Neural ablation of the PARK10 candidate Plpp3 leads to dopaminergic transmission deficits without neurodegeneration

Parkinson’s disease (PD) is a multifactorial neurodegenerative disorder, characterised by the progressive loss of midbrain dopaminergic neurons and a variety of motor symptoms. The gene coding for the phospholipid phosphatase 3, PLPP3 (formerly PPAP2B or LPP3), maps within the PARK10 locus, a region that has been linked with increased risk to late-onset PD. PLPP3 modulates the levels of a range of bioactive lipids controlling fundamental cellular processes within the central nervous system. Here we show that PLPP3 is enriched in astroglial cells of the adult murine ventral midbrain. Conditional inactivation of Plpp3 using a Nestin::Cre driver results in reduced mesencephalic levels of sphingosine-1-phosphate receptor 1 (S1P1), a well-known mediator of pro-survival responses. Yet, adult PLPP3-deficient mice exhibited no alterations in the number of dopaminergic neurons or in the basal levels of striatal extracellular dopamine (DA). Potassium-evoked DA overflow in the striatum, however, was significantly decreased in mutant mice. Locomotor evaluation revealed that, although PLPP3-deficient mice exhibit motor impairment, this is not progressive or responsive to acute L-DOPA therapy. These findings suggest that disruption of Plpp3 during early neural development leads to dopaminergic transmission deficits in the absence of nigrostriatal degeneration, and without causing an age-related locomotor decline consistent with PD.

We previously showed that mice lacking PLPP3 in the neural lineage exhibit impaired motor performance from one month of age. Abnormalities in cerebellar foliation in mutant animals suggested that structural or electrophysiological alterations of the cerebellum, the brain structure that regulates skilled movement, may be causal to these defects 19 . Here we investigated whether the motor deficit of PLPP3-deficient mice may be related to parkinsonism. We found that, although PLPP3 is expressed in the ventral midbrain, Nestin::Cre mediated inactivation of Plpp3 does not substantially affect the development and maintenance of the nigrostriatal pathway. Additionally, we show that the motor phenotype of PLPP3-deficient mice is not progressive or responsive to acute L-DOPA therapy. However, potassium-evoked dopamine (DA) overflow in the striatum is markedly reduced in mice lacking PLPP3. Collectively, these data suggest that disruption of Plpp3 during embryonic central nervous system (CNS) development does not cause ventral midbrain DA neuron loss or locomotor impairments consistent with PD, but results in dopaminergic transmission deficits that might contribute to the pathology under specific contexts.

Results
Plpp3 is expressed in the ventral midbrain. PARK10 spans up to a 9.5 megabase region on chromosome 1p32 (50700000-62300000). Alterations in one or more genes within this locus may confer susceptibility to late-onset PD 4,5 . To investigate whether PLPP3 (NC_000001.11: 56494747-56579584) may be implicated in PD pathogenesis, we analysed its expression in the mouse midbrain. Using a Plpp3 lacZ reporter allele and immunostaining, we previously showed that both β -galactosidase reporter activity and PLPP3 protein are detected in the developing midbrain between embryonic days (E) 10.5 and 12.5 20 , when the great majority of DA neurons of the SN arise 21 . X-gal staining on brain sections from juvenile (4-week-old) and adult (12-week-old) Plpp3 lacZ mice revealed moderate but well-defined β -galactosidase expression along the nigrostriatal pathway, from the SN to the caudate putamen (CPu) (Fig. 1a-d). Additionally, strong reporter expression was seen in the cerebellum, olfactory bulb and ventricular system, as previously reported 19 , as well as in the hippocampal dentate gyrus and olfactory tubercle (Fig. 1a,c). Immunostaining for PLPP3, using a previously characterised antibody 19,22 , showed clear staining in the adult SN ( Fig. 1e and Fig. S1), demonstrating that during adulthood, protein distribution in the ventral midbrain correlates with reporter activity.
Transcriptional profiling analyses of astroglial, neuronal and oligodendroglial cell populations acutely purified from the postnatal and adult mouse forebrain indicate that Plpp3 transcripts are significantly enriched in astrocytes in comparison to neurons and oligodendrocytes 23,24 . In the cerebellum, expression of PLPP3 protein is restricted to a specialised type of GLAST + and S100β + glial cell, the Bergmann glia, and is absent from Purkinje cells, interneurons and granule cells 19 . To determine the identity of the cells expressing PLPP3 in the adult ventral midbrain, we performed immunofluorescence analyses for PLPP3 in combination with antibodies detecting either tyrosine hydroxylase (TH), the rate-limiting enzyme in the synthesis of dopamine, or the glutamate-aspartate transporter GLAST, which is enriched at the plasma membrane of astrocytes. We found that in the SN and ventral tegmental area (VTA), PLPP3 shows a similar expression pattern to that of GLAST, and while being in close contact with TH + neurons, no obvious immunoreactivity was detected in these cells (Fig. 1f,g). This suggests that PLPP3 might modulate paracrine signals between astroglia and neurons by regulating the extracellular availability of bioactive lipid phosphates, such as S1P.

Development of midbrain DA neurons in the absence of PLPP3. To investigate whether PLPP3
is required for the development of midbrain DA neurons, we used a mouse line bearing floxed Plpp3 alleles (Plpp3 F/F ) 25 to conditionally deplete Plpp3 in CNS progenitor cells. Since Plpp3 is expressed in the prospective midbrain 20 , its ablation at early embryonic stages could potentially affect the proper specification of this structure. Therefore, we inactivated Plpp3 using a Nestin::Cre transgenic mouse line 26 that exhibits Cre recombinase activity in the ventral midbrain from E10.5 27 . Efficient excision of Plpp3 in the CNS of Plpp3 F/F ; Nestin::Cre mice (hereafter referred to as Plpp3 Δ/Δ ) has been previously demonstrated 19 and was confirmed by PCR of genomic DNA, western blot and immunostaining (Fig. 1e, Figs S1 and S2). Expression of TH in E14.5 Plpp3 Δ/Δ and control Plpp3 F/F mouse brains was analysed by immunostaining of whole-mount preparations and tissue sections across the mesencephalon. There were no substantial differences throughout the TH + domains between genotypes (Fig. 2). However, we found that in the PLPP3-deficient brains the third ventricle was enlarged when compared with non-excised controls (Fig. 2b).
Lack of PLPP3 in the adult nigrostriatal pathway alters S1P/S1P 1 signalling without affecting DA neuron survival. S1P/S1P 1 signalling has been shown to exert a neuroprotective effect on DA neurons 16,18 and aberrant sphingolipid metabolism has been implicated in distinct neurodegenerative disorders 28,29 . Moreover, S1PR1 was recently proposed as a candidate gene for a newly identified PD susceptibility locus 30 , suggesting a potential role for S1P 1 signalling in regulating midbrain DA neuron survival and/or function. We previously showed that depletion of PLPP3 in the neural lineage leads to downregulation of S1P 1 in the cerebellum 19 . Immunofluorescence analysis of brain sections from 16-week-old Plpp3 Δ/Δ and control mice revealed that neural deficiency of PLPP3 also causes a pronounced decrease of S1P 1 levels in the ventral midbrain ( Fig. 3a and Fig. S1). To determine if the absence of PLPP3 and the concomitant reduction of mesencephalic S1P 1 levels affect the integrity of the adult nigrostriatal pathway, we performed TH immunostaining. We found no obvious differences in the distribution of ventral midbrain DA neurons between Plpp3 Δ/Δ and non-excised animals (Fig. 3b).
Stereological quantitation of TH + cell bodies in the SN and VTA showed comparable numbers of DA neurons between genotypes (Fig. 3d). Moreover, densitometric analyses of TH + innervation in the striatum revealed no significant differences among genotypes (Fig. 3c,e). Thus, despite displaying a motor deficit from one month of age 19 , the integrity of the nigrostriatal pathway in Plpp3 Δ/Δ mice seems unaffected at least up to 16 weeks of age.
Since loss of midbrain DA neurons in patients with PD occurs progressively, we measured expression levels of TH in the ventral midbrain of 10-to 12-month-old Plpp3 Δ/Δ and Plpp3 F/F mice. Consistent with our observations in 16-week-old PLPP3-deficient mice, western blot analyses showed that at 10-to 12-months of age, Plpp3 Δ/Δ mice exhibit a considerable reduction of S1P 1 levels in the ventral midbrain, but no differences in TH expression when compared to age-matched controls (Fig. 4). These data suggest that, despite leading to S1P 1 dysregulation in the ventral midbrain, PLPP3 deficiency does not cause a substantial loss of TH + cells in this region.
Unchanged levels of astrocyte reactivity and Fluoro-Jade + cells in the SN of adult Plpp3 Δ/Δ mice. Astrocytes become reactive in response to different CNS insults, including pathological conditions and neurodegenerative disorders. Although this cell state -characterised by hypertrophy and upregulation of glial fibrillary acidic protein (GFAP)-is considered a general feature of neuroinflammation, the role of reactive astrocytes in PD is still debatable 31 . To determine whether neural depletion of Plpp3 may lead to astrocytic reactivity, we analysed GFAP expression in the ventral midbrain of 16-week-old Plpp3 Δ/Δ and control mice by immunofluorescence (Fig. 5a). There were no significant differences in the density of GFAP immunoreactivity between genotypes (Fig. 5b). However, we found that in mice with neural deficiency of PLPP3, there seemed to be accumulation of GFAP + astrocytes around the blood vessels (Fig. 5c).
To confirm the lack of midbrain dopaminergic neurodegeneration in adult PLPP3-deficient mice, we stained tissue sections from 20-month-old animals with Fluoro-Jade to label all possible degenerating cells 32 . No substantial amount of Fluoro-Jade + cells was detected in the ventral midbrain of either control or mutant mice (Fig. 5d). Accordingly, there were no considerable differences in the TH + domains at this age between genotypes ( Fig. 5e and Fig. S3). Together with our previous results, these data suggest that CNS-depletion of PLPP3 does not lead to ventral midbrain DA neuron loss, even in aged individuals.
Non-progressive locomotor impairment in PLPP3-deficient mice. Using the rotarod and the raised beam tests we previously found that mice with neural deficiency of PLPP3 exhibit a motor dysfunction when compared with age-matched Plpp3 F/F mice. Such deficits were apparent at one month of age and showed no signs of worsening up to 12 months of age 19 . To investigate whether this motor phenotype could be related to parkinsonism, we evaluated the basal activity of 6-, 12-and 18-month-old Plpp3 Δ/Δ mice. At 6 months of age, Plpp3 Δ/Δ animals displayed decreased total locomotion, rearing and stereotypic movements (movement without displacement) when compared with age-matched controls (Fig. 6a,c,d and Fig. S4). These motor deficits became less evident at 12 and 18 months of age, as all animals exhibited an age-related reduction of spontaneous activity ( Fig. 6e,g,h and Fig. S5). Yet, a significant impairment of vertical activity was still detected in 12-and 18-month-old mutant mice during distinct phases of the test session (Fig. S5). Collectively, these data indicate that the motor deficit of Plpp3 Δ/Δ mutant mice occurs in the absence of midbrain dopaminergic neurodegeneration and is not progressive, as would be expected for a parkinsonian phenotype. Deficits in dopaminergic transmission in Plpp3-ablated mice. Studies with transgenic mouse models of PD suggest that DA neuron dysfunction can occur prior to 33 or in the absence of neurodegeneration [34][35][36] .
To determine whether the motor symptoms of Plpp3 Δ/Δ mice may result from physiological alterations in the nigrostriatal pathway, we first analysed their acute locomotor response to L-DOPA. We found that L-DOPA administration had no effect on the motor performance of Plpp3 Δ/Δ mice (Fig. 7).
Next, we investigated whether neural depletion of PLPP3 may affect dopaminergic transmission. We used in vivo microdialysis and capillary electrophoresis to assess basal and high potassium-evoked DA overflow in the striatum of 14 to 17-month-old Plpp3 F/F and Plpp3 Δ/Δ mice. Whereas there were no differences in the basal levels of striatal extracellular DA between genotypes (Fig. 8a), we found that potassium-evoked DA release was significantly reduced in the striatum of PLPP3-deficient mice (Fig. 8b).

Discussion
Several studies have found an association between the PARK10 locus and an increased risk for late-onset PD [4][5][6][7]37 . However, the biological mechanisms underlying such susceptibility have not been established. Although genetic studies have nominated a set of candidate genes for the PARK10 locus 7,8,[38][39][40] , some of these findings have been controversial [41][42][43] , and until now, further characterisation of these genes in the context of PD had been limited to expression analyses 9,44 .
We have performed a functional evaluation of Plpp3 as a candidate gene for the PARK10 locus. SAGE profiling of the SN of patients with PD and age-matched controls previously revealed decreased PLPP3 transcripts in PD tissues 9 . Subsequent work with PD model systems implicated S1P, one of the substrates of PLPP3, in promoting survival of midbrain DA neurons via S1P 1 activation 16,18 . Our studies show that Plpp3 is expressed in astrocytes throughout the murine nigrostriatal pathway and that its inactivation in CNS progenitor cells causes a severe downregulation of S1P 1 in the adult ventral midbrain. Yet, we found that PLPP3-deficiency, and the accompanying reduction of S1P 1 levels, have no significant effect on the number of midbrain DA neurons. Neither do they induce a progressive age-related motor phenotype consistent with PD. This suggests that the changes in PLPP3 levels detected in patients with PD 9 , and the dysregulation of S1P/S1P 1 observed in PD murine models 16 , could be a consequence rather than a cause of neuronal loss.
Alternatively, given that PD may result from a complex combination of environmental and genetic factors, the alterations described in this work might increase the vulnerability of midbrain dopaminergic neurons under particular physiological or pathological conditions, so that their effects would only become apparent in specific experimental settings. For instance, dysfunction in the blood-brain barrier (BBB) inevitably leads to inflammatory changes, which in turn contribute to the process of neurodegeneration 45 . In this regard, it is worth noting that although no signs of astrocyte reactivity were detected in the SN of PLPP3-deficient animals, astrocytes seemed to accumulate around blood vessels, being endothelial cells the only cells within the brain that still express PLPP3, and therefore where S1P 1 was not dysregulated (Fig. S1). This is especially relevant in the light of the recently described role of S1P/S1P 1 signalling in reinforcing the integrity and function of the BBB by acting on both endothelial cells and astrocytes 46 . We previously showed that targeted ablation of Plpp3 in endothelial and haematopoietic cells leads to heightened vascular inflammation and permeability 47 . Since in this work we only ablated Plpp3 in neural tissue, we cannot rule out the possibility that inflammation and increased vascular permeability produced by reduction of PLPP3 levels in endothelial cells could be factors contributing to PD development. In this sense, it would be interesting to explore whether individuals with the PLPP3 polymorphism associated with increased risk to coronary artery disease 48 , which is susceptible to epigenetic regulation 49 , may also have an increased risk of developing PD.
Despite the lack of any clear signs of neurodegeneration in the ventral midbrain, we found that high potassium-evoked striatal dopamine overflow was significantly decreased in adult mice lacking PLPP3, suggesting alterations in DA release and/or reuptake in the presynaptic nigral terminals. This is particularly interesting since S1P, as well as LPA, another PLPP3 substrate, have been proposed to regulate neurotransmission at different levels [11][12][13] . In cultured rat hippocampal neurons, for example, S1P has been shown to promote glutamate secretion both in a depolarisation-independent manner and by potentiating depolarisation-evoked release 50 . LPA, on the other hand, has been found to reduce glutamate release by decreasing the number of synaptic vesicles at active zones in rat presynaptic hypoglossal motoneurons, and to reduce GABA signalling by reducing the number of receptors at postsynaptic inhibitory synapses 51 . It would be important to explore whether PLPP3 substrates may exert some effect on DA transport, vesicular release from nigral terminals or on D2 autoreceptors, which modulate the balance between DA release and reuptake.
We previously found that neural inactivation of Plpp3 leads to abnormalities in cerebellar foliation 19 . Since the cerebellum coordinates sensory and motor functions, it was proposed that the motor phenotype observed in Plpp3 Δ/Δ mice was likely due to defects in cerebellar architecture and/or physiology. However, considering our new findings regarding the deficits in stimulus-evoked striatal DA overflow in PLPP3 mice, at this point we cannot discard a potential contribution of altered dopaminergic transmission to motor impairment. Indeed, it may be possible that the locomotor phenotype detected in this model arises from complex defects in the modulation of neurotransmission in several brain structures, like in the striatum. Further investigation will be required to establish how PLPP3 contributes to this process.
Scientific RepoRts | 6:24028 | DOI: 10.1038/srep24028 Methods Animals. Studies involving animal use were approved by the Committee of Animal Use and Care of the Instituto de Fisiología Celular and conformed Mexican guidelines (protocols DEA38-14 and LMT01-14). Mice were grouped-housed in an environment with controlled 12 h light/dark cycles and access to food and water ad libitum. For Plpp3 expression analyses, we used CD1 mice heterozygous for a reporter allele of Plpp3 (Plpp3 lacZ ) 20 . To deplete PLPP3 in the neural lineage, female 129 mice homozygous for a conditional null allele of Plpp3 (Plpp3 F/F ) 25 were bred to Plpp3 F/F males expressing Cre recombinase under the control of the Nestin promoter (Tg(Nes-Cre)1Kln) 26 in a 129 background. The day of vaginal plug was considered E0.5. All experimental animals were genotyped by PCR analysis of genomic DNA as described 25 . In all analyses, non-excised littermates or agematched Plpp3 F/F mice were used as controls.
X-gal staining. Adult mice were anaesthetized by intraperitoneal (i.p.) administration of Avertin   Whole-mount immunohistochemistry. Methods for whole-mount antibody staining were adapted from previously established protocols 52 . Dissected brains from E14.5 embryos were fixed in phosphate buffered 4% PFA overnight at 4 °C, rinsed in PBS and sequentially dehydrated in methanol. After 6 h incubation in methanol/ DMSO/H 2 O 2 (4:1:1) at RT, tissues were washed twice with 100% methanol, rehydrated in descending concentrations of methanol/PBS and rinsed twice with PBS. Samples were incubated in an ice-cold blocking solution containing 2% skim milk and 0.5% Triton X-100 in PBS (PBSTM) for 2.5 h on a rocker. Primary polyclonal rabbit anti-TH antibody (1:250, Millipore AB152) was applied in PBSTM overnight at 4 °C with gentle rocking. Following five 1 h washes with PBSTM and constant rocking, tissues were incubated with HRP-conjugated anti-rabbit secondary antibody (1:500, Santa Cruz sc-2030) in PBSTM overnight at 4 °C on a rocker. After five 1 h washes with PBSTM and two PBS rinses, tissues were incubated with 3-3′ -diaminobenzidine/NiCl 2 (DAB Kit, Vector) for 30 min in the dark, before adding H 2 O 2 . Stained brains were rinsed with dH 2 O, post-fixed in 4% PFA, rinsed with PBS, dehydrated in ascending methanol concentrations, cleared in benzyl alcohol/benzyl benzoate (1:1) and photographed under a stereomicroscope (Nikon SMZ800).
Fluoro-Jade B (FJB) staining was performed as previously described 53 . One drop of a 1% NaOH solution diluted in 80% ethanol was added to brain sections adhered onto slides and replaced 2 min later by 70% ethanol. Sections were incubated in 0.06% potassium permanganate for 10 min, washed, and incubated 20 min with 0.0004% FJB solution in 0.1% acetic acid. Sections were washed, dried at 50 °C, cleared in xylene for 5 min and mounted in Permount.
Confocal images were acquired using a FV10i (Olympus), a FV1000 (Olympus) or a LSM710 NLO microscope (Zeiss), and processed using ImageJ (NIH, USA) and Photoshop (Adobe). For mosaic assembly, we used ImageJ Stitching plugin 54 . Cell counts and measurement of striatal TH + fibre and midbrain GFAP + astrocyte densities. week-old animals, the number of TH + somata in the entire extent of the SN-VTA was counted by unbiased stereology according to the optical fractionator principle, using an Olympus BX51WI spinning disc confocal microscope and the StereoInvestigator software (MicroBrightField Biosciences). Number of TH + somata in Plpp3 Δ/Δ samples is given as a percentage of the controls. Density of dorsolateral striatal TH + fibres was measured at two different coronal levels in each hemisphere using ImageJ. Mean grey values obtained from the striatum were normalised using corresponding measurements from the adjacent cortex. Assessment of GFAP + immunofluorescence intensity in the SN was performed using ImageJ, as previously reported 55 . Measurements were done on confocal images of each hemisphere acquired at two different coronal levels, and corrected for not specific background staining. Average values for each mouse were used for statistical analyses.
In vivo microdialysis and capillary electrophoresis. Microdialysis experiments were performed following the methodology reported by Camacho and colleagues 57 , adapted to the mouse. Adult mice were anaesthetized with 1.5% isoflurane and maintained under low (0.8%) anaesthesia throughout the procedure. Animals were fixed on a stereotaxic frame for unilateral implantation of a 1 mm long and 0.5 mm diameter microdialysis probe (CMA 12, CMA Microdialysis) in the dorsal striatum at the coordinates: AP 0.6 mm, ML 1.8 mm and DV 3.2 mm, relative to bregma and the skull surface and according to Paxinos  /3.3 mM glucose) and three fractions were collected in these conditions. Solution was switched back to KRB medium before collecting the last four fractions. Dialysates were collected in tubes containing 1 μL of antioxidant mixture (0.25 mM ascorbic acid/0.27 mM Na 2 EDTA/0.1 M acetic acid) and immediately frozen at − 80 °C.
Dialysate DA levels were determined by micellar capillary electrophoresis 59 . For derivatisation, 6 μL of 2.5 mg/mL 3-(2-furoyl)quinoline-2-carboxaldehyde (Molecular Probes, Eugene, OR, USA) were added to each striatal microdialysate sample in the presence of 2 μL of KCN 25 mM/10 mM borate buffer pH 9.2 and 1 μL of 0.075 mM O-methyl-L-threonine as internal standard. The derivatisation reaction was carried out in the dark at 65 °C for 15 min. Separation and analysis were conducted on a capillary electrophoresis system (Beckman-Coulter P/ACE MDQ Glycoprotein System, Beckman Coulter, Brea, CA, USA) with laser induced fluorescence detection. An argon ion laser was used to excite the 3-(2-furoyl)quinoline-2-carboxaldehyde-labeled analytes with light at 488 nm. The separation of compounds was based on a micellar electrokinetic chromatography buffer system that included 35 mM borates/25 mM sodium dodecyl sulphate/13% (vol/vol) methanol HPLC grade, pH 9.6. The dialysate samples were injected hydrodynamically at 0.5 psi for 5 seconds in a 75 μm i.d. capillary (Beckman Coulter); the separation was performed at 25 kV. After each sample separation, the capillary was flushed with 0.1 M NaOH, H 2 O and running buffer. In order to identify dopamine, migration electropherograms were matched with a spiked sample. Samples were corrected by relating the area under the curve of the unknown sample with the area under the curve of the internal standard. Data were analysed using Karat System Gold (Beckman Coulter). Basal dialysates DA levels are given as percentage of controls. Results for all the perfusion period are expressed as percentage of baseline concentration (analyte concentration × 100/mean of the three first fractions).
Locomotor tests. Spontaneous locomotor activity was evaluated in an automatic activity monitor (Med Associates Inc.) during the light phase of the light/dark cycle. Mice were individually placed in the centre of a 43 × 43 cm open-field arena and distance travelled, stereotypic movements, rearing and time spent in the centre of the arena (45% of total surface) were measured in sequential 5 min intervals during a 30 min test period. The arena was thoroughly cleaned with 10% isopropanol after each test session. For the analysis of motor phenotype progression, each animal was tested for 3 consecutive days at 6, 12 and 18 months of age. Mean values of each three-day trial were used for statistical analyses. For assessment of time spent in the centre of the arena, only values of the first test session at 6 months of age were used.
To evaluate response to L-DOPA, 7-old-month mice were tested for 4 consecutive days. On the first 2 days, each animal was given a saline injection and their activity was subsequently monitored for 30 min as indicated above. On the next 2 days, mice were treated with L-DOPA combined with the peripheral DOPA-decarboxylase inhibitor benserazide (20 mg/kg; Sigma) prior to locomotor evaluation. All locomotor tests were carried out with male mice.
Statistical Analyses. For TH + neuron counts, densitometric measurements of TH + striatal fibres and GFAP immunoreactivity, immunoblot densitometry and basal extracellular dopamine levels within striatum, means comparisons were done using two-tailed t tests. For spontaneous locomotor activity, data for each locomotor parameter were analysed within each age group using a repeated measures (time intervals) analysis of variance (ANOVA), followed by Bonferroni post hoc tests (when there was a significant time/genotype interaction) or the Fisher's least significant difference test (when there was no time/genotype interaction). Two-tailed t tests were used to evaluate differences in average total distance, mean speed, total rearing and total stereotypic movements between genotypes within a given age group. The effects of L-DOPA treatment on each locomotor parameter in both control and Plpp3 Δ/Δ mice were assessed using a two-way ANOVA, with genotype and treatment as the two factors. Genotype effects on changes in striatal dopamine overflow induced by high potassium were analysed using a repeated measures (dialysates) ANOVA, followed by the Fisher's least significant difference test. All analyses were performed with GraphPad Prism (GraphPad Software). The level of significance was p ≤ 0.05.