afterwards washed and harvested for flow cytometry analysis. Fluorescence data acquisition was performed by recording 10,000 events in an LSR Fortessa flux of pedunculopontine cholinergic neurons in the vulnerability of nigral cytometer (BD Biosciences) and analyzed with FCS Express Version 4 software. dopaminergic neurons in Parkinson’s disease. Exp. Neurol. 275(Pt 1), 209–219 Intact cells were gated in an FSC/SSC plot to exclude small debris. Cell metabolic (2016). activity and ROS production at different time points are presented as mean 5. Szego, E. M., Gerhardt, E., Outeiro, T. F. & Kermer, P. Dopamine-depletion fluorescence intensity, and represent the mean of three independent experiments and increased alpha-synuclein load induce degeneration of cortical cholinergic each performed in triplicates. fibers in mice. J. Neurol. Sci. 310, 90–95 (2011). 6. Fedorow, H. et al. Neuromelanin in human dopamine neurons: comparison with peripheral melanins and relevance to Parkinson’s disease. Prog. Transmission electron microscopy. (i) Animals: rats were perfused transcardially Neurobiol. 75, 109–124 (2005). with 4% PFA and 0.1% glutaraldehyde (Merck Millipore) in Dulbecco’s phosphate- 7. Fedorow, H. et al. Evidence for specific phases in the development of human buffered saline (DPBS, Gibco). Brains were removed and post-fixed for 4 h at 4 °C neuromelanin. Neurobiol. Aging 27, 506–512 (2006). in the same fixative, washed in DPBS and sectioned into 100-μm-thickness slices 8. Driver, J. A., Logroscino, G., Gaziano, J. M. & Kurth, T. Incidence and using a vibratome (Leica, Germany). Free-floating sections were blocked in 0.1% remaining lifetime risk of Parkinson disease in advanced age. Neurology 72, Triton X100 and 10% NGS in PBS for 1 h at RT, incubated with primary antibody 432–438 (2009). rabbit anti-TH (1:1000, Calbiochem) in PBS, a secondary goat anti-rabbit antibody 9. Hirsch, E. C., Graybiel, A. M. & Agid, Y. Selective vulnerability of pigmented conjugated to biotin (1:1000, Vector Laboratories), and visualized using Immu- dopaminergic neurons in Parkinson’s disease. Acta Neurol. Scand. Suppl. 126, nopure ABC Peroxidase staining kit (Thermo Fisher Scientific) and DAB Perox- 19–22 (1989). idase Substrate kit (Vector laboratories). The sections were post-fixed in 3% 10. Kastner, A. et al. Is the vulnerability of neurons in the substantia nigra of glutaraldehyde and in 1% osmium tetroxide in 0.1 M cacodylate buffer (pH 7.4), patients with Parkinson’s disease related to their neuromelanin content? dehydrated in ethanol, and embedded in Spurr’s epoxy resin (Sigma-Aldrich). J. Neurochem. 59, 1080–1089 (1992). Survey sections (2 μm) and ultra-thin sections (around 70 nm) were cut with a 11. Braak, E. et al. alpha-synuclein immunopositive Parkinson’s disease-related diamond knife (Diatome) using an ultramicrotome Ultracut E (Reichert-Jung). inclusion bodies in lower brain stem nuclei. Acta Neuropathol. 101, 195–201 Ultrathin sections were collected onto formvar-coated copper/gold grids, coun- (2001). terstained with 2% uranyl acetate and led citrate, and examined in a JEOL 1010 12. Halliday, G. M. et al. Alpha-synuclein redistributes to neuromelanin lipid in (tungsten filament) transmission electron microscope. Images were acquired with the substantia nigra early in Parkinson’s disease. Brain 128, 2654–2664 (2005). an Orius CCD camera using Digital Micrograph software (GATAN). (ii) Post- 13. Fasano, M., Giraudo, S., Coha, S., Bergamasco, B. & Lopiano, L. Residual mortem human brain: human control samples were initially fixed in formaldehyde substantia nigra neuromelanin in Parkinson’s disease is cross-linked to alpha- and post-fixed in 3% glutaraldehyde, embedded in epoxy resin. Semi-thin and synuclein. Neurochem. Int. 42, 603–606 (2003). ultra-thin sections were processed as above. Sections were examined at a Zeiss 10 A 14. Braak, H. et al. Staging of brain pathology related to sporadic Parkinson’s transmission electron microscope. (iii) TR5TY6 cells: confluent 10-cm culture disease. Neurobiol. Aging 24, 197–211 (2003). plates containing differentiated TR5TY6 cells induced for hTyr expression for 0, 1, 15. Barden, H. & Levine, S. Histochemical observations on rodent brain melanin. 3, and 6 days were washed with a fixative solution containing 2.5% glutaraldehyde Brain Res. Bull. 10, 847–851 (1983). (Merck Millipore) in 0.1 M DPBS and then fixed for 1 h at RT in the same fixative. 16. Marsden, C. D. Pigmentation in the nucleus substantiae nigrae of mammals. Cells were afterwards gently scraped and centrifuged at 300 × g to get a visible pellet J. Anat. 95, 256–261 (1961). that was kept in fixative at 4 °C until forward processing. Cell pellets were post- 17. Sulzer, D. et al. Neuromelanin biosynthesis is driven by excess cytosolic fixed in 1% osmium tetroxide in 0.1 M cacodylate buffer (pH 7.4), dehydrated in catecholamines not accumulated by synaptic vesicles. Proc. Natl Acad. Sci. ethanol, and embedded in Spurr’s epoxy resin (Sigma-Aldrich). Semi-thin and USA 97, 11869–11874 (2000). ultra-thin sections, counterstain, and image acquisition were done as described above for rat sections. 18. Simon, J. D., Peles, D., Wakamatsu, K. & Ito, S. Current challenges in understanding melanogenesis: bridging chemistry, biological control, morphology, and function. Pigment Cell Melanoma Res. 22, 563–579 (2009). Statistics. All values are expressed as the mean ± standard error of the mean 19. Sanchez-Ferrer, A., Rodriguez-Lopez, J. N., Garcia-Canovas, F. & Garcia- (SEM). Statistical comparisons were performed with SigmaStat software (v4, Systat Carmona, F. Tyrosinase: a comprehensive review of its mechanism. Biochim. Software Inc, USA) using the appropriate statistical tests, as indicated in each figure Biophys. Acta 1247, 1–11 (1995). legend. No statistical methods were used to pre-determine sample size but our 20. Miranda, M., Botti, D., Bonfigli, A., Ventura, T. & Arcadi, A. Tyrosinase-like sample sizes are equivalent to those reported in previous similar publications81. activity in normal human substantia nigra. Gen. Pharmacol. 15, 541–544 Outlier values were identified by the Grubbs’ test (i.e. Extreme Studentized Deviate, (1984). ESD, method) using GraphPad Prism software (v6, GraphPad Software Inc, USA) 21. Greggio, E. et al. Tyrosinase exacerbates dopamine toxicity but is not and excluded from the analyses when applicable. Selection of the pertinent sta- genetically associated with Parkinson’s disease. J. Neurochem. 93, 246–256 tistical test for each experiment was determined after formally testing for nor- (2005). mality. Accordingly, differences among means were analyzed either by 1- or 2-way 22. Xu, Y. et al. Tyrosinase mRNA is expressed in human substantia nigra. analysis of variance (ANOVA), Kruskal–Wallis ANOVA on ranks, two-tailed t-test Brain Res. Mol. Brain Res. 45, 159–162 (1997). or Mann–Whitney rank sum test, as appropriate. When ANOVA showed sig- 23. Sulzer, D. et al. Neuronal pigmented autophagic vacuoles: lipofuscin, nificant differences, pairwise comparisons between means were subjected to neuromelanin, and ceroid as macroautophagic responses during aging Student–Newman–Keuls, Dunn’s or Holm-Sidak post-hoc testing for multiple and disease. J. Neurochem. 106, 24–36 (2008). comparisons, as appropriate. In all analyses, the null hypothesis was rejected at 24. Halliday, G. M. et al. Evidence for specific phases in the development of the 0.05 level. human neuromelanin. J. Neural Transm. 113, 721–728 (2006). NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 17 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y 25. Lundblad, M., Decressac, M., Mattsson, B. & Bjorklund, A. Impaired 55. Roberts, R. F., Wade-Martins, R. & Alegre-Abarrategui, J. Direct visualization neurotransmission caused by overexpression of alpha-synuclein in nigral of alpha-synuclein oligomers reveals previously undetected pathology in dopamine neurons. Proc. Natl Acad. Sci. USA 109, 3213–3219 (2012). Parkinson’s disease brain. Brain 138, 1642–1657 (2015). 26. Pifl, C. et al. Is Parkinson’s disease a vesicular dopamine storage disorder? 56. Martinez-Vicente, M. & Cuervo, A. M. Autophagy and neurodegeneration: Evidence from a study in isolated synaptic vesicles of human and nonhuman when the cleaning crew goes on strike. Lancet Neurol. 6, 352–361 (2007). primate striatum. J. Neurosci. 34, 8210–8218 (2014). 57. Kaushik, S. & Cuervo, A. M. Proteostasis and aging. Nat. Med. 21, 1406–1415 27. Sawamoto, N. et al. Cognitive deficits and striato-frontal dopamine release (2015). in Parkinson’s disease. Brain 131, 1294–1302 (2008). 58. Sardiello, M. et al. A gene network regulating lysosomal biogenesis and 28. Iranzo, A. et al. Dopamine transporter imaging deficit predicts early transition function. Science 325, 473–477 (2009). to synucleinopathy in idiopathic rapid eye movement sleep behavior disorder. 59. Settembre, C. et al. TFEB links autophagy to lysosomal biogenesis. Science 332, Ann. Neurol. 82, 419–428 (2017). 1429–1433 (2011). 29. Iranzo, A. et al. Neurodegenerative disease status and post-mortem pathology 60. Medina, D. L. et al. Transcriptional activation of lysosomal exocytosis in idiopathic rapid-eye-movement sleep behaviour disorder: an observational promotes cellular clearance. Dev. Cell 21, 421–430 (2011). cohort study. Lancet Neurol. 12, 443–453 (2013). 61. Song, W. et al. TFEB regulates lysosomal proteostasis. Hum. Mol. Genet. 22, 30. Iranzo, A. et al. Serial dopamine transporter imaging of nigrostriatal function 1994–2009 (2013). in patients with idiopathic rapid-eye-movement sleep behaviour disorder: 62. Magalhaes, J., Gegg, M. E., Migdalska-Richards, A. & Schapira, A. H. Effects a prospective study. Lancet Neurol. 10, 797–805 (2011). of ambroxol on the autophagy-lysosome pathway and mitochondria in 31. Bergareche, A. et al. DAT imaging and clinical biomarkers in relatives at primary cortical neurons. Sci. Rep. 8, 1385 (2018). genetic risk for LRRK2 R1441G Parkinson’s disease. Mov. Disord. 31, 335–343 63. Napolitano, G. & Ballabio, A. TFEB at a glance. J. Cell Sci. 129, 2475–2481 (2016). (2016). 32. Sossi, V. et al. Dopamine turnover increases in asymptomatic LRRK2 64. Murer, M. G. et al. Chronic levodopa is not toxic for remaining dopamine mutations carriers. Mov. Disord. 25, 2717–2723 (2010). neurons, but instead promotes their recovery, in rats with moderate 33. Artzi, M. et al. DaT-SPECT assessment depicts dopamine depletion among nigrostriatal lesions. Ann. Neurol. 43, 561–575 (1998). asymptomatic G2019S LRRK2 mutation carriers. PLoS ONE 12, e0175424 65. Yamamuro, Y. & Ogura, S. Regional expression of tyrosinase in central (2017). catecholaminergic systems of colored mice. Exp. Anim. (2018). https://doi.org/ 34. Sierra, M. et al. Prospective clinical and DaT-SPECT imaging in premotor 10.1538/expanim.18-0053 [Epub ahead of print]. LRRK2 G2019S-associated Parkinson disease. Neurology 89, 439–444 66. Tribl, F., Arzberger, T., Riederer, P. & Gerlach, M. Tyrosinase is not detected (2017). in human catecholaminergic neurons by immunohistochemistry and Western 35. Nandhagopal, R. et al. Longitudinal evolution of compensatory changes in blot analysis. J. Neural Transm. Suppl. 51–55 (2007). striatal dopamine processing in Parkinson’s disease. Brain 134, 3290–3298 67. Ikemoto, K. et al. Does tyrosinase exist in neuromelanin-pigmented (2011). neurons in the human substantia nigra? Neurosci. Lett. 253, 198–200 36. Kotagal, V. et al. Symptoms of rapid eye movement sleep behavior disorder (1998). are associated with cholinergic denervation in Parkinson disease. Ann. Neurol. 68. Zucca, F. A. et al. Neuromelanin of the human substantia nigra: an update. 71, 560–568 (2012). Neurotox. Res. 25, 13–23 (2014). 37. Adams, J. R. et al. PET in LRRK2 mutations: comparison to sporadic 69. Karlsson, O. & Lindquist, N. G. Melanin and neuromelanin binding of drugs Parkinson’s disease and evidence for presymptomatic compensation. Brain and chemicals: toxicological implications. Arch. Toxicol. 90, 1883–1891 128, 2777–2785 (2005). (2016). 38. Fu, J. F. et al. Investigation of serotonergic Parkinson’s disease-related 70. de Lau, L. M. & Breteler, M. M. Epidemiology of Parkinson’s disease. covariance pattern using [(11)C]-DASB/PET. NeuroImage. Clin. 19, 652–660 Lancet Neurol. 5, 525–535 (2006). (2018). 71. Fernandez-Santiago, R. et al. Aberrant epigenome in iPSC-derived 39. Beach, T. G. et al. Marked microglial reaction in normal aging human dopaminergic neurons from Parkinson’s disease patients. EMBO Mol. Med. 7, substantia nigra: correlation with extraneuronal neuromelanin pigment 1529–1546 (2015). deposits. Acta Neuropathol. 114, 419–424 (2007). 72. Lubbe, S. J. et al. Rare variants analysis of cutaneous malignant melanoma 40. Langston, J. W. et al. Evidence of active nerve cell degeneration in the genes in Parkinson’s disease. Neurobiol. Aging 48, 222e.1–222.e7 (2016). substantia nigra of humans years after 1-methyl-4-phenyl-1,2,3,6- 73. Cano-Jaimez, M. et al. Vulnerability of peripheral catecholaminergic neurons tetrahydropyridine exposure. Ann. Neurol. 46, 598–605 (1999). to MPTP is not regulated by alpha-synuclein. Neurobiol. Dis. 38, 92–103 41. Sulzer, D. & Zecca, L. Intraneuronal dopamine-quinone synthesis: a review. (2010). Neurotox. Res. 1, 181–195 (2000). 74. Piedra, J. et al. Development of a rapid, robust, and universal picogreen-based 42. Hastings, T. G., Lewis, D. A. & Zigmond, M. J. Role of oxidation in the method to titer adeno-associated vectors. Hum. Gene Ther. Methods 26, 35–42 neurotoxic effects of intrastriatal dopamine injections. Proc. Natl Acad. Sci. (2015). USA 93, 1956–1961 (1996). 75. Paxinos, G. & Watson, C. The Rat Brain In Stereotaxic Coordinates. 43. Caudle, W. M. et al. Reduced vesicular storage of dopamine causes progressive (Academic Press, United States of America, 1982). nigrostriatal neurodegeneration. J. Neurosci. 27, 8138–8148 (2007). 76. Bortolozzi, A., Diaz-Mataix, L., Scorza, M. C., Celada, P. & Artigas, F. The 44. Burbulla, L. F. et al. Dopamine oxidation mediates mitochondrial and activation of 5-HT receptors in prefrontal cortex enhances dopaminergic lysosomal dysfunction in Parkinson’s disease. Science 357, 1255–1261 activity. J. Neurochem. 95, 1597–1607 (2005). (2017). 77. Diaz-Mataix, L. et al. Involvement of 5-HT1A receptors in prefrontal cortex 45. Mor, D. E. et al. Dopamine induces soluble alpha-synuclein oligomers and in the modulation of dopaminergic activity: role in atypical antipsychotic nigrostriatal degeneration. Nat. Neurosci. 20, 1560–1568 (2017). action. J. Neurosci. 25, 10831–10843 (2005). 46. Kordower, J. H. et al. Disease duration and the integrity of the nigrostriatal 78. Bortolozzi, A. & Artigas, F. Control of 5-hydroxytryptamine release in system in Parkinson’s disease. Brain 136, 2419–2431 (2013). the dorsal raphe nucleus by the noradrenergic system in rat brain. Role of 47. Kuusisto, E., Parkkinen, L. & Alafuzoff, I. Morphogenesis of lewy bodies: alpha-adrenoceptors. Neuropsychopharmacology 28, 421–434 (2003). dissimilar incorporation of a-synuclein, ubiquitin, and p62. J. Neuropathol. 79. Lemos-Amado, F. et al. Electrospray tandem mass spectrometry of Exp. Neurol. 62, 1241–1253 (2003). aminochromes. Rapid Commun. Mass Spectrom. 15, 2466–2471 (2001). 48. Beach, T. G. et al. Substantia nigra Marinesco bodies are associated with 80. Kaushik, S. & Cuervo, A. M. Methods to monitor chaperone-mediated decreased striatal expression of dopaminergic markers. J. Neuropathol. Exp. autophagy. Methods Enzymol. 452, 297–324 (2009). Neurol. 63, 329–337 (2004). 81. Decressac, M. et al. TFEB-mediated autophagy rescues midbrain dopamine 49. Tessari, I. et al. The reaction of alpha-synuclein with tyrosinase: possible neurons from alpha-synuclein toxicity. Proc. Natl Acad. Sci. USA 110, implications for Parkinson disease. J. Biol. Chem. 283, 16808–16817 (2008). E1817–E1826 (2013). 50. Pan, T., Zhu, J., Hwu, W. J. & Jankovic, J. The role of alpha-synuclein in melanin synthesis in melanoma and dopaminergic neuronal cells. PLoS ONE 7, e45183 (2012). Acknowledgements 51. Klionsky, D. J. et al. Guidelines for the use and interpretation of assays This work was supported by funds from the Fondo de Investigación Sanitaria-Instituto for monitoring autophagy (3rd edition). Autophagy 12, 1–222 (2016). de Salud Carlos III (FIS-ISCIII, Spain)-European Regional Development Fund (FEDER, 52. Dehay, B. et al. Pathogenic lysosomal depletion in Parkinson’s disease. E.U.) (PI13/01897, to M.V.), Parkinson’s U.K. (to M.V.), Ministry of Economy and J. Neurosci. 30, 12535–12544 (2010). Competitiveness (MINECO, Spain) (SAF2016-77541-R and RTC-2014-2812-1, to M.V.), 53. Cook, C. & Petrucelli, L. A critical evaluation of the ubiquitin-proteasome The Michael J. Fox Foundation (U.S.A) (ID15291, to M.V.), La Caixa Banking Foun- system in Parkinson’s disease. Biochim. Biophys. Acta 1792, 664–675 (2009). dation (Health Research Project HR17-00513, to M.V.) and CIBERNED (to M.V.). In 54. Hara, T. et al. Suppression of basal autophagy in neural cells causes addition, the Authors would like to acknowledge additional support from the Fundación neurodegenerative disease in mice. Nature 441, 885–889 (2006). Tatiana Pérez de Guzmán el Bueno (Spain, to A.L.), Michael J. Fox Foundation (U.S.A.) 18 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-08858-y ARTICLE (ID11580, to A.L.), MINECO (SAF2016-75797-R, to A.Bo.) and FIS-ISCIII-FEDER Additional information (PI15/01937 to J.B. and PI13/01390 to A.Bo.). A.L. was the recipient of a post-doctoral Supplementary Information accompanies this paper at https://doi.org/10.1038/s41467- fellowship Beatriu de Pinós (2013 BP-DGR B 00043) from the Agència de Gestió d’Ajuts 019-08858-y. Universitaris i de Recerca (AGAUR, Spain) with co-funding from the Marie Curie COFUND program (E.U.) and a postdoctoral contract SAF2015-73997-JIN from Competing interests: VHIR has applied for a use patent for the modulation of MINECO (Spain) with co-funding from FEDER (E.U.). N.P. is the recipient of a pre- neuromelanin levels in the treatment of Parkinson's disease and/or brain aging. doctoral fellowship FPI (BES-2017-080191) from MINECO (Spain). A.T. is the recipient of a pre-doctoral fellowship PFIS (FI14/00613) from the FIS-ISCIII (Spain). We are Reprints and permission information is available online at http://npg.nature.com/ grateful to: E. Ruiz-Bronchal (IIBB-CSIC-IDIBAPS, Barcelona, Spain) for technical reprintsandpermissions/ assistance in the performance of microdialysis/HPLC experiments; the Neurological Tissue Bank of the Biobanc-Hospital Clinic-IDIBAPS (Barcelona, Spain) for human data Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in and sample procurement; S. Lope-Piedrafita (UAB, Barcelona, Spain) for technical published maps and institutional affiliations. assistance in the performance of MRI experiments in rodents; the Unit 20 of CIBER in Bioengineering, Biomaterials & Nanomedicne (CIBER-BBN) at the VHIR (Barcelona, Spain) for assistance in histological processing; D. Sebastián and A. Zorzano (IRB-PCB, Open Access This article is licensed under a Creative Commons Barcelona, Spain) for assistance in the SeaHorse mitochondrial respirometry experi- Attribution 4.0 International License, which permits use, sharing, ments; I. Fariñas (University of Valencia-CIBERNED) for providing aSyn-deficient adaptation, distribution and reproduction in any medium or format, as long as you give mice; A. Rovira (Neuroradiology section, Vall d’Hebron Hospital, Barcelona, Spain) appropriate credit to the original author(s) and the source, provide a link to the Creative for providing the MRI image from a human control subject and S. Dovero (University Commons license, and indicate if changes were made. The images or other third party of Bordeaux, France) for drawing the rodent brain atlas illustrations. material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory Author contributions regulation or exceeds the permitted use, you will need to obtain permission directly from I.C.-C., A.L., J.R-G., T.C., J.B., M.M-V., A.P., M.G-S., N.P., A.T., B.R-G., A.Bo. and E.G. the copyright holder. To view a copy of this license, visit http://creativecommons.org/ performed the research and analyzed the data; A.Ba. produced the AAV-TFEB viral licenses/by/4.0/. vector; T.H. produced the hTyr-expressing cell line; A.L. and M.V. supervised experi- ments; M.V. conceived and planned the study, acquired the funding, analyzed the data and wrote the manuscript, which was reviewed by all authors. © The Author(s) 2019 NATURE COMMUNICATIONS | (2019)10:973 | https://doi.org/10.1038/s41467-019-08858-y | www.nature.com/naturecommunications 19