OUP user menu

Impact of grafted serotonin and dopamine neurons on development of l-DOPA-induced dyskinesias in parkinsonian rats is determined by the extent of dopamine neuron degeneration

Thomas Carlsson, Manolo Carta, Ana Muñoz, Bengt Mattsson, Christian Winkler, Deniz Kirik, Anders Björklund
DOI: http://dx.doi.org/10.1093/brain/awn305 319-335 First published online: 27 November 2008

Summary

Previous studies have shown that serotonin neurons play an important role in the induction and maintenance of l-DOPA-induced dyskinesia in animals with lesion of the nigrostriatal dopamine system. Patients with Parkinson's disease that receive transplants of foetal ventral mesencephalic tissue, the graft cell preparation is likely to contain, in addition to dopamine neurons, serotonin neurons that will vary in number depending on the landmarks used for dissection. Here, we have studied the impact of grafted serotonin neurons—alone or mixed with dopamine neurons—on the development of l-DOPA-induced dyskinesia in rats with a partial 6-hydroxydopamine lesion of the host nigrostriatal projection. In these rats, which showed only low-level dyskinesia at the time of transplantation, serotonin grafts induced a worsening in the severity of dyskinesia that developed during continued l-DOPA treatment, while the dopamine-rich graft had the opposite, dampening effect. The detrimental effect seen in animals with serotonin neuron grafts was dramatically increased when the residual dopamine innervation in the striatum was removed by a second 6-hydroxydopamine lesion. Interestingly, rats with grafts that contained a mixture of dopamine and serotonin neurons (in ∼2:1) showed a marked reduction in l-DOPA-induced dyskinesia over time, and the appearance of severe dyskinesia induced by the removal of the residual dopamine innervation, seen in the animals with transplants of serotonin neurons alone, was blocked. FosB expression in the striatal projection neurons, which is associated with dyskinesias, was also normalized by the dopamine-rich grafts, but not by the serotonin neuron grafts. These data indicate that as long as a sufficient portion, some 10–20%, of the dopamine innervation still remains, the increased host serotonin innervation generated by the grafted serotonin neurons will have limited effect on the development or severity of l-DOPA-induced dyskinesias. At more advanced stages of the disease, when the dopamine innervation of the putamen is reduced below this critical threshold, grafted serotonin neurons are likely to aggravate l-DOPA-induced dyskinesia in those cases where the dopamine re-innervation derived from the grafted neurons is insufficient in magnitude or do not cover the critical dyskinesia-inducing sub-regions of the grafted putamen. We conclude that it is not the absolute number of serotonin neurons in the grafts, but the relative densities of dopamine and serotonin innervations in the grafted striatum that is the critical factor in determining the long-term effect of foetal tissue graft, beneficial or detrimental, on dyskinesia in grafted Parkinson's disease patients.

  • 5-HT
  • cell transplantation
  • levodopa
  • ventral mesenchephalon
  • Parkinson's disease

Introduction

Dyskinesia is a problematic side-effect of l-DOPA therapy in patients with Parkinson's disease and has also emerged as a potentially serious adverse effect induced by dopamine (DA) neuron transplants in patients that have undergone cell replacement therapy (Freed et al., 2001; Hagell et al., 2002; Olanow et al., 2003). In the clinical trials conducted so far, the effect of foetal DA neuron transplants on l-DOPA-induced dyskinesia has been variable. In some patients a significant reduction has been observed, while in others the dyskinesias have been unaffected or even made worse (Widner et al., 1992; Defer et al., 1996; Hagell et al., 1999; Hauser et al., 1999; Brundin et al., 2000). Moreover, in two placebo-controlled double-blind trials (Freed et al., 2001; Olanow et al., 2003), a significant number of grafted patients have developed off-state dyskinesias not present before transplantation. These results have come as a surprise to scientists working with foetal cell transplantation in animal models of Parkinson's disease since the experimental work has indicated that functional DA neuron grafts should be quite effective in dampening l-DOPA-induced dyskinesias in the rat Parkinson's disease model (Lee et al., 2000; Steece-Collier et al., 2003; Carlsson et al., 2006, 2007; Lane et al., 2006; Maries et al., 2006). Prior to the reports from the double-blind trials, no signs of graft-induced dyskinesias had ever been observed or reported in either rodent or primate models of Parkinson's disease. Notably, however, none of the transplant studies in rat and monkey Parkinson's disease models that formed the basis for the clinical trials had been performed in l-DOPA-treated and dyskinetic recipients.

Previous studies have suggested that discrepancies in the composition of the grafts, the inclusion of serotonin neurons in the graft cell preparation, in particular, may contribute to the variable outcome in grafted Parkinson's disease patients (Carlsson et al., 2007). In addition to DA neurons, the ventral mesencephalic (VM) tissue grafts are known to contain also serotonin neurons, and the number of these cells can vary greatly depending on the landmarks used for dissection. When present in larger numbers serotonin neurons are effective in inducing a widespread serotonergic hyperinnervation of the host striatum (Takeuchi et al., 1991; Wright et al., 1991; Ishida et al., 1998; Carlsson et al., 2007), and in rats with already established l-DOPA-induced dyskinesias. We have shown that serotonin neuron transplants will induce a progressive worsening of dyskinesia over time, without any improvement in the underlying parkinsonian motor impairments (Carlsson et al., 2007).

Serotonin neurons are particularly interesting since they have the ability to synthesize, store and release DA, formed from exogenous l-DOPA, but due to the lack of any autoregulatory feedback control the DA released from serotonin terminals will show excessive swings in response to repetitive, intermittent l-DOPA treatment (Arai et al., 1994, 1995; Maeda et al., 2005; Carta et al., 2007). Such dysregulated release of l-DOPA-derived DA is likely to be the main trigger of dyskinesia in l-DOPA-primed animals (Carta et al., 2007), and may also play a role in Parkinson's disease patients undergoing long-term l-DOPA therapy (de la Fuente-Fernandez et al., 2001, 2004). Serotonin neurons, if present in large numbers in the graft cell preparation, may thus have a detrimental effect on the development of l-DOPA-induced dyskinesias in grafted patients, and possibly also make the patients more prone to develop graft-induced off-state dyskinesias.

The present study was designed to clarify the impact of serotonin neuron-containing foetal tissue grafts, in the presence or absence of DA neurons, on the evolution of dyskinesia in l-DOPA-primed rats with partial 6-hydroxydopamine (6-OHDA) lesions of the nigrostriatal DA system, i.e. in rats that exhibit no or low-level l-DOPA-induced dyskinesia at the time of transplantation. After a 3-month period of continued l-DOPA treatment these same rats received a second 6-OHDA lesion aimed at removing the remaining forebrain DA projection, and in some cases also a lesion of the host serotonin system, so as to simulate the situation when grafted patients experience a severe loss of residual DA neurons, in combination with or without a concomitant loss of serotonin neurons. The results show that grafted serotonin neurons indeed can induce a severe worsening of l-DOPA-induced dyskinesias in previously non-dyskinetic or mildly dyskinetic rats, but that the effect is critically dependent on the presence or absence of a spared DA innervation, and the extent of the graft-derived DA innervation, in the host striatum.

Materials and Methods

Animals

A total of 198 adult female Sprague–Dawley rats (225–250 g; B&K Universal, Stockholm, Sweden) were used, and housed under standard 12:12 h light:dark cycle and ad libitum access to food and water. All surgical and behavioural procedures were approved by the Ethical Committee for use of laboratory animals in the Lund-Malmö Region.

Experimental design

The experiment was conducted in three stages, as outlined in Fig. 1.

Fig. 1

Experimental design. In Stage I all animals received a unilateral intrastriatal 6-OHDA lesion in order to obtain a partial lesion of the nigrostriatal DA pathway. Animals displayed >6 fullbody turns/min, towards the lesioned side in the amphetamine (amph) rotation test were selected. These animals, excluding five rats that were included as a saline control group [Les-Ctrl (Drug-Naïve)], were treated daily with l-DOPA (6 mg/kg; in combination with 10 mg/kg benserazide hydrochloride) for 25 days to induce stable dyskinesia (Induction phase). Sixty animals that displayed low levels of dyskinesia (total mean integrated AIM scores: 152 ± 10) were finally included in the study and allocated into four balanced groups. At 12–14 weeks after the first intrastriatal DA lesion, three of the groups received single cell suspension grafts (Trpl), containing serotonin neuroblasts (5-HT, n = 16), DA neuroblasts (DA narrow, n = 15), or a mixture of DA and serotonin neuroblasts (DA wide, n = 15). The fourth group was sham-operated and included as a dyskinetic control group [Les-Ctrl (l-DOPA), n = 14)]. In Stage II twice-weekly l-DOPA treatment was reinstated 1 week after transplantation (Maintenance phase), in the four dyskinetic groups, while the Les-Ctrl (Drug-Naïve) animals receive vehicle. In this stage dyskinesias were evaluated at 3, 6 and 10 weeks after transplantation surgery. In Stage III, half of the animals in the DA narrow, DA wide, 5-HT and Les-Ctrl (l-DOPA) groups received a second 6-OHDA lesion, in the medial forebrain bundle (MFB), in order to completely lesion the nigrostriatal pathway, while the other half [and all animals in the Les-Ctrl (Drug-Naive)] received only a sham-operation. After a 2-week drug-free period daily l-DOPA treatment (6 mg/kg) was resumed for an additional 14 days and dyskinesias were evaluated at days 1, 7 and 14. l-DOPA-, apomorphine (apo)- and amphetamine (amph)-induced rotation tests, as well as forelimb use in the cylinder test, were performed at different time-points after lesion and transplantation, as indicated.

In Stage I (induction phase) all animals received a unilateral intrastriatal 6-OHDA lesion in order to achieve a partial depletion of the nigrostriatal DA pathway (Kirik et al., 1998). About 5–6 weeks later the animals were screened using amphetamine-induced rotation (2.5 mg/kg) and 129 rats that exhibited more than six fullbody turns/min, towards the lesioned side, were selected and treated daily with l-DOPA (6 mg/kg, s.c., Research Organics, Cleveland, OH, USA) in combination with the decarboxylase inhibitor benserazide hydrochloride (10 mg/kg; Sigma-Aldrich) for 25 days to induce stable abnormal involuntary movements (AIMs). Out of 80 animals that exhibited AIMs, 60 animals with a generally low level of dyskinesia (total mean integrated AIM scores: 152 ± 10) were included in the study and allocated into four balanced groups. In addition, five l-DOPA-naïve rats [Les-Ctrl (Drug-Naïve)] were included as a second control group. About 12–14 weeks post-lesion, three of the groups received single cell suspension grafts containing either serotonin neuroblasts (5-HT, n = 16), DA neuroblasts (DA narrow, n = 15) or a mixture of DA and serotonin neuroblasts (DA wide, n = 15), obtained from the VM region of E14 rat embryos, as described (Carlsson et al., 2007). The fourth group was sham-operated and included as control [Les-Ctrl (l-DOPA, n = 14)]. In addition, 10 animals that had not developed any dyskinesia during the l-DOPA treatment (AIM score = 0) were included in the study; half of them received grafts of serotonin neuroblasts, as above [Non-Dys (5-HT)], while the other five were sham-operated [Non-Dys (Sham)].

In Stage II the l-DOPA treatment was reinstated 1 week after transplantation at a maintenance regimen of twice-weekly injections (maintenance phase), in the four dyskinetic groups and the two ‘non-dyskinetic’ animal groups, while the Les-Ctrl (Drug-Naïve) animals received saline injections. AIMs were evaluated at 3, 6 and 10 weeks after transplantation, followed by assessment of graft-induced functional recovery in the cylinder and amphetamine-induced rotation tests.

In Stage III, 14 weeks after transplantation, half of the animals in the DA narrow, DA wide, 5-HT and Les-Ctrl (l-DOPA) groups and all 10 ‘non-dyskinetic’ animals in the Non-Dys (5-HT) and Non-Dys (Sham) groups, received a second 6-OHDA lesion, this time in the medial forebrain bundle (MFB), in order to completely remove the nigrostriatal pathway. The other half, and all animals in the Les-Ctrl (Drug-Naive) group, received only a sham-operation, as a burr hole in the skull. After a 2-week drug-free period daily l-DOPA treatment (6 mg/kg) was resumed for an additional 14 days and occurrence of dyskinesias was analysed at days 1, 7 and 14. Forelimb use in the cylinder test, and tests of rotation and dyskinesia in response to either l-DOPA or apomorphine, were performed at 5–6 weeks after the MFB lesion (at 19–20 weeks after grafting). The rats in the Non-Dys (5-HT) and Non-Dys (Sham) groups received, at 8 weeks after the MFB injection, an injection of 5,7-dihydroxytryptamine (5,7-DHT) into the MFB, in order to deplete the intrinsic serotonin system in the striatum on the lesion and grafted side, and tested for l-DOPA-and apomorphine-induced dyskinesia and rotation. The brains were processed for immunohistochemistry at 21 or 25 weeks post-grafting.

Toxin lesions

Toxin injections were conducted under general anaesthesia (20:1 mixture of Fentanyl and Dormitor, i.p.). For the first partial DA lesion 28 µg of 6-OHDA (3.5 µg/µl, free base, in 0.2 mg/ml ascorbic acid; Sigma–Aldrich), was distributed over four sites in the striatum (7 µg/site) at (i) AP: +1.3 mm, ML: −2.6 mm; (ii) AP: +0.4 mm, ML: −3.2 mm; (iii) AP: −0.4 mm, ML: −4.2 mm; (iv) AP: −1.3 mm, ML: −4.5 mm, all −5.0 mm below the dura and tooth-bar set at 0.0. For the second, complete DA lesion, 14 µg of 6-OHDA (3.5 µg/µl, free base) was injected into the MFB at AP: −4.4 mm, ML: −1.2 mm from bregma, −7.8 mm below dura; tooth-bar: −2.4. The serotonin lesions were performed by injecting 10 μg 5,7-DHT creatine sulphate (5 μg/μl, free base; Sigma-Aldrich), dissolved in 0.02% ascorbic acid into the MFB at AP: −3.0 mm and ML: −1.6 mm, from bregma, −7.8 mm below dura; tooth-bar: −3.3. All injections were made at a rate of 1 µl/min, and the needle was kept in place for additional 3 min before retracted.

Transplantation surgery

The graft cell preparations were obtained from E14 rat embryos and prepared as described in Carlsson et al. (2007), with the exception that the rostal and caudal landmarks for the DA narrow grafts were moved about 0.5 mm rostral, in order to minimize the inclusion of serotonin neuroblasts. The DA narrow, DA wide and 5-HT cell suspensions had a final adjusted concentration of 128 000, 130 000 and 120 000 cells/μl, respectively. The rats were mounted in the stereotaxic frame with a 5 μl Hamilton syringe fitted with a glass capillary (ø 60–80 μm). About 120 000–130 000 cells were injected as two 0.5 μl deposits, at DV −5.0 mm and −4.0 mm (below dura), through a single needle penetration, at AP: +0.2 mm and ML: −3.5 mm; tooth-bar: 0.0. After the second deposit the syringe was kept in place for an additional 3 min before retracted. Using trypan blue viability staining in a counting chamber, the cell viability of all cell suspensions after transplantation was estimated to >95%.

Behavioural analyses

Drug-induced rotation

Rotation in response to amphetamine (2.5 mg/kg, i.p.), apomorphine (0.05 mg/kg, s.c.) or l-DOPA (6 mg/kg, s.c.) was recorded in automated rotometer bowls (AccuScan Instrument Inc., Columbus, OH, USA) over 90 min (for amphetamine and l-DOPA) or 60 min (for apomorphine) and expressed as net full body turns per minute, with negative values indicating rotation away from the lesion side.

Cylinder test

Spontaneous forelimb use was evaluated in the cylinder test (Schallert and Tillerson, 1999; Kirik et al., 2001). The rats were placed individually in a glass cylinder (20 cm in diameter), and video filmed while performing 20–25 weight-shifting movements of the forepaws in contact with the cylinder wall. The numbers of 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 rats will score 50% in this test.

l-DOPA-induced dyskinesia

AIMs, resembling peak-dose dyskinesias in patients, were induced by daily injections of l-DOPA methyl ester (6 mg/kg, s.c., with 10 mg/kg benserazide hydrochloride; Sigma-Aldrich), and shifted to a maintenance regimen of twice-weekly l-DOPA injections at the same dose (6 mg/kg) for 16 weeks (Stages I–II). A new 2-week daily l-DOPA challenge was performed after the second 6-OHDA lesion (Stage III). Pre-grafting scores were calculated as means of three consecutive tests (3 days apart) and the 3, 6 and 10 weeks post-grafting scores as means of two test sessions (4 days apart). The post-MFB lesion scores, at days 1, 7 and 14, are from single observations.

The AIMs were scored every 20 min (for at least 3 h, until no AIMs were observed) according to a rat dyskinesia scale, where severity scores from 0 to 4 were given for four dyskinetic subtypes according to topographic distribution as forelimb, orolingual, axial and locomotive behaviours, as described by Lee et al. (2000), with minor modifications (Winkler et al., 2002; Carlsson et al., 2006). In the rat model, the first components to appear after l-DOPA treatment are the limb and orolingual dyskinesias. As the chronic treatment continuous the axial dyskinesias get more and more apparent. In animals with intrastriatal lesions, where the dyskinesia is less severe as compared to a MFB lesion, the limb and orolingual dyskinesias are the main behaviours observed after l-DOPA treatment. Animals with MFB lesions (where the DA denervation is more complete) develop more severe dyskinesias and the axial component becomes more prominent. While the limb and oroligual movements are viewed as hyperkinesia, the axial component is clearly of a different, dystonic type. Enhanced normal behaviours, such as grooming, gnawing, rearing and sniffing, could be observed, but these were 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.

Apomorphine-induced dyskinesia

Dyskinesias induced by apomorphine (0.05 mg/kg; Sigma–Aldrich) were assessed at 20 weeks after transplantation and for the ‘non-dyskinetic’ groups [Non-Dys (5-HT) and Non-Dys (Sham)] at 3 weeks after the 5,7-DHT lesion. AIMs were scored every 10 mins for 90 mins, using the same rating scale as for l-DOPA-induced dyskinesia, and presented as the integrated AIM score for the whole observation period.

Histological analyses

At 21–25 weeks post-transplantation (5 weeks after the 5,7-DHT lesion in the ‘non-dyskinetic’ groups) the animals were anesthetized by an overdose of sodium pentobarbital, and perfused transcardially with 50 ml saline, followed by 250 ml of ice-cold 4% PFA in 0.1 M phosphate buffer (pH 7.2–7.6). The brains were postfixed in the same fixative for 2–3 h, placed in 25% sucrose for at least 24 h and sectioned on a freezing microtome at 35 μm. The sections were divided into eight series and stored (at −20°C) until further processing.

Immunohistochemistry

Immunostainings were performed as described (Carlsson et al., 2007) using primary antibodies against tyrosine hydroxylase [TH; mouse anti-TH (MAB318, 1:2000; Chemicon)], serotonin transporter [SERT; mouse anti-SERT (MAB1564, 1:1000; Chemicon)], serotonin [rabbit anti-serotonin (1:10 000; Immunostar Inc.)] and FosB [goat anti-FosB (SC-48X, 1:15 000; Santa Cruz)], with appropriate secondary antibodies (for TH and SERT: horse anti-mouse BA2001; for serotonin: goat anti-rabbit BA1000; and for FosB: horse anti-goat BA9500; Vector Laboratories). SERT staining was intensified by addition of nickel ammonium (2.5 mg/ml) in the 3,3′-diaminobenzidine solution. For FosB staining, 0.25% Triton X100 was added in all steps, including rinses. To minimize background, the FosB antibody was diluted 1:50 and pre-incubated with untreated forebrain sections (from control rats) for 3–5 h.

Stereology

The number of TH- and serotonin-positive cells present in the transplants were estimated using the Computer Assisted Stereological Toolbox (CAST) module in VIS software (Visiopharm A/S, Denmark) and a 60× magnification oil lens (numerical aperture: 1.40) as described (Kirik et al., 1998; Carlsson et al., 2007). The total numbers of cells were estimated according to the optical fractionator (West, 1999).

Estimation of TH and SERT-positive fibre densities

TH- and SERT-positive fibre densities were determined by optical density measurements from the stained sections using the Image J software platform version 1.38× for Mac OSX (National Institutes of Health, USA, http://rsb.info.nih.gov/ij/) as described (Carlsson et al., 2007). Measurements were made from the rostral (Ros) and caudal (Cau) parts of the head of the caudate–putamen, including two sub-areas, the rostro–dorsolateral (Rdl) and caudal-lateral (Cl) regions, as illustrated in Fig. 8A and F, and expressed as per cent of the control side. The Ros and Rdl values give the means of two sections from the rostral striatum (AP: +1.60 and +1.00), and the Cau and Cl values of two sections from the caudal striatum (AP: +0.20 mm and −0.30) (Paxinos and Watson, 2005).

Estimations of FosB-positive cell numbers

To evaluate changes in FosB expression in the striatal target neurons a single l-DOPA injection was given at 48 h before sacrifice. The total number of FosB-positive cells was evaluated by Image J software. Two high-resolution images were captured, corresponding to the rostral and caudal aspect of the head of striatum (1.00 and −0.30 mm from bregma) using Scanscope GL system with Imagescope v8.2 software. The same four striatal regions as above (Ros, Rdl, Cau and Cl) were analysed. A threshold was set to exclude basal grey matter background staining, and the total number of cells that appeared with higher optical density was automatically calculated by the software.

Statistical analysis

All statistical group comparisons were conducted using one-way ANOVAs, Two-way repeated measures ANOVA or un-paired t-test where appropriate. Post hoc analyses were performed using Dunnett's test or Tukey HSD analysis. For all comparisons statistical significance was set at P ≤ 0.05, and the significant statistics are reported in respective figure legend. All comparisons were performed using JMP Statistical software version 5.0.1.2 (SAS Institute Inc., Cary, NC, USA).

Results

The experiment was conducted in three stages (Fig. 1). In Stage I, 129 rats with partial, intrastriatal 6-OHDA lesions were given daily injections of l-DOPA (6 mg/kg, plus 10 mg/kg benserazide) for 25 days to develop a stable level of l-DOPA-induced dyskinesia. All animals exhibited >6 turns/min in the amphetamine rotation test (2.5 mg/kg), which is consistent with a 70–90% loss of the nigral DA neurons (Kirik et al., 1998; Winkler et al., 2002; Breysse et al., 2007). In Stage II, 60 of the l-DOPA-treated rats that had developed a low level of l-DOPA-induced dyskinesia (total AIM scores 152 ± 10) were allocated to four balanced groups that received intrastriatal grafts containing either DA neurons only (DA narrow, n = 15); serotonin neurons only (5-HT, n = 16); a mixture of DA and serotonin neurons (DA wide, n = 15); or sham surgery [Les-Ctrl (l-DOPA), n = 14]. An additional 10 rats that showed no signs of dyskinesia at the end of the 25-days l-DOPA treatment period (AIM score = 0) were given serotonin neuron transplants [Non-Dys (5-HT), n = 5] or sham surgery [Non-Dys (Sham), n = 5]. All rats received continued l-DOPA treatment (6 mg/kg, twice weekly) and the changes in l-DOPA-induced AIMs were monitored over the subsequent 10 weeks. In Stage III half of the animals in each of the four first groups, and all 10 rats in the ‘Non-dyskinetic’ groups, received a second 6-OHDA injection, this time in the MFB, in order to remove the remaining neurons of the nigrostriatal DA pathway. The impact of this second lesion on the severity of the l-DOPA-induced dyskinesias was monitored during 14 days of daily l-DOPA treatment, starting 2 weeks after the MFB lesion.

At sacrifice, one animal in the DA narrow and one in the 5-HT graft group were found to have large necrotic tissue damages in the cortex, and one animal in the DA wide group had a misplaced graft extending into the ventral forebrain. These animals were excluded from the study.

Development of l-DOPA-induced dyskinesia in rats with advancing degeneration of the nigrostriatal DA pathway

The rats were selected in the pre-test (in Stage I) to exhibit a low level of l-DOPA-induced dyskinesia prior to transplantation. In Stage II the non-transplanted lesioned controls [Les-Ctrl (l-DOPA)] remained mildly dyskinetic after continued l-DOPA treatment, with AIMs limited to repetitive limb and orolingual movements, without any signs of axial (torsive head, neck and trunk) movements (filled circles in Fig. 2A-A′′). In animals with DA-rich grafts dyskinesias were almost completely suppressed (79 and 57% reduction in limb and orolingual AIMs at 10 weeks after transplantation in the DA narrow and DA wide groups, respectively) (open and filled squares in Fig. 2A-A′′). Rats with serotonin neurons grafts, by contrast, showed a marked, 2.5-fold, increase in the magnitude of l-DOPA-induced dyskinesia (open triangles in Fig. 2A). This increase was most prominently due to the appearance of the more severe, dystonia-like axial component, not seen in any of the other groups (Fig. 2A′′).

Fig. 2

Impact of DA and serotonin grafts with advancing DA neurodegeneration (Stages II and III). l-DOPA-induced dyskinesia was evaluated before (pre) and at 3, 6 and 10 weeks after the transplantation surgery (A–A′′). The Les-Ctrl (l-DOPA) group maintained the low magnitude of dyskinesias throughout. In the DA-rich grafts (DA narrow and DA wide) the total AIMs scores were progressively reduced over time, reaching 79 and 57% reduction at 10 weeks respectively, while the serotonin-only graft animals (5-HT) showed an overall 2.5-fold increase (A–A′′). In the 5-HT graft group the more severe axial, dystonia-like, component, not observed in any of the other groups, became evident at 3 weeks and further increase over the following weeks (A′′). After the second 6-OHDA lesion (in the MFB) daily l-DOPA treatment was re-instated for 2 weeks, and l-DOPA-induced dyskinesia was evaluated at days 1, 7 and 14 (B–B′′). In the Les-Ctrl (l-DOPA) group total AIMs score were increased, up to 4-fold (B), seen as an increase in the limb and orolingual AIMs, and as the appearance of the more severe, dystonia-like, axial component (B′ and B′′). Although dyskinesia showed some increase over time in the DA narrow and DA wide graft groups, the overall AIMs scores remained on a low-to-moderate level. The 5-HT graft group, in contrast, showed a dramatic increase after the MFB lesion, with AIMs scores reaching levels exceeding 1000, which was marked higher than seen in any of the lesion controls. The groups receiving only sham-operation at the time of the second 6-OHDA lesion, showed no change in dyskinesia compared to their own 10-week post-transplantation scores (C–C′′). Asterisk indicates the different from Les-Ctrl (l-DOPA) group (a sum of dyskinesias at 6 and 10 weeks were calculated). A, A, A′′, B, B and B′′: One-way ANOVAs F(3,56) = 14.35, F(3,56) = 12.56, F(3,56) = 11.76, F(3,27) = 15.71, F(3,56) = 7.21 and F(3,27) = 22.15, respectively. For all comparisons P ≤ 0.0004. All ANOVAs were followed by Dunnett's control test.

In Stage III the residual dopaminergic forebrain projections were removed by an additional injection of 6-OHDA placed in the MFB. After a 2-week recovery period daily l-DOPA treatment was resumed for 14 days. In the non-grafted animals [Les-Ctrl (l-DOPA)] this resulted in a marked increase not only in the overall magnitude of l-DOPA-induced dyskinesia, as reflected in about 4-fold increase in the total integrated AIM scores (filled circles in Fig. 2B and B′), but also in increased severity, as shown by the appearance of axial dyskinetic movements (Fig. 2B′′). In the DA narrow and DA wide graft groups the dyskinesia scores also increased over the 2-week l-DOPA treatment period, but the magnitude remained much below that seen in the non-transplanted animals, and the development of the dystonia-like axial dyskinesias was effectively suppressed (open and filled squares in Fig. 2B-B′′). In the rats with serotonin grafts (5-HT) dyskinesias were dramatically increased, from a level of 398 ± 112 in total integrated AIM scored prior to the second lesion, to 1122 ± 144 after the continued l-DOPA treatment (open triangles in Fig. 2B-B′′). This increase was accompanied by the development of severe axial dyskinesia in all serotonin-grafted animals (Fig. 2B′′). In contrast, in the animals that received a sham-operation instead of the second 6-OHDA lesion, dyskinesias remained unchanged over the l-DOPA treatment period, as compared with the 10-weeks time-point (Fig. 2C-C′′).

Development of dyskinesia in non-dyskinetic rats

In Stage II of the experiment ten l-DOPA-treated but non-dyskinetic 6-OHDA lesioned rats received either serotonin neuron grafts [Non-Dys (5-HT), n = 5] or sham surgery [Non-Dys (Sham), n = 5] (Fig. 3A). These rats developed no signs of dyskinesia during the 25-day l-DOPA priming period (in Stage I), and they all remained essentially non-dyskinetic during the second period of l-DOPA treatment (Post Trpl, Fig. 3B). After the second 6-OHDA lesion, in the MFB, performed 2-weeks post-transplantation, both groups developed severe l-DOPA-induced dyskinesia with mean integrated AIM scores of 918 ± 220 and 786 ± 152 for the Non-Dys (5-HT) and Non-Dys (Sham) groups, respectively (Post MFB lesion in Fig. 3B), as well as marked turning behaviour in response to low doses of either apomorphine or l-DOPA (Fig. 3D and E).

Fig. 3

Effect of serotonin grafts, and lesion of the intrinsic serotonin system, in l-DOPA-treated but non-dyskinetic animals. In order to investigate the effect of serotonin grafts in the presence or absence of an intrinsic serotonin innervation, ‘non-dyskinetic’ animals, i.e. 6-OHDA lesioned rats that did not show any signs of dyskinesia after the 25-day induction phase, were transplanted with serotonin grafts or sham-operated and later the intrinsic serotonin fibres were removed by a 5,7-DHT injection in the MFB (A: Experimental design; B: Post Intrastriatal Lesion). After the transplantation [Non-Dys (5-HT)] or sham surgery [Non-Dys (Sham)] none of the animals displayed any notable increase in dyskinesia (B: Post Trpl). However, after the second 6-OHDA lesion, placed in the MFB, both groups displayed severe AIMs of all sub-categories (B: Post MFB lesion). Following the 5,7-DHT lesion dyskinesias in the sham-operated group were completely abolished, while the dyskinesias seen in the serotonin graft group [Non-dys (5-HT)] remained unchanged (B: Post 5,7-DHT lesion). In agreement, l-DOPA-induced rotation was completely abolished in the sham-operated rats, but unchanged in the serotonin transplanted animals (E). In contrast to the dyskinesias and rotation induced by l-DOPA, apomorphine-induced dyskinesias and rotation were both unaffected by the serotonin lesion (C and D). SERT immunohistochemistry revealed an almost complete denervation of the serotonin fibres in the striatum after the 5,7-DHT lesion (7.2 ± 4.4% of intact side, F), while the transplanted animals showed extensive re-innervation (108.4 ± 18.2% of intact side, G). Scale bar in F represents 3 mm. Asterisk indicates the different from post MFB. B: One-way ANOVAs: F(3,19) = 21.22, P < 0.0001 and F(3,19) = 26.96, P < 0.0001 for and Non-Dys (Sham) and Non-Dys (MFB), respectively. E: t-test, P = 0.045.

To assess the role of the intrinsic serotonin neurons for the induction of dyskinesia in these animals, the selective serotonin neurotoxin 5,7-DHT was injected into the MFB 8 weeks after the second 6-OHDA lesion, in order to remove the intrinsic forebrain serotonin afferents on the lesion and grafted side (Fig. 3F and G). Consistent with previous data (Carta et al., 2007) l-DOPA-induced (but not apomorphine-induced) dyskinesia was completely abolished in the Non-Dys (Sham) animals, but remained unchanged in the animals with serotonin neuron grafts in the Non-Dys (5-HT) group (Post 5,7-DHT lesion in Fig. 3B and C). In the rotation test l-DOPA-induced turning was completely abolished in the non-grafted Non-Dys (Sham) animals (0.1 ± 0.1 turns/min) but remained at a high level (−14.0 ± 3.5 turns/min) in the serotonin graft group (Fig. 3E). Postsynaptic DA receptor supersensitivity, as reflected in the apomorphine-induced turning response (Fig. 3D), was unaffected by the 5,7-DHT lesion.

Impact on postsynaptic receptor supersensitivity

In order to monitor changes in postsynaptic DA receptor supersensitivity in the lesioned and grafted animals, dyskinesia and rotational behaviour in response to a low dose of the mixed D1/D2 agonist apomorphine (0.05 mg/kg, s.c.) were assessed at 19–20 weeks post-transplantation (Fig. 4). In rats with partial lesions, apomorphine-induced dyskinesias (limb and orolingual AIMs) were reduced by 81–86% in the two DA neuron containing graft groups (DA narrow and DA wide), as compared to the non-grafted Les-Ctrl group, while the 5-HT graft group remained unchanged (Fig. 4A, Intrastriatal + Sham). In the animals that had received the second 6-OHDA lesion, the dyskinesias increased 2- to 3-fold in both the lesion controls and in the 5-HT graft group, but remained unchanged, or on a low level, in the DA neuron-containing graft groups (Fig. 4A, Intrastriatal + MFB).

Fig. 4

Impact of DA and serotonin grafts on DA receptor supersensitivity with advancing DA neurodegeneration. Apomorphine-induced dyskinesia (A) and rotation (B) as well as l-DOPA-induced rotation (C) were assessed in order to monitor the development of postsynaptic DA receptor supersensitivity in the lesioned and grafted animals. In rats with partial lesions (Intrastriatal + Sham), apomorphine-induced dyskinesias (limb and orolingual AIMs) were reduced by 81–86% in the two DA neuron containing graft groups (DA narrow and DA wide), as compared with the non-grafted Les-Ctrl group, while the 5-HT graft group remained unchanged (A). In the animals that had received the second DA lesion (Intrastriatal + MFB), dyskinesias increased several-fold in both the lesion controls and in the 5-HT graft group, but remained unchanged, or on a low level, in the DA neuron-containing graft groups. In animals with intrastriatal lesion only (Intrastriatal + Sham) rotation induced by apomorphine or l-DOPA were not observed in either of the groups after the intrastriatal 6-OHDA lesion, however the serotonin graft group (5-HT) showed a tendency to increase in rotation away from the lesion side (B and C). After the second DA lesion (Intrastriatal + MFB), apomorphine- and l-DOPA-induced rotations increased in both the Les-Ctrl (l-DOPA) and 5-HT groups, but effectively repressed in the DA narrow and DA wide groups (B and C). Interestingly, l-DOPA-induced rotation showed a dramatic, 6.5-fold increase in the 5-HT graft group compared to the lesion only controls (C). Asterisk indicates different from Les-Ctrl (l-DOPA) group. A–C: One-way ANOVAs F(3,28) = 11.82/F(3,27) = 24.35, P < 0.0001 and F(3,28) = 10.37/F(3,27) = 21.05, P < 0.0001 and F(3,28) = 3.77, P < 0.023/F(3,27) = 22.91, P < 0.0001 for Intrastriatal + Sham/Intrastriatal + MFB and A–C, respectively. All ANOVAs were followed by Dunnett's control test.

Similar changes were evident in the apomorphine-induced rotation test (Fig. 4B). Apomorphine-induced rotation, in the direction contralateral to the lesion, was minimal or absent in the partially lesioned animals (Intrastriatal + Sham), but became apparent after the second 6-OHDA lesion (Intrastriatal + MFB). The increased rotation seen after the second lesion in the lesion controls, from 0.0 ± 0.1 to −7.0 ± 1.0 turns/min, was blocked by the DA neuron containing grafts. The 5-HT graft group, by contrast, showed a 2-fold higher apomorphine-induced rotation, compared with the lesioned controls [Les-Ctrl (l-DOPA)], after the second 6-OHDA lesion.

In the l-DOPA-induced rotation test, the animals with partial lesions showed only minimal response to the l-DOPA (Fig. 4C, Stage II: Intrastriatal + Sham). After the second lesion the animals in the Les-Ctrl (l-DOPA) group displayed an increased rotation, from 0.1 ± 0.1 to −1.8 ± 1.0 turns/min. Strikingly, the 5-HT graft animals showed an average 6.5-fold higher rotation rate (from −0.5 ± 0.3 to −11.4 ± 2.1 turns/min) as compared to the lesion control group, while the response was completely suppressed in the two DA graft groups (Fig. 4C).

The apomorphine-induced rotation indicates that the increase in postsynaptic DA receptor supersensitivity that developed in the lesioned controls after the second lesion was differentially affected by the two types of graft: normalized in the rats with DA neuron containing grafts, but made worse by the serotonin neuron transplants.

Graft-induced functional improvement

The turning bias seen in the amphetamine rotation test, present in all animals before transplantation, was completely reversed in the two DA graft groups, but not in the 5-HT graft group after grafting (Fig. 5A). Similarly, the left forelimb use in the cylinder test, which was reduced to <10% in the lesion control animals, was restored to normal performance in the DA narrow and DA wide graft groups (54.7 ± 6.5% and 55.6 ± 6.7% of intact side, respectively) but unaffected by the 5-HT grafts (9.9 ± 4.7% of intact side; Fig. 5B). Together, these data indicate that functional DA release, accompanied by reversal of DA receptor supersensitivity was restored in the grafted striatum in the two groups that had received DA neuron-containing transplants. However, the functional recovery induced by the DA-rich grafts, seen in Stage II (Fig. 5B), was partially lost after the second 6-OHDA lesion (Fig. 5C), which is consistent with previous findings (Breysse et al., 2007). Amphetamine-induced rotation after the second 6-OHDA lesion was not evaluated in the current study, however, previous data have shown that recovery in rotation by DA neuron grafts in partial DA lesioned rats is not effected by a second complete lesion placed in the MFB (Breysse et al., 2007).

Fig. 5

Functional impact of DA and serotonin grafts. Changes in amphetamine-induced rotation were assessed at 11 weeks post-grafting (A), and forelimb use in the cylinder test at 12 weeks post-grafting (B) and at 5 weeks after the second DA lesion (C). After the first, intrastriatal, lesion all groups displayed a strong amphetamine-induced rotational bias ipsilateral to the lesion side. Following transplantation the rotation in the DA narrow (−0.2 ± 0.8 turn/min) and DA wide (0.7 ± 0.9 turn/min) groups were completely reversed back to normal, while the 5-HT and the Les-Ctrl (l-DOPA) groups were unaffected (19.3 ± 2.3 and 18.5 ± 1.8 turn/min, respectively). In the cylinder test, which evaluates the rats’ spontaneous forelimb use, all groups were severely impaired and balanced prior to transplantation (Les-Ctrl (l-DOPA): 11.1 ± 2.0%, DA narrow: 12.5 ± 3.9%, DA wide: 11.5 ± 2.6% and 5-HT: 8.7 ± 2.9% of intact side). After transplantation the DA neuron graft groups were reversed back to normal performance (DA narrow: 54.7 ± 6.5% and DA wide: 55.6 ± 5.7% of intact). The 5-HT (9.9 ± 4.7% of intact) and the Les-Ctrl (l-DOPA) (6.1 ± 1.9% of intact) groups were unaffected. The second 6-OHDA lesion, however, partly removed the beneficial effect seen in the cylinder test in the DA grafted animals. Dashed lines in B and C represent the performance of a normal, unbiased, animal. Asterisk indicates different from respective pre-value. (AB) Two-way repeated measures ANOVAs: Time interactions, F(3,53) = 97.99, P < 0.001 and F(3,53) = 62.40, P < 0.0001, respectively.

Changes in striatal FosB expression

In order to monitor postsynaptic changes in the striatal target neurons in the l-DOPA-treated dyskinetic animals, FosB-protein expression in the striatum was evaluated 48 h after the last l-DOPA injection. The number of FosB-positive cells, which is known to be strongly correlated with the magnitude of l-DOPA-induced dyskinesia in the rat model (Andersson et al., 1999; Winkler et al., 2002), were counted in the rostral (Ros) and caudal (Cau) striatum, as well as in two sub-regions, the rostro–dorsolateral (Rdl) and Caudal-lateral (Cl) areas of the striatal head (Fig. 8A).

In the animals with partial lesions of the nigrostriatal DA pathway, i.e. the animals that had received intrastriatal 6-OHDA followed by a sham injection in the second surgery (Intrastriatal + Sham in Fig. 6A–D), FosB was expressed in a large number of cells on the lesion side, ranging from 926 ± 177 to 1579 ± 168 cells in all four regions analysed (Fig. 6A–D, open bars). The animals in the DA graft groups (DA narrow and DA wide) had significantly lower numbers of FosB-positive cells in all four regions (grey bars in Fig. 6A–D), ranging from 136 ± 32 to 517 ± 103, similar to the number observed on the non-lesioned side (averages of number of cells on the intact side, ±SD, are indicated by the full and dashed horizontal lines in Fig. 6A–D). In the 5-HT graft group the FosB-positive cell number tended to exceed that seen in the dyskinetic Les-Ctrl (l-DOPA) controls (from 1381 ± 242 to 2164 ± 316 cells in the four regions analysed; black bars in Fig. 6A–D). These differences, however, did not reach significance.

Fig. 6

Effect of DA and serotonin grafts on striatal FosB expression. FosB-positive cells were estimated in the striatum in four areas corresponding to the rostral, rostro–dorsolateral (Rdl), Caudal and Caudal-lateral (Cl) striatum (Fig. 8A). In the dyskinetic lesion control animals [Les-Ctrl (l-DOPA)] extensive FosB expression were observed in all regions after the first intrastriatal 6-OHDA lesion (Intrastriatal + Sham; A–D) and was further increased after the second DA lesion (Intrastriatal + MFB; A–D). Increased FosB expression was mainly observed in DA denervated areas, as seen in the rostro–dorsolateral and caudal regions in the lesion control group after the intrastriatal lesion (E), and further also in the rostral region after the second 6-OHDA lesion (F). The serotonin neuron-containing grafts (5-HT) did not show an overall significant increase in FosB expression as compared to the lesion-only controls [Les-Ctrl (l-DOPA)]; individual cases, however, showed a clear tendency to an increase in both number, distribution and staining intensity of the FosB-expressing cells, as illustrated in I and J. The DA graft groups (DA narrow and DA wide), in contrast, reversed FosB expression in all regions after the first lesion (Intrastriatal + Sham), and expression in the rostro–dorsolateral and caudal areas, close to the grafted DA cells, were effectively repressed also after the second 6-OHDA lesion (Intrastriatal + MFB; B–D, G and H). However, in the rostral striatum, which were poorly innervated by TH-positive fibres from the graft, the protein expression were not different from the Les-Ctrl (l-DOPA) group (Intrastriatal + MFB; A). The horizontal lines in A–D represent the average cell number (full) ± 1 SD (dashed) on the intact side (all animals in respective region and lesion). Insets give the total AIM scores for the illustrated animals. Scale bars in J represent 1 mm. Asterisk indicates different from Les-Ctrl (l-DOPA) group. A: One-way ANOVA F(3,28) = 10.71, P = 0.0001 intrastriatal + MFB; B-D: One-way ANOVAs F(3,28) = 12.53, P < 0.0001, F(3,28) = 18.61, P < 0.0001 and F(3,28) = 19.03, P < 0.0001/F(3,27) = 9.29, P = 0.0003, F(3,27) = 5.95, P = 0.004 and F(3,27) = 10.03, P = 0.0002 for Intrastriatal + Sham/Intrastriatal + MFB and B–D, respectively. All ANOVAs were followed by Dunnett's control test.

In the animals that received the second 6-OHDA injection (Intrastriatal + MFB in Fig. 6A–D), FosB expression was further increased in the non-grafted controls [Les-Ctrl (l-DOPA)], from 1471 ± 241 to 2674 ± 526 cells. Similar high expression levels were observed in all regions in the 5-HT graft group (from 1838 ± 311 to 2962 ± 613 cells). The animals in the DA graft groups displayed significantly lower expression of FosB protein, in particular in the Rdl (524 ± 183 and 390 ± 170 cells) and Cl (372 ± 122 and 431 ± 179 cells) regions, which were not different from that seen on the contralateral intact side (DA narrow: P > 0.11, in all regions; DA wide: P > 0.14, in all regions, un-paired t-tests).

Interestingly, in the lesioned control animals FosB-expressing cells were found mainly in the completely DA denervated areas, as determined from adjacent TH stained sections (Fig. 6E). Following the second 6-OHDA lesion the head of striatum was almost completely devoid of TH-positive innervation and the FosB expressing cells were widely distributed over large areas of the striatum (Fig. 6F). In the DA narrow and DA wide groups FosB expression was completely normalized in the areas reinnervated by the graft-derived TH-positive fibres (Fig. 6G and H). In the 5-HT graft group the graft-derived serotonin innervation was unable to reverse FosB expression; both the number and staining intensity of the FosB-positive cells tended to be increased, above that seen in the lesion-only controls, in the areas innervated by the graft-derived serotonin fibres (compare Fig. 6I and J with E and F).

In the non-dyskinetic animals, that received an additional unilateral serotonin lesion, the serotonin transplanted group [Non-Dys (5-HT)] tended to increase in FosB expression in all regions in striatum (ranging from 1893 ± 620 to 4149 ± 1034), as compared to the sham-operated group [Non-Dys (Sham)], from 1038 ± 134 to 2761 ± 669. This was observed in particular in the Rdl part [2499 ± 477 versus 1265 ± 158 in the Non-dys (5-HT) and Non-dys (Sham), respectively, P = 0.06, un-paired t-test]. However, the cell numbers in both groups were comparable to respective dyskinetic group, Non-Dys (5-HT) versus 5-HT group (P > 0.28 for all regions, un-paired t-test) and Non-Dys (Sham) versus Les-Ctrl (l-DOPA) group (P ≥ 0.14 for all regions, un-paired t-test).

Histological analysis of graft survival and fibre outgrowth

TH and serotonin immunohistochemistry revealed surviving grafts in all transplanted animals (Fig. 7A and F). The DA narrow grafts contained high numbers of TH-positive cells, 3621 ± 459 (3822 ± 825 and 3419 ± 466 in Intrastriatal + Sham and Intrastriatal + MFB respectively, P = 0.68, un-paired t-test), but virtually devoid of any serotonin-positive neurons. Conversely, the 5-HT graft group contained high numbers of serotonin neurons, 2256 ± 308 (1860 ± 298 and 2709 ± 540 in Intrastriatal + Sham and Intrastriatal + MFB, respectively, P = 0.20, un-paired t-test), with very few TH-positive cells, 19 ± 4 (11 ± 4 and 29 ± 6 in Intrastriatal + Sham and Intrastriatal + MFB, respectively, P = 0.04, un-paired t-test). The DA wide grafts, in contrast, included high numbers of both DA, 3633 ± 501 (2510 ± 524 and 4355 ± 798 in Intrastriatal + Sham and Intrastriatal + MFB, respectively, P = 0.16, un-paired t-test), and serotonin neurons, 2307 ± 325 (1929 ± 260 and 2686 ± 585 in Intrastriatal + Sham and Intrastriatal + MFB, respectively, P = 0.27, un-paired t-test), with an average DA:serotonin neuron ratio of 1.8:1.

Fig. 7

TH- and serotonin-positive cell numbers and TH- and SERT-positive fibre innervation in the striatum. The total number of grafted TH- and serotonin-positive cell in the striatum was estimated using stereological technique (A and F). The DA narrow grafts contained high numbers of TH-positive cells (3621 ± 459) without inclusion of serotonin-positive neurons, and the 5-HT graft group contained high numbers of serotonin neurons (2256 ± 308), with very few TH-positive cells (19 ± 4). In the DA wide grafts, high numbers of both TH-positive (3633 ± 501) and serotonin-positive neurons (2307 ± 325), were found. The extent of TH- and SERT-positive fibre innervation was determined, by optical density measurements from the TH and SERT stained sections, for the rostral and the caudal aspect of the head of the striatum, and two sub-regions, representing Rostro–dorsolateral (Rdl) and Caudal-lateral (Cl) areas (illustrated in A and F in Fig. 8). The initial intrastriatal 6-OHDA lesion depleted the TH-positive fibres in the rostral striatum with ∼80%, while in the caudal striatum ∼90% of the DA fibres were lesioned (B and D, Intrastriatal+Sham). The spared fibres in the rostral region were mainly located in the medial and ventral parts of the striatum, and the rostro–dorsolateral regions of the striatum showed less than 10% spared fibres (C, Intrastriatal+Sham). The second 6-OHDA lesion removed the remaining DA fibres and <4% of normal innervation remained in the four regions evaluated (BEIntrastriatal+MFB). In the DA narrow and DA wide groups significant TH-positive fibre reinnervation (∼12–14% of normal) was observed around the graft, in the caudal and caudal-lateral regions (D and E). The first 6-OHDA lesion resulted in a marginal loss of the serotonin fibres (∼20%; GJ, Intrastriatal+Sham). The second lesion further removed an additional ∼10% of the SERT-positive fibres (to ∼70% of intact side) in the striatum (G–JIntrastriatal+MFB). In the 5-HT and DA wide groups an extensive SERT-positive hyperinnervation was evident in the striatum at all levels (143–238% and 139–246% of intact side in the intrastriatal+Sham and intrastriatal+MFB groups respectively; GJ). In contrast, the DA narrow group, which contained no serotonin neurons, showed no increased SERT-positive innervation at any level. Ash indicate different from other groups; Asterisk indicate different from Les-Ctrl (L-DOPA) group. A and F: One-way ANOVAs F(2,42) = 30.30, P < 0.0001 and F(2,42) = 25.13, P < 0.0001 respectively; D and E: One-way ANOVAs F(3,28) = 2.91, P = 0.05 and F(3,28) = 3.55, P = 0.03/F(3,28) = 4.85, P = 0.009 and F(3,28) = 5.19, P = 0.007 for Intrastriatal+Sham/Intrastriatal+MFB and D and E, respectively; (GJ) One-way ANOVAs F(3,28) = 11.73, P < 0.0001, F(3,28) = 9.04, P = 0.0003, F(3,28) = 16.69, P < 0.0001 and F(3,28) = 13.59, P < 0.0001/F(3,27) = 8.47, P = 0.0005, F(3,27) = 13.83, P < 0.0001, F(3,27) = 9.05, P = 0.0003 and F(3,27) = 8.28, P = 0.0006 for Intrastriatal+Sham/Intrastriatal+MFB and G–J, respectively. All ANOVAs were followed by Tukey HSD analysis (A and F) or Dunnett's control test (BE, GJ).

The intrastriatal 6-OHDA lesion depleted the TH-positive fibres in the caudal striatum by more than 90%, while the rostral region was partly spared (∼20% of normal; Figs 7B–E and 8A). The second 6-OHDA lesion removed the spared fibres in the whole striatum (Figs 7B–E and 8B). In the DA narrow and DA wide graft groups, the transplanted DA neurons restored TH-positive fibre density in the caudal striatum by 12–14% of normal, while the rostral parts were overall poorly innervated (Ros: 3–5%, Rdl: 6–9% of intact side; Figs 7B–E, 8C and D).

Fig. 8

TH and SERT immunohistochemistry. After the first intrastriatal 6-OHDA lesion (A) the caudal part of the striatum was severely depleted, while spared TH-positive innervation could be observed in the rostral (in particular in medial and ventral) regions. The second 6-OHDA lesion, placed in the MFB, (B) removed the remaining DA fibres and left the striatum completely denervated. In the animals receiving DA-rich grafts (DA narrow and DA wide) a TH-positive reinnervation could be observed around the graft core in the caudal aspect of the head of the striatum (C and D), while in the group receiving serotonin grafts (5-HT), with only few included TH-positive neurons in the grafts, the striatum remained completely denervated (E). In the animals with serotonin-rich grafts (5-HT and DA wide), strong hyperinnervation was observed in the whole striatum, in particular in the Rdl region (I and J). A distinct, localized increase in SERT-positive fibres was observed in the ventrolateral striatum in many of the non-grafted lesion controls, and in the DA narrow graft group (marked by asterisks in F, G and H). Scale bar in J represents 3 mm.

SERT immunohistochemistry showed that the intrastriatal 6-OHDA lesion had caused a minor damage (19–23%) to the serotonin innervation in the striatum (Figs 7G, I and 8F). The second DA lesion led to an additional loss of about one-third of the SERT-positive striatal serotonin innervation, to 66–69% of intact side in the rostral and caudal striatum (Figs 7G–J and 8G). In the 5-HT and the DA wide groups, which contained high numbers of serotonin neurons in their grafts, a marked SERT-positive hyperinnervation was observed at all levels investigated (Figs 7G–J, 8I and J). This hyperinnervation was particularly prominent in the Rdl region, where the density of SERT-positive fibres was increased to 178–237% and 171–260% of intact side in the 5-HT and DA wide graft groups, respectively (Fig. 7H).

Interestingly, a localized SERT-positive serotonin hyperinnervation was observed in the ventrolateral region of the caudal striatum in a significant number (9 of 14) of the non-transplanted animals in the dyskinetic lesion control groups (asterisk in Fig. 8F and G), both in the animals receiving the second 6-OHDA lesion (Intrastriatal + MFB; 2 of 7) as well as in the animals receiving only the partial, intrastriatal lesion (Intrastriatal + Sham; 7 of 7). This effect was also observed in the DA narrow group, (10 of 14; Fig. 8H) as well as the Drug-Naïve (i.e. non-l-DOPA-treated) lesion control animals (3 of 5). This serotonin hyperinnervation, however, did not correlate with either the magnitude of dyskinesia, or with the pattern of FosB expression, suggesting that this sprouting response may be induced by the 6-OHDA lesion, independent of any l-DOPA treatment. Indeed, close comparison of adjacent SERT and FosB stained sections showed that regions of increased density of SERT-positive fibres displayed only low numbers of FosB-expressing cells.

Discussion

The study was designed to investigate the influence of serotonin neuron-containing grafts on the development of l-DOPA-induced dyskinesia in chronically l-DOPA-treated rats with partial 6-OHDA lesions of the nigrostriatal DA pathway, that had developed no or only low-level dyskinesias at the time of transplantation. Three months later, the rats were subjected to a second 6-OHDA lesion, in order to simulate the situation in grafted, mildly dyskinetic patients when the degeneration of the intrinsic DA system has progressed further. In Parkinson's disease patients, disease progression is accompanied by a progressive worsening of the l-DOPA-induced peak-dose dyskinesias and a reduction of the therapeutic window of l-DOPA medication (Mouradian et al., 1990; Obeso et al., 2000). Similarly, the low-level dyskinesia seen in the partially lesioned rats (in Stages I and II of the experiment) increased both in magnitude and severity when the residual striatal DA innervation was removed by the second lesion (Stage III). The DA neuron-rich grafts had a marked dampening effect on dyskinesia present at the time of transplantation, as seen earlier (Lee et al., 2000; Steece-Collier et al., 2003; Carlsson et al., 2006, 2007; Lane et al., 2006; Maries et al., 2006), and they were also able to block the increase in dyskinesia induced by the second DA lesion, as well as the increase in striatal FosB expression that is closely linked to the development of dyskinesia in l-DOPA-treated animals. In the DA narrow and DA wide graft groups dyskinesias remained on a low-to-moderate level, also after removal of the remaining DA afferents, without any clear signs of the more severe, dystonia-like axial dyskinesias (for further explanation of the axial component see Materials and Methods section). Dyskinesia induced by direct stimulation of postsynaptic DA receptors by a low dose of apomorphine was almost completely eliminated. The results obtained in the amphetamine-induced rotation and forelimb-use tests, as well as the normalization of DA receptor supersensitivity in the apomorphine rotation test, indicate that a functional DA innervation had been re-established from the grafted DA neurons in these rats.

The changes induced by serotonin-rich grafts were quite different: The low-level dyskinesia present in the partially lesioned rats at the time of transplantation increased after transplantation in the 5-HT group, as reflected in an average 2-fold increase in total AIM score and the appearance of the more severe, dystonia-like, axial dyskinesia during Stage II of the experiment. Further, there was a dramatic increase in both the magnitude and severity of dyskinesia after the second DA lesion (Stage III). The serotonin transplants did not induce any functional improvement in the amphetamine-induced rotation and cylinder tests, the postsynaptic DA receptors remained supersensitive (as assessed by the apomorphine- and l-DOPA-induced rotation tests), and both apomorphine-induced dyskinesia and cellular FosB expression remained as high, or increased, compared with the dyskinetic lesion-only controls.

Interaction of the serotonin neurons with the spared DA innervation of the host

In the rat Parkinson model the magnitude, and also the expression pattern, of l-DOPA-induced dyskinesia is critically dependent on an interaction between the spared DA innervation and the intrinsic serotonin afferents (Carta et al., 2007). In the lesioned controls studied here the dyskinesia scores recorded in the animals with partial intrastriatal 6-OHDA lesions increased 3-fold when the residual DA innervation was removed. Consistent with previous findings (Carta et al., 2007) all components of AIMs (orolingual, limb, axial and locomotive) were completely eliminated by the 5,7-DHT lesion, indicating that the dyskinetic movements were triggered almost exclusively by DA formed and released by the intrinsic serotonin afferents.

In rats with serotonin-rich transplants the grafted serotonin neurons provided a widespread serotonergic hyperinnervation of the host striatum, with a terminal density that was up to 2.5-fold higher than normal. In the partially lesioned rats the impact of this graft-derived serotonergic hyperinnervation was rather modest: dyskinesia was increased above that seen in the lesion-only controls, but remained at a low-to-moderate level, i.e. at a total AIM score of below 400. Removal of the residual intrinsic DA innervation, however, caused a dramatic increase in both the magnitude and severity of l-DOPA-induced dyskinesia, with the appearance of intense axial movements, and total AIM scores that exceeded 1100. These data indicate that the spared intrinsic DA innervation has an important dampening effect on the serotonin-dependent dyskinesias, and that DA released from serotonin terminals is detrimental in animals where the striatum is completely denervated. At that stage of the disease, the serotonin terminals may be the only source of DA release in the striatum, at least when low doses of l-DOPA are administered as in the present experiment.

The mechanism underlying this dampening effect has most probably both a pre and a postsynaptic component. On the presynaptic side, the swings in extracellular levels of l-DOPA-derived DA, released as a false transmitter from the serotonin terminals, will be counteracted by an active re-uptake of DA from the extracellular space by the remaining DA terminals. In this situation the interplay between the serotonin and DA terminals may, in fact, be functionally beneficial in that the serotonin terminals can provide additional sites for synthesis and storage of l-DOPA-derived DA that can be utilized by the residual DA terminals. On the postsynaptic side, the residual DA innervation, with the help of the exogenously supplied l-DOPA, will be able to maintain the postsynaptic DA receptors in a normo-sensitive state, as indicated by the absence of apomorphine-induced rotation in the partially lesioned rats. The postsynaptic DA receptor supersensitivity that developed after the removal of the remaining DA afferents by the MFB lesion (as reflected by the marked increase in apomorphine-induced turning and cellular FosB expression) is likely to make the striatal neurons prone to elicit abnormal behavioural responses.

The dysregulated, pulsatile release of DA from the graft-derived serotonin terminals was clearly non-functional (in the amphetamine-induced rotation and cylinder tests) and, interestingly, also unable to normalize DA receptor supersensitivity (in the apomorphine-induced rotation test) and striatal FosB expression. In situations when the residual DA innervation has degenerated, therefore, DA released from serotonin terminals will trigger more severe dyskinetic responses by acting on supersensitive DA receptor sites, in the absence of any beneficial effects of the l-DOPA treatment.

Interaction between DA and serotonin neurons in grafts containing a mixture of both neuron types

In one of the transplant groups, the DA wide group, the dissection of the VM tissue was made so as to obtain grafts that contained both DA and serotonin neurons in large numbers. As in our previous study (Carlsson et al., 2007) these mixed grafts restored the TH-positive innervation in the caudal striatum to 12–14% of normal, and increased the host serotonin innervation to levels up to 2-fold, above normal, in particular, in the rostro–dorsolateral part of the striatum, and they were as efficient as the DA narrow grafts (i.e. the transplants containing DA neurons only) in restoring DA-dependent motor behaviour in the cylinder and amphetamine-induced rotation tests. Most notably, however, these mixed grafts were as efficient as the DA neuron only grafts in the dampening of l-DOPA-induced dyskinesia, both in partially lesioned animals (Stage II) and after removal of the remaining host DA afferents (Stage III).

Interestingly, the dampening effect on dyskinesia in the rats with DA wide grafts, i.e. grafts with mixed DA and serotonin neurons, was opposite to the worsening seen in rats with grafts containing serotonin neurons only (i.e. the 5-HT graft group). It seems unlikely that this difference was due to any major differences in number of serotonin neurons in the grafts, or the extent of graft-derived serotonin innervation of the host striatum since several of the animals in the DA wide group, that showed marked reductions in l-DOPA-induced dyskinesia, had similar numbers of serotonin neurons in the graft and similar levels of SERT-positive striatal innervation as animals in the 5-HT graft group that showed the opposite effect. This indicates that the effect of the graft-derived serotonin innervation is detrimental only in the absence of a significant functional dopaminergic input. Similar to the spared host DA innervation, as discussed above, we propose that the DA innervation derived from the DA neurons in the graft will be effective in blocking the pro-dyskinetic effect of grafted serotonin neurons by a combination of efficient re-uptake of l-DOPA-derived DA from the extracellular space (which will help to dampen the excessive swings in DA released from the serotonin terminals), and reversal of postsynaptic DA receptor supersensitivity in the re-innervated areas. This is supported by the observations that dyskinesia induced by direct stimulation of the postsynaptic receptors, by a low dose of apomorphine, was much reduced in the DA wide group, and that apomorphine- and l-DOPA-induced turning, and increased FosB expression in the striatal projection neurons, were almost completely eliminated in the DA wide graft group (but increased in the 5-HT graft group).

Clinical implications

The results show that the development of l-DOPA-induced dyskinesia in the rat Parkinson's disease model depends on a close interaction between the dopaminergic and serotonergic afferents. In non-grafted 6-OHDA lesioned rats systemic l-DOPA administration induces AIMs only when the loss of DA neurons in the SN is greater than ∼80% and the loss of DA innervation in the lateral part of the striatum exceeds ∼90% (Winkler et al., 2002). This suggests that when dyskinesias is induced in the rat model by low l-DOPA doses (6–12 mg/kg; which corresponds to the doses used clinically), as little as 10% of the normal DA innervation density in the critical, lateral dyskinesia-triggering part of the caudate–putamen may be sufficient to buffer the swings in DA released from the serotonin terminals. Since the DA innervation in the intact striatum is some 15- to 20-times higher in density than the serotonin ones (as judged by the tissue levels of DA and serotonin; see Carta et al., 2007) this suggests that a DA:serotonin terminal ratio of about 1:1, or higher, would be sufficient to keep the dyskinesias in check. Indeed, when this situation is met DA synthesized and released from the serotonin terminals might be beneficial and contribute positively to the therapeutic efficacy of l-DOPA treatment. It is interesting to compare these values, obtained in rats, with the situation in Parkinson's disease patients. Similar to the rat striatum the DA:serotonin ratio in the putamen of control human cases is about 20:1 (Kish et al., 2008). In advanced cases of Parkinson's disease the putaminal DA levels are typically reduced by more than 90% (by 98% in the Kish et al., 2008 study), while the loss of putaminal serotonin innervation (as assessed by levels of serotonin, SERT and tryptophan hydroxylase) is less, and also more variable (30–65%). As a result, the DA:serotonin ratio is reduced to an average of about 1:1 (Kish et al., 2008), i.e. within the range where dyskinesias would be expected to be held in check by the residual DA innervation. Kish et al. did not observe any difference in putaminal DA:serotonin ratio in patients with or without recorded dyskinesias, which may be explained by the fact that the tissue samples used for analysis included the whole putamen. If the induction of l-DOPA-induced dyskinesias is critically dependent on DA released in a particular, dyskinesia-prone sub-region of the putamen (as is the case in rodents; see Cenci et al., 1998) then the overall putaminal tissue levels may mask important regional differences in the DA and serotonin innervation densities that may determine the patient's sensitivity to develop dyskinetic side-effects.

In grafted animals, we propose that the impact of serotonin neurons in the transplant will depend on the density of graft-derived and/or spared intrinsic DA innervation in the striatum, and whether the relative density of the serotonin innervation in the critical part of the striatum will be kept below this critical threshold. In animals with partial 6-OHDA lesions (with an overall 80–90% reduction in striatal DA innervation) the serotonin-only transplants (5-HT graft group) produced a 2- to 3-fold increase in the SERT-positive innervation in the caudate–putamen, corresponding to a DA:serotonin innervation ratio in the range of 1:1 to 1:2. In these previously mildly dyskinetic animals the AIMs scores increased about 2.5-fold, but were still kept in the low-to-moderate range. In the DA wide group, transplants contained DA and serotonin neurons in a ratio of ∼2:1. In these animals, the TH-positive innervation density was ∼15% of normal, while the SERT-positive innervation was increased about 2-fold above normal, suggesting that the relative densities of the DA and serotonin innervations of the host striatum was maintained within the critical 1:1 range. Consequently, these animals displayed low or no dyskinesias also after the second 6-OHDA lesion.

The implications of these findings for the appearance of dyskinesia in grafted Parkinson's disease patients are 3-fold: (i) In patients with transplants that contain DA neurons but no or very few serotonin neurons the grafts should retard the development of l-DOPA-induced dyskinesia, and reduce already established dyskinesias, provided that the graft-derived DA innervation of the host putamen is extensive enough to exceed that of the intrinsic serotonin innervation; (ii) In patients with grafts that contain serotonin neurons in excess of the dopaminergic ones the impact on dyskinesia will depend on the extent of degeneration of the host nigrostriatal system or the extent of reinnervation of the host striatum by the graft-derived DA neurons. The excess innervation provided by the grafted serotonin neurons will promote the development of l-DOPA-induced dyskinesias, and also make already established dyskinesias worse, in case the DA innervation of the dyskinesia-triggering sub-region of the putamen is reduced below the critical threshold and (iii) In patients with more advanced disease (i.e. extensive degeneration of the intrinsic putaminal DA innervation) the impact of mixed DA and serotonin grafts will depend on the ability of the grafted DA neurons to provide a widespread innervation of the host putamen, in particular of those sub-regions of the putamen that are primarily involved in the induction of AIMs in dyskinetic patients. Interestingly, in a recent study of the brains of two VM-transplanted Parkinson's disease patients (Mendez et al., 2008) these long-term surviving grafts were seen to contain DA and serotonin neurons in roughly equal numbers (as assessed by TH and tryptophan hydroxylase immunohistochemistry; O. Isacson, personal communication). Consistent with the observations in the DA wide graft group in the present study, it seems that l-DOPA-induced dyskinesia had remained low in these patients still by 21 months after transplantation.

Although we did not observe any spontaneous graft-induced dyskinesias in the present study, this experimental design does not provide evidence to rule out the possibility that serotonin neurons may play a role in the development of graft-induced off-state dyskinesia in grafted patients. Observations in transplanted patients (Freed et al., 2001; Ma et al., 2002) and 6-OHDA lesioned rats (Steece-Collier et al., 2003; Carlsson et al., 2006, 2007; Lane et al., 2006; Maries et al., 2006) indicate that this type of dyskinesia is induced by dysregulated release of DA, and/or excessive activation of supersensitive DA receptors, caused by an uneven, patchy graft-derived DA innervation of the host striatum. It seems possible that an unfavorable DA:serotonin innervation ratio could contribute to this effect by maintaining the postsynaptic DA receptors in a supersensitive state. Moreover, the excessive swings in extracellular DA generated from graft-derived serotonin terminals could be the driving force to induce a hypersensitive state of the postsynaptic receptors that will make the striatal targets prone to generate AIMs even in the absence of any exogenous l-DOPA. Investigations of the presence and extent of serotonin innervation in the striatum of patients with graft-induced dyskinesias, e.g. by PET imaging with ligands for the serotonin transporter, may help to clarify this point.

Funding

Michael J. Fox Foundation (Community Fast Track 2005 to A.B.); the Swedish Research Council (04X-3874 to A.B. and 61X-14552 to D.K.). Parkinson Disease Foundation (to M.C.); the Michael J. Fox Foundation (to M.C.).

Acknowledgements

We thank Anneli Josefsson and Ulla Jarl for expert technical assistance and Tomas Björklund for statistical assistance.

Footnotes

  • *Present address: Experimental Neurology Unit, Department of Neurology, Phillips University Marburg, Hans Meerwein Strasse, 350 43 Marburg, Germany.

  • These authors contributed equal to this work.

  • Abbreviations:
    Abbreviations
    AIMs
    abnormal involuntary movements
    DA
    dopamine
    5,7-DHT
    5,7-dihydroxytryptamine
    MFB
    medial forebrain bundle
    6-OHDA
    6-hydroxydopamine
    SERT
    serotonin transporter
    TH
    tyrosine hydroxylase
    VM
    ventral mesencephalic

References

View Abstract