Striatal tyrosine hydroxylase-positive neurons are associated with l -DOPA-induced dyskinesia in hemiparkinsonian mice

Neuroscience 298 (2015) 302–317

STRIATAL TYROSINE HYDROXYLASE-POSITIVE NEURONS ARE ASSOCIATED WITH L-DOPA-INDUCED DYSKINESIA IN HEMIPARKINSONIAN MICE U. KEBER, a M. KLIETZ, a,b T. CARLSSON, a,c W. H. OERTEL, a E. WEIHE, b M. K.-H. SCHA¨FER, b G. U. HO¨GLINGER a,d,e AND C. DEPBOYLU a*

contrast, identified accumbal and cortical TH+ cells did not contribute to the generation of LID. Thus, striatal TH+ cells and serotonergic terminals may cooperatively synthesize DA and subsequently contribute to supraphysiological synaptic DA concentrations, an accepted cause in LID pathogenesis. Ó 2015 IBRO. Published by Elsevier Ltd. All rights reserved.

a Experimental Neurology, Department of Neurology, Philipps University Marburg, Marburg, Germany b Department of Molecular Neuroscience, Institute of Anatomy and Cell Biology, Philipps University Marburg, Marburg, Germany c

Section of Pharmacology, Institute for Neuroscience and Physiology, The Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden  d

German Center for Neurodegenerative Diseases (DZNE), Munich, Germany  e

Key words: Parkinson’s disease, dopamine, involuntary movement, striatum, accumbens, TH.

abnormal

Department of Neurology, Technical University, Munich, Germany

Abstract—L-3,4-Dihydroxyphenylalanine (L-DOPA) is the therapeutic gold standard in Parkinson’s disease. However, long-term treatment is complicated by the induction of debilitating abnormal involuntary movements termed L-DOPA-induced dyskinesias (LIDs). Until today the underlying mechanisms of LID pathogenesis are not fully understood. The aim of this study was to reveal new factors, which may be involved in the induction of LID. We have focused on the expression of striatal tyrosine hydroxylasepositive (TH+) neurons, which are capable of producing either L-DOPA or dopamine (DA) in target areas of ventral midbrain DAergic neurons. To address this issue, a daily L-DOPA dose was administered over the course of 15 days to mice with unilateral 6-hydroxydopamine-induced lesions of the medial forebrain bundle and LIDs were evaluated. Remarkably, the number of striatal TH+ neurons strongly correlated with both induction and severity of LID as well as DFosB expression as an established molecular marker for LID. Furthermore, dyskinetic mice showed a marked augmentation of serotonergic fiber innervation in the striatum, enabling the decarboxylation of L-DOPA to DA. Axial, limb and orolingual dyskinesias were predominantly associated with TH+ neurons in the lateral striatum, whereas medially located TH+ neurons triggered locomotive rotations. In

INTRODUCTION Parkinson’s disease (PD) is a progressive neurodegenerative disorder and clinically characterized by akinesia, rigidity and resting tremor. These motor symptoms are related to reduced dopamine (DA) levels in the striatum (Bernheimer et al., 1973), resulting from a progressive loss of terminals of degenerating neuromelanin-containing DAergic neurons in the substantia nigra pars compacta (SNpc) (Hirsch et al., 1988). Since its discovery in the 1960s, DA replacement therapy with L-3,4dihydroxyphenylalanine (L-DOPA) has remained the gold standard symptomatic treatment for PD (Birkmayer and Hornykiewics, 1961; Smith et al., 2012). Unfortunately, its chronic administration leads to debilitating abnormal involuntary movements (AIMs) termed L-DOPA-induced dyskinesia (LID) that occur in nearly 90% of PD patients within a decade (Ahlskog and Muenter, 2001). Several pre- and postsynaptic mechanisms are supposed to contribute to LID development and expression. As a pivotal factor in pathogenesis, fluctuations in central DA levels cause aberrant plasticity in DAergically innervated brain structures, with the striatum playing a crucial role among them (de la Fuente-Ferna´ndez et al., 2004; Lindgren et al., 2010; Feyder et al., 2011; Huang et al., 2011). Interestingly, recent evidence has implicated raphe-striatal serotonergic neurons as an additional source of striatal DA following peripheral administration of L-DOPA, as they contain the requisite transport and enzymatic machinery to take up, convert and release L-DOPA-derived DA (Arai et al., 1995; Tanaka et al., 1999; Maeda et al., 2005). Unfortunately, these neurons lack DA transporters and D2 autoreceptors and are thus incapable of regulating

*Corresponding author. Address: Department of Neurology, Philipps University Marburg, Baldingerstrasse, 35043 Marburg, Germany. Tel: +49-6421-5866428; fax: +49-6421-5865384. E-mail address: [email protected] (C. Depboylu).   Present address. Abbreviations: 6-OHDA, 6-hydroxydopamine; AADC, aromatic acid decarboxylase; AIMs, abnormal involuntary movements; DA, dopamine; L-DOPA, L-3,4-dihydroxyphenylalanine; LID, L-DOPAinduced dyskinesia; MFB, medial forebrain bundle; NHS, normal horse serum; OD, optical density; PD, Parkinson’s disease; SNpc, substantia nigra pars compacta; TH, tyrosine hydroxylase; TH+, THpositive; DFosB, delta isoform of oncogene FosB. http://dx.doi.org/10.1016/j.neuroscience.2015.04.021 0306-4522/Ó 2015 IBRO. Published by Elsevier Ltd. All rights reserved. 302

U. Keber et al. / Neuroscience 298 (2015) 302–317

the exaggerated DA efflux that triggers LID (Carta et al., 2007; Eskow et al., 2009; Navailles et al., 2010). In addition to the ‘‘classical’’ catecholaminergic brainstem nuclei synthesizing DA, norepinephrine or epinephrine as neurotransmitters, neurons partially expressing individual enzymes of DA biosynthesis, namely tyrosine hydroxylase (TH) and/or aromatic acid decarboxylase (AADC), are found in different areas of the central nervous system (Weihe et al., 2006; Ugrumov, 2013). TH-positive (TH+) neurons exist widely spread throughout the brain and are reactive to DAergic perturbations. Following DAergic denervation in PD patients (Porritt et al., 2000) and in corresponding animal models (Tashiro et al., 1989; Betarbet et al., 1997; Meredith et al., 1999; Lopez-Real et al., 2003) there was a compensatory increase in TH+ neurons in the striatum. This increment does not result from neurogenesis of TH+ cells but from a phenotypic shift of pre-existing GABAergic intrastriatal neurons (Tande´ et al., 2006; Darmopil et al., 2008). Additionally, stereotactic injection of 6-hydroxydopamine (6-OHDA) into the medial forebrain bundle (MFB) of mice lead to an increment of TH+ cell numbers and fiber sprouting in target areas (Depboylu, 2014; Depboylu et al., 2014). Thus, they might provide an additional source of L-DOPA and possible DA in these areas which may aggravate dyskinesia. Here, we aim to elucidate a possible functional involvement of these striatal, accumbal and cortical TH+ neurons in LID pathogenesis in mice.

EXPERIMENTAL PROCEDURES Animals 50 male wildtype C57Bl/6 mice (10–12 weeks old, weighing 25–30 g) were obtained from Charles River (Sulzfeld, Germany) and housed at the animal facility of the Biomedical Research Centre at Philipps University Marburg under standardized conditions with 12:12-h light/dark cycle, room temperature 23 ± 1 °C and ad libitum access to food and water. All animal experiments were performed according to the EU Council Directive 2010/63/EU and approved by the local animal care committee (Regierungspra¨sidium Giessen, Germany). Experimental design The experimental design is outlined in Fig. 1. Initially, all animals received a unilateral 6-OHDA lesion of the MFB in order to achieve a complete depletion of the nigrostriatal DAergic pathway. After three weeks the mice were screened using cylinder test and amphetamine-induced rotation test and the animals with the severest lesion were selected for the experiment. Inclusion criteria were a forelimb use asymmetry showing <20% left forelimb touches of total, as well as amphetamine-induced exhibition of >3.5 net full body turns per minute ipsilateral to the lesioned side. Six weeks post-lesion animals were treated daily with LDOPA in combination with the DOPA decarboxylase inhibitor benserazide hydrochloride for 15 days to induce

303

stable AIMs. Mice receiving daily injections of saline served as a control group. AIMs were evaluated at days 1, 7 and 15 of the treatment period. 48 h after the last L-DOPA dose animals were sacrificed and the brains were processed for immunohistochemistry. 6-OHDA lesion Stereotactic injections were conducted under general anesthesia induced by intraperitoneal (i.p.) injection of 8 ml/kg ketamine (2%; Pfizer, Karlsruhe, Germany) with 2 ml/kg xylazine (Bayer Healthcare, Leverkusen, Germany). At 30 min before cerebral injections of 6-OHDA, the mice received an i.p. injection of 25 mg/kg desipramine (Sigma-Aldrich, Munich, Germany) to protect norepinephrinergic neurons and fibers. To induce a severe DAergic lesion, 2-ll 6-OHDA (SigmaAldrich; 2 lg/ll in 0.9% NaCl with 0.2% ascorbic acid) was injected into the MFB at following coordinates: anterior–posterior 0.82 mm and medio-lateral 1 mm from bregma, 4.8 mm ventral to the dura. All injections were made at a rate of 200 nl/min, and the needle (33 gauge; Hamilton, Bonaduz, Switzerland) was kept in place for additional 5 min before retracted. Stereological coordinates were determined according to the mouse atlas of Paxinos and Franklin (2001). To minimize the post-operative mortality of 80% reported in mice (Lundblad et al., 2002), an intensive care was performed for two weeks following lesion by daily subcutaneous (s.c.) injections of 1–2 ml 0.9% NaCl against dehydration and by soaking the food pellets in water to allow eating for weak mice. This preventive protocol led to a survival rate of 76%. Behavioral analysis Cylinder test. To evaluate the extension of the lesion, at three weeks after the 6-OHDA lesion spontaneous forelimb use was evaluated in the cylinder test according to the previously described test paradigm for rats (Schallert and Tillerson, 1999; Carlsson et al., 2007). The mice were placed individually in a glass cylinder (10.5-cm diameter, 14-cm height) and video recorded while performing 20 weight-shifting movements of the forepaws in contact with the cylinder wall. The numbers of the left or right forepaw contacts were scored by an observer blinded to the animals’ identity and presented as left (impaired) touches in percentage of total touches. Normal mice will score 50% in this test. Amphetamine-induced rotation test. At 3 weeks after the 6-OHDA lesion rotational behavior induced by amphetamine (5 mg/kg, i.p.; Sigma-Aldrich) was observed in an open-field arena with an area of 52 cm  52 cm. The testing sessions were performed and video recorded over 30 min and the animals’ right and left full body turns were checked by an experimentally blinded investigator. The data were expressed as net full body turns per minute, with negative values indicating rotation contralateral to the lesion side.

304

U. Keber et al. / Neuroscience 298 (2015) 302–317

Fig. 1. (A) Experimental design. Adult male wildtype C57Bl/6 mice (n = 50) received a unilateral 6-OHDA lesion of the medial forebrain bundle in order to induce a depletion of the nigrostriatal dopaminergic pathway. 24 animals that displayed <20% left forelimb touches of total in the cylinder test (B) and >3.5 net full body turns/min in the amphetamine-induced rotation test (C) were included in further analysis. 18 Mice received daily treatment of L-DOPA (6 mg/kg with 12 mg/kg benserazide hydrochloride) for 15 days to induce stable dyskinesia. 6 Mice receiving saline treatment served as a control group. Both treatment groups did not significantly differ (n.s., p > 0.05) in left paw use and net full body turns/min in the cylinder test (B) and amphetamine-induced rotation test (C), respectively. Abnormal involuntary movements (AIM) were evaluated on days 1, 7 and 15 of the treatment period. Afterward, animals were sacrificed and brains processed for histological analysis.

L-DOPA treatment and AIM scoring. Six weeks after the 6-OHDA lesion AIMs, resembling peak dose dyskinesias in PD patients, were induced in mice by daily s.c. injections of 6 mg/kg L-DOPA in combination with benserazide hydrochloride (12 mg/kg; SigmaAldrich) for 15 days. The route of administration has been adopted after the findings of Lindgren et al. (2007) in rats suggesting that the s.c. injections gave more consistent and reliable L-DOPA response compared with the i.p. route. On days 1, 7 and 15 of treatment, the AIMs were scored individually for 1 min every 20 min during a 3-h period following L-DOPA or saline administration according to a mouse dyskinesia scale previously described by Lundblad et al. (2004). Briefly, severity scores from 0 to 4 were given for four dyskinetic subtypes according to topographic distribution as forelimb, orolingual, axial and locomotive behaviors as follows: 0, absent; 1, present during <50% of the observation time; 2, present during >50% of the observation time; 3, continuous, but interruptible by repeated sensory stimuli (e.g. sudden noise, opening of the cage lid); 4, continuous, not interrupted by repeated sensory stimuli. Enhanced manifestation of normal behaviors such as grooming, gnawing,

rearing and sniffing was not included in the rating. The data are presented as integrated AIM scores, calculated as the area under the curve over the whole test session. Thus, the theoretical maximum score for each mouse in one test session was 144 (maximal severity of 16 at 9 time points).

Tissue preparation Following 48 h after the last L-DOPA or saline application mice were euthanized with 100 mg/kg pentobarbital (Sigma-Aldrich) and perfused transcardially with 50 mL saline followed by 100 mL 4% paraformaldehyde (Sigma-Aldrich). Brains were carefully removed, postfixed in 4% paraformaldehyde for 24 h, cryoprotected by incubation in 30% sucrose-containing phosphatebuffered saline (PBS) for 48 h and snap-frozen on dry ice. Coronal 30-lm-thick sections were cut on a cryomicrotome and collected in 10 regularly spaced series in antifreeze buffer solution (containing 30% ethylene glycol and 30% glycerin) and stored at - 20 °C until use.

305

U. Keber et al. / Neuroscience 298 (2015) 302–317

Immunohistochemistry Free-floating sections were permeabilized with 3% Triton X-100 (Sigma-Aldrich) in 0.1 M PB. Non-specific binding sites and endogenous peroxidases were blocked with 5% normal horse serum (NHS; Sigma-Aldrich; diluted in 0.1 M PB) and with 30% H2O2/methanol (Sigma-Aldrich), respectively. After several washes in 0.1 M PB, the sections were incubated with primary antibodies (diluted in 0.1 M PB containing 1% NHS) for 14 h at 4 °C. The following primary antibodies were used: sheep polyclonal anti-TH (AB1542; Chemicon International, Temecula, CA, USA; 1/500), rabbit polyclonal anti-serotonin transporter (SERT; 24330; ImmunoStar, Hudson, WI, USA; 1/2000), mouse monoclonal anti-delta isoform of oncogene FosB (DFosB; sc-48; Santa Cruz Biotechnology, Santa Cruz, CA, USA; 1/2000). Bound primary antibodies were visualized using biotinylated secondary antibodies from donkey (Dianova, Hamburg, Germany; 1/200), standard avidin–biotin-peroxidase techniques (Vectastain Elite ABC kit, Boehringer, Germany; 1/500), and finally with 3,30 -diaminobenzidine (Sigma-Aldrich) and nickel ammonium sulfate (Fluka, Buchs, Switzerland), resulting in a dark blue/black staining (Depboylu et al., 2011). Additionally, TH (1/500) was co-stained with neuronal nuclear antigen (NeuN; mouse monoclonal; MAB377; Chemicon International; 1/ 1000) by immunofluorescence using Cy3- and Alexa488conjugated secondary antibodies (Dianova; 1/500), respectively. Double immunofluorescence labeling of TH and NeuN was analyzed by laser confocal microscopy (Zeiss Axiovert 200 M, Jena, Germany).

landmarks were determined according to the mouse brain atlas of Paxinos and Franklin (2001). Estimation of TH+ and SERT+ fiber densities. Images were captured by a digital camera (Olympus E330, Tokyo, Japan) and TH+ and SERT+ fiber densities of the striatum (seven sections; +1.1 to 0.7 mm from bregma), nucleus accumbens (five sections; +1.1 to +0.2 mm from bregma) and cortex (five sections; +0.8 to 0.4 mm from bregma) were determined by optical density (OD) measurements from the stained sections using the Image J software platform version 1.43 (National Institute of Health, Bethesda, MD, USA; http://rsbweb.nih.gov/ij/) and corrected for non-specific background density on defined non-immunoreactive areas of the sections (Depboylu et al., 2011). The data are expressed as OD in percentage (%) of the intact control side which was set 100%. All immunohistochemical parameters and procedures relevant for TH- and SERT-immunostaining quantifications were kept constant (e.g. antibody concentrations, reaction and washing times, temperatures, illumination and image analysis settings). Estimation of DFosB+ cell numbers. The total number of DFosB+ cells was evaluated by Image J software. High-resolution images were captured of the seven striatal sections as described before. The same four striatal subregions (DL, DM, VL, VM) and the nucleus accumbens were analyzed by counting all DFosB+ cells manually. Statistics

Histological analysis Histological analysis was performed in order to evaluate the extent of 6-OHDA-induced lesions and to characterize cellular parameters with possible associations to LID. Only mice showing a severe lesion with nigral TH+ cell numbers <25% of intact side and with striatal TH+ fiber densities <25% of intact side were included in final statistical analyses. Stereological counts of TH+ cells. Unbiased stereological estimations of total numbers of TH+ cells were determined on nine coronal sections throughout the rostrocaudal extent of SNpc (2.57 to 3.92 mm from bregma) and seven throughout the ventral tegmental area (VTA; 2.87 to 3.92 mm from bregma) using the optical fractionator method (Stereoinvestigator software, MicroBrightField, Magdeburg, Germany) (West, 1999). For TH+ cell counts in the striatum (seven sections; +1.1 to 0.7 mm from bregma), nucleus accumbens (five sections; +1.1 to +0.2 mm from bregma) and cortex (five sections; +0.8 to 0.4 mm from bregma) ipsilateral to the 6-OHDA lesion side the counting frame was moved systematically over the entire delineated brain structure on each section to allow analysis of 100% of the regions. The striatum was further divided into four subregions (dorsolateral, DL; dorsomedial, DM; ventrolateral, VL; ventromedial, VM). All anatomical

All statistical group comparisons of L-DOPA-treated dyskinetic, L-DOPA-treated non-dyskinetic and salinetreated mice were conducted using a one-way Analysis of Variance followed by post hoc Student–Newman– Keuls test. Comparisons of two groups regarding behavioral and cellular lesion parameters were performed with unpaired t-test analysis. Correlations between variables were examined by linear regression. Statistical significance was set at p < 0.05 and all data were expressed as group mean ± standard error of the mean (SEM). The statistical comparisons were performed using GraphPad Prism software version 5.0 (GraphPad Software Inc., La Jolla, USA).

RESULTS Effects of 6-OHDA lesion Behavioral analysis: Lesion-induced motor deficits in spontaneous forelimb use and rotational behavior. Three weeks after surgery lesion-induced motor deficits of the mice were evaluated in order to select animals with a high DAergic lesion for treatment (Fig. 1A). Inclusion criteria were <20% left forelimb touches of total in the cylinder test and >3.5 net full body turns/min to the lesioned side in the amphetamine-induced rotation test. In the cylinder test, the selected animals showed a

306

U. Keber et al. / Neuroscience 298 (2015) 302–317

severe forelimb use asymmetry, performing 18.1 ± 2.2% wall contacts with the paw contralateral to the lesion (Fig. 1B). Treatment groups were allocated in a way showing no difference regarding the degree of forelimb akinesia (L-DOPA 19.9 ± 3.3% vs. saline 15.8 ± 3.7% left paw use of total). Recordings of amphetamineinduced rotations in an open field displayed rotational asymmetry in all 6-OHDA-lesioned mice and showed 3.9 ± 0.6 net full body turns/min to the lesioned side of those animals included in the experiment (Fig. 1C). The turning behavior in L-DOPA-treated mice did not differ significantly from saline-treated mice (3.9 ± 0.7 vs. 3.9 ± 1.2 net full body turns/min).

midbrain neurons and densitometric analysis of striatal, accumbal and cortical DAergic target fibers were performed. Two animals (one saline-treated and one LDOPA-treated non-dyskinetic mouse) did not show a severe lesion with <25% nigral TH+ cell number and <25% striatal TH+ fiber density of intact side and thus were excluded from all statistical analyses. In the remaining 22 animals, the number of TH+ neuronal cell bodies was markedly decreased by about 85% to 14.8 ± 1.0% of the number found in the intact side in the SNpc (5142 ± 283 vs. 750 ± 62) and by approximately 50% to 48.7 ± 1.9% of the intact side in the VTA (3442 ± 181 vs. 1638 ± 85; Fig. 2G, H). Accordingly, OD measurements showed a pronounced reduction in TH+ target fiber innervations by nearly 90% to 12.2 ± 1.0% of the intact side in the striatum, to 28.6 ± 2.3% of the intact side in the accumbens and to 63.8 ± 3.2% of the intact side in the cortex (Fig. 2A–F). There was no significant difference

Histological analysis: lesion-induced loss of DAergic midbrain neurons and their striatal, accumbal and cortical projections // increment of striatal serotonergic fibers. In order to evaluate the extent of the unilateral 6-OHDA lesion of the MFB, stereological counts of DAergic

intact

6-OHDA lesioned

Str + Acc

A

B

Str Acc 400 µm

C

400 µm

D

50 µm

F

Cortex

E

50 µm

SN + VTA

50 µm

G SNpr

SNpc

VTA

300 µm

50 µm

H

300 µm

Fig. 2. Immunohistochemical staining of tyrosine hydroxylase (TH) six weeks following unilateral 6-OHDA lesion of the medial forebrain bundle. 6-OHDA induced a severe loss of TH+ dopaminergic fibers in the striatum (Str), nucleus accumbens (Acc) and cortex of the lesioned side (B, D, F) compared to the intact side (A, C, E). C and D illustrate the detailed frame of A and B, respectively. Note the two TH+ neuron-like cells in the dopamine-depleted Str (D). Accordingly, the lesioned ventral midbrain (H) showed a distinct reduction in TH+ neurons and fibers in the substantia nigra pars compacta/reticulata (SNpc/pr) and ventral tegmental area (VTA) as compared to the intact side (G).

307

U. Keber et al. / Neuroscience 298 (2015) 302–317

in the extent of DA denervation between both treatment groups. In addition to DAergic projections, the serotonergic innervation in the striatum following unilateral 6-OHDA lesion was examined by OD measurements of SERT+ fibers. The SERT+ fiber density in the whole striatum rose to 151.4 ± 4.7% of the intact side and in the posterior striatum nearly doubled to 189.5 ± 12.3% of the intact side, showing no significant difference between L-DOPA- and saline-treated mice, respectively. LID To induce stable dyskinesia, mice received daily s.c. LDOPA injections for 15 days and scoring of AIMs was performed on days 1, 7 and 15 of the treatment period (Fig. 1A). As axial, limb and orolingual AIMs in rats and mice have shown to be a functional equivalent of LID in PD (Andersson et al., 1999; Lundblad et al., 2002, 2005; Winkler et al., 2002), the scores of these three AIMs were summed and used for correlations with histological criteria, whereas locomotive AIM, which is considered a less sensitive behavioral measure of dyskinesia, was presented separately. During the application period, twelve of 17 L-DOPA-treated animals (71%) exhibited dyskinesias, beginning immediately after L-DOPA injection, reaching highest AIM scores between 20 and 80 min and lasting up to 120 min (Fig. 3A–C). Nearly 50% of the AIMs comprised hyperkinetic limb and orolingual AIMs and one third consisted of dystonic axial movements, while locomotive rotations accounted for 20% of

limb

L-DOPA-induced dyskinesia per session (day 15)

*

*

*

C

*

Integrated AIM scores

Integrated AIM scores

B

orolingual

100

* 50

*

*

*

*

*

* *

0

Striatal serotonergic fiber innervation and LID An association between LID and striatal serotonergic projections has been described in different mammals (Carlsson et al., 2007; Carta et al., 2007; Rylander et al., 2010; Zeng et al., 2010). Axial, limb and orolingual AIMs were accompanied by a sprouting of striatal serotonergic fibers. Here, SERT+ densitometric analysis of denervated striata demonstrated a markedly increased serotonergic innervation in dyskinetic mice compared to non-dyskinetic animals (163.4 ± 5.8% vs. 128.1 ± 3.6% of intact side, F(2,21) = 7.99; p < 0.001), with a maximal increment of 150% in the posterior striatum

axial

locomotive

L-DOPA-induced dyskinesia per treatment period

*

* 600

* 400

200

*

*

*

0

D

L-DOPA-induced dyskinesia distribution of subtypes Integrated AIM Scores

A

dyskinesias (Fig. 3D). In contrast, saline-treated mice did not develop any dyskinesia. In order to examine a possible relationship between dyskinesias and the 6-OHDA lesion, statistical comparison of dyskinetic (LD-dys, n = 12) and nondyskinetic (LD-non-dys, n = 5) mice was performed and revealed no significant difference regarding the extent of nigral and tegmental DAergic cell body depletion (SNpc 13.5 ± 1.5 vs. 17.1 ± 1.4 of intact side; VTA 50.3 ± 2.6 vs. 49.6 ± 4.4 of intact side; p > 0.05, respectively). However, TH+ target fiber densities in the denervated striatum, nucleus accumbens and cortex were significantly lower in the LD-dys group than in the LD-non-dys group (striatum 10.8 ± 0.4 vs. 17.4 ± 2.5 of intact side, p < 0.001; accumbens 25.1 ± 2.4% vs. 37.7 ± 3.8% of intact side; cortex 58.4 ± 4.4% vs. 75.9 ± 3.7% of intact side; p < 0.05, respectively).

400

48%

300

32% limb

200

20%

100 ol

saline

L-DOPA axial+limb+ol

lo co

ax ia l

+ lim b

Treatment period

ol

15 da y

7 da y

1 da y

0

0 16

min after L-DOPA injection

18

0

0

0 14

12

10

60

80

40

0

20

0

L-DOPA loco

Fig. 3. L-DOPA induced dyskinesia in mice with unilateral 6-OHDA lesion of the medial forebrain bundle. (A) Different L-DOPA-induced dyskinesia subtypes. (B) As early as 20 min after L-DOPA injection abnormal involuntary movements (AIM) could be observed, lasting up to 120 min with plateau of about 60 min as a maximum (data of session day 15 are shown representatively). Saline-treated mice did not show any dyskinesias. (C) The AIM scores in L-DOPA-treated mice did not significantly change during treatment period. (D) Relational comparison of the different L-DOPAinduced dyskinesia subtypes as observed in the experiment. ⁄p < 0.05 vs. saline-treated group.

308

U. Keber et al. / Neuroscience 298 (2015) 302–317

A

saline

LD-non-dys

10 μm

10 μm

Striatal SERT+ fiber innervation

Striatal SERT+ fibers vs. AIMs

##

150 100

(axial + limb + orolingual)

##

#

Integrated AIM Score

200

C

10 μm

1000

saline LD-non-dys LD-dys

250

R2 = 0.46 p = 0.003

500

LD-dys LD-non-dys 0

rio

r io

os rp

100

150

200

250

Optical density SERT+ fibers (% lesioned / intact side)

St

St

ra

nt

te

er

ho rw St

r

50

le

Optical density SERT+ fibers (% lesioned / intact side)

B

LD-dys

Fig. 4. Serotonergic alterations in L-DOPA-induced dyskinesias following unilateral 6-OHDA lesion of the medial forebrain bundle. (A) Representative images demonstrate that the density of serotonergic fibers, as identified by immunostaining of serotonin transporter (SERT), in the dopamine-depleted striatum was increased in mice with L-DOPA-induced dyskinesia (LD-dys) as compared to L-DOPA-treated mice without dyskinesia (LD-non-dys) and saline-treated mice. (B) Bars show the quantification of SERT+ innervation of the whole, the anterior and the posterior parts of the striatum (Str) in the different mouse groups. Note that the optical density of SERT+ fibers of the lesioned side (presented as% of intact side which was set 100%) was significantly higher in LD-dys mice than in LD-non-dys mice. #p < 0.05; ##p < 0.001. (C) The optical density of striatal SERT+ fibers correlated with abnormal involuntary movement (AIM) scores of L-DOPA-treated mice.

(213.2 ± 18.4% vs. 139.6 ± 7.1% of intact side, F(2,21) = 3.64; p < 0.05; Fig. 4). Neither the LD-dys nor the LD-non-dys group differed significantly from salinetreated mice regarding striatal serotonergic fiber densities (whole striatum F(2,21) = 3.64; posterior striatum F(2,21) = 3.64; p > 0.05, respectively). Striatal TH+ and DFosB+ cells: phenotype, association with LID and striatal cell distribution The striatal TH+ cell number correlated positively with LID and its molecular marker DFosB. In the DAergic denervated striatum TH+ cells could be detected (Fig. 2D). To examine their neuronal phenotype, confocal double immunofluorescence of TH with NeuN was performed. All TH+ cells showed a co-expression with the neuronal nuclear marker NeuN and could hence be specified as neurons (Fig. 5). In order to investigate a possible association between striatal TH+ cells and LID expression, TH+ cells were counted stereologically in the denervated striatum. Notably, the number of TH+ neurons in the lesioned striatum was increased in dyskinetic mice by 41% as compared to non-dyskinetic mice and by 29% as compared to saline-treated mice (F(2,21) = 14.69, p < 0.001, respectively), while the LD-non-dys group

did not differ significantly from the saline-treated group (F(2,21) = 14.69, p > 0.05; Fig. 6A–C, G). Thus, the increase by approximately one third in all L-DOPAtreated mice compared to saline-treated animals (721 ± 52 vs. 512 ± 38; p < 0.05) is mainly conditioned by the amount of dyskinesia produced by the treatment. In addition to the group differences, the striatal TH+ cell number correlated strongly positively with the severity of axial, limb and orolingual AIMs in L-DOPA-treated animals (R2 = 0.76; p < 0.001; Fig. 7A), as well as with the striatal DFosB expression as a valid molecular marker of LID (R2 = 0.83; p < 0.0001; Fig. 6I). Counts of striatal DFosB+ cells revealed a pronounced increase in DFosB+ cells by approximately one third in dyskinetic animals as compared to nondyskinetic ones (13450 ± 910 vs. 7996 ± 1895; F(2,21) = 9.68, p < 0.001; Fig. 6D–F, H), but showed no difference between LD-non-dys and saline-treated animals. The increment by 44% in all L-DOPA-treated animals compared to saline-treated animals is predominantly based upon the increased number of DFosB+ cells in dyskinetic mice. Likewise, the striatal expression of DFosB+ correlated positively with the integrated AIM score (R2 = 0.64; p < 0.001; Fig. 7B). The striatum was further divided into four subregions to examine the distribution of TH+ and DFosB+ cells

309

U. Keber et al. / Neuroscience 298 (2015) 302–317

TH

NeuN

TH + NeuN

Fig. 5. Representative confocal double immunofluorescence images from the mouse striatum six weeks following 6-OHDA lesion of the medial forebrain bundle. The tyrosine hydroxylase (TH; green) expressing cell is a neuron as its cell body is also positive for the neuronal nuclear antigen (NeuN; red). Single color images are merged (yellow). Note that a TH-/NeuN + neuron is most likely innervated by a TH+ fiber (yellow arrow). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

R2

Fig. 6. Striatal neuronal plasticity in L-DOPA-induced dyskinesias 6 weeks following unilateral 6-OHDA lesion of the medial forebrain bundle. (A–C) The number of tyrosine hydroxylase (TH) expressing neurons in the dopamine-depleted striatum increased only in mice exhibiting L-DOPA-induced dyskinesia (LD-dys; n = 12; A) as compared to mice without dyskinesia (LD-non-dys; n = 5; B) and saline-treated mice (n = 5; C). (D–F) Accordingly, the number of striatal DFosB+ nuclei was increased only in LD-dys animals (F). G and H, Bars demonstrate the quantitative analysis of TH+ and DFosB+ cells in the dopamine-depleted striatum in the different groups. ⁄p < 0.05; ⁄⁄p < 0.001. (I) The number of TH+ neurons in the striata of L-DOPA-treated mice correlated strongly positively with the number of DFosB+ cells.

in relation with the occurrence of dyskinesia. Fig. 8 illustrates the obtained data and the levels of significance in group differences. An increase in TH+

cell numbers in dyskinetic animals (including axial, limb and orolingual AIMs) compared to LD-non-dys and saline-treated mice could most distinctly be seen in the

310

U. Keber et al. / Neuroscience 298 (2015) 302–317

Fig. 7. The number of tyrosine hydroxylase (TH) expressing neurons (A) as well as DFosB+ cells (B) in the dopamine-depleted striatum positively correlated with the occurrence and severity of abnormal involuntary movements (AIM) in L-DOPA-treated mice.

A

Distribuon of TH+ cells in striatal subregions saline

LD-non-dys

dorsal

Str

DM

LD-dys

DL

lateral

medial VM VL

ventral

Acc

B

C

TH+ cells in striatal subregions *#

ΔFosB+ cells in striatal subregions *

250

*#

*#

200 150 100

LD-dys

50

LD-non-dys saline

0

VL

DL lateral

DM

VM

medial

Subregions

Number of ΔFosB+ cells

Number of TH+ cells

#

*#

*#

4000 3000 2000

LD-dys

1000

LD-non-dys saline

0

VL

DL lateral

DM VM medial

Subregions

Fig. 8. (A) Distribution of tyrosine hydroxylase (TH) expressing neurons in the subregions of the dopamine-depleted striatum (Str) and accumbens (Acc) shown representatively for saline-treated mice and L-DOPA-treated mice without dyskinesia (LD-non-dys) and with dyskinesia (LD-dys). (B and C) Quantitative analysis of the number of TH+ neurons and DFosB+ cells in the different subregions of the Str in saline-treated mice, LD-nondys and LD-dys mice. Note that the cell numbers predominantly increased in the lateral subregions of LD-dys animals. ⁄p < 0.05 vs. saline group; # p < 0.05 vs. LD-non-dys group. F-values TH+ cell number comparisons: VL F(2,21) = 12.23, DL F(2,21) = 21.85, DM F(2,21) = 11.6, VM F(2,21) = 7.3; F-values DFosB+ cell number comparisons: VL F(2,21) = 8.76, DL F(2,21) = 17.76, DM F(2,21) = 8.60, VM F(2,21) = 0.34. DL, dorsolateral; DM, dorsomedial; VL, ventrolateral; VM, ventromedial.

lateral subregions with only moderate increments in the medial subregions. LD-non-dys mice and saline-treated animals showed similar TH+ cell numbers in all four subregions. A comparable distribution of DFosB+ expression was observed, showing the strongest increment of cells in the lateral subregions of dyskinetic mice compared to LDnon-dys and saline-treated mice, while there was little

and no difference in the DM and VM subregions, respectively. TH+ neurons and striatal TH+ and SERT+ fiber innervations In order to determine further correlations, the number of TH+ neurons in the denervated striatum was analyzed

U. Keber et al. / Neuroscience 298 (2015) 302–317

with respect to striatal TH+ and SERT+ fiber innervations. We found low positive correlations of an increasing number of striatal TH+ neurons with both decreasing striatal TH+ fiber density (R2 = 0.58; p < 0.05) and increasing striatal SERT+ fiber density (R2 = 0.45; p < 0.05; Fig. 9).

311

Expression of AIM subtypes depends on the distribution of TH+ neurons in striatal subregions To examine the relation of TH+ cell numbers in the striatal subregions to the expression of specific AIM subtypes, linear regression analysis was performed (Fig. 10). Strikingly, increased TH+ cell numbers in the

Fig. 9. The number of tyrosine hydroxylase (TH) expressing neurons correlated with the degree of TH+ fiber loss (A) and the increase of serotonergic fibers, as identified by immunostaining of serotonin transporter (SERT; B) in the lesioned striata of all L-DOPA-treated mice.

Fig. 10. Correlative analysis of the number of tyrosine hydroxylase (TH) expressing neurons in the four dopamine-depleted striatal subregions with occurrence of abnormal involuntary movement (AIM) subtypes in all L-DOPA-treated mice. Note that the number of TH+ neurons in the lateral striatum (DL, dorsolateral; VL, ventrolateral) correlated stronger with axial (A), limb (B) and orolingual (C) dyskinesias than TH+ cell numbers in the medial striatum (DM, dorsomedial; VM, ventromedial) did. The number of medially located TH+ neurons showed a stronger correlation with AIM scores for locomotive dyskinesia (D). ⁄p < 0.05; ⁄⁄p < 0.001.

312

U. Keber et al. / Neuroscience 298 (2015) 302–317

lateral subregions were accompanied by the development of axial (R2 = 0.79 DL; R2 = 0.77 VL; p < 0.001, respectively), limb (R2 = 0.83 DL; R2 = 0.77 VL; p < 0.001, respectively) and to a lower extent of orolingual (R2 = 0.48 VL; R2 = 0.45 DL; p < 0.05, respectively) dyskinesias. In contrast, locomotive rotations were predominantly associated with TH+ cell increments in the medial subregions (R2 = 0.54 DM with p < 0.001; R2 = 0.48 VM with p < 0.05). Accumbal and cortical TH+ neurons and LID In accordance with the striatum, a marked decrease of DAergic target fibers could be observed in the nucleus accumbens and cortex following unilateral 6-OHDA lesion of the MFB. TH+ neurons could be detected in the denervated accumbens and cortex, as recently reported (Depboylu, 2014; Depboylu et al., 2014). Here, we aimed to examine whether TH+ neurons in these two denervated regions were also associated with the occurrence of LID. Stereological counts and statistical analysis of TH+ cell numbers revealed no quantitative difference between LD-dys and LD-non-dys mice, neither in the nucleus accumbens (126 ± 12 vs. 77 ± 17; F(2,21) = 2.37, p > 0.05) nor in the cortex (379 ± 23 vs. 411 ± 20; F(2,21) = 0.39, p > 0.05). Additionally, the number of TH+ cells in both L-DOPA-treated groups did not differ significantly from the number in saline-treated animals (accumbens 123 ± 22; cortex 388 ± 21; p > 0.05 each vs. control).

DISCUSSION Our results reveal that striatal TH+ neurons may be functionally associated with the expression of LID in mice. In contrast, accumbal and cortical TH+ neurons are likely to be not involved in LID. The number of striatal TH+ neurons correlated positively with AIMs, striatal DFosB expression and striatal SERT+ fiber density. Additionally, the extent of AIM subtypes correlated with the distribution of TH+ neurons in striatal subregions. Development of LIDs in MFB-lesioned mice In order to induce stable dyskinesia and to subsequently investigate neurochemical changes between dyskinetic and non-dyskinetic animals, we performed daily s.c. injections of L-DOPA to MFB-lesioned mice. Approximately 70% of the animals exhibited dyskinesia with maximal AIM scores beginning on day one. This immediate initiation of LID in contrast to PD patients, who develop dyskinesias steadily in the course of disease and L-DOPA treatment, may likely be attributed to the severe DA denervation of MFBlesioned mice, mimicking the end stage of PD in humans. Indeed, the degree of nigrostriatal depletion has been shown to reflect an important risk factor for the development of LID in patients and animal models (Di Monte et al., 2000; Winkler et al., 2002; Linazasoro et al., 2004; Ulusoy et al., 2010).

Accordingly, MFB-lesioned mice show a much stronger propensity to develop LID compared to those with less severe intrastriatal or nigral 6-OHDA-lesions at the same L-DOPA dose (Francardo et al., 2011; Smith et al., 2012) and a lower L-DOPA-dose is needed to induce stable AIMs (Lundblad et al., 2004). In this study, dyskinetic mice suffered from a higher average TH+ fiber loss in the striatum than non-dyskinetic mice but they did not differ in terms of nigral and tegmental degeneration of TH+ cell bodies. This constellation particularly highlights a decreasing striatal DAergic fiber innervation as an important parameter for the development of LID, as the buffering capacity for exogenously administered L-DOPA is lost progressively (Carta and Bezard, 2011). However, as even strongly lesioned mice fail to develop LID (own experiment; Smith et al., 2012), pathogenesis of LID seem to rely on multiple factors and on variable susceptibilities of affected species (Linazasoro, 2005). Interestingly, other groups investigating AIMs in mice following chronic L-DOPA treatment initially observed lower AIM scores with an increment over days (Francardo et al., 2011; Lundblad et al., 2004; Smith et al., 2012). A possible explanation for the severe LID we already observed at day 1 may be the relatively high first L-DOPA dose of 6 mg/kg, whereas treatments in the cited studies were performed with distinctly lower initial doses of only 2 or 3 mg/kg for the first 4–14 days. Following subsequent elevation to 6 mg/kg, AIM scores increased considerably in those experiments. Another important point is that except Smith et al. authors administered the drug i.p. and therefore evoked the risk of dose failure episodes due to erratic absorption of L-DOPA (Lindgren et al., 2007). Plasticity of TH+ neurons in LID and following L-DOPA treatment In the predecessor work of the current study, it was shown that TH+ neurons exist in target areas of DAergic ventral midbrain neurons - striatum, nucleus accumbens and cortex – in mice, and that they increase in number following DAergic deafferentiation induced by 6-OHDA lesion of the MFB (Depboylu, 2014; Depboylu et al., 2014). Such compensatory numerical increment upon DAergic denervation has previously only been reported for TH+ neurons in the striatum (Tashiro et al., 1989; Betarbet et al., 1997; Meredith et al., 1999; Lopez-Real et al., 2003), based on initiated TH expression in GABAergic interneurons or projection neurons of the direct and indirect pathways (Tande´ et al., 2006; Darmopil et al., 2008). In this study we further reported a significant increase of striatal TH+ cell numbers in L-DOPA-treated mice when compared to saline-treated animals. This finding is consistent with observations of previous studies on parkinsonian rodents (Lopez-Real et al., 2003; Jollivet et al., 2004; Darmopil et al., 2008; Espadas et al., 2012) and non-human primates (DiCaudo et al., 2012; Lundblad et al., 2005). In contrast, Huot and Parent (2007) and Huot et al. (2007, 2008) reported a numerical decrease of striatal TH+ neurons in PD patients and non-human primates following

U. Keber et al. / Neuroscience 298 (2015) 302–317

L-DOPA treatment. One possible explanation for this discrepancy may be found in the different treatment paradigms: the L-DOPA doses administered to rodents as well as the cumulative (but interestingly not the daily) L-DOPA dose orally given to macaques by DiCaudo et al. highly exceeded the doses applied in the works of Huot et al. From these data it can be hypothesized that L-DOPA in high concentrations exerts a trophic effect via growth factors (Okazawa et al., 1992; Du et al., 1995), whereas lower levels might control TH expression via a negative feedback mechanism (Kumer and Vrana, 1996). However, further investigations are needed to entirely understand the regulative mechanisms of TH expression. Importantly, none of those studies have examined the occurrence of dyskinesia in the L-DOPA-treated animals or patients. Here, we investigated whether TH+ neurons of the striatum, nucleus accumbens and cortex are associated with the expression of LID in 6OHDA-mediated MFB-lesioned mice. During the experimental phase of our study, two reports were published analyzing striatal TH+ neurons in terms of LID in MFB-lesioned mice (Francardo et al., 2011; Smith et al., 2012). In line with the data obtained by Francardo et al. (2011) and Smith et al. (2012), our results show that the number of striatal TH+ neurons correlated positively with the occurrence and severity of AIMs and with the striatal expression of DFosB, a molecular marker of dyskinesia in mice (Lundblad et al., 2004; Pavo´n et al., 2006; Fasano et al., 2010), rats (Andersson et al., 1999; Winkler et al., 2002), non-human primates (Berton et al., 2009; Fasano et al., 2010) and PD patients (Lindgren et al., 2011). In this context it may be worth mentioning that acute or chronic L-DOPA treatment itself has been reported to elevate striatal DFosB levels in 6-OHDA-lesioned rodents (Valastro et al., 2007) and PD patients (Calon et al., 2004), so we cannot entirely rule out that the drug itself may have affected DFosB expression. However, the cited studies did not distinguish between occurrence and absence of dyskinesia and the observed average increase may be predominantly due to dyskinetic animals, as seen in our results. Furthermore, there is convincing evidence that above a critical threshold DFosB is causally linked to LID, as its striatal overexpression in rats induced by viral vectors provoked dyskinesia even without L-DOPA administration (Cao et al., 2010) and as molecular interventions reducing DFosB activity attenuated the severity of already established LID in different animal models of PD (Berton et al., 2009; Engeln et al., 2014). Conclusively, our data suggest that striatal TH+ neurons may be functionally involved in the expression of LID. A possible underlying mechanism might be an increased synthesis of L-DOPA via TH. Simultaneous application of external L-DOPA may lead to pulsatile and high extracellular concentrations of L-DOPAconverted DA, which has been reported to play a key role in LID pathogenesis. Further work will be necessary to replicate these results in primates, as the phenotype of TH+ neurons slightly varies between species

313

(Ugrumov, 2013) and we cannot exclude different responses to similar stimuli. Alteration of striatal serotonergic innervation in LID Previous works suggested that the raphe-striatal serotonergic system might be essential in the generation of LID as an association has been reported in rats (Carlsson et al., 2007; Carta et al., 2007), non-human primates and in human post-mortem material (Rylander et al., 2010; Zeng et al., 2010). Intrastriatal transplantation of embryonic serotonergic neurons potentiated LID (Carlsson et al., 2007), whereas a serotoninspecific lesion or certain serotonin receptor agonists reduced dyskinesia (Carta et al., 2007; Eskow et al., 2009; Mun˜oz et al., 2009). In our study, striatal serotonergic innervation was increased in dyskinetic mice as compared to LD-non-dys and to saline-treated mice. This is consistent with investigations of previous reports in rodents (Carlsson et al., 2007; Carta et al., 2007). In contrast, Smith et al. (2012) found an inverse correlation of striatal serotonin fibers with AIMs when using nigral cell loss as a covariate, while Francardo et al. (2011) observed no difference in striatal serotonin levels between the groups. Differences between those and our results might be caused by differences in sensitivity of the methods used to analyze serotonergic projections in the striatum. While Smith et al. (2012) performed immunostaining directly against serotonin, we visualized serotonergic fibers by SERT-immunoreactivity. Immunostaining of monoamines like serotonin is still a matter of debate of specificity and highly depends on the quality of tissue fixation and histochemical handling. Francardo et al. (2011) measured serotonin levels by high-pressure liquid chromatography in striatal tissue homogenates. This procedure might further reduce the low striatal serotonin concentration with the consequence that no significant differences could be detected. Another crucial point is that striatal serotonin levels may not directly reflect the ‘‘real’’ density of serotonergic projections, as production of serotonin can even be reduced when its AADC is used for conversion of exogenous L-DOPA to DA (Arai et al., 1995; Tanaka et al., 1999; Maeda et al., 2005). Thus, we consider immunostaining of serotonergic neuron- and projectionspecific SERT the better alternative. Serotonergic neurons possess the enzymatic machinery for DA synthesis as well as the ability to vesicular storage and DA release (Arai et al., 1995; Tanaka et al., 1999; Maeda et al., 2005). However, they lack a feed-back control mechanism to regulate synaptic DA levels, therefore leading to sensitization of postsynaptic DA receptors (Carta et al., 2007; Eskow et al., 2009; Navailles et al., 2010). Those lines of evidence clearly support the important role of striatal serotonergic projections in LID pathogenesis. In agreement, Politis et al. (2014), using a series of in vivo nuclear imaging assessments with radioligand markers of serotonergic and DAergic functions, most recently found that serotonergic terminals contributed to abnormal L-DOPA-induced short-term increases in synaptic DA levels in PD patients with LID.

314

U. Keber et al. / Neuroscience 298 (2015) 302–317

synergistically perform a cooperative DA synthesis (Ugrumov, 2013) (Fig. 11). The integration of other cell types with AADC expression in this process has not been delineated definitively (Hardebo et al., 1980; Melamed et al., 1980; Juorio et al., 1993). Importantly, it has been demonstrated that some TH+ neurons in the striatum additionally express AADC, qualifying them for the entire DA synthesis (Lopez-Real et al., 2003; Weihe et al., 2006; Darmopil et al., 2008).

Interaction of striatal TH+ cells and serotonergic fibers in dyskinetic mice None of the cited studies correlated the number of striatal TH+ neurons with the serotonergic fiber density. Here, we found a positive correlation between serotonergic innervation and TH+ cell number in the striatum. This observation allows the assumption that AADCcontaining serotonergic terminals and TH+ neurons

A physiological

B

DA terminal

PD + non-dys

degenerated DA terminal

L-tyrosine

5-HT terminal

TH+ neuron

L-DOPA

L-tyrosine

AADC

L-DOPA

TH L-DOPA

dopamine

D₁/₂ receptor

D₁-receptor

MSN

MSN

C

PD + dys

degenerated DA terminal

5-HT terminals

TH+ neurons DOPAerg

L-DOPA AADC

AADC

dopaminerg L-tyrosine

L-Tyrosin

TH

TH

AADC

dopamine

L-DOPA

increased level of extracellular dopamine

hypersensitive D₁ receptors

MSN ∆ FosB

↑ transcription

↑ activation of direct pathway

Fig. 11. Neuronal plasticity in L-DOPA-induced dyskinesia. Based on our results and on recent literature, this scheme illustrates in a simplified way striatal pre- and postsynaptic plasticity in L-DOPA-treated Parkinson’s disease without dyskinesia (PD + non-dys; B) and with dyskinesia (PD + dys; C) as compared to the physiological state (A). Under normal conditions, nigrostriatal dopamine (DA) terminals produce an adequate amount of dopamine (yellow dots) (A). PD patients suffering DAergic degeneration require substitution of oral L-DOPA (red dot; B). Tyrosine hydroxylase-positive (TH+) neurons can produce additional L-DOPA or DA. L-DOPA is converted to DA by the aromatic acid decarboxylase (AADC) of serotonin (5-HT) terminals and released in an uncontrolled manner. Partly, other AADC containing neurons and non-neuronal cells might contribute to this conversion (not depicted). In the dyskinetic state (C), the number of TH+ neurons and the density of serotonergic fibers are significantly increased. Their cooperative DA synthesis in combination with external L-DOPA administration leads to pulsatile excessive DA concentrations and consecutive sensitization of postsynaptic D1 receptors of striatal medium-sized spiny projection neurons (MSN). Some TH+ cells may be capable to the autonomic production of DA. The resulting overstimulation of postsynaptic signaling cascades and expression of the transcription factor DFosB ultimately leads to overactivity of the direct pathway with L-DOPA-induced dyskinesia. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

315

U. Keber et al. / Neuroscience 298 (2015) 302–317

The role of nucleus accumbens and cortex in LID Our quantitative analysis revealed that accumbal and cortical TH+ neurons were not involved in LID in our study, although we observed a significant reduction of TH+ fibers in the nucleus accumbens and cortex in LD-dys mice when compared to LD-non-dys and salinetreated mice. Interestingly, Halje et al. (2012) described cortical 80-Hz oscillation that appears to be present only in association with dyskinetic symptoms. This suggests a direct causal pathophysiological link, which has also been confirmed in pharmacological experiments (Brown et al., 2002; Brown and Williams, 2005; Alonso-Frech et al., 2006). The authors proposed that a loss of cortical DAergic innervation represents a key predisposing factor for dyskinesia and that the ensuing supersensitivity to DA makes cortical circuits prone to network resonance when exposed to systemic L-DOPA concentrations that are needed to obtain pro-kinetic effects in the basal ganglia. Consequently, future studies should aim to investigate approaches to prevent or interfere with the mechanisms underlying cortical resonant oscillations, e. g. by pharmacological interventions or electrical stimulation of selected neuronal elements. In our experiments, plasticity in the nucleus accumbens was evident as the DFosB+ cell number was decreased in this area in dyskinetic mice as compared to LD-non-dys and saline-treated mice. Smith et al. (2012) showed that DFosB expression was increased in the shell region of the nucleus accumbens in dyskinetic mice. Unfortunately, we did not perform the subregional analysis of the nucleus accumbens as we did for the striatum. As an explanation for the discrepancy of our data with those of Smith et al. (2012), oral behavior, which is at least in part controlled by the nucleus accumbens (Prinssen et al., 1994; Zahm, 1999), might have been false-positively interpreted as orolingual dyskinesia and consecutively increased the AIM score. In concordance with our data, other authors ruled out any role of accumbal DFosB+ cells in LID in rats (Andersson et al., 1999). Interestingly, it has been demonstrated that psychostimulant-induced sensitization can differently regulate DFosB expressions in accumbal subregions (Brenhouse and Stellar, 2006). Here, we also showed an increment of DFosB expression following L-DOPAtreatment. This observation lends support to the hypothesis that L-DOPA stimulates the activity of the nucleus accumbens which is a center of mood and limbic functions and may therefore account for the well-known neuropsychiatric side effects of this drug. Regions of origin for LID subtype exhibition In terms of TH+ cell distribution in the subregions of the striatum, increased TH+ neuronal numbers in lateral compartments were accompanied by the development of axial, limb and to a lower extent of orolingual LID. In contrast, locomotive rotations were associated with a numerical increment of TH+ cells in medio-striatal regions. In line with these results, similar distributions were previously obtained in mice (Lundblad et al., 2004; Pavo´n et al., 2006), rats (Andersson et al., 1999;

Winkler et al., 2002) and dyskinetic PD patients (Lindgren et al., 2011) when analyzing markers of LID, e.g. DFosB immunoreactivity, opioid peptide mRNA expression or externally regulated kinase phosphorylation. These compelling pieces of evidence highlight the rodent lateral striatum, which corresponds to the human putamen, as an accepted region of LID origin and furthermore implicates the essential role of TH+ cells in LID pathogenesis. On the other hand, locomotive rotations seem to be mediated by the medial striatal subregion, which has previously been acknowledged for DA agonist-induced locomotion in rodents (Dickson et al., 1994). Conclusion Our data suggest that TH+ neurons in the denervated murine striatum may be functionally involved in LID pathogenesis. In particular, the lateral striatum represents a key region of presynaptic neuronal plasticity correlating with LID. Fig. 11 illustrates the presumptive mechanisms by which TH+ neurons either autonomically or in cooperation with serotonergic fibers might contribute to the development of LID. Striatal TH+ neurons may therefore serve as a promising target for new antidyskinetic treatment strategies in PD. Future work should be performed to replicate these results in primates. Apart from the striatum, accumbal and cortical TH+ neurons are likely not to be associated with LID, although they might be involved in development of L-DOPA-induced limbic and cortical side effects like psychosis or dementia. Acknowledgments—This study was supported by the University Clinics Giessen and Marburg (UKGM), Germany, and the German Parkinson Society (DPG). We thank Susanne Stei, Sabine Anfimov und Silke Caspari for technical help. WHO is Hertie Senior Research Professor of the Charitable Hertie Foundation, Frankfurt/Main, Germany. GUH is funded by the German Research Foundation (DFG; HO2402/6-2).

REFERENCES Ahlskog JE, Muenter MD (2001) Frequency of levodopa-related dyskinesias and motor fluctuations as estimated from the cumulative literature. Mov Disord 16(3):448–458. Alonso-Frech F, Zamarbide I, Alegre M, Rodrı´ guez-Oroz MC, Guridi J, Manrique M, Valencia M, Artieda J, Obeso JA (2006) Slow oscillatory activity and levodopa-induced dyskinesias in Parkinson’s disease. Brain 129(Pt 7):1748–1757. Andersson M, Hilbertson A, Cenci MA (1999) Striatal fosB expression is causally linked with l-DOPA-induced abnormal involuntary movements and the associated upregulation of striatal prodynorphin mRNA in a rat model of Parkinson’s disease. Neurobiol Dis 6(6):461–474. Arai R, Karasawa N, Geffard M, Nagatsu I (1995) L-DOPA is converted to dopamine in serotonergic fibers of the striatum of the rat: a double-labeling immunofluorescence study. Neurosci Lett 195(3):195–198. Bernheimer H, Birkmayer W, Hornykiewicz O, Jellinger K, Seitelberger F (1973) Brain dopamine and the syndromes of Parkinson and Huntington. Clinical, morphological and neurochemical correlations. J Neurol Sci 20(4):415–455. Berton O, Guigoni C, Li Q, Bioulac BH, Aubert I, Gross CE, Dileone RJ, Nestler EJ, Bezard E (2009) Striatal overexpression of

316

U. Keber et al. / Neuroscience 298 (2015) 302–317

DeltaJunD resets L-DOPA-induced dyskinesia in a primate model of Parkinson disease. Biol Psychiatry 66(6):554–561. Betarbet R, Turner R, Chockkan V, DeLong MR, Allers KA, Walters J, Levey AI, Greenamyre JT (1997) Dopaminergic neurons intrinsic to the primate striatum. J Neurosci 17(17):6761–6768. Birkmayer W, Hornykiewics O (1961) The L-3,4-dioxyphenylalanine (DOPA)-effect in Parkinson-akinesia. Wien Klin Wochenschr 73:787–788. Brenhouse HC, Stellar JR (2006) C-Fos and deltaFosB expression are differentially altered in distinct subregions of the nucleus accumbens shell in cocaine-sensitized rats. Neuroscience 137(3):773–780. Brown P, Williams D (2005) Basal ganglia local field potential activity: character and functional significance in the human. Clin Neurophysiol 116(11):2510–2519. Brown P, Kupsch A, Magill PJ, Sharott A, Harnack D, Meissner W (2002) Oscillatory local field potentials recorded from the subthalamic nucleus of the alert rat. Exp Neurol 177(2):581–585. Calon F, Dridi M, Hornykiewicz O, Be´dard PJ, Rajput AH, Di Paolo T (2004) Increased adenosine A2A receptors in the brain of Parkinson’s disease patients with dyskinesias. Brain 127:1075–1084. Cao X, Yasuda T, Uthayathas S, Watts RL, Mouradian MM, Mochizuki H, Papa SM (2010) Striatal overexpression of DeltaFosB reproduces chronic levodopainduced involuntary movements. J Neurosci 30:7335–7343. Carlsson T, Carta M, Winkler C, Bjo¨rklund A, Kirik D (2007) Serotonin neuron transplants exacerbate L-DOPA-induced dyskinesias in a rat model of Parkinson’s disease. J Neurosci 27(30):8011–8022. Carta M, Bezard E (2011) Contribution of pre-synaptic mechanisms to L-DOPA-induced dyskinesia. Neuroscience 198:245–251. Carta M, Carlsson T, Kirik D, Bjo¨rklund A (2007) Dopamine released from 5-HT terminals is the cause of L-DOPA-induced dyskinesia in parkinsonian rats. Brain 130(Pt 7):1819–1833. Darmopil S, Mun˜eto´n-Go´mez VC, de Ceballos ML, Bernson M, Moratalla R (2008) Tyrosine hydroxylase cells appearing in the mouse striatum after dopamine denervation are likely to be projection neurones regulated by L-DOPA. Eur J Neurosci 27(3):580–592. de la Fuente-Ferna´ndez R, Sossi V, Huang Z, Furtado S, Lu JQ, Calne DB, Ruth TJ, Stoessl AJ (2004) Levodopa-induced changes in synaptic dopamine levels increase with progression of Parkinson’s disease: implications for dyskinesias. Brain 127(Pt. 12):2747–2754. Depboylu C (2014) Non-serine-phosphorylated tyrosine hydroxylase expressing neurons are present in mouse striatum, accumbens and cortex that increase in number following dopaminergic denervation. J Chem Neuroanat 56C:35–44. Depboylu C, Schorlemmer K, Klietz M, Oertel WH, Weihe E, Ho¨glinger GU, Scha¨fer MK (2011) Upregulation of microglial C1q expression has no effects on nigrostriatal dopaminergic injury in the MPTP mouse model of Parkinson disease. J Neuroimmunol 236(1–2):39–46. Depboylu C, Klietz M, Maurer L, Oertel WH, Kobayashi K, Weihe E, Ho¨glinger GU, Scha¨fer MK-H (2014) Transcriptional and structural plasticity of tyrosine hydroxylase expressing neurons in both striatum and nucleus accumbens following dopaminergic denervation. J Chem Neuroanat 61–62:169–175. DiCaudo C, Riverol M, Mundinano IC, Ordonez C, Herna´ndez M, Marcilla L, Luquin MR (2012) Chronic levodopa administration followed by a washout period increased number and induced phenotypic changes in striatal dopaminergic cells in MPTPmonkeys. PLoS One 7(11):e50842. Dickson PR, Lang CG, Hinton SC, Kelley AE (1994) Oral stereotypy induced by amphetamine microinjection into striatum: an anatomical mapping study. Neuroscience 61(1):81–91. Di Monte DA, McCormack A, Petzinger G, Janson AM, Quik M, Langston WJ (2000) Relationship among nigrostriatal denervation, parkinsonism, and dyskinesias in the MPTP primate model. Mov Disord 15:459–466.

Du X, Stull ND, Iacovitti L (1995) Brain-derived neurotrophic factor works coordinately with partner molecules to initiate tyrosine hydroxylase expression in striatal neurons. Brain Res 680:229–233. Engeln M, Bastide MF, Toulme´ E, Dehay B, Bourdenx M, Doudnikoff E, Li Q, Gross CE, Boue´-Grabot E, Pisani A, Bezard E, Fernagut PO (2014) Selective Inactivation of Striatal FosB/DFosBExpressing Neurons Alleviates L-Dopa-Induced Dyskinesia. Biol Psychiatry. Eskow KL, Dupre KB, Barnum CJ, Dickinson SO, Park JY, Bishop C (2009) The role of the dorsal raphe nucleus in the development, expression, and treatment of L-DOPA-induced dyskinesia in hemiparkinsonian rats. Synapse 63(7):610–620. Espadas I, Darmopil S, Vergan˜o-Vera E, Ortiz O, Oliva I, VicarioAbejo´n C, Martı´ n ED, Moratalla R (2012) L-DOPA-induced increase in TH-immunoreactive striatal neurons in parkinsonian mice: insights into regulation and function. Neurobiol Dis 48(3):271–281. Fasano S, Bezard E, D’Antoni A, Francardo V, Indrigo M, Qin L, Dovero´ S, Cerovic M, Cenci MA, Brambilla R (2010) Inhibition of Ras-guanine nucleotide-releasing factor 1 (Ras-GRF1) signaling in the striatum reverts motor symptoms associated with L-dopa-induced dyskinesia. Proc Natl Acad Sci U S A 107(50): 21824–21829. Feyder M, Bonito-Oliva A, Fisone G (2011) L-DOPA-Induced dyskinesia and abnormal signaling in striatal medium spiny neurons: focus on dopamine D1 receptor-mediated transmission. Front Behav Neurosci 5:71. Francardo V, Recchia A, Popovic N, Andersson D, Nissbrandt H, Cenci MA (2011) Impact of the lesion procedure on the profiles of motor impairment and molecular responsiveness to L-DOPA in the 6-hydroxydopamine mouse model of Parkinson’s disease. Neurobiol Dis 42(3):327–340. Halje P, Tamte` M, Richter U, Mohammed M, Cenci MA, Petersson P (2012) Levodopa-induced dyskinesia is strongly associated with resonant cortical oscillations. J Neurosci 32(47):16541–16551. Hardebo JE, Emson PC, Falck B, Owman C, Rosengren E (1980) Enzymes related to monoamine transmitter metabolism in brain microvessels. J Neurochem 35(6):1388–1393. Hirsch E, Graybiel AM, Agid YA (1988) Melanized dopaminergic neurons are differentially susceptible to degeneration in Parkinson’s disease. Nature 334(6180):345–348. Huang YZ, Rothwell JC, Lu CS, Chuang WL, Chen RS (2011) Abnormal bidirectional plasticity-like effects in Parkinson’s disease. Brain 134(Pt 8):2312–2320. Huot P, Parent A (2007) Dopaminergic neurons intrinsic to the striatum. J Neurochem 101:1441–1447. Huot P, Le´vesque M, Parent A (2007) The fate of striatal dopaminergic neurons in Parkinson’s disease and Huntington’s chorea. Brain 130(Pt 1):222–232. Huot P, Le´vesque M, Morissette M, Calon F, Dridi M, Di Paolo T, Parent A (2008) L-DOPA treatment abolishes the numerical increase in striatal dopaminergic neurons in parkinsonian monkeys. J Chem Neuroanat 35:77–84. Jollivet C, Montero-Menei CN, Venier-Julienne MC, Sapin A, Benoit JP, Menei P (2004) Striatal tyrosine hydroxylase immunoreactive neurons are induced by L-dihydroxyphenylalanine and nerve growth factor treatment in 6-hydroxydopamine lesioned rats. Neurosci Lett 362(2):79–82. Juorio AV, Li XM, Walz W, Paterson IA (1993) Decarboxylation of LDOPA by cultured mouse astrocytes. Brain Res 626(1– 2):306–309. Kumer SC, Vrana KE (1996) Intricate regulation of tyrosine hydroxylase activity and gene expression. J Neurochem 67:443–462. Linazasoro G (2005) New ideas on the origin of L-dopa-induced dyskinesias: age, genes and neural plasticity. Trends Pharmacol Sci 26:391–397. Linazasoro G, Antonini A, Maguire RP, Leenders KL (2004) Pharmacological and PET studies in patient’s with Parkinson’s disease and a short duration-motor response: implications in the

U. Keber et al. / Neuroscience 298 (2015) 302–317 pathophysiology of motor complications. J Neural Transm 111:497–509. Lindgren HS, Rylander D, Ohlin KE, Lundblad M, Cenci MA (2007) The ‘‘motor complication syndrome’’ in rats with 6-OHDA lesions treated chronically with L-DOPA: relation to dose and route of administration. Behav Brain Res 177:150–159. Lindgren HS, Andersson DR, Lagerkvist S, Nissbrandt H, Cenci MA (2010) L-DOPA-induced dopamine efflux in the striatum and the substantia nigra in a rat model of Parkinson’s disease: temporal and quantitative relationship to the expression of dyskinesia. J Neurochem 112(6):1465–1476. Lindgren HS, Rylander D, Iderberg H, Andersson M, O’Sullivan SS, Williams DR, Lees AJ, Cenci MA (2011) Putaminal upregulation of FosB/DFosB-like immunoreactivity in Parkinson’s disease patients with dyskinesia. J Parkinsons Dis 1(4):347–357. Lopez-Real A, Rodriguez-Pallares J, Guerra MJ, Labandeira-Garcia JL (2003) Localization and functional significance of striatal neurons immunoreactive to aromatic L-amino acid decarboxylase or tyrosine hydroxylase in rat Parkinsonian models. Brain Res 969(1–2):135–146. Lundblad M, Andersson M, Winkler C, Kirik D, Wierup N, Cenci MA (2002) Pharmacological validation of behavioural measures of akinesia and dyskinesia in a rat model of Parkinson’s disease. Eur J Neurosci 15(1):120–132. Lundblad M, Picconi B, Lindgren H, Cenci MA (2004) A model of LDOPA-induced dyskinesia in 6-hydroxydopamine lesioned mice: relation to motor and cellular parameters of nigrostriatal function. Neurobiol Dis 16(1):110–123. Lundblad M, Usiello A, Carta M, Hakansson K, Fisone G, Cenci MA (2005) Pharmacological validation of a mouse model of l-DOPAinduced dyskinesia. Exp Neurol 194:66–75. Maeda T, Nagata K, Yoshida Y, Kannari K (2005) Serotonergic hyperinnervation into the dopaminergic denervated striatum compensates for dopamine conversion from exogenously administered l-DOPA. Brain Res 1046(1–2):230–233. Melamed E, Hefti F, Wurtman RJ (1980) Nonaminergic striatal neurons convert exogenous L-DOPA to dopamine in parkinsonism. Ann Neurol 8(6):558–563. Meredith GE, Farrell T, Kellaghan P, Tan Y, Zahm DS, Totterdell S (1999) Immunocytochemical characterization of catecholaminergic neurons in the rat striatum following dopamine-depleting lesions. Eur J Neurosci 11(10):3585–3596. Mun˜oz A, Carlsson T, Tronci E, Kirik D, Bjo¨rklund A, Carta M (2009) Serotonin neuron-dependent and -independent reduction of dyskinesia by 5-HT1A and 5-HT1B receptor agonists in the rat Parkinson model. Exp Neurol 219(1):298–307. Navailles S, Bioulac B, Gross C, De Deurwaerde`re P (2010) Serotonergic neurons mediate ectopic release of dopamine induced by L-DOPA in a rat model of Parkinson’s disease. Neurobiol Dis 38(1):136–143. Okazawa H, Murata M, Watanabe M, Kamei M, Kanazawa I (1992) Dopaminergic stimulation up-regulates the in vivo expression of brain-derived neurotrophic factor (BDNF) in the striatum. FEBS Lett 313(2):138–142. Pavo´n N, Martı´ n AB, Mendialdua A, Moratalla R (2006) ERK phosphorylation and FosB expression are associated with LDOPA-induced dyskinesia in hemiparkinsonian mice. Biol Psychiatry 59(1):64–74. Paxinos G, Franklin KB (2001) The mouse brain in stereotaxic coordinates. San Diego, California: Academic Press. Politis M, Wu K, Loane C, Brooks DJ, Kiferle L, Turkheimer FE, Bain P, Molloy S, Piccini P (2014) Serotonergic mechanisms responsible for levodopa-induced dyskinesias in Parkinson’s disease patients. J Clin Invest 124(3):1340–1349.

317

Porritt MJ, Batchelor PE, Hughes AJ, Kalnins R, Donnan GA, Howells DW (2000) New dopaminergic neurons in Parkinson’s disease striatum. Lancet 356(9223):44–45. Prinssen EP, Balestra W, Bemelmans FF, Cools AR (1994) Evidence for a role of the shell of the nucleus accumbens in oral behavior of freely moving rats. J Neurosci 14(3 Pt 2):1555–1562. Rylander D, Parent M, O’Sullivan SS, Dovero S, Lees AJ, Bezard E, Descarries L, Cenci MA (2010) Maladaptive plasticity of serotonin axon terminals in levodopa-induced dyskinesia. Ann Neurol 68(5):619–628. Schallert T, Tillerson JL. Intervention strategies for degeneration of dopamine neurons in Parkinsonism: optimizing behavioral assessment of outcome. In: Central Nervous System Disease: Innovative Animal Models from Lab to Clinic (Emerich, D. F. et al., eds), pp 131–151 Totowa, New Jersey 07512: Humana Press Inc., 1999. Smith GA, Heuer A, Dunnett SB, Lane EL (2012a) Unilateral nigrostriatal 6-hydroxydopamine lesions in mice II: predicting l-DOPA-induced dyskinesia. Behav Brain Res 226(1): 281–292. Smith Y, Wichmann T, Factor SA, DeLong MR (2012b) Parkinson’s disease therapeutics: new developments and challenges since the introduction of levodopa. Neuropsychopharmacology 37(1):213–246. Tanaka H, Kannari K, Maeda T, Tomiyama M, Suda T, Matsunaga M (1999) Role of serotonergic neurons in L-DOPA-derived extracellular dopamine in the striatum of 6-OHDA-lesioned rats. NeuroReport 10(3):631–634. Tande´ D, Ho¨glinger G, Debeir T, Freundlieb N, Hirsch EC, Franc¸ois C (2006) New striatal dopamine neurons in MPTP-treated macaques result from a phenotypic shift and not neurogenesis. Brain 129(Pt 5):1194–1200. Tashiro Y, Sugimoto T, Hattori T, Uemura Y, Nagatsu I, Kikuchi H, Mizuno N (1989) Tyrosine hydroxylase-like immunoreactive neurons in the striatum of the rat. Neurosci Lett 97(1–2):6–10. Ugrumov MV (2013) Brain neurons partly expressing dopaminergic phenotype: location, development, functional significance, and regulation. Adv Pharmacol 68:37–91. Ulusoy A, Sahin G, Kirik D (2010) Presynaptic dopaminergic compartment determines the susceptibility to L-DOPA-induced dyskinesia in rats. Proc Natl Acad Sci USA 107:13159–13164. Valastro B, Andersson M, Lindgren HS, Cenci MA (2007) Expression pattern of JunD after acute or chronic L-DOPA treatment: comparison with deltaFosB. Neuroscience 144:198–207. Weihe E, Depboylu C, Schu¨tz B, Scha¨fer MK, Eiden LE (2006) Three types of tyrosine hydroxylase-positive CNS neurons distinguished by dopa decarboxylase and VMAT2 co-expression. Cell Mol Neurobiol 26(4–6):659–678. West MJ (1999) Stereological methods for estimating the total number of neurons and synapses: issues of precision and bias. Trends Neurosci 22(2):51–61. Winkler C, Kirik D, Bjo¨rklund A, Cenci MA (2002) L-DOPA-induced dyskinesia in the intrastriatal 6-hydroxydopamine model of Parkinson’s disease: relation to motor and cellular parameters of nigrostriatal function. Neurobiol Dis 10(2):165–186. Zahm DS (1999) Functional-anatomical implications of the nucleus accumbens core and shell subterritories. Ann N Y Acad Sci 877:113–128. Zeng BY, Iravani MM, Jackson MJ, Rose S, Parent A, Jenner P (2010) Morphological changes in serotoninergic neurites in the striatum and globus pallidus in levodopa primed MPTP treated common marmosets with dyskinesia. Neurobiol Dis 40(3):599–607.

(Accepted 12 April 2015) (Available online 16 April 2015)