RGS9‐2 rescues dopamine D2 receptor levels and signaling in DYT1 dystonia mouse models

Abstract Dopamine D2 receptor signaling is central for striatal function and movement, while abnormal activity is associated with neurological disorders including the severe early‐onset DYT1 dystonia. Nevertheless, the mechanisms that regulate D2 receptor signaling in health and disease remain poorly understood. Here, we identify a reduced D2 receptor binding, paralleled by an abrupt reduction in receptor protein level, in the striatum of juvenile Dyt1 mice. This occurs through increased lysosomal degradation, controlled by competition between β‐arrestin 2 and D2 receptor binding proteins. Accordingly, we found lower levels of striatal RGS9‐2 and spinophilin. Further, we show that genetic depletion of RGS9‐2 mimics the D2 receptor loss of DYT1 dystonia striatum, whereas RGS9‐2 overexpression rescues both receptor levels and electrophysiological responses in Dyt1 striatal neurons. This work uncovers the molecular mechanism underlying D2 receptor downregulation in Dyt1 mice and in turn explains why dopaminergic drugs lack efficacy in DYT1 patients despite significant evidence for striatal D2 receptor dysfunction. Our data also open up novel avenues for disease‐modifying therapeutics to this incurable neurological disorder.


Introduction
Striatal dopaminergic transmission is central to movement control and several disease conditions (Redgrave et al, 2010;Gittis & Kreitzer, 2012). Dopaminergic dysfunction has been implicated in early-onset generalized DYT1-TOR1A dystonia, a highly disabling and incurable neurological disease typically manifesting in childhood, which generalizes within a few years causing involuntary movements and abnormal postures (Balint et al, 2018). This disorder is most frequently caused by an autosomal dominant Δgag mutation in the TOR1A gene, causing loss of function of the gene product torsinA, a member of the AAA+ (ATPases associated with cellular activities) family of proteins (Ozelius et al, 1997). Mutant ΔE-torsinA is mislocalized from the endoplasmic reticulum to the nuclear envelope, causing abnormalities in folding, assembly, and trafficking of proteins targeted for secretion or to membranes (Torres et al, 2004;Burdette et al, 2010;Granata et al, 2011).
Clinical neuroimaging studies have revealed decreased caudateputamen dopamine D2 receptor (DRD2) availability in DYT1 patients compared to controls (Asanuma et al, 2005;Carbon et al, 2009). Reduced striatal DRD2 binding and protein level have also been reported in several different DYT1 experimental models (Napolitano et al, 2010;Yokoi et al, 2011;Dang et al, 2012). Notably, multiple lines of evidence demonstrated reduced coupling between the DRD2 and its cognate G proteins and severely altered receptor function (Pisani et al, 2006;Napolitano et al, 2010;Sciamanna et al, 2011Sciamanna et al, , 2012aMartella et al, 2014;Scarduzio et al, 2017).
DRD2 signaling via G proteins inhibits cAMP production and in turn PKA activity. However, there is also accumulating evidence for G protein-independent DRD2 signaling functions , as well as G protein-independent regulation of the GPCR activity of DRD2. The reciprocal interactions of DRD2 with spinophilin or arrestin represent a regulatory mechanism for finetuning receptor-mediated signaling. Indeed, b-arrestin 2 (b-Arr2) is involved in internalization and G protein-independent signaling of DRD2 Del'guidice et al, 2011), while spinophilin antagonizes arrestin actions (Wang et al, 2004). In addition, the striatal-enriched regulator of G protein signaling 9-2 (RGS9-2) regulates the amplitude of the behavioral responses to DRD2 activation (Rahman et al, 2003;Gold et al, 2007;Traynor et al, 2009), inhibits DRD2 internalization (Celver et al, 2010), and specifically modulates DRD2 signaling in striatal neuronal subtypes (Cabrera-Vera et al, 2004). On the other hand, the receptor can target the RGS protein to the plasma membrane (Kovoor et al, 2005;Celver et al, 2012), and exposure to DRD2 ligands can alter RGS9-2 level in wildtype animals (Seeman et al, 2007) indicating a reciprocal modulation.
In the present work, we investigated the molecular mechanisms underlying DRD2 reduced levels and altered signaling in the striatum of DYT1 dystonia models, Tor1a +/À -knock-out and Tor1a Δgag/+knock-in mice (Goodchild et al, 2005). Our findings shed new light on DRD2 dysfunction in DYT1 striatum and show that in vivo delivery of RGS9-2 is able to rescue DRD2 expression levels and to recover striatal D2DR signaling. These findings might explain the paradox of the lack of efficacy of dopaminergic drugs in DYT1-TOR1A dystonia patients, despite strong evidence that abnormal dopamine signaling is central to disease pathophysiology. Further, they also define a potential therapeutic target that restores dopaminergic responses.

Results
DRD2 and RGS9-2 protein levels are simultaneously downregulated in DYT1 striatum In order to analyze the molecular mechanisms of DRD2 dysfunction, we utilized the Tor1a +/À mouse model that mimics the loss of function effect of the DYT1 dystonia TOR1A mutation.

DRD2 downregulation is mediated by endolysosomal trafficking and mimicked by RGS9-2 silencing
Our experiments provide evidence of normal DRD2 mRNA levels ( Fig EV1), on one hand, but reduced protein stability and levels, on the other. We therefore investigated the pathway of protein quality control targeting plasma membrane proteins to endocytic trafficking and lysosomal degradation (MacGurn, 2014).
Source data are available online for this figure.
WB analysis of markers of endocytic trafficking and degradative pathway in the dorsal striatum of Tor1a +/À mice.
Data information: The graphs in (A, B, D, E, G) report mean AE SEM of the ratio of protein vs. loading control intensity level, normalized to the Tor1a +/+ controls of the same experiment. Source data are available online for this figure.
Immunoblotting and immunohistochemical data jointly suggest an increased trafficking of DRD2 from the plasma membrane to the endolysosomal pathway, causing its downregulation. Nevertheless, reduced torsinA levels in DYT1 mice may cause DRD2 retention into the endoplasmic reticulum (ER), where torsinA resides (Cascalho et al, 2017) and may influence receptor protein maturation and export processes. However, our confocal microscopy analysis (Fig 6A, B 1 and B 2 ) seems to rule out this possibility, as we did not observe co-localization of DRD2 with the ER marker protein disulfide isomerase (PDI; Fig 6B 1 and B 2 ).

Discussion
GPCRs are central mediators of neurotransmission and are proven therapeutic targets for disease (Hauser et al, 2017). The mechanisms by which cells control GPCR trafficking and signaling are therefore topics of intensive study, and a large number of interacting pathways and partners have been uncovered. However, it is also clear that these do not yet represent the full picture of GPCR regulation. This also includes our relatively weak understanding of how the striatal D2DR is regulated, despite the fact that this receptor is central to movement control and its dysfunction is a key event in the pathogenesis of several neurological and psychiatric diseases. DYT1 dystonia is an incurable movement disorder, strongly associated with abnormal striatal dopaminergic responses, but its symptoms are nonresponsive to dopaminergic drugs. We show in vivo that: In wild-type striatum, changes in DRD2 and RGS9-2 levels are correlated during postnatal development, and RGS9-2 silencing causes DRD2 downregulation; in DYT1 striatum: (i) DRD2 downregulation is determined by an altered receptor stability and mediated by lysosomal degradation; (ii) accordingly, changes in endosomal and autophagy-lysosomal markers support an enhanced DRD2 trafficking through the degradative pathway; (iii) reduced spinophilin and RGS9-2 levels favor b-Arr2-mediated actions; (iv) hence, RGS9-2 upregulation rescues DRD2 level and function.
Our data demonstrate in wild-type striatum a parallel increase in DRD2 and RGS9-2 protein level during postnatal development, further pointing to their close functional relationship. Though in Tor1a +/À mice the two proteins undergo a similar increase during early postnatal development (P7-P21), however, both exhibit a simultaneous abrupt reduction at P60. Notably, the comparable reduction of DRD2 and RGS9-2 observed in the striatum of Tor1a +/À and Tor1a Dgag/+ mice strongly suggests that these alterations are caused by torsinA loss of function (Torres et al, 2004;Goodchild et al, 2005). Indeed, we found similarly reduced torsinA levels in P60 Tor1a +/À and Tor1a Dgag/+ striatum, in accordance with the notion that the Dgag mutation destabilizes torsinA protein (Giles et al, 2008).
Surprisingly, our data demonstrate that different mechanisms underlie the similar reduction of DRD2 and RGS9-2 levels. Wild-type DRD2 is a stable long-lived protein, whereas RGS9-2 possesses specific degradation determinants targeting the protein for constitutive lysosomal breakdown (Hara et al, 2006). In Tor1a +/À mice, we found a decreased DRD2 protein half-life, whereas the sensitivity of RGS9-2 protein to lysosomal degradation was unaffected. Indeed, our data show that the reduction of RGS9-2 level in mutant mice is attributable to an enhanced localization to the DRM, in line with the increased level of its specific membrane anchor R7BP. The observation that striatal levels of a different R7 RGS family member, RGS7, and of the RGS9-2 binding partner Gb5 are unaffected further rules out a generalized dysregulation of protein turnover in DYT1 striatum.
Functional or structural modifications may cause a selective reduction of DRD2 protein stability in Tor1a +/À mice. Integral membrane proteins with limited structural/conformational defects, such as an altered post-translational modification, may escape the ER and reach the plasma membrane. This seems to be the case of DRD2 in Tor1a +/À striatum, since our confocal analysis shows the absence of co-localization of DRD2 and PDI signals. Commonly, these proteins are in turn identified and targeted for lysosomal proteolysis by the quality control system (Tansky et al, 2007;Apaja et al, 2010). Interestingly, modifications at the N-terminal or C-terminal regions of DRD2, altering the level of post-translational modifications, extensively affect receptor internalization rate, plasma membrane expression, and protein stability (Cho et al, 2012;Ebersole et al, 2015). Alternatively, a failure in the developmental switch from the hypersensitive "juvenile" activity to the mature state of reduced sensitivity (Kim et al, 2002;McDougall et al, 2015) may cause an increased rate of desensitization of DRD2 through the endocytic-degradative pathway after P21. The increase in Rab4 protein level observed in Tor1a +/À striatum is indeed suggestive of an accelerated trafficking of the receptor through the constitutive recycling pathway, which acts as a quality control system, to redirect internalized receptors toward either the plasma membrane or the degradation pathways (Apaja et al, 2010;Li et al, 2012). Endosomal recycling depends on arrestin-dependent receptor endocytosis from the plasma membrane (Kim et al, 2001;Zheng et al, 2016). Though we did not observe changes in the level of b-Arr2 in Tor1a +/À striatum, we found a significant decrease in spinophilin, a scaffolding protein known to interact with DRD2 (Smith et al, 1999) and to antagonize b-Arr2-dependent signaling and trafficking of GPCRs (Wang et al, 2004). Thus, spinophilin downregulation would favor b-Arr2-dependent internalization and signaling of DRD2 in Tor1a +/À striatum. Further, the observation that RGS9-2 co-immunoprecipitates with b-Arr2 and inhibits DRD2 internalization in vitro indicates antagonistic actions between RGS9-2 and b-Arr2 as well (Celver et al, 2010;Zheng et al, 2011). Our in vivo data strongly support this notion. Indeed, we demonstrate that RGS9-2 silencing causes DRD2 downregulation in wild-type striatum, while restoration of striatal RGS9-2 level in Tor1a +/À and Tor1a Dgag/+ striatum rescues both protein expression and signaling of the receptor.
Thus, we hypothesize that a genetic modification resulting in torsinA loss of function might alter DRD2 maturation in the ER, causing disruption of the receptor signaling complex at the plasma membrane. The ensuing increased interaction of the receptor with b-Arr2 in turn leads to an increased DRD2 internalization rate, redirecting receptor either to a variety of alternative signaling pathways  or to degradation. In support of this view, uncoupling of DRD2 from its cognate Gproteins has been reported in a DYT1 mouse model (Napolitano et al, 2010). Our autoradiography data showing a 15% reduction of DRD2 binding density in Tor1a +/À striatal sections are in striking resemblance with DRD2 binding reduction reported in caudate and putamen of DYT1 mutation carriers (Asanuma et al, 2005) and support the hypothesis of an enhanced degradation of the receptor, in accordance with previous data from different DYT1 dystonia models (Yokoi et al, 2011;Dang et al, 2012). This hypothesis is further supported by our confocal microscopy analysis, ruling out DRD2 retention into the ER, and by immunoblotting experiments showing an increase in markers of the autophagy-lysosomal pathway, which operates constitutively at low rate to ensure quality control and turnover of long-lived proteins (Hara et al, 2006;Komatsu & Ichimura, 2010;Johansen & Lamark, 2011;Lilienbaum, 2013). In particular, we observed an increased lysosomal turnover of LC3-II (Tanida et al, 2005) in Tor1a +/À striatum. The observation that changes in basal LC3-II, ª 2018 The Authors EMBO Molecular Medicine 11: e9283 | 2019 p62, and LAMP-2 levels did not reach statistical significance rules out an overt dysregulation of autophagy and supports a more confined impairment involving targeting of DRD2 to degradation. In line with this view, the mCherry-GFP-LC3 reporter (Castillo et al, 2013) was increased both in the autolysosomes and in other compartments of ChIs (Matus et al, 2014). Our findings may explain why, despite a clear DRD2 involvement, dopaminergic drugs do not provide clinical benefit in DYT1 dystonia. Furthermore, they suggest that strategies targeting b-Arr2biased signaling or DRD2 interacting proteins, like RGS9-2, can effectively rescue striatal DRD2 function.
Alterations of DRD2-mediated dopaminergic transmission have been implicated in different neurologic and neuropsychiatric disorders. Interestingly, the clinical efficacy of antipsychotics and moodstabilizing drugs has been ascribed to blockade of b-Arr2-biased signaling of DRD2 (Peterson & Luttrell, 2017). Furthermore, a recent study demonstrated an ameliorating effect of b-Arr2 upregulation on levodopa-induced dyskinesia in pre-clinical animal models (Urs et al, 2015). On the other hand, manipulation of RGS9-2 levels in animal models of levodopa-induced or tardive dyskinesia was shown to affect the manifestation of these abnormal movements (Kovoor et al, 2005;Gold et al, 2007).
With such premises, b-Arr2 and RGS9-2, as well as DRD2-biased drugs, have been proposed as drug targets (Traynor et al, 2009;Peterson & Luttrell, 2017;Urs et al, 2015;Sjögren, 2017). Accordingly, RGS9 has now been added to the IUPHAR/BPS Guide to Pharmacology database of drug targets. Such novel therapeutic approaches hold great promise for an effective treatment of DYT1 dystonia.

Experimental design
Age-and sex-matched wild-type and mutant mice were randomly allocated to experimental groups and processed by a blinded investigator. Sample size for any measurement was based on the ARRIVE recommendations on refinement and reduction of animal use in research, as well as on our previous studies. For data analysis of WB, confocal and binding images, and electrophysiology experiments, the investigator was blinded to genotype/treatment. Each observation was obtained from an independent biological sample. The number of biological replicates is represented with N for number of animals and n for number of cells. For electrophysiology, each cell was recorded from a different brain slice.

Receptor autoradiography
The radioligand binding protocol has been modified from Fasano et al (2009). Briefly, fresh-frozen dissected mouse brains were sliced coronally at 16 lm on a cryostat at the level of the caudate-putamen (plate 15-36 of Franklin & Paxinos, 2008). Unfixed slides were preincubated in assay buffer for 60 min [50 mM Tris-Cl (pH 7.4), 120 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , and 40 nM ketanserin]. Sections were then incubated for 60 min at room temperature in the same buffer with the addition of 5 nM 3 H-spiperone (PerkinElmer, Boston, MA) and 40 nM ketanserin. Cold competition of tritiated spiperone with 10 lM cold spiperone was used to assess nonspecific signal. After ligand binding, slides were rinsed twice for 10 min in the same buffer, allowed to dry, and exposed to Hyperfilm MP film (Merck) for 4 weeks.

Image analysis and quantification
Quantification analyses were performed in blind, and sample identity was not revealed until correlations were completed. Densitometric analyses and ROD measurements were performed as previously described (Pratelli et al, 2017). Briefly, 10 sections per animal along the whole rostro-caudal extent of the striatum were used. Optical density (OD) was evaluated in the caudate-putamen, and background OD value was determined in structures of the same section devoid of specific signal and subtracted for correction to obtain the relative OD (ROD) value. Results were expressed as percentage increase/decrease in 3 H-spiperone density. Data were analyzed by Student's t-test.

Confocal imaging
Slices processing and confocal image acquisition were performed as previously described (Vanni et al, 2015;Ponterio et al, 2018). Mice were deeply anesthetized and perfused with cold 4% paraformaldehyde in 0.12 M phosphate buffer (pH 7.4). The striatum was dissected, post-fixed for at least 3 h at 4°C and equilibrated with 30% sucrose overnight. Coronal striatal sections (30 lm thick) were cut with a freezing microtome. Slices were dehydrated with serial alcohol dilutions (50-70-50%) and then incubated 1 h at RT in 10% donkey serum solution in PBS 0.25%-Triton X-100 (PBS-Tx). The following primary antibodies were utilized (3 days at 4°C; Appendix Table S2): goat anti-ChAT (1:500, NBPI30052; Novus Biologicals); rabbit anti-DRD2 (1:500, AB5084P; Millipore); goat anti-RGS9 (1:1,000, sc-8142; Santa Cruz), mouse anti-PDI (1:500, SPA891; StressGene); guinea pig anti-P62 (1:300, GP62-C; Progen). The following secondary antibodies were used (1:200, RT, 2 h): Alexa 488 and Alexa 647 (Invitrogen), and cyanine 3 (cy3)-conjugated secondary antibodies (Jackson ImmunoResearch). After washout, slices were mounted on plus polarized glass slides with Vectashield mounting medium (Super Frost Plus; Thermo Scientific) and coverslipped. Images were acquired with a LSM700 Zeiss confocal laser scanning microscope (Zeiss, Germany), with a 5×, a 20× objective, or a 63× oil immersion lens (1.4 numerical aperture) with an additional digital zoom factor (1×-1.5×-2×). Single-section images (1,024 × 1,024) or z-stack projections in the z-dimension (z-spacing, 1 lm) were collected. Z-stack images were acquired to analyze the whole neuronal soma, which spans multiple confocal planes. The confocal pinhole was kept at 1, the gain and the offset were adjusted to prevent saturation of the brightest signal and sequential scanning for each channel was performed. The confocal settings, including laser power, photomultiplier gain, and offset, were kept constant for each marker. For quantitative analysis, images were collected from at least 3-4 slices processed simultaneously from each striatum (n ≥ 3 mice/genotype) and exported for analysis with ImageJ software (NIH). Software background subtraction was utilized to reduce noise. To quantify the density of a specific marker on a defined area, we calculated the "overlapping signal" by utilizing a region of interest (ROI), as previously described (Vanni et al, 2015). When neurons were loaded with lucifer yellow (0.2%) through the patch pipette during the electrophysiological recordings, slices were fixed overnight with 4% paraformaldehyde in 0.12 M phosphate buffer (pH 7.4), washed in PBS, mounted with Vectashield mounting medium on plus polarized glass slides (Super Frost Plus Thermo Scientific), and coverslipped (Ponterio et al, 2018).

Viral constructs and preparations
For RGS9-2 silencing by lentiviral (LV) particles, two candidate shRNA sequences were designed to specifically silence mouse RGS9-2 gene (NCBI GenBank AF125046), according to the clone sequences provided by Sigma-Aldrich (clones shRNA_RGS9-2: 37134 e 37135). Each sequence was inserted into the miR30a backbone sequence (formed by the stem loop and by 100-bp downstream and upstream the premiR30a) to increase the expression efficiency of the shRNA (Denti et al, 2004). shRNA37134: TTATCTTTTCCACCCAAGCTTTGTAAAAATAAATC  AAAGAGAAAGCAAGGTATTGGTTTCAGCCAACAAGATAATTACTC  CCTTGAAGTTGGAGGCAGTAGGCACCCAAGTGCATTAGGATAATA  CCCATTTGTGGCTTCACAGTATTATCCTAATGCACTTGGGGTCGCT  CACTGTCAACGTTGATATGCCTTCTTCAGCATTCTGTCTTACTGAC  CTGAGAAGTGCTCTGCGGGAGTTTCTGAAATGTACAGGCAACATT  CTGTAAAC  shRNA37135: GTTTACAGAATGTTGCCTGTACATTTCAGAAACTCC  CGCAGAGCACTTCTCAGGTCAGTAAGACAGAATGCTGAAGAAGGC  ATATCAACGTTGACAGTGAGCGACCCGATTTCAGACGCCATATTTC  TGTGAAGCCACAAATGGGAAATATGGCGTCTGAAATCGGTGCCTA  CTGCCTCCAACTTCAAGGAGTAATTATCTTGTTGGCTGAAACCAAT  ACCTTGCTTTCTCTTTGATTTATTTTTACAAAGCTTGGGTGGAAAA  GATAA The sequences were cloned into the p207.pRRLsinPPTs.hCMV. GFP.WPREp207 plasmid (kindly provided by L. Naldini, Milan, Italy) under the U1 promoter for shRNA expression (p207 was modified as p207-U1: p207_U1 shRNA-34 and p207_U1 shRNA-35) upstream to the CMV promoter. Both plasmids carrying each sequence were used for a single LV preparation. After cloning, the VSV-G-pseudotyped lentiviral vectors were generated by calcium phosphate transfection of HEK293T cells with a mixture of the four plasmids required to produce third-generation lentiviruses (kindly provided by L. Naldini, San Raffaele Scientific Institute, Milan, Italy). The LV particles were prepared and purified according to previously published protocols (Mandolesi et al, 2009;Sciamanna et al, 2012b).
The herpes simplex viral (HSV) vector used for overexpressing mouse RGS9-2 (NCBI GenBank AF125046) has been previously described and validated (Rahman et al, 2003;Gold et al, 2007). The description of the control HSV-LacZ (b-galactosidase) vector is provided elsewhere (Carlezon et al, 1997). After cloning, HSV constructs were grown and purified according to previously published protocols (Coopersmith & Neve, 1999;Carlezon et al, 2000). To monitor in vivo autophagy flux in ChIs, we utilized an adenoassociated virus (AAV2/9) carrying a monomeric tandem mCherry-GFP-LC3 construct (AAV2/9-mCherry-GFP-LC3; kindly provided by C. Hetz) as a reporter (Castillo et al, 2013). The GFP fluorescent signal of the reporter is sensitive to acidic conditions; thus, co-localization of green and red fluorescence (yellow puncta) indicates that the tandem protein is not localized in compartments fused with a lysosome, while detection of red puncta indicates that the protein is located in the autolysosome (Matus et al, 2014).

Stereotactic injection of viral particles
Viral injection into the dorsal striatum was performed as previously described (Mandolesi et al, 2009;Fasano et al, 2010;Sciamanna et al, 2012b;Bourdenx et al, 2015;Urs et al, 2015). Briefly, male Tor1a +/À and Tor1a Dgag/+ mice were anesthetized with tiletamine/ zolazepam (80 mg/kg) and xylazine (10 mg/kg). Viral preparation suspensions were injected into the dorsal striatum, using a glass capillary (Sutter Instruments) connected to a picospritzer (Parker Inst, USA) as previously described (e.g., see Mandolesi et al, 2009;Fasano et al, 2010;Sciamanna et al, 2012b;Bourdenx et al, 2015;Urs et al, 2015). The following coordinates from Bregma were used for dorsal striatum: anteroposterior +0.4 mm; lateral AE2.5 mm; and ventral at À1.7 and À2 mm from the dura. Mice received 2 ll of viral preparation administered at 0.1 ml/min rate. At the end of the injection, the capillary was left in place for 5 min before being slowly removed. The skin was sutured, and mice were allowed to recover. In line with preliminary immunofluorescence experiments assessing the time-course and extent of viral infection, mice were sacrificed either 3-5 days after injection of HSV particles, or 3-4 weeks following LV or AAV particles.

Electrophysiology
Electrophysiological patch-clamp recordings were performed as previously described from individual ChIs in striatal coronal slices, prepared as previously described from P60 to P90 old mice (Sciamanna et al, 2011(Sciamanna et al, , 2012bPonterio et al, 2018). ChIs were visualized using standard IR-DIC microscopy and identified based on their large somatic size and distinctive electrophysiological properties. Electrophysiological signals were detected using Multiclamp 700B and AxoPatch 200 amplifiers (Molecular Devices), using borosilicate glass pipettes pulled on a P-97 Puller (Sutter Instruments). For cell-attached recordings, the electrodes were filled with a solution containing the following (in mM): 140.5 KMeSO 4 , 0.2 EGTA, 7.5 NaCl, 10 HEPES, 2 NaATP, and 0.2 NaGTP, adjusted to a pH of 7.3 with KOH. For perforated-patch recordings, gramicidin was added at a final concentration of 20 lg/ml and the perforation process was considered complete when both the amplitude of the action potentials and electrode resistance were steady. Action potential firing frequency was analyzed with Clampfit 10 (pClamp 10, Molecular Devices); neurons with frequencies outside the 0.5-4.5 range were excluded from further analysis.

Statistics
All data were obtained from at least two independent experiments. Sample size for any measurement was based on the ARRIVE recommendations on refinement and reduction of animal use in research, as well as on our previous studies. Age-and sex-matched wild-type and mutant mice were randomly allocated to experimental groups and processed by a blinded investigator. For data analysis of WB, confocal and binding images, and electrophysiology experiments, the investigator was blinded to genotype/treatment. Each observation was obtained from an independent biological sample. The number of biological replicates is represented with N for number of animals and n for number of cells. For electrophysiology, each cell was recorded from a different brain slice. Data analysis was performed with Clampfit 10 (pClamp 10), ImageJ (NIH), and Prism 5.3 (GraphPad). Data are reported as mean AE SEM. Statistical significance was evaluated as indicated in the text, with Pearson's r correlation test, one-way ANOVA with post hoc tests between groups corrected for multiple comparisons, and two-tailed two-sample t-test (parametric or nonparametric, unpaired, or paired) as appropriate according to each test assumptions. For example, normality tests were used to assess Gaussian distribution. F-test was used to compare variances between groups; when variance was different, Welch's correction was used. Statistical tests were two-tailed, the confidence interval was 95%, and the alpha-level used to determine significance was set at P < 0.05.

Study approval
Animal breeding and handling were performed in compliance with the ethical and safety rules and guidelines for the use of animals in biomedical research provided by the European Union's directives and Italian laws (2010/63EU, D.lgs. 26/2014; 86/609/CEE, D.Lgs 116/1992). The experimental procedures were approved by The paper explained Problem DYT1 dystonia is a progressive and highly disabling disease, with symptom onset frequently between childhood and adolescence. DYT1 causative mutation has been identified in the Tor1a gene. However, the molecular mechanisms underlying symptom manifestation are far from being clarified and medical treatments are unavailable. Dysfunction of dopamine D2 receptors in the striatum, a brain region involved primarily in motor control and motor learning, has been reported in DYT1 patients and rodent models. Therefore, restoring striatal D2 receptor function represents a potential therapeutic strategy.

Results
We investigated the molecular mechanisms of striatal dopamine D2 receptor dysfunction in multiple mouse models of DYT1 dystonia. We found that the striatal levels of D2 receptor, and of its regulatory proteins spinophilin and RGS9-2, are reduced. We show that D2 receptor downregulation is mediated by an abnormal, selective trafficking to lysosomal degradation. Furthermore, we present evidence that viral-induced expression of RGS9-2 is able to restore both the striatal level and the function of dopamine D2 receptor.

Impact
These results provide an explanation for the lack of effectiveness of dopaminergic drugs in DYT1 dystonia, despite a clear involvement of dopamine receptors in disease pathophysiology. More importantly, our work identifies therapeutic targets that are effective in rescuing the dopaminergic dysfunction in DYT1 dystonia.
Fondazione Santa Lucia Animal Care and Use Committee and the Italian Ministry of Health (authorization # 223/2017-PR).
Expanded View for this article is available online.