Neurobiology of Disease 43 (2011) 576–587

Contents lists available at ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r.com/ locate /ynbd i

Impact of dopamine to serotonin cell ratio in transplants on behavioral recovery andL-DOPA-induced dyskinesia

Joanna García a,b, Thomas Carlsson c,1, Máté Döbrössy b, Guido Nikkhah b, Christian Winkler a,⁎a Department of Neurology, University Hospital Freiburg, 79106 Freiburg, Germanyb Department of Stereotactic Neurosurgery, Laboratory of Molecular Neurosurgery, University Hospital Freiburg, 79106 Freiburg, Germanyc Experimental Neurology, Department of Neurology, Philipps-University Marburg, 35043 Marburg, Germany

⁎ Corresponding author at: Department of NeurologyBreisacher Str. 64, D-79106 Freiburg, Germany. Fax: +4

E-mail address: christian.winkler@uniklinik-freiburg1 Present address: Institute for Neurophysiology,

University Frankfurt, 60590 Frankfurt am Main, GermanAvailable online on ScienceDirect (www.scienced

0969-9961/$ – see front matter © 2011 Elsevier Inc. Aldoi:10.1016/j.nbd.2011.05.004

a b s t r a c t

a r t i c l e i n f o

Article history:Received 23 February 2011Revised 30 April 2011Accepted 5 May 2011Available online 13 May 2011

Keywords:Parkinson's diseaseCell transplantationMotor behaviorL-DOPA-induced dyskinesiaSerotonin

Fetal dopamine (DA) cell transplantation has shown to be efficient in reversing behavioral impairmentsassociated with Parkinson's disease. However, the beneficial effects on motor behavior and L-DOPA-induceddyskinesia have varied greatly in between clinical trials and patients within the same trial. Recently, theinclusion of serotonin (5-HT) neurons in the grafted tissue has been suggested to play an important negativerole, in particular, on the effect of L-DOPA-induced dyskinesia. In the present study we have evaluated theinfluence of different ratios of DA neurons in relation to 5-HT neurons in the graft on spontaneous motorbehavior and L-DOPA-induced dyskinesia in a rat model of Parkinson's disease. We show that using thestandard dissection method that gives rise to a DA:5-HT ratio in the graft of 2:1 to 1:2 there is significant andconsistent improvement in spontaneous motor behavior and reversal of L-DOPA-induced dyskinesia.Increasing the ratio of 5-HT neurons in the graft, to a DA:5-HT ratio of in between 1:3 and 1:10, still inducessignificant reduction of L-DOPA-induced dyskinesia, suggesting that the detrimental effect of 5-HT neurons onL-DOPA-induced dyskinesia is prevented even by small numbers of DA neurons in the graft. Nonetheless,while the post-synaptic responses were normalized following peripheral L-DOPA delivery in animals with lowDA:5-HT ratio, we observed a pharmacological indication of hyperactive pre-synaptic response in theseanimals. These data suggests that 5-HT cells within a graft are neither detrimental nor beneficial for functionaleffects of DA-rich transplants; however, in absence of sufficient numbers of DA neurons, the 5-HT neuronsmay induce negative effects following L-DOPA therapy. In summary, our data indicate that for future clinicaltrials the inclusion of 5-HT neurons in grafted tissue is not critical as long as there are sufficient numbers of DAcells in the graft.

, University Hospital Freiburg,9 761 27053900..de (C. Winkler).Neuroscience Center, Goethey.irect.com).

l rights reserved.

© 2011 Elsevier Inc. All rights reserved.

Introduction

Transplantation of dopamine (DA) neurons, derived from embry-onic ventral mesencephalon (VM) has shown to be efficient inreversing behavioral deficits and to reduce the need for pharmaco-therapy in patients with Parkinson's disease (PD) (Winkler et al.,2005). The results have, however, highly varied in between patientsand clinical trials, and two double-blind trials have shown improve-ments in subgroups of patients but failed to reach an overallsignificant behavioral recovery (Freed et al., 2001; Ma et al., 2010;Olanow et al., 2003, 2009). Fetal cell transplantation has also shownvarious effects on the expression of L-DOPA-induced dyskinesia, oneof the most common side effects of pharmacotherapy in PD. Some

patients have thus shown good benefit from cell therapy, while inothers no change or even worsening in L-DOPA-induced abnormalinvoluntary movements (AIMs) has been reported (Hagell and Cenci,2005).

Several parameters of the transplant protocols have been putforward to play a role for the functional outcome after fetal DA celltransplantation and to explain the discrepancies within patients andtrials, including tissue preparation and graft composition (Brundinet al., 2010; Lane et al., 2010; Winkler et al., 2005). Recently, theinclusion of serotonin (5-HT) neurons in the grafts has emerged as acritical factor for the effect on treatment-induced dyskinesia. 5-HTneurons have been shown to possess the cellularmachinery to take upperipheral L-DOPA, convert it to DA, store it in vesicles and release it inan activity dependent manner (Arai et al., 1994; Carta et al., 2010;Maeda et al., 2005; Navailles and De Deurwaerdère, 2011; Ng et al.,1970, 1971; Tanaka et al., 1999). However, in contrast to DA neuronsthese cells lack the auto-regulatory feedback by activation of D2

receptors in order to control the release of its “false transmitter”DA. Infact, the serotonin system has shown to be a major player in inductionand maintenance of L-DOPA-induced dyskinesias in animal models of

577J. García et al. / Neurobiology of Disease 43 (2011) 576–587

PD (Carta et al., 2010). Thus, Carta et al.(2007) have shown that L-DOPA-induced dyskinesia can be blocked in the rat PD model byremoving the intrinsic 5-HT system by toxic lesions or by pharma-cologically blocking the 5-HT release. In addition, it has recently beenshown that serotonin grafts can significantly worsen L-DOPA-induceddyskinesia in the rat PDmodel (Carlsson et al., 2007, 2009). This effectdeveloped over time as the graft matured and the detrimental effectwas present both as an increase in severity of dyskinesias as well as aprolonged dyskinetic response after a single dose of L-DOPA (Carlssonet al., 2007). In addition, dyskinesia in the 5-HT grafted animals wasevident already after partial DA-denervating lesions and drasticallyworsened as the DA lesion extended suggesting a protective role ofthe DA system on the 5-HT-effect on dyskinesia (Carlsson et al., 2009).Since fetal grafts in PD patients may contain large numbers 5-HTneurons (Mendez et al., 2008; Politis et al., 2010), and sincedegeneration of the DA system will progress in these patients, theseexperimental data clearly indicate the importance of the 5-HT systemin the outcome of fetal cell transplantation, and the need for betterunderstanding the role of grafted 5-HT cells.

The effects of transplanted 5-HT neurons in animal PD models,however, have only been evaluated using “pure” 5-HT grafts andcompared to “regular” grafts, containing DA and 5-HT neurons at a ratioof approximately 1:1. Using the rat model of PD, we investigated theeffects of increasing the ratio of 5-HT neurons in relation to DAneurons inthe graft from approximately 1:1 to 10:1 on functional changes inspontaneous motor behavior and L-DOPA-induced dyskinesia. We showthat grafts containing sufficient numbers of DA neurons (more than 600cells) inducing DA-reinnervation to at least 20% of normal in thecaudolateral part of the striatum will always improve spontaneousmotor functionand reduce L-DOPA-induceddyskinesia bymore than50%,irrespective of the 5-HT-neuron number or the DA:5-HT ratio. Also in thegroup of animals with a DA:5-HT ratio between 1:3 and 1:10 there wassignificant reduction of L-DOPA-induced dyskinesia despite these animalsnot improving spontaneous motor function due to low DA neuronnumbers, thus indicating that small numbers of grafted DA neurons aresufficient to block the effects of 5-HT neurons on the expressiondyskinesia. Nonetheless, an increased ratio of 5-HT neurons versus DAcells in the grafts may aggravate hyperactive pre-synaptic activity, afterpharmacological manipulation, manifested as a strong tendency to

Fig. 1. Experimental design and individual animal rank order. (A) After the unilateral 6-hydanimals according to performance in amphetamine-induced rotation and cylinder test, animarea). Animals with moderate to severe dyskinesia were allocated and balanced in 3 groups pinduced rotation and cylinder test. Animals received single cell suspension grafts containing(VM) and the raphe nucleus of E14 rat embryos (High DA ratio and Low DA ratio groups), oregimen of twice-weekly L-DOPA (maintenance phase, light-gray area). Grafts effects oparaformaldehyde (PFA) for histological analysis. (B) Animals with VM grafts showed a TH:5cells but additional cells from the raphe nucleus showed a TH:5-HT ratio between 1:3 and 1:animals in the High DA ratio group rather had a low TH:5-HT ratio between 1:3 and 1:10 a

increase in L-DOPA-induced rotation, and an unchanged apomorphine-induced response.

Materials and methods

Subjects

Adult female Sprague–Dawley rats (Charles River, Sulzfeld,Germany) weighing 225 g at the beginning of the experiment wereused in this study. Animals were housed under a 12 h light/dark cyclewith water and food ad libitum. The experiments were carried outaccording to the ethical guidelines set by the ethical board of theAlbert-Ludwigs University of Freiburg and the RegierungspräsidiumFreiburg, Germany.

Experimental design

All animals received unilateral injections of 6-OHDA into themedialforebrain bundle (MFB) in order to obtain a complete DA depletion ofthe nigrostriatal pathway (see experimental timeline in Fig. 1A). Theextent of the lesion was characterized by amphetamine-inducedrotation and in the cylinder test at 4 and 5 weeks after the lesion,respectively. Animals exhibiting more than 5 full body turns/mintowards the side of DA deficiency in amphetamine rotation and b30%contralateral paw use in the cylinder test were selected for the study,and subsequently treated chronically with 3,4-dihydroxyphenylalanine(L-DOPA) in combination with the peripheral decarboxylase inhibitorbenserazide hydrochloride daily for 4 weeks. During this period (L-DOPA-induced dyskinesia induction phase in Fig. 1A), the animals wererepeatedly tested for the expression of abnormal involuntary move-ments (AIMs) until dyskinesia scores had reached a plateau. Animalswith moderate to severe dyskinesia were selected and allocated in 3groups, before grafting, according to their dyskinesia scores andperformance in amphetamine-induced rotation and cylinder test. Twoof the groups received single cell suspension grafts containing amixtureof DA and 5-HT neuroblasts, obtained from the ventral mesencephalon(VM) and the raphe nucleus of 14 day old (E14) rat embryos. The firstgrafted group (High DA ratio) received cells of a regular VM dissection,where according to our previous studies part of the developing upper

roxydopamine (6-OHDA) lesion of the medial forebrain bundle (MFB) and selection ofals received daily injections of L-DOPA to induce dyskinesia (induction phase, dark-grayrior to grafting according to their dyskinesia scores and performance in amphetamine-different ratios of DA and 5-HT neuroblasts obtained from the ventral mesencephalonr sham operation (Les-Ctrl group). After grafting animals were kept on a maintenancen motor behaviors and dyskinesia were repeatedly assessed before perfusion with-HT cell ratio between 3:1 and 1:3 (High DA ratio group), whereas animals with less VM10 (postmortem analysis of numbers of grafted DA and 5-HT neurons showed that twond were therefore moved to the Low DA ratio group for further analysis).

578 J. García et al. / Neurobiology of Disease 43 (2011) 576–587

pontine raphe nucleus is included, and the ratio between DA and 5HTcells in the graft would be expected to range between 2:1 and 1:2. Thesecond group (Low DA ratio) received more 5-HT neuroblasts ascompared to DA neuroblasts by adding additional 5-HT neuroblastsfrom the pontine raphe nucleus to the cell suspension. Afterpostmortem analysis of numbers of grafted DA and 5-HT neurons, thegroups were finally defined according to DA:5-HT ratio with a cutbetween the groups at more or less than a 1:3 DA:5-HT cell ratio. Thus,two animals that were originally assigned to the High DA ratio groupwere moved to the Low DA ratio group because of DA:5-HT cell ratio ofless than 1:3, giving a total of n=19 in the High DA ratio, and n=21 inthe Low DA ratio group (see individual rank order in Fig. 1B). The thirdgroup was sham-operated and included as lesioned but non-graftedcontrol animals.

From one week following transplantation and throughout thefurther experiment, L-DOPA was injected twice per week in order tomaintain L-DOPA-induced dyskinesia (L-DOPA-induced dyskinesiamaintenance phase in Fig. 1A) and AIMs were repeatedly evaluated.Spontaneous forelimb use in the cylinder test was evaluated at 3 and15 weeks after transplantation. L-DOPA-induced rotation was per-formed at 16 weeks and apomorphine-induced rotation and dyski-nesias were evaluated at 19 and 20 weeks after transplantation,respectively. Finally, the animals were perfused for histologicalanalysis at 25 weeks post grafting.

6-OHDA lesion

Lesion surgery was performed under anesthesia with ketamine (10%;0.1 ml/kg bodyweight; Essex, Munich, Germany) and rompun (2%;0.01 ml/kg bodyweight; Bayer, Leverkusen, Germany). In order to obtaina complete lesion of the nigrostriatal DA system, all animals received twoinjections of 2.5 μl and 3 μl 6-OHDA (3.6 μg/μl in 0.2% L-ascorbic acid-saline, Sigma Aldrich, Steinheim, Germany) into the right MFB using astereotactic frame(Stoelting,WoodDale, IL)witha10 μl Hamilton syringeattached (Winkler et al., 1999). The anteroposterior (AP) andmediolateral(ML) coordinates were set according to bregma, and the dorsoventral(DV) coordinate according to the dura (Paxinos and Watson, 2005): (1),AP: −4.4 mm, ML: −1.2 mm, DV: −7.8 mm, toothbar: −2.3 mm; and(2), AP: −4.0 mm, ML: −0.8 mm, DV: −8.0 mm, toothbar: +3.4 mm.The injection ratewas 1.0 μl/min and the cannula was kept in place for anadditional 4 min before it was slowly retracted.

Dissection and transplantation procedure

Fetal cells for all transplantations were dissected from E14 ratembryos. To obtain a cell suspension with a high DA:5HT ratio aregular VM dissection was performed with more stringent definitionof the caudal cut (Carlsson et al., 2007): A first rostral cut wasperformed caudal of the thalamus followed by a midline cut throughthe tectum. The neural tube was folded open so the inside was facingupwards and a caudal cut was made slightly caudal of therhombencephalic isthmus in order to include the entire VM and themost rostral part of the pontine raphe region. According to ourprevious studies in a regular VM dissection even without opening ofthe neural tube, this caudal cut is always slightly caudal of therhombencephalic isthmus. Two lateral cuts were then performed toinclude only the ventral third of the neural tube into the dissectionand the standard butterfly-shape of the VMwas obtained. For the LowDA ratio group, animals received grafts of VM as described above pluscells obtained from the pontine raphe nucleus. For dissection of thepontine raphe nucleus a dorsal midline cut was performed throughthe pontine tectum, and the rhombencephalic isthmus and thepontine flexure were identified. The pontine raphe region was furtherdissected by cutting out a rectangular block of 0.5 mm on either sideof the ventral midline between these two landmarks and the rostralhalf was used for preparation of the cell suspension.

Tissue pieces were processed by incubation in Dulbecco's ModifiedEagle Medium (DMEM) containing 0.1% trypsin and 0.05% DNase(both Sigma Aldrich, Steinheim, Germany) for 30 min at 37 °C,mechanically dissociated into single-cell suspensions and concentrat-ed by centrifugation at 500 rpm for 5 min (Nikkhah et al., 1994, 2000).The total amount of neuroblasts in the cell suspensions was countedby trypan blue viability staining using a counting chamber andvolume of the suspensions was adjusted so that the final concentra-tion was 130,000 cells/μl. The viabilities of the cell suspensions priorto transplantation were N98% for all cell suspensions. For transplan-tation in the Low DA ratio group cells from VM and raphe dissectionwere mixed at a 1:1 ratio.

Transplantation surgery was performed under anesthesia withketamine and rompun, using a 5 μl Hamilton syringe fitted to a glasscapillary (outer diameter of 50–70 μm) and a stereotactic frame (Carlssonet al., 2007). The animals received an injection of 130,000 cells in thestriatum at the following coordinates: AP: +0.2 mm, ML:−3.5 mm andtoothbar: 0.0 mm. The injection was made as two 0.5 μl deposits at DV:−5.0 mm and −4.0 mm below dura. Cells were injected at a speed of0.5 μl/min and after the injection the needle was kept in place for anadditional 4 min before it was slowly retracted. At the end of thetransplantation session, the viability of all cell suspensions was estimatedto be N95% using trypan blue dye exclusion.

Behavioral analysis

Drug-induced rotationRotational behavior induced by amphetamine (2.5 mg/kg i.p.;

Sigma Aldrich, Steinheim, Germany), apomorphine (0.05 mg/kg s.c.;plus 0.2% L-ascorbic acid-saline; Sigma Aldrich, Steinheim, Germany),or L-DOPA (6 mg/kg s.c.; Research Organics, Cleveland, Ohio; plus10 mg/kg benserazide hydrochloride s.c.; Sigma Aldrich, Steinheim,Germany) was monitored in automated rotometer boxes (Ungerstedtand Arbuthnott, 1970). The tests were performed over 90 min. Alldata is expressed as total net full-body turns per minute, and apositive value indicates rotations ipsilateral to the lesion side andnegative value contralateral rotations. Amphetamine-induced rota-tion was performed at 4 weeks after the 6-OHDA lesion to evaluatethe extension of the lesion, and at 18 weeks after transplantation toconfirm the presence of surviving functional DA neurons in the grafts.L-DOPA-induced rotation was performed at 16 weeks and apomor-phine-induced rotation at 19 weeks after transplantation. Apomor-phine-induced rotation was performed late in the course of theexperiment in order not to interfere with graft-induced functionalrecovery or L-DOPA treatment by stimulation of DA receptors.

Cylinder test (spontaneous forelimb use)Spontaneous forelimb use in the cylinder test was analyzed at

5 weeks after the 6-OHDA lesion and at 3 and 15 weeks aftertransplantation (Schallert and Tillerson, 1999; Breysse et al., 2007;Winkler et al., 2006). The animals were placed individually in aplexiglass cylinder (diameter: 20 cm), in which they could movefreely while they were videotaped for 4 min, and later scored by arater blinded to the identity of the animals. The number of touchesusing either the left or right forelimb was counted for each animal,and the data is presented as use of the contralateral (impaired) paw inpercentage of total touches.

DyskinesiaDyskinesia was induced by daily injections of L-DOPA (6 mg/kg) in

combination with benserazide hydrochloride (10 mg/kg) for 4 weeks.L-DOPA and benserazide were mixed and dissolved in physiologicalsaline and given to each rat as an s.c. injection. The animals wereplaced individually in empty transparent plastic cages and evaluatedevery 20 min for 180 min. L-DOPA-induced AIMs were classifiedaccording to their topographic distribution as limb, orolingual, axial,

579J. García et al. / Neurobiology of Disease 43 (2011) 576–587

and locomotive behaviors as described previously (Cenci et al., 1998;Winkler et al., 2002). The frequency of the dyskinetic behaviors wasassessed using scores of 0–4 as follows: 0: absent; 1: occasional (i.e.,present during b50% of the observation time); 2: frequent (i.e.,present during N50% of the observation time); 3: continuous, butinterrupted by repeated sensory stimuli (e.g., sudden noise, openingof the cage lid); 4: continuous, not interrupted by repeated sensorystimuli.

In addition, the amplitude of axial and limb dyskinesia was alsoassessed. The axial component was rated according to the torsion ofneck and trunk from the longitudinal axis of the body: 1: consistentlateral deviation of head and neck up to 30°; 2: lateral deviation ofhead and neck between 30° and 60°; 3: lateral deviation and/ortorsion of head, neck and upper trunk between 60° and 90°; 4: torsionof head, neck and trunk at an angle N90°, causing the rat to losebalance. Limb amplitude was rated based on the magnitude of paw/limb translocation and visible involvement of distal versus proximalmuscle groups: 1: tiny oscillatory movements of the paw and thedistal forelimb around a fixed position; 2: movements of lowamplitude but causing visible translocation of both distal andproximal limb; 3: translocation of the whole limb with visiblecontraction of shoulder muscles; 4: vigorous limb and shouldermovements of maximal amplitude. Normal behaviors such asgrooming, gnawing, rearing, and sniffing were not included in therating.

AIM scores were assessed every 4 days during the induction periodand the pre-grafting score is the average of the last two consecutive tests.Post grafting, dyskinesiawas evaluated at 2, 4, 12, 16 and24 weeks,wherescores at 2, 4 and 12 weeks are from single observations and scores at 16and24 weeks are the average of two test sessions, 4 days apart. Thedata ispresented as integratedAIMscores, calculated as the area under the curveover the whole test session.

In addition, AIMs were also assessed after the injection of apomor-phine (0.05 mg/kg; plus 0.2% L-ascorbic acid-saline) at 20 weeks aftergrafting. The AIMs were scored every 10 min for 60 min, using the ratingscale for L-DOPA-induced dyskinesia as described above.

Immunohistochemistry

At 25 weeks after grafting, animals were deeply anesthetized by anoverdose of a mixture of 10% ketamine and 2% rompun and perfusedwith 50 ml of phosphate buffer saline (PBS), followed by 250 ml ofice-cold 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. Thebrains were removed and post-fixed in the same fixative for anadditional 1.5 h and transferred into 30% sucrose for cryoprotectionfor at least 24 h. The brains were cut using a freezing microtome(Frigomobil Leica, Nussloch, Germany) at 40 μm, divided into eightseries, and stored in an antifreeze solution at −20 °C until furtherprocessing.

For immunohistochemistry, free-floating sections were first rinsedin PBS. Endogenous peroxidase activity was quenched by incubationin 3% H2O2 and 10% methanol in PBS for 10 min. Non-specific bindingwas blocked by 1 h pre-incubation in 5% of the appropriate serumcontaining 0.25% Triton X-100 in PBS, followed by incubation with theappropriate primary antibody in the same mixture of serum andTriton X-100 overnight at room temperature: mouse anti-tyrosinehydroxylase (TH; 1:2500; Sigma-Aldrich, Steinheim, Germany);mouse anti-serotonin transporter (SERT; 1:1000; Chemicon, Temecula,CA); and rabbit anti-serotonin (5-HT; 1:10,000; Immunostar, Hudson,WI). Sections were then rinsed 3 times with PBS, followed by 1 hincubation in the appropriate biotinylated secondary antibody(horse anti-mouse for TH and SERT; goat anti-rabbit for 5-HTstaining; all from Vector Laboratories, Burlingame, CA), 3 rinses and1 h incubationwith avidin–biotin peroxidase solution (ABC Elite; VectorLaboratories, Burlingame, CA). Finally 3,3′-diaminobenzidine (DAB;Merk, Darmstadt, Germany) and 0.01% H2O2 were used to develop the

color reaction. For SERT staining, nickel ammonium sulfate (2.5 mg/ml)was used to intensify the staining. The sections were mounted onsuper frost plus slides (Langenbrinck, Emmendingen, Germany),dehydrated in ascending alcohol solutions, and cleared in xylenebefore theywere coverslippedwith Histofluid (Marienfeld Laborglas,Lauda-Königshofen, Germany).

Estimation of TH- and 5-HT-positive cell numbers in the striatum

TH- and 5-HT-positive cells in the striatum were counted underbright-field illumination using a microscope with an X–Y motor stage(Leica, Nussloch, Germany). Analysis of cells was then performedusing the Stereoinvestigator software (Microbrightfield, Magdeburg,Germany) with outlining of the graft area at 2.5× magnification andcounting all grafted TH- and 5-HT-positive cells in the striatumthroughout each section of the graft using 40× magnification. Thetotal number of neurons was finally estimated using Abercrombie'sformula (Abercrombie, 1946).

Striatal TH- and 5-HT-positive fiber density measurements

Striatal TH- and SERT-positive fiberdensities were evaluated at 5rostrocaudal levels (AP: +1.6 mm, +1.0 mm, +0.2 mm, −0.3 mmand −0.9 mm) in order to analyze graft integration and fiberoutgrowth. Images were taken by a digital camera (Olympus E330)on a constant light illumination table and analyzed using the ImageJsoftware platform (version 1.43; National Institutes of Health;http://rsb.info.nih.gov/ij/). The entire striatum was then delineatedon the computer screen with exclusion of the graft core. In addition, atsections 1 and 2, the rostrodorsolateral region (Rdl) was delineated byan oblique line passing through the striatum, and in sections 3 and 4 thecaudolateral region (Cl) was delineated by a vertical line dividing thestriatum in the middle (Carlsson et al., 2007). For all sections, opticaldensity of the corpus callosum was used as a background value andsubtracted from the densities of the lesioned and unlesioned striatum.Fiber density is expressed as optical density in percentage of the controlside. Average densities are calculated from all five coronal sections forthewhole striatum, sections 1 and 2 for Rdl region, and sections 3 and 4for the Cl region.

Statistical analysis

For statistical evaluation the data was subjected to One-wayANOVAs or Two-way ANOVA with repeated measurements followedby post hoc Tukey–Kramer test using StatView 4.5 (Abacus ConceptsInc., Berkeley). Level of significance was set at pb0.05. For betterreadability, F-values and p-values are given in the legends of thefigures.

Results

Graft survival and fiber outgrowth into the host striatum

Immunohistochemistry for TH and 5-HT showed surviving DA andserotonin grafts in all transplanted animals (Figs. 2G, H and J, K). Anaverage of 989±93 TH-positive neurons and 1334±177 5-HT-positiveneurons was counted in the High DA ratio group. There was a significantreduction of TH-positive neurons in the Low DA ratio group to 264±34cells, whereas 5-HT-positive cell numbers did not differ between thegraftedgroups (1569±156cells in the LowDAratiogroup; Fig. 3A). TH-or5-HT-positive cells were not observed in any of the striatal sections in theLes-Ctrl group (Figs. 2A, B, D, E).

TH- and SERT-positive fiber densities were evaluated throughoutthe whole striatum in five coronal sections. Additional analysis wasperformed for the rostrodorsolateral (Rdl) and caudolateral (Cl)striatum. TH-staining showed an almost complete DA denervation of

Fig. 2. Graft morphology. Immunohistochemistry for tyrosine hydroxylase (TH) revealed an almost complete loss of TH-positive striatal fibers after the lesion (A, D). The serotonin(5-HT) and serotonin transporter (SERT) immunohistochemistry showed no significant change in the striatum caused by the same lesion (B, E and C, F). TH and 5-HT-stainingsshowed surviving DA and 5-HT neurons in all grafted animals (G, J and H, K). Numbers of DA neurons were higher in the High DA ratio group as compared to the Low DA ratio group(G, J), while both groups showed similar numbers of 5-HT-positive neurons (H, K). In theHigh DA ratio group the TH-positive fiber density in the striatumwas increased as comparedto the Les Ctrl group, in particular in the caudolateral region (G), while animals in the Low DA ratio group showed an increase only in closer vicinity to the grafts (H). After graftingthere was a slight hyperinnervation of SERT-positive fibers in the striatum (I, L). High-power views in panels G–L illustrate single TH- and 5-HT-positive cells and fibers. Scale bars inL=1 mm and 150 μm (inset).

580 J. García et al. / Neurobiology of Disease 43 (2011) 576–587

Fig. 3. Quantification of grafted cell numbers and striatal reinnervation. The total number of grafted TH- and 5-HT-positive cells in the striatum (A)was assessed using stereology. TH- andSERT-positive fibers in the striatum (B, C) were assessed using optical density measurements. (A) The number of TH-positive cells was significantly reduced in the Low DA ratio group ascompared to the High DA ratio group [* = different from High DA ratio group; One-way ANOVA, (F(1,38)=57.610 and pb0.0001)] but there were no differences for 5-HT-positive cellsbetween the two groups [One-way ANOVA, (F(1,38)=1.004 and p=0.323)]. (B) TH-positive fibers in the striatum were reduced by N95% in the Les-Ctrl group as compared to the intactcontrol side in all regions analyzed. In grafted animals therewere significant increases in TH-positive fibers in thewhole and the caudolateral striatum in theHigh DA ratio group,whereastherewereonly slightbut non-significant increases offiber density in the in the LowDAratio group [*=different fromnon-grafted Les-Ctrlgroup.Whole:One-wayANOVA, (F(2,45)=8.803and p=0.0006); Rdl: One-way ANOVA, (F(2,45)=2.747 and p=0.0749); Cl: (F(2,45)=15.595 and pb0.0001)]. (C) SERT-positive fibers in the striatum were slightly increased in bothgraftedgroups, but this didnot reach significance [Whole:One-wayANOVA, (F(2,45)=0.270 andp=0.765); Rdl:One-wayANOVA, (F(2,45)=2.978 andp=0.0610); Cl: (F(2,45)=1.206 andp=0.309)].

581J. García et al. / Neurobiology of Disease 43 (2011) 576–587

the striatum in non-transplanted animals (Les-Ctrl group) in allregions measured (Whole: 4.0±1.1%; Rdl: 4.1±1.6% and Cl: 3.5±1.6% of intact side; Figs. 2A, D,3B). In grafted animals TH-positivefiber densities in the whole striatum were significantly increased to13.7±1.9% in the High DA ratio group, but not in the Low DA ratiogroup (6.3±1.3% of intact side). Sub-regional analysis in the High DAratio group showed significant increases in TH-positive fiberdensities in the rostrodorsolateral striatum and, in particular, in thecaudolateral striatum (Rdl: 10.2±2.4%; Cl: 21.8±2.8% of intact side;Fig. 3B). In the Low DA ratio group there were visible but non-significant increases of TH-positive fiber densities in the caudolateralstriatum (9.1±1.5% of intact side only), whereas fiber densities inthe rostrodorsolateral striatum remained unchanged as compared tocontrol animals (4.7±1.5% of intact side).

Analysis of serotonin fibers in the striatum using SERT-immunohis-tochemistry did not show a significant lesion effect in the Les-Ctrl groupin any of the regions studied (Whole: 108.2.0±5.1%; Rdl: 103.8±6.3%andCl: 103.2±4.8% of intact side; Figs. 2C, F, 3C). In contrast, the graftedgroups showed a moderate but non-significant increase of SERT-positive fiber innervation, in particular in the rostrodorsolateralstriatum (Low DA ratio group: Rdl: 125.6±5.4% and Cl: 113.6±3.2%of intact side;HighDA ratiogroup: Rdl: 113.9±5.1% andCl: 116.3±6.1%of intact side; Figs. 2I, L and 3C).

Functional recovery in amphetamine-induced rotation

Rotational asymmetry was assessed at 18 weeks after transplantationwith amphetamine (Fig. 4A). Prior to transplantation, all groups showed astrong rotational bias towards the side ipsilateral of the lesion (groupaverages from8.9±0.7 to 9.6±1.2 ipsilateral full body turns/min).Whileturning behavior didnot change in the Les Ctrl groupover time (11.2±1.5turns/minat18 weeks after shamoperation), bothgraftedgroups showedsignificant reduction of rotational bias. In the Low DA ratio group, therewas reversal to 1.1±1.0 turns/min, and in the High DA ratio group therewas an overcompensation to−5.8±1.2 contralateral turns/min.

Performance in amphetamine-induced rotation was significantlycorrelated with the TH-positive cell number in the striatum (Fig. 4C)and TH-positive innervation density in the caudolateral region(Fig. 4E). Here a striatal TH-positive cell number of more thanapproximately 600 cells or a TH-positive fiber density in thecaudolateral striatum of more than approximately 20% of normal

indicated particularly good graft effects on rotation (see dashed linesin Figs. 4C and E), while the rotational response was more variablewith lower cell or fiber scores. The number of 5-HT cells in thestriatum also correlated with rotation, however, less strong (R=0.4),while SERT-positive fiber innervations showed no correlation withamphetamine-induced rotation.

Functional recovery in spontaneous forelimb use

Graft-induced functional recovery was assessed at 3 and 15 weeksafter transplantation in the cylinder test (Fig. 4B). In the pre-transplantation test all animals showed a clear deficit in the use ofthe impaired paw contralateral to the lesion (group averages from11.2±1.8% to 14.2±2.8% of total paw use). At 3 weeks post grafting,none of the grafted groups showed any improvement in use of theimpaired paw (group averages from 13.5±1.3% to 15.0±1.9%).However, at 15 weeks after grafting, the High DA ratio group showedsignificant improvements in paw use (28.7±2.7% of total paw use),while the Low DA ratio group remained impaired (17.5±1.9%).

Similar to amphetamine-induced rotation, improvement in thecylinder test was significantly correlated with the TH-positive cellnumber in the striatum (Fig. 4D), and there was a non-significant trendto correlate with the TH-positive fiber innervation in the Cl region(Fig. 4F). As for amphetamine-induced behavioral improvement,substantial improvement in spontaneous forelimb use could beexpected, when striatal TH-positive cell numbers exceeded approxi-mately 600, or when the striatal fiber innervation in the caudolateralstriatumwashigher than20%of normal (seedashed lines in Figs. 4D andF). Improvement in spontaneous forelimb use were not correlated with5-HT cells or SERT-positive fiber innervation in the striatum (5HT-cells:F(1,38)=0.126; R=0.057, p=0.7246; SERT fibers: F(1,38)=0.747;R=0.139, p=0.39).

Modulation of L-DOPA-induced dyskinesia with increasing 5-HT:DA cellratio

Stable baseline AIMs were induced by daily injections of low doseL-DOPA (6 mg/kg) over 4 weeks (induction phase in Fig. 1A). AIMsincluded prominent movements in the forelimb and orolingualmuscles (predominantly choreatiform hyperkinesia) and in thetrunk muscles (predominantly dystonia). Moreover, some animals

Fig. 4. Graft-induced behavioral improvements. (A) Rotational behavior after injection of 2.5 mg/kg amphetamine was improved in both grafted groups, with normalization ofrotational bias in the Low DA ratio group and overcompensation to contralateral rotations in theHigh DA ratio group. [*= different from Les-Ctrl group and their own pre-graft values.Two-way ANOVA (F(2,45)=22.446 and pb0.0001)]. (B) Forepaw asymmetry in the cylinder test was assessed after the lesion (pre) and at 3 and 15 weeks post transplantation. At15 weeks after grafting, the High DA ratio group showed significant improvements in the use of the lesioned paw. [* = different from the Les-Ctrl group and their own pre-graftvalues. Two-way ANOVA (F(4,90)=6.415 and p=0.0001)]. (C, E) Performance in amphetamine-induced rotation was significantly correlated with the number of grafted TH-positivecells in the striatum (F(1,38)=32.831; R=0.681 and pb0.0001), and with the TH-positive fiber density in the caudolateral striatum (F(1,38)=24.332; R=0.624 and pb0.0001).(D, F) Improvement in the cylinder test was significantly correlated with the TH-positive cell number in the striatum (F(1,38)=12.990; R=0.505 and p=0.0009), but less well withthe TH-positive fiber density in the caudolateral striatum (F(1,38)=2.920; R=0.266 and p=0.0969). The dashed lines in C–F indicate cell numbers or fiber densities, which have tobe exceeded to guarantee good behavioral improvement.

582 J. García et al. / Neurobiology of Disease 43 (2011) 576–587

also showed increased locomotion contralateral to the lesioned side.The onset and progression of the severity of dyskinesia variedbetween rats, but in the end of the 4th week, all rats displayed stabledyskinetic responses. After an individual L-DOPA injection, AIMswerefirst observed at 10–15 min, peaked at 60–90 min, and return tobaseline by 120–180 min.

Following the induction phase, AIMs were evaluated at 2, 4, 12, 16and 24 weeks after transplantation to evaluate the effect of differentDA:5-HT cell ratios on L-DOPA-induced dyskinesia. Les-Ctrl animalsstably expressed AIMs throughout the full 24-week observationperiod (96.5±15.2% of pre-scores at 24 weeks; Figs. 5A, B). Asexpected, theHigh DA ratio group, which included high numbers of DA

Fig. 5. Graft-derived changes in L-DOPA-induceddyskinesia. Dyskinesias inducedby L-DOPA (6mg/kg)were evaluatedover24 weeks after transplantation (A for groups, B shown for individualanimals at pre-graft versus24 weeksafter transplantation). Thegraftedgroups,HighDA ratio (circles) and LowDAratio (triangles), showedaprogressive reduction in L-DOPA-induceddyskinesiain contrast to the Les-Ctrl group (boxes),which became significant at 12 weeks and reached amaximumof 27–42% reductions of ownpre-scores at 24 weeks after transplantation [*=differentfromLes-Ctrl group,#=different fromLowDAratiogroup.Two-wayANOVA(F(10,225)=6.80andpb0.001) followedbyTukey–Kramerposthoc test]. (C,E,G)Analysisof individual componentsof dyskinesia showed significant reduction of frequency of axial (C), limb (E), and orolingual dyskinesia in both grafted groups [*= different from ownpre-graft values; axial: two-way ANOVA(F(2,45)=7.477 and p=0.0016); limb: (F(2,45)=12.444 and pb0.0001); orolingual: (F(2,45)=10.212 and p=0.0002); all followed by Tukey–Kramer post hoc test]. Frequency of orolingualdyskinesiawasalso slightly reduced in the Les-Ctrlgroup. (D, F)Analysis of amplitude changesof axial and limbdyskinesia showed reductions inbothgraftedgroups [*=different fromownpre-graft values; axial: two-way ANOVA (F(2,45)=2.985 and p=0.0606); limb: (F(2,45)=6.430 and p=0.0035); all followed by Tukey–Kramer post hoc test].

583J. García et al. / Neurobiology of Disease 43 (2011) 576–587

Fig. 6. Regression analysis between TH- and 5-HT-positive cell numbers and TH- and SERT-positive fibers densities and L-DOPA-induced dyskinesia. Regression analysis showed asignificant inverse correlation between number of TH-positive cells in the striatum and severity of L-DOPA-induced dyskinesias (A; F(1,38)=7.746; R=0.411 and p=0.0083) orbetween TH-positive fiber reinnervation in the caudolateral (Cl) region of the striatum and dyskinesia (B; F(1,38)=4.553; R=0.327 and p=0.0394). The dashed lines in A and B indicatecell numbers orfiber densities, which have to be exceeded to guarantee good reduction of dyskinesia. Therewere no correlations betweendyskinesia and number of striatal 5-HT-positivecells (C; F(1,38)=0.380; R=0.099 and p=0.541) or striatal SERT-positive fibers (D; F(1,38)=0.447; R=0.108 and p=0.508).

584 J. García et al. / Neurobiology of Disease 43 (2011) 576–587

neurons and substantial TH-positive fiber re-innervation in thestriatum, significantly reduced their total AIMs to 26.6±6.5% of pre-scores (Figs. 5A, B). Interestingly, the Low DA ratio group, whichincluded animals with low numbers of DA neurons andmarginal, non-significant TH-positive fiber re-innervation, also showed significantreduction of AIMs to 42.4±5.0% of pre-transplantation scores despitehigh numbers of 5-HT-positive neurons in the graft and a ratio ofDA:5-HT sometimes as low as 1:10 (Figs. 5A, B).

Further analysis of the individual components of AIMs showed thatthe frequency of axial, limb and orolingual dyskinesia was signifi-cantly reduced in both grafted groups, although this was morepronounced in the High DA ratio group. Thus, axial dyskinesia wasdecreased to 23% and 54% of pre-graft scores in the High DA ratio andLow DA ratio groups, respectively (Fig. 5C). Similarly, the frequency inlimb and orolingual AIMs decreased to 35% and 16% of pre-graft valuein the High DA ratio group, and to 54% and 20% in the Low DA ratiogroup (Figs. 5E, G). In the Les-Ctrl group, frequency of axial and limbdyskinesia did not show a significant change, while the orolingualcomponent showed only a minor, but significant, reduction (72% ofpre-graft value; Figs. 5C, E, G). The amplitude of axial and limbdyskinesia was also significantly reduced in the High DA ratio group to46% and 61% of pre-graft scores (Figs. 5D, F). In the Low DA ratio groupa non-significant trend was observed for axial amplitude (62% of pre-graft scores; Fig. 5D), while limb amplitude was significantly reducedto 78% of pre-graft scores (Fig. 5F).

Regression analysis showed that L-DOPA-induced dyskinesia(axial, limb and orolingual components) was significantly inversely

correlated with the TH-positive cell number in the striatum (Fig. 6A).Thus, animals with a TH-positive cell number of more thanapproximately 600 cells in the striatum would always reduce theirdyskinesia score by more than 50%, while animals with lower striatalcell numbers showedmore variable reduction in dyskinesia. Similarly,their dyskinesia scores were also significantly correlated with the TH-positive innervation density in the caudolateral striatum. Here a fiberdensity of approximately 20% or more of normal in the Cl regionindicated profound reduction of dyskinesia by more than 50% of pre-grafting values, while animals with lower reinnervation showedmorevariable reduction of dyskinesia (Fig. 6B). Striatal TH-positive fiberdensities in the whole striatum or the Cl region, or 5-HT-positive cellnumbers or SERT-positive striatal fiber densities were not correlatedwith reduction of L-DOPA-induced dyskinesia (Figs. 6C, D).

Pre- and postsynaptic modulation with increasing 5-HT:DA cell ratio

L-DOPA- and apomorphine-induced rotationswere performed at 16and 19 weeks, respectively, to evaluate the role of pre- and postsynapticDA response in the striatum. Effects of L-DOPA are dependent on both,the pre-synaptic release of DA as well as post-synaptic sensitivity of theDA receptors in the striatum. Apomorphine on the other hand is a D1/D2

receptor agonist acting directly on the postsynaptic receptors. L-DOPA-induced rotationwas reduced from−8.8±2.0 turns/min in the Les-Ctrlgroup to −4.7±1.3 turns/min in the High DA ratio group (Fig. 7A).Interestingly, the Low DA ratio group showed a strong tendency toincrease in the rotational response to−14.5±1.8 turns/min suggesting

Fig. 7. Graft-derived effects on pre- and postsynaptic responses. Rotations induced by L-DOPA (A) and the direct DA-agonist apomorphine (0.05 mg/kg) (B) were evaluated tomeasure pre- and postsynaptic DA responses. L-DOPA-induced rotation showed a strong albeit non-significant reduction in the High DA ratio group as compared to the non-graftedLes-Ctrl group. In contrast, the Low DA ratio group showed a strong albeit non-significant increase of rotational response suggestive of a dysregulation of either pre- and/orpostsynaptic type [One-way ANOVA, (F(2,45)=10.357 and p=0.0002)]. Since apomorphine-induced rotation (B) and -dyskinesia (C) were significantly reduced in the High DA ratiogroup, and there was a clear trend of reduction in the Low DA ratio group [* = different from Les-Ctrl group. One-way ANOVA for apomorphine-induced rotation, (F(2,45)=5.613 andp=0.0067); One-way ANOVA for apomorphine-induced dyskinesia, (F(2,45)=7.182 and p=0.0020)], this suggests partial normalization of the postsynaptic DA receptor response.Thus, the effect seen in the Low DA ratio group for L-DOPA-induced rotation with an increase in rotation as compared to animals with high DA ratio suggests a hyperactivepresynaptic response.

585J. García et al. / Neurobiology of Disease 43 (2011) 576–587

a dysregulated response in this group that could either be pre- and/orpostsynaptic.

In the apomorphine-induced rotation test, the High DA ratio groupshowed significantly reduced scores (−5.3±1.0 turns/min; Fig. 7B) ascompared to the Les-Ctrl group (−11.3±1.9 turns/min). The Low DAratio group showed no worsening in rotational response to apomor-phine as compared to L-DOPA, but rather a tendency to a decrease(−9.2±1.1 turns/min). Inorder to get additional informationon theDAreceptor supersensitivity in the striatum, we performed an apomor-phine-induced dyskinesia test at 20 weeks post grafting (Fig. 7C). TheHigh DA ratio group showed similar reduction in dyskinesia as seen inapomorphine-induced rotation when compared to the Les-Ctrl group(dyskinesia score of 117.1±20.3 versus 262.5±38.5, respectively).Moreover, in the Low DA ratio group, a strong tendency to reduce inapomorphine-induced dyskinesia was also observed.

Discussion

Understanding the factors influencing the functional outcome offetal DA cell transplantation is highly warranted, in particular, sincetransplantation in PD patients will be re-initiated in several Europeancenters in the coming years. We have in the present study, therefore,further evaluated the influence of DA and 5-HT neurons in graftedfetal tissue in the rat PD model. We show in a large cohort of 40grafted animals that significant improvement in drug-induced andspontaneous behaviors, assessed by amphetamine-induced rotationand cylinder test, and profound reduction of L-DOPA-induceddyskinesia of more than 50% from pre-graft values, is obtainedwhen the graft containsmore than 600 surviving DA cells or a DA fiberreinnervation in the caudolateral striatum of at least 20% of normalvalues. In contrast, the number of 5-HT neurons within the DA graftsdid not correlate with motor function or expression of L-DOPA-induced dyskinesia suggesting that their appearance within a DA graftis not critical as long as there are sufficient numbers of DA neuronswithin the graft. We also show that 5-HT neurons within the graftmay induce a disturbance of pre-synaptic release of DA fromgrafted DA neurons when the number of DA neurons in the graftis low. This disturbance, however, does not influence the reductionof L-DOPA-induced dyskinesia induced by the DA neurons withinthe graft.

In the present study we have used the rat model of PD withunilateral complete degeneration of the host DA system in order tomimic advanced PD in patients. In this model graft-inducedbehavioral recovery is more difficult to obtain (Breysse et al., 2007;Kirik et al., 2001; Winkler et al., 1999) and L-DOPA-induceddyskinesia is more pronounced (Carlsson et al., 2009; Winkler et al.,2002), as compared to animals with partial sparing of the host DAsystem. Similar effects are also seen in older patients with moreadvanced disease and degeneration of extrastriatal DA fibers (Freedet al., 2001; Olanow et al., 2003; Piccini et al., 2005) suggesting thatthe complete lesion model is a good model to obtain further insightsinto functional regeneration in the brain and to obtain good safetydata, because all patients irrespective of the disease stage at the timeof transplantation will advance in their disease and show continuousdegeneration of the remaining host DA system.

The number of grafted DA neurons and their localization haspreviously been shown to be important for graft-induced functionalrecovery. Thus, we can confirm prior findings that transplants as smallas 100–200 DA neurons can be sufficient to normalize amphetamine-induced rotation when placed in the dorsal striatum (Brundin et al.,1988; Dunnett et al., 1983;Mandel et al., 1990) but it appears that largernumbers are required and that a certain threshold of DA reinnervationhas to be obtained before more complexmotor behaviors are improved(Kirik et al., 2001;Winkler et al., 1999).Herewe add to thesefindings byshowing that improvement in spontaneous motor behavior in thecylinder test is always obtained when the number of DA neurons in thegraft exceeds 600, and that this improvement is dependent on DAreinnervation specifically in the caudolateral striatum. Thus, whileanimalswith less than 600 survivingDAneurons in the graft or less than20% DA fiber reinnervation in the caudolateral striatum improved theirrotational scores induced by amphetamine, the majority of theseanimals did not improve spontaneous motor behavior in the cylindertest. Extending these findings to the clinical setting, it is thus suggestedthat sufficient numbers of DA neurons have to be placed in thecaudolateral part of the putamen to achieve sufficient reinnervation ofthis striatal substructure and behavioral recovery.

Graft effects on L-DOPA-induced dyskinesia, however, have beenvariable in PD patients and details on graft composition and placementare still under debate in order to obtain reduction in L-DOPA-induceddyskinesia. We know from previous studies in animal PD models thatgrafts containing DA neurons can reduce L-DOPA-induced dyskinesia

586 J. García et al. / Neurobiology of Disease 43 (2011) 576–587

(Carlsson et al., 2006, 2007, 2009; Kuan et al., 2007; Lane et al., 2006,2008, 2009, 2010; Lee et al., 2000; Maries et al., 2006; Soderstrom et al.,2008; Steece-Collier et al., 2009; Vinuela et al., 2008). In two previousstudies we have shown that very small grafts of less than 300 DAneurons did not affect L-DOPA-induced dyskinesia while very largegrafts of more than 10.000 DA neurons did (Lane et al., 2006) and that amore caudolateral instead of rostromedial placement of the graft wasbeneficial for reduction of dyskinesia (Carlsson et al., 2006). Here weextend thesefindings to show that graftswithmore thanapproximately600 DA neurons will always reduce L-DOPA-induced dyskinesia bymore than 50% and that DA fiber reinnervation in the caudolateralstriatum of more than 20% of normal, but not the rostral or wholestriatum, is a predictor for a profound reduction of dyskinesia. Overall,the impact of number of grafted DA cells or localization of striatal DAreinnervation was similar for improvement of spontaneous motorbehavior in the cylinder test and reduction of L-DOPA-induceddyskinesia suggesting that similar graft mechanisms may affect thesedifferent symptoms of PD.

In a next step we analyzed the role of 5-HT neurons within the DAgrafts on expression of behavior and L-DOPA-induced dyskinesia.Within the two groups of animals (High versus Low DA ratio group)there were large differences in the number of surviving DA neurons inthe striatum, approximately 1000 and 250 neurons in the High versusLow DA ratio groups, respectively, while the number of 5-HT neuronswas similar in both groups. Regression analysis between behavioralscores in amphetamine-induced rotation and cylinder test did notcorrelate with either number of grafted 5-HT neurons or SERT-positive fiber density in the different striatal subregions. Therefore,the improvement in amphetamine-induced rotation in both graftedgroups, but also the lack of improvement in spontaneous motorbehavior in the cylinder test in the Low DA ratio group in contrast toimprovement in the High DA ratio group in this test, is most likely aresult of differences in number of surviving DA neurons and striatalDA reinnervation and not dependent on the inclusion of 5-HT neuronsin the graft. Indeed, further analysis of ratios between DA:5-HTneurons showed, that there was no correlation between DA:5-HT ratioand functional parameters in this study. Thus, animals with low DA cellnumbers in the graft and DA:5-HT ratios ranging from 1:3 to 1:10 didnot show any differences in graft-induced changes in drug-induced andspontaneous motor behavior suggesting that 5-HT neurons within astriatal DA graft are neither beneficial nor detrimental for behavioralrecovery, at least not as long as DA neurons are present within the graft.

DA neurons within a mixed DA:5-HT graft also appear to be theprime parameter predicting changes in the expression of L-DOPA-induced dyskinesia. In previous studies regular grafts of the VMcontaining DA and 5-HT neurons at a ratio of approximately 2:1 to 1:1have been shown to reduce L-DOPA-induced dyskinesia (Carlssonet al., 2007, 2009; Lane et al., 2006, 2009). We show here thatsubstantial improvement of L-DOPA-induced dyskinesia was obtainedin both grafted groups and irrespective of the ratio of DA:5-HTneurons. Reduction of dyskinesia was slightlymore pronounced in theHigh DA ratio group with 75% reduction of pre-graft scores ascompared to a 50–60% reduction in the Low DA ratio group, buteven low DA numbers in the graft and a DA:5-HT ratio of up to 1:10would not be sufficient to worsen L-DOPA-induced dyskinesia. This isin contrast to “pure” 5-HT grafts, which have clearly shown thepotential detrimental effect of 5-HT neurons in a graft by causingsignificant worsening of L-DOPA-induced dyskinesia (Carlsson et al.,2007, 2009). Comparing these data, DA neurons numbers as low asapproximately 250 cells in the Low DA ratio group appear to besufficient to reduce L-DOPA-induced dyskinesia in the present study,irrespective of the number of 5-HT neurons in the graft. This suggeststhat dynamics of transmitter release are reinstated by rather smallnumbers of DA neurons within a graft. Thus, while animals in ourstudy are suggested to suffer from a complete loss of autoregulatoryfeedback control of DA release due to a complete DA-denervating

lesion, thus facilitating excessive fluctuations of DA release fromserotonin neurons in response to pulsatile administration of L-DOPA,autoregulation is reinstated by low numbers of grafted DA neuronsdue to synaptic connections between grafted cells and host and byremoving part of the DA/5-HT-transmitter competition withinserotonin neurons. To further substantiate these findings, additionalstudies would be required using microdialysis or electrophysiology.

Interestingly, in animals with low DA cell numbers there wasvariability in reduction of spontaneous motor behavior in the cylindertest and reduction of L-DOPA-induced dyskinesia. Thus while someanimals improved their motor behavior and reduced dyskinesiasimilar to those with larger DA cell numbers in the High DA ratiogroup, other animals in the low DA ratio group showed lessimprovement. We used rotational tests induced by L-DOPA andapomorphine and a test of apomorphine-induced dyskinesia tofurther characterize pre- and post-synaptic DA responses in thebrain after grafting. The expectation is that a reduction in rotation anddyskinesia induced by the DA agonist apomorphine would indicateimprovement of DA-receptor responses, i.e. post-synaptic effects, whilea reduction in L-DOPA-induced rotation could indicate improvements inrelease of DA and/or DA-receptor responses, i.e. pre- and/or post-synaptic responses. In animals with large DA grafts in the High DA ratiogroup we found reduction in L-DOPA- and apomorphine-inducedrotation and in apomorphine-induced dyskinesia suggesting improve-ments in pre- and post-synaptic responses. In contrast, animals in thelow DA ratio group showed a strong tendency to improve inapomorphine-induced rotation and -dyskinesia, but a clear, but non-significant, increase in L-DOPA-induced rotation, which may suggestaberrant pre-synaptic responses. Further analysis of multiple parame-ters including number of 5-HT neurons in the graft or SERT-positivefibers in different striatal regions, however, did not show differences inanimals of the Low DA ratio group with good versus less prominentbehavioral recovery. Nonetheless, this does not exclude a prominentrole of grafted 5-HTneurons in these changes of pre-synaptic responses,since some changes may occur on a subcellular level such as a changein SERT-immunoreactive varicosities making contact with striatalneurons (Rylander et al., 2010). In addition, hyperactive pre-synapticDA responses have also been observed in “pure” 5-HT grafts, andpersisted after the removal of the intrinsic 5-HT system (Carlssonet al., 2009). Moreover, the postsynaptic receptor supersensitivity,causedbypulsatile and intermittent L-DOPA treatment prior to grafting,was not reversed but tended to increase after grafting of “pure” 5-HTneurons (Carlsson et al., 2007, 2009), indicating the inability of DAreleased from 5-HT to normalize the DA receptor sensitivity in theseanimals.

In conclusion, our experimental data shed further light on how toprepare an optimized transplant protocol for the clinical situation. Weknow from rare clinical cases that clinical DA grafts do include 5-HTneurons, most likely at a 1:1 ratio (Mendez et al., 2008), and that suchgrafts can induce a 5-HT-hyperinnervation in the striatum (Politis et al.,2010). Dissection of sparse fetal material goes along with the risk toextend the borders of dissection to include as many DA progenitors aspossible, thereby also increasing the ratio of 5-HT neurons in the cellpreparation.Our experimental data indicates that a certainnumber of 5-HT neuronswithin a DA neuron graft will not be critical for either graft-induced behavioral recovery or reduction of L-DOPA-induced dyskine-sia, as long as sufficient numbers of DA neurons are present. Even at aDA:5-HT ratio of 1:10wedidnot observe anynegative behavioral effectssuggesting that a regular dissection in the clinical setting giving rise to aDA:5-HT ratio of 2:1 to 1:2 should be safe. Indeed, those patientscharacterized in the studies by Mendez et al. (2008) and Politis et al.(2010) showed improvement inmotor function and either stable levelsof L-DOPA-induced dyskinesia (Mendez et al., 2008) or even reduction(Politis et al., 2010). This suggests that the issue of inclusion of serotoninneurons into the graft suspension should not be a major concern whenre-initiating neural transplantation for Parkinson's disease. However,

587J. García et al. / Neurobiology of Disease 43 (2011) 576–587

our studypre-synaptic activity, inducedby L-DOPA treatment, showedaclear tendency to be aberrant in animalswith lowDA cell numbers and aDA:5-HT ratio of 1:3 to 1:10 suggesting that the inclusion of 5-HTneurons into a graft suspension is not a negligible phenomenon, andthat numbers of 5-HT neurons should be restricted and borders fordissection be clearly defined.

Acknowledgments

We would like to thank Brunhilde Baumer, Donata Maciaczyk,Beate Schmitt and Johanna Wessolleck for their experimental andtechnical assistance in the process of this manuscript. The currentwork was supported by a grant of the European Commission withinthe 7th Framework Programme (TRANSEURO, to CW and GN) and theDeutsche Parkinson Vereinigung e.V. (to CW). JG was supported bythe Graduiertenkolleg Neuroscience Freiburg, and TC by the EuropeanCommunity's 7th Framework Programme (FP7/2007–2013) undergrant agreement nr. 220656.

References

Abercrombie, M., 1946. Estimation of nuclear population from microtome sections.Anat. Rec. 94, 239–247.

Arai, R., Karasawa, N., Geffard, M., Nagatsu, T., Nagatsu, I., 1994. Immunohistochemicalevidence that central serotonin neurons produce dopamine from exogenous L-DOPA in the rat, with reference to the involvement of aromatic L-amino aciddecarboxylase. Brain Res. 667, 295–299.

Breysse, N., Carlsson, T., Winkler, C., Björklund, A., Kirik, D., 2007. The functional impactof intrastriatal dopamine neuron grafts in parkinsonian rats is reduced withadvancing disease. J. Neurosci. 27, 5849–5856.

Brundin, P., Barbin, G., Strecker, R.E., Isacson, O., Prochiantz, A., Björklund, A., 1988.Survival and function of dissociated rat dopamine neurones grafted at differentdevelopmental stages or after being cultured in vitro. Brain Res. 467 (2), 233–243.

Brundin, P., Barker, R.A., Parmar, M., 2010. Neural grafting in Parkinson's disease:problems and possibilities. Prog. Brain Res. 184, 265–294.

Carlsson, T., Winkler, C., Lundblad, M., Cenci, M.A., Björklund, A., Kirik, D., 2006. Graftplacement and uneven pattern of reinnervation in the striatum is important fordevelopment of graft-induced dyskinesia. Neurobiol. Dis. 21, 657–668.

Carlsson, T., Carta, M., Winkler, C., Björklund, A., Kirik, D., 2007. Serotonin neurontransplants exacerbate L-DOPA-induced dyskinesias in a rat model of Parkinson'sdisease. J. Neurosci. 27, 8011–8022.

Carlsson, T., Carta, M., Muñoz, A., Mattsson, B., Winkler, C., Kirik, D., et al., 2009. Impactof grafted serotonin and dopamine neurons on development of L-DOPA-induceddyskinesias in parkinsonian rats is determined by the extent of dopamine neurondegeneration. Brain 132, 319–335.

Carta, M., Carlsson, T., Kirik, D., Björklund, A., 2007. Dopamine released from 5-HTterminals is the cause of L-DOPA-induced dyskinesia in parkinsonian rats. Brain130, 1819–1833.

Carta, M., Carlsson, T., Muñoz, A., Kirik, D., Björklund, A., 2010. Role of serotonin neuronsin the induction of levodopa- and graft-induced dyskinesias in Parkinson's disease.Mov. Disord. 25 (Suppl 1), S174–S179.

Cenci, M.A., Lee, C.S., Björklund, A., 1998. L-DOPA-induced dyskinesia in the rat isassociated with striatal overexpression of prodynorphin- and glutamic aciddecarboxylase mRNA. Eur. J. Neurosci. 10, 2694–2706.

Dunnett, S.B., Björklund, A., Schmidt, R.H., Stenevi, U., Iversen, S.D., 1983. Intracerebralgrafting of neuronal cell suspensions. IV. Behavioural recovery in rats withunilateral 6-OHDA lesions following implantation of nigral cell suspensions indifferent forebrain sites. Acta Physiol. Scand. Suppl. 522, 29–37.

Freed, C.R., Greene, P.E., Breeze, R.E., Tsai, W.Y., DuMouchel, W., Kao, R., et al., 2001.Transplantation of embryonic dopamine neurons for severe Parkinson's disease. N.Engl. J. Med. 344, 710–719.

Hagell, P., Cenci, M.A., 2005. Dyskinesias and dopamine cell replacement in Parkinson'sdisease: a clinical perspective. Brain Res. Bull. 68 (1–2), 4–15.

Kirik, D., Winkler, C., Björklund, A., 2001. Growth and functional efficacy of intrastriatalnigral transplants depend on the extent of nigrostriatal degeneration. J. Neurosci.21 (8), 2889–2896.

Kuan, W.L., Lin, R., Tyers, P., Barker, R.A., 2007. The importance of A9 dopaminergicneurons in mediating the functional benefits of fetal ventral mesencephalontransplants and levodopa-induced dyskinesias. Neurobiol. Dis. 25 (3), 594–608.

Lane, E.L., Winkler, C., Brundin, P., Cenci, M.A., 2006. The impact of graft size on thedevelopment of dyskinesia following intrastriatal grafting of embryonic dopamineneurons in the rat. Neurobiol. Dis. 22, 334–345.

Lane, E.L., Soulet, D., Vercammen, L., Cenci, M.A., Brundin, P., 2008. Neuroinflammationin the generation of post-transplantation dyskinesia in Parkinson's disease.Neurobiol. Dis. 32 (2), 220–228.

Lane, E.L., Brundin, P., Cenci, M.A., 2009. Amphetamine-induced abnormal movementsoccur independently of both transplant- and host-derived serotonin innervationfollowing neural grafting in a rat model of Parkinson's disease. Neurobiol. Dis. 35(1), 42–51.

Lane, E.L., Björklund, A., Dunnett, S.B., Winkler, C., 2010. Neural grafting in Parkinson'sdisease unraveling the mechanisms underlying graft-induced dyskinesia. Prog.Brain Res. 184, 295–309.

Lee, C.S., Cenci, M.A., Schulzer, M., Björklund, A., 2000. Embryonic ventral mesence-phalic grafts improve levodopa-induced dyskinesia in a rat model of Parkinson'sdisease. Brain 123, 1365–1379.

Ma, Y., Tang, C., Chaly, T., Greene, P., Breeze, R., Fahn, S., et al., 2010. Dopamine cellimplantation in Parkinson's disease: long-term clinical and (18)F-FDOPA PEToutcomes. J. Nucl. Med. 51 (1), 7–15.

Maeda, T., Nagata, K., Yoshida, Y., Kannari, K., 2005. Serotonergic hyperinnervation intothe dopaminergic denervated striatum compensates for dopamine conversionfrom exogenously administered L-DOPA. Brain Res. 1046, 230–233.

Mandel, R.J., Brundin, P., Björklund, A., 1990. The importance of graft placement andtask complexity for transplant-induced recovery of simple and complexsensorimotor deficits in dopamine denervated rats. Eur. J. Neurosci. 2 (10),888–894.

Maries, E., Kordower, J.H., Chu, Y., Collier, T.J., Sortwell, C.E., Olaru, E., Shannon, K.,Steece-Collier, K., 2006. Focal not widespread grafts induce novel dyskineticbehavior in parkinsonian rats. Neurobiol. Dis. 21, 165–180.

Mendez, I., Viñuela, A., Astradsson, A., Mukhida, K., Hallett, P., Robertson, H., et al., 2008.Dopamine neurons implanted into people with Parkinson's disease survive withoutpathology for 14 years. Nat. Med. 14, 507–509.

Navailles, S., De Deurwaerdère, P., 2011. Presynaptic control of serotonin on striataldopamine function. Psychopharmacology (Berl.) 213 (2–3), 213–242.

Nikkhah, G., Olsson, M., Eberhard, J., Bentlage, C., Cunningham, M.G., Björklund, A.,1994. A microtransplantation approach for cell suspension grafting in the ratParkinson model: a detailed account of the methodology. Neuroscience 63 (1),57–72.

Nikkhah, G., Winkler, C., Rödter, A., Samii, M., 2000. Microtransplantation of nigraldopamine neurons: a step-by-step recipe. In: Dunnett, S.B., Boulton, A.A., Baker,G.B. (Eds.), Neuromethods: Cell and Tissue Transplantation in the CNS. TheHuman Press, Totowa, pp. 207–231.

Ng, K.Y., Chase, T.N., Colburn, R.W., Kopin, I.J., 1970. L-DOPA-induced release of cerebralmonoamines. Science 170, 76–77.

Ng, K.Y., Colburn, R.W., Kopin, I.J., 1971. Effects of L-DOPA on efflux of cerebralmonoamines from synaptosomes. Nature 230, 331–332.

Olanow, C.W., Goetz, C.G., Kordower, J.H., Stoessl, A.J., Sossi, V., Brin, M.F., et al., 2003. Adouble-blind controlled trial of bilateral fetal nigral transplantation in Parkinson'sdisease. Ann. Neurol. 54, 403–414.

Olanow, C.W., Kordower, J.H., Lang, A.E., Obeso, J.A., 2009. Dopaminergic transplanta-tion for Parkinson's disease: current status and future prospects. Ann. Neurol. 66(5), 591–596.

Paxinos, G., Watson, C., 2005. The Rat Brain in Stereotaxic Coordinates, 5 Ed. Academic,San Diego.

Piccini, P., Pavese, N., Hagell, P., Reimer, J., Björklund, A., Oertel, W.H., et al., 2005.Factors affecting the clinical outcome after neural transplantation in Parkinson'sdisease. Brain 128 (Pt 12), 2977–2986.

Politis, M., Wu, K., Loane, C., Quinn, N.P., Brooks, D.J., Rehncrona, S., et al., 2010.Serotonergic neurons mediate dyskinesia side effects in Parkinson's patients withneural transplants. Sci. Transl. Med. 2 (38), 38–46.

Rylander, D., Parent, M., O'Sullivan, S.S., Dovero, S., Lees, A.J., Bezard, E., et al., 2010.Maladaptive plasticity of serotonin axon terminals in levodopa-induced dyskinesia.Ann. Neurol. 68 (5), 619–628.

Schallert, T., Tillerson, J., 1999. Intervention strategies for degeneration of dopamine neuronsin parkinsonism: optimising behavioral assessment of outcome. Innovative Models ofCNS Disease: FromMolecule to Therapy. Humana, Clifton, NJ, pp. 131–151.

Soderstrom, K.E., Meredith, G., Freeman, T.B., McGuire, S.O., Collier, T.J., Sortwell, C.E., etal., 2008. The synaptic impact of the host immune response in a parkinsonianallograft rat model: influence on graft-derived aberrant behaviors. Neurobiol. Dis.32 (2), 229–242.

Steece-Collier, K., Soderstrom, K.E., Collier, T.J., Sortwell, C.E., Maries-Lad, E., 2009. Effect oflevodopa priming on dopamine neuron transplant efficacy and induction of abnormalinvoluntary movements in parkinsonian rats. J. Comp. Neurol. 515 (1), 15–30.

Tanaka, H., Kannari, K., Maeda, T., Tomiyama, M., Suda, T., Matsunaga, M., 1999. Role ofserotonergic neurons in L-DOPA-derived extracellular dopamine in the striatum of6-OHDA-lesioned rats. Neuroreport 10 (3), 631–634.

Ungerstedt, U., Arbuthnott, G.W., 1970. Quantitative recording of rotational behavior inrats after 6-hydroxy-dopamine lesions of the nigrostriatal dopamine system. BrainRes. 24 (3), 485–493.

Vinuela, A., Hallett, P.J., Reske-Nielsen, C., Patterson, M., Sotnikova, T.D., Caron, M.G.,Gainetdinov, R.R., Isacson, O., 2008. Implanted reuptake-deficient or wild-typedopaminergic neurons improve ON L-dopa dyskinesias without OFF-dyskinesias ina rat model of Parkinson's disease. Brain 131, 3361–3379.

Winkler, C., Bentlage, C., Nikkhah, G., Samii, M., Björklund, A., 1999. Intranigraltransplants of GABA-rich striatal tissue induce behavioral recovery in the ratParkinson model and promote the effects obtained by intrastriatal dopaminergictransplants. Exp. Neurol. 155 (2), 165–186.

Winkler, C., Kirik, D., Björklund, A., Cenci, M.A., 2002. L-DOPA-induced dyskinesia in theintrastriatal 6-hydroxydopamine model of Parkinson's disease: relation to motorand cellular parameters of nigrostriatal function. Neurobiol. Dis. 10, 165–186.

Winkler, C., Kirik, D., Björklund, A., 2005. Cell transplantation in Parkinson's disease:how can we make it work? Trends Neurosci. 28, 86–92.

Winkler, C., Georgievska, B., Carlsson, T., Lacar, B., Kirik, D., 2006. Continuous exposureto glial cell line-derived neurotrophic factor to mature dopaminergic transplantsimpairs the graft's ability to improve spontaneous motor behavior in parkinsonianrats. Neuroscience 141, 521–531.