Rosiglitazone is a commonly prescribed insulin-sensitizing drug with a selective agonistic activity on the peroxisome proliferator-activated receptor-gamma (PPAR-γ). PPAR-γ can modulate inflammatory responses in the brain, and agonists might be beneficial in neurodegenerative diseases. In the present study we used a chronic 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine plus probenecid (MPTPp) mouse model of progressive Parkinson’s disease (PD) to assess the therapeutic efficacy of rosiglitazone on behavioural impairment, neurodegeneration and inflammation. Mice chronically treated with MPTPp displayed typical features of PD, including impairment of motor and olfactory functions associated with partial loss of tyrosine hydroxylase (TH)-positive neurons in the substantia nigra pars compacta (SNc), decrease of dopamine (DA) and 3,4-dihydroxyphenylacetic acid (DOPAC) content and dynorphin (Dyn) mRNA levels in the caudate-putamen (CPu), intense microglial and astroglial response in the SNc and CPu. Chronic rosiglitazone, administered in association with MPTPp, completely prevented motor and olfactory dysfunctions and loss of TH-positive cells in the SNc. In the CPu, loss of striatal DA was partially prevented, whereas decreases in DOPAC content and Dyn were fully counteracted. Moreover, rosiglitazone completely inhibited microglia reactivity in SNc and CPu, as measured by CD11b immunostaining, and partially inhibited astroglial response assessed by glial fibrillary acidic protein immunoreactivity. Measurement of striatal MPP+ levels 2, 4, 6 h and 3 days after chronic treatment indicated that MPTP metabolism was not altered by rosiglitazone. The results support the use of PPAR-γ agonists as a putative anti-inflammatory therapy aimed at arresting PD progression, and suggest that assessment in PD clinical trials is warranted.
Several studies have shown that the chronic administration of low doses of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) to primates and mice provides an experimental model that closely reproduces human PD in many aspects. Chronic MPTP delivery in association with the clearance inhibitor probenecid (MPTPp), as in the present study, or through osmotic minipumps, or chronic intracerebral delivery of MPP+, all induce a persistent dopamine (DA) neuron degeneration, associated with motor deficits and typical histological markers of the disease, including chronically activated microglia (Bezard et al., 1997; Petroske et al., 2001; Meredith et al., 2002; Fornai et al., 2005; Novikova et al., 2006; Yazdani et al., 2006). Therefore, at variance with acute MPTP protocols, which fail to reproduce the chronic development of symptomatic and neuropathological aspects of PD, the chronic MPTP protocol used here provides a valuable experimental model to investigate mechanisms of neurodegeneration and to test neuroprotective strategies.
In the present study, we assessed the neuroprotective effect of the PPAR-γ agonist rosiglitazone and its relevance to PD symptomatology in the chronic MPTPp mouse model of PD. Rosiglitazone, currently approved for use in type II diabetes, was recently shown to cross the blood–brain barrier in mice, suggesting that it might be suitable for testing neuroprotective effects of PPAR-γ agonists in the CNS (Strum et al., 2007). A dose of drug near the dose range used for diabetes therapy was selected for the study. Rosiglitazone was administered chronically in the MPTPp model, and motor deficits, olfactory dysfunction, neuronal damage in the substantia nigra pars compacta (SNc), pre- and postsynaptic dysfunctions in the caudate-putamen (CPu) were evaluated as PD hallmarks. Moreover, based on evidences that ascribe to an anti-inflammatory activity, the neuroprotective outcome of PPAR-γ stimulation in models of neurodegenerative disorders, we evaluated the effect of rosiglitazone on chronic MPTPp-induced glial response in the SNc and CPu, through an evaluation of reactive microglia and astroglia by CD11b and glial fibrillary acidic protein (GFAP) immunostaining. The present study, by providing evidence of neuroprotective efficacy of a PPAR-γ agonist in a chronic MPTP model, suggests a novel therapeutic use of these compounds in PD.
Materials and methods
MPTP-HCl (Sigma, Italy) was dissolved in water, probenecid (Sigma, Italy) in 5% NaHCO3. The PPAR-γ agonist rosiglitazone (GlaxoSmithKline) was suspended in 0.5% methylcellulose.
Male C57Bl/6J mice, 3 months old (Charles River, Italy), received 10 doses of MPTP (25 mg/kg i.p.) and the clearance inhibitor probenecid (250 mg/kg i.p.), given to prolong MPTP neurotoxicity, twice a week for 5 weeks. Rosiglitazone (10 mg/kg i.p.) was administered daily, 1 h before MPTPp and until death. After 1 or 10 MPTPp injections, olfactory function and motor performance were evaluated.
Three days after discontinuation of MPTPp, one group of mice was anaesthetized with chloral hydrate and transcardially perfused for immunohistochemistry. Another group was CO2-anaesthetized and brains rapidly removed. Fresh tissue from one hemisphere was used for measurement of striatal DA-3,4-dihydroxyphenylacetic acid (DOPAC) content, whereas the contralateral hemisphere was flash frozen in isopentane/dry ice for in situ hybridization of dynorphin (Dyn) mRNA.
For evaluation of striatal MPP+, mice chronically treated with MPTPp ± rosiglitazone or with vehicle, as described above, were CO2-anaesthetized 2, 3, 6 h and 3 days after the last MPTP administration and brains rapidly removed for mass spectrometer (MS) analysis.
All animal experimentations have been conducted in accordance with the guidelines for care and use of experimental animals of the European Communities Council Directive of 24 November 1986 (86/609/EEC) and approved by the ethical committee of University of Cagliari.
Beam traversal test
The beam was constructed as described (Fleming et al., 2004). Mice were trained for two consecutive days to traverse the beam. On the test day a grid (1 cm2) of corresponding width was placed 1 cm above the beam, mice were videotaped while traversing it for a total of five trials, and errors were calculated (when a limb slipped through the grid). By scoring each limb slip individually, the severity of the error could be measured (Fleming et al., 2004).
Mice were food-deprived for 20 h before test. The olfactory test was conducted in a plastic cage (24 w, 42 l, 15 h cm). A smelly pellet was buried under the bedding (1 cm) in a corner of the cage, the mouse was positioned in the centre, and time to retrieve the pellet and bite it was measured.
CPu and SNc sections (50 μm thick) were vibratome-cut and immunoreacted with tyrosine hydroxylase (TH), GFAP and CD11b antibodies (polyclonal rabbit anti-TH, 1 : 1000, Biomol; monoclonal mouse anti-GFAP, 1 : 400, Sigma, Italy; monoclonal rat anti-mouse CD11b, 1 : 1000, Serotec, UK) and proper secondary antibodies. For visualization, the Avidin–peroxidase protocol (ABC, Vector, UK) was applied, using 3,3′-diaminobenzidine (Sigma) as chromogen. Adjacent SNc sections were Nissl-stained to evaluate actual cell loss. Adjacent sections from SNc or CPu were stained in the following order: TH immunoreactivity, Nissl-staining (for SNc only), CD-11b and GFAP immunoreactivity. For each animal, a total of three sections from the SNc (anterior: −2.92 mm; medial: −3.28 mm; posterior: −3.64 mm from bregma, accordingly to Mouse Brain Atlas, Paxinos & Franklin, 2001) and three sections from the CPu (anterior: 1.10 mm; medial: 0.74 mm; posterior: 0.38 mm from bregma) were analysed for each protein evaluated in the study and for Nissl-stained cells.
Images were digitized under constant light conditions (videocamera Pixelink PL-A686). Sections were captured at 100× real magnification (for SNc analysis, A-PLAN Zeiss, numerical aperture 0.25) or at 200× real magnification (for CPu analysis, A-PLAN Zeiss, numerical aperture 0.45). In each section, the whole left and right SNc were analysed, whereas for CPu evaluation one portion from the dorsolateral CPu and one from the ventromedial CPu (520 × 380 μm), left and right, were analysed.
Analysis of TH-, GFAP- and Nissl-positive cells
The absolute number of TH-, GFAP- or Nissl-positive cells, was obtained separately for each SNc and CPu level (Table 1). Thereafter, in order to obtain an average value from all levels analysed, the number of cells/level from each mouse was normalized with respect to the vehicle. Individual values from the three levels were than averaged to generate a mean.
Table 1. Numbers of TH-positive cells and Nissl-stained cells in the SNc of mice chronically treated with vehicle or MPTPp, with or without rosiglitazone
MPTPp + rosiglitazone
Cells in three different SNc levels were counted and expressed as absolute values (mean ± SEM). *P <0.05 as compared with vehicle-treated mice; †P <0.05 as compared with MPTPp-treated mice, by Tukey’s post hoc test. MPTPp, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine plus probenecid; SNc, substantia nigra pars compacta; TH, tyrosine hydroxylase.
Anterior SNc (−2.92 mm from bregma)
314.80 ± 9.54
245.00 ± 20.64*
304.00 ± 11.32†
329.40 ± 8.73
240.50 ± 19.69*
311.20 ± 12.65†
Medial SNc (−3.28 mm from bregma)
280.20 ± 7.69
194.70 ± 11.23*
276.00 ± 7.73†
275.40 ± 8.86
201.75 ± 6.55*
282.00 ± 3.65†
Posterior SNc (−3.64 mm from bregma)
173.40 ± 8.96
116.25 ± 8.82*
173.20 ± 8.95†
151.80 ± 6.04
109.50 ± 6.18*
158.00 ± 6.94†
Analysis of CD11b immunoreactivity
Images were digitized in a grey-scale, and CD11b immunostaining was evaluated with the analysis program scion image. A threshold, whose value was set above the mean value ± SEM of the background, was applied for background-correction. Inside each frame, the area occupied by grey values above the threshold was automatically calculated. For each level of SNc or CPu, the obtained value was first normalized with respect to the vehicle, values from different levels were averaged thereafter.
High-pressure liquid chromatography (HPLC)
Striatal tissue was sonicated in 250 μL of 0.2 m perchloric acid, then centrifuged at 9391.2 g for 15 min at 4°C. Supernatant was filtered and diluted 1 : 62.5. Twenty microlitres was injected into an HPLC apparatus, equipped with a reverse-phase column (LC -18 DB, 15 cm, 5 μm particle size; Supelco, Milano, Italy) and a coulometric detector (ESA Coulochem IIm, Bedford) to quantitate DA and DOPAC. Electrodes were set at +150 mV (oxidation) and −250 mV (reduction). The mobile phase (in mm: NaH2PO4, 100; NA2EDTA, 0.1; n-octyl sodium sulphate, 0.5; 7.5% methanol; pH 5.5) was pumped (Jasco Europe, Italy) at 1 mL/min flow rate. The assay sensitivity for DA and DOPAC was 5 and 10 fmol/sample, respectively (Carboni et al., 2006).
In situ hybridization
Glass-mounted cryostat sections (12 μm) were postfixed as described and hybridized with 100 μL of buffer containing 2 × 106 cpm of radioactive probe complementary to Dyn mRNA, at 55 °C overnight, washed and apposed to X-ray film (Carta et al., 2005). For Dyn mRNA evaluation, three sections from each CPu (4.9 mm anterior to the interaural line accordingly to Mouse Brain Atlas, Paxinos & Franklin, 2001) were analysed. The average density of autoradiogram grey values was background-corrected and quantitated by the image analysis program scion image.
HPLC/electrospray ionization source (ESI)-MS/MS analysis of MPP+
Tissue preparation was carried out as for HPLC studies. A Varian 1200 L triple quadrupole tandem MS (Palo Alto) coupled with a ProStar 410 autosampler and two ProStar 210 pumps was used with an ESI. The Varian MS workstation 6.9 software was used for data acquisition and processing. Chromatographic separation was performed on a Varian Pursuit C18 (2.0 × 50 mm I.D., 3 μm; Lake Forest, CA, USA). The mobile phase was pumped at 0.3 mL/min flow rate and 10 μL volume. MPP+ retention time was 2.7 min. ESI was operated in the positive ion mode. Full-scan spectra were obtained in the range of 100–300 amu, scan time 0.5 amu, scan width 0.70 amu, detector at 1200 V. Parent compounds were subjected to collision-induced dissociation using argon at 2.20 mTorr in the multiple reaction monitoring-positive mode. The observed mass transitions and collision energy were 170→168.8 (−38 V) and 170→127.9 (−38 V) used for quantitation of MPP+. The scan time was 1 s and the detector multiplier voltage was 1200 V, with an isolation width of m/z 1.2 for the quadrupole 1 and m/z 2.0 for the quadrupole 3.
Behavioural results were statistically compared with a two-way anova, followed by Newman–Keuls post hoc test, for comparison between experimental groups.
For immunohistochemistry, HPLC and in situ hybridization experiments, results were statistically analysed with a one-way anova, followed by Tukey’s post hoc test.
Rosiglitazone prevents behavioural impairments induced by chronic MPTPp
Beam traversal test
Mice treated acutely or chronically with MPTPp plus vehicle or plus rosiglitazone were tested for motor performance on the beam traversal test.
Two-way anova detected a significant effect of treatment on the number of errors per step (F2,32 = 20.15; P = 2.2E−06). Newman–Keuls post hoc test indicated that chronic treatment with MPTPp, but not acute administration, produced a significant increase in the number of errors as compared with vehicle (P = 0.0004). After chronic treatment with MPTPp, performance was significantly worse (P = 0.007) as compared with acute administration, suggesting a progression in neuropathology. Combined treatment with rosiglitazone plus MPTPp totally prevented the increase in number of errors induced by chronic MPTPp (P = 0.002; Fig. 1A).
The retrieval time of a buried smelly pellet was assessed for each mouse. Two-way anova indicated a significant effect of treatment on the retrieval time (F2,54 = 11.5; P = 8.1E−05). Newman–Keuls post hoc test showed that both acute and chronic treatment with MPTPp produced a significant delay in retrieval time as compared with vehicle (P < 0.05). In contrast, upon acute or chronic treatment with rosiglitazone plus MPTPp, retrieval time was significantly lower than upon MPTPp treatment (P <0.05) and similar to vehicle (Fig. 1B).
Rosiglitazone prevents neurodegeneration induced by chronic MPTPp in the SNc
Chronic MPTPp treatment induced a 30% loss of TH-positive cells in the SNc, as measured by TH immunoreactivity and Nissl staining (Fig. 2 and Table 1). Combined treatment with rosiglitazone plus MPTPp completely prevented cell loss (Fig. 2 and Table 1). One-way anova indicated a significant effect of treatment on the number of TH-positive cells (F2,12 = 21.9; P = 0.001). Post hoc analysis by Tukey’s test showed a significant decrease (P = 0.0006) in TH immunoreactivity in mice receiving chronic MPTPp, but not in mice receiving rosiglitazone plus MPTPp, which displayed TH levels significantly lower than MPTPp (P = 0.0007) and similar to vehicle.
Rosiglitazone prevents functional changes induced by chronic MPTPp in the CPu
Measurement of striatal DA by HPLC showed that after chronic MPTPp DA levels were significantly decreased by 80% compared with vehicle-treated mice, whereas after rosiglitazone–MPTPp co-administration a reduction of 60% was detected (Fig. 3A). One-way anova indicated a significant effect of the treatment on DA content (F2,16 = 18.1; P = 1.4E−05). Post hoc analysis by Tukey’s test showed a significant decrease in DA content following both MPTPp and rosiglitazone plus MPTPp co-administration (P <0.05); however, striatal DA levels in rosiglitazone plus MPTPp-treated mice were significantly higher than in MPTPp-treated mice (P = 0.04).
Moreover, DOPAC levels after chronic MPTPp were significantly decreased by 65% as compared with vehicle-treated mice, whereas after rosiglitazone plus MPTPp co-administration DOPAC was reduced by 32% only (Fig. 3A). One-way anova indicated a significant effect of the treatment on DOPAC content (F2,16 = 9.7; P = 0.0008). Post hoc analysis by Tukey’s test showed a significant decrease in DOPAC content in mice receiving chronic MPTPp (P = 0.0006), but not in mice receiving rosiglitazone plus MPTPp, which displayed a DOPAC content significantly higher than MPTPp (P = 0.03).
Chronic administration of MPTPp induced a significant decrease in Dyn mRNA levels, as measured by in situ hybridization, compared with vehicle-treated mice, which was completely prevented by rosiglitazone co-administration (Fig. 3B). One-way anova indicated a significant effect of the treatment on Dyn levels (F2,16 = 15.7; P = 0.0005). Post hoc analysis by Tukey’s test showed a significant decrease in Dyn mRNA level following MPTPp (P = 0.01), but not in mice receiving rosiglitazone plus MPTPp, which displayed Dyn levels significantly lower than MPTPp (P = 0.001) and similar to vehicle.
Rosiglitazone counteracts glial activation induced by chronic MPTPp in the SNc and CPu
In sections from vehicle-treated mice, low CD11b immunostaining and few GFAP-positive cells were detected in the SNc and CPu (Figs 4 and 5). MPTPp chronic treatment induced a significant increase in CD11b immunoreactivity in both SNc and CPu (Fig. 4A and B). In vehicle-treated mice, microglia was extensively ramified, with a small body and tiny processes, corresponding to a resting morphology as described by Colburn et al. (1997). In contrast, in MPTP-treated mice microglia was more densely distributed, and displayed a large, intensely labelled body, and short bold processes, corresponding to an intense microglial response according to the Colburn scale (Colburn et al., 1997). Combined rosiglitazone plus MPTPp treatment completely prevented the increase in CD11b immunostaining in both SNc and CPu (Fig. 4A and B). Microglia displayed a morphology similar to vehicle-treated mice. One-way anova revealed a significant effect of the treatment in both SNc (F2,12 = 7.3; P = 0.01) and CPu (F2,12 = 5.7; P = 0.02). Tukey’s post hoc test showed a significant increase in CD11b immunostaining following MPTPp (P <0.05) but not in mice receiving rosiglitazone plus MPTPp, which displayed, in both areas, levels of CD11b immunostaining significantly lower than MPTPp (P <0.05) and similar to vehicle.
In vehicle-treated mice, astroglial cells displayed a highly branched morphology with tiny processes and a small body, indicative of a resting state (Colburn et al., 1997; Fig. 5A and B). After MPTPp chronic treatment, astroglia became densely arranged and hypertrophic in both the SNc and CPu, displaying a large body intensely immunolabelled, short and thick projections, indicative of an intense astroglial response (Colburn et al., 1997; Fig. 5A and B). Rosiglitazone only partially prevented the MPTPp-induced increase in GFAP immunoreactivity in the SNc (Fig. 5A), whereas it did not modify MPTPp-induced response in the CPu (Fig. 5B). One-way anova revealed a significant effect of the treatment in both SNc (F2,12 = 72.3; P = 2.02E−07) and CPu (F2,12 = 410.1; P = 9.0E−12). Tukey’s post hoc test showed a significant increase in GFAP immunoreactivity in both areas following MPTPp and rosiglitazone–MPTPp co-administration (P <0.05); however, in the SNc rosiglitazone plus MPTPp induced levels of GFAP immunoreactivity significantly lower than MPTPp (P = 0.0002).
Effect of rosiglitazone on striatal MPP+ level
Analysis by MS showed that MPP+ levels were similar in mice treated with MPTPp or rosiglitazone plus MPTPp, as measured 2, 3, 6 h and 3 days after the chronic treatment, suggesting that rosiglitazone did not alter the conversion of MPTP to the toxic metabolite MPP+ (Fig. 6; F1,8 = 1.23; P = 0.33). The identity of the peak assigned to MPP+ was confirmed by both the retention time and the specific transition m/z 170.4 to 127.9.
The present study shows that rosiglitazone prevented the neurodegeneration of DA neurons in the SNc and the development of behavioural symptoms in a chronic mouse model of progressive PD, suggesting that rosiglitazone might be beneficial as PD therapy aimed at arresting disease progression. Moreover, evaluation of the glial response indicated that reactive microglia was a target of the PPAR-γ agonist effect, suggesting that the anti-inflammatory activity of this drug might be a mechanism of neuroprotection.
With clinical trials in phase III ongoing for the evaluation of rosiglitazone as a symptomatic therapy in patients with Alzheimer’s disease, the present study suggests that this drug might have a broader benefit in neurodegenerative diseases, and highlights the need to test this class of compounds in patients with PD.
Effects of rosiglitazone in symptomatic and biochemical PD features
The aim of the present study was the assessment of rosiglitazone neuroprotective properties in a validated experimental PD model. The chronic MPTPp paradigm used here has been shown to reproduce several features of PD, allowing the evaluation of changes in both symptomatic and biochemical PD parameters upon treatment with neuroprotective drugs. In this model, rosiglitazone prevented the development of motor impairment induced by chronic MPTPp, as measured by the ‘beam traversal’ test. Because the MPTPp protocol used here induced a partial loss of nigral dopaminergic neurons, and the severity of motor impairment depends upon the extent of DA neurons degeneration, we used a validated behavioural task highly sensitive for motor deficits associated with subtle alterations of the nigrostriatal pathway (Di Monte et al., 2000; Fernagut et al., 2003; Fleming et al., 2004; Meredith & Kang, 2006). Interestingly, whereas MPTPp-treated mice displayed postural abnormalities as, for instance, truncal dystonia, mice treated with rosiglitazone plus MPTPp displayed a posture similar to vehicle-treated mice (data not shown).
Moreover, the development of olfactory deficits was fully prevented by rosiglitazone co-administration, suggesting that rosiglitazone protective effects improved both motor and non-motor symptoms in the PD model. A number of studies have provided evidence to consider olfactory dysfunction a non-motor symptom of PD, which occurs early in the disease and might precede the appearance of motor symptoms (Braak et al., 2003; Haehner et al., 2007). Interestingly, in the present study olfactory dysfunction but not motor deficits was recorded after the first MPTPp administration, whereas both behaviours were impaired after the chronic treatment, suggesting a gradual development of symptoms, in line with the progressive nature of PD.
Rescue of PD symptoms by rosiglitazone reflected a full protection of dopaminergic cells in the SNc. TH immunoreactivity displayed a decrease in immunostaining following chronic MPTPp treatment, which was fully prevented by rosiglitazone. As it has been shown that TH staining by itself does not represent an accurate measure of neuronal cell death after MPTP intoxication, analysis of Nissl-staining was performed in adjacent sections. Results indicate that the MPTPp model used here caused a consistent loss of neurons in the SNc, whereas rosiglitazone protected the SNc from neurodegeneration.
In contrast to a full protection of DA neurons in the SNc, rosiglitazone only partially prevented the MPTPp-induced decrease of striatal DA content, suggesting a partial protection of nigrostriatal fibres. Interestingly, a decrease of DA metabolite DOPAC induced by chronic MPTPp was not observed after rosiglitazone plus MPTPp treatment, which induced DOPAC levels similar to controls. Moreover, the ratio DOPAC/DA was inverted upon MPTPp or rosiglitazone plus MPTPp chronic administration (1.14 and 1.01, respectively) as compared with control mice (0.59), suggesting an increased turnover rate of striatal DA, likely occurring to compensate loss of dopaminergic terminals. As a result, mice chronically treated with rosiglitazone plus MPTPp, despite the partial rescue of striatal DA, displayed DOPAC levels similar to control mice. Accordingly, in the early stages of PD, functional compensatory mechanisms develop in the CPu, delaying the appearance of motor symptoms (Blanchard et al., 1995; Anglade et al., 1996; Bezard et al., 2000).
Rosiglitazone also prevented the decrease in Dyn mRNA levels induced by chronic MPTPp in the CPu. In PD models, as well as in human, neuroadaptive changes in the expression of postsynaptic molecules, including an abnormal decrease in Dyn mRNA levels, play a crucial role in the development of PD motor symptoms (Chase & Oh, 2000;Gerfen, 2003).
Taken together, results show that in a chronic MPTP mouse model, stimulation of PPAR-γ by rosiglitazone prevented the development of behavioural and biochemical changes that typically characterize PD. Neuroprotection of SNc neurons was associated with a partial inhibition of neurotoxic insult to dopaminergic terminals and with compensatory mechanisms in DA transmission, which prevented downstream postsynaptic changes and development of motor symptoms.
Effects of rosiglitazone on chronic glial response
Based on previous evidence that suggests that PPAR-γ agonist-mediated neuroprotection in neurodegenerative models might rely on their anti-inflammatory activity, we also evaluated the effect of rosiglitazone on chronic MPTPp-induced glial activation. Chronic MPTPp treatment resulted in intense CD11b immunostaining in SNc and CPu, as measured 3 days after neurotoxin administration. Microglia was highly branched at rest, but displayed an amoeboid morphology following exposure to MPTPp. Remarkably, upon rosiglitazone co-administration, protection of DA neurons was associated with levels of CD11b and a microglial morphology similar to controls, in both SNc and CPu, indicating a complete inactivation of microglial response. Activated microglia have been found in the SNc of parkinsonian as well as MPTP-intoxicated humans and monkeys, even after several years of neurotoxin exposure, suggesting that neuroinflammation plays a crucial role in the aetiology of PD (McGeer et al., 1988, 2003; Langston et al., 1999; Hirsch et al., 2003; Barcia et al., 2004). Chronically activated microglia would contribute to sustain and amplify neurodegeneration in PD by repeatedly releasing toxic factors as inflammatory cytokines, NO, glutamate. PPAR-γ can regulate gene expression by binding to the DNA promoter region of several genes, including inflammatory cytochines, TNF-α, COX, iNOS, whose expression is decreased by PPAR-γ agonists (Heneka et al., 2000; Park et al., 2003; Bernardo & Minghetti, 2006; Jung et al., 2007; Woster & Combs, 2007; Xu & Drew, 2007; Chaturvedi & Beal, 2008; Chung et al., 2008). Therefore, microglia inactivation reported here might represent a mechanism of neuroprotection by rosiglitazone from chronic MPTPp intoxication. In line with the present result, a number of studies have proposed the anti-inflammatory action of PPAR-γ agonists as the main neuroprotective mechanism in several chronic neurodegenerative pathologies (Kiaei et al., 2005; Klotz et al., 2005; Bernardo & Minghetti, 2006; Drew et al., 2006; Luo et al., 2006; Heneka et al., 2007; Park et al., 2007). Moreover, blockade of microglial activation or administration of anti-inflammatory drugs have been proven to be neuroprotective against an acute MPTP insult (Teismann & Ferger, 2001; Wu et al., 2002).
Chronic MPTPp treatment induced an intense astrogliosis in both the SNc and CPu, as indicated by GFAP immunostaining. An increased GFAP immunoreactivity has also been detected in the SNc of monkeys 1 year after chronic MPTP intoxication and in post mortem PD brains (Langston et al., 1999; Barcia et al., 2003; McGeer & McGeer, 2008). MPTPp-induced reactive astroglia was only partially inhibited by rosiglitazone in the SNc and unaffected in the CPu, suggesting that rosiglitazone did not have a direct effect on these cells. Whereas the detrimental role played by microglia in neurodegenerative disorders has been extensively documented, the role of reactive astroglia remains uncertain (Sofroniew, 2005). Increasing evidence, however, ascribes a beneficial role to reactive astroglial cells in several neurodegenerative conditions in vivo, by restricting inflammation and limiting tissue degeneration (Bush et al., 1999; Myer et al., 2006). Our results show that in the SNc reactive astroglia was associated with complete neuronal protection by rosiglitazone, suggesting that it did not exert any harmful effect on neurons. Taken together, the results point to microglia as the main target of rosiglitazone in neuroinflammation, and suggest that a direct microglia inactivation might contribute to neuroprotection, rather than being a consequence of it.
The present results, while showing an inhibition of the MPTP-induced inflammatory response in the presence of rosiglitazone, do not exclude the contribution of other inflammation-independent mechanisms to the neuroprotective outcome. PPAR-γ, through their DNA-binding activity, are involved in the regulation of a number of pathways that directly or indirectly contribute to neurodegeneration in PD. In vitro studies have reported an inhibition of lipid peroxidation, increased expression of antioxidant enzymes and modulation of the expression of pre-apoptotic and anti-apoptotic genes by rosiglitazone (Jung et al., 2006, 2007). Moreover, it was recently demonstrated that rosiglitazone increases levels of the glial glutamate transporter and reduces extracellular glutamate in stress or ischaemic rat models (García-Bueno et al., 2007; Romera et al., 2007). Therefore, rosiglitazone might exert neuroprotection in experimental PD by targeting several mechanisms involved in PD pathogenesis, including oxidative stress, apoptosis, excitotoxicity and inflammation (Chaturvedi & Beal, 2008). Finally, PPAR-γ agonists are insulin-sensitizing compounds, which increase glucose uptake and metabolism via a regulation of glucose transporter expression. Therefore, a counteraction of MPTP-induced oxidative damage might lead to a decreased vulnerability of dopaminergic neurons (el-Kebbi et al., 1994; Smith, 2001; García-Bueno et al., 2007).
It has been recently reported that pioglitazone might prevent degeneration of DA neurons, induced by acute exposure to MPTP, by blockade of MAO-B-mediated MPTP conversion to the toxic metabolite MPP+ (Quinn et al., 2008). The results of the present study show that brain levels of MPP+ were similar in mice treated with MPTPp or rosiglitazone plus MPTPp, as analysed at different time-points after the chronic treatment, suggesting that rosiglitazone-induced neuroprotective effects were not mediated by blockade of MAO-B activity.
Relevance of the chronic MPTPp model
The present study provides the first evidence of neuroprotective efficacy of a PPAR-γ agonist in a chronic MPTP mouse model of PD. In order to thoroughly investigate neuroprotective strategies, experimental models should be used that, as much as possible, mimic clinical features of PD. Whereas MPTP is the most widely used toxin to model PD in experimental animals, several data have highlighted that, upon MPTP intoxication, development of symptomatic and histopathological features characterizing idiopathic PD strictly depends on the mode and schedule of neurotoxin administration (Schmidt & Ferger, 2001; Meredith et al., 2002; Manning-Bog & Langston, 2007). Chronic administration of low MPTP doses in a continuous mode, or in a semi-continuous mode as obtained by probenecid co-administration, induces an enduring degeneration of DA neurons both in primates and in mice, in association with motor and, as shown here, olfactory impairment, reproducing typical PD symptomatology. Moreover, histological hallmarks of the disease, such as Lewy bodies-like inclusions and, most importantly for the aim of the present study, chronic neuroinflammation, have been reported after this MPTP regimen, suggesting that it might be a valuable model to assess the neuroprotective outcome of anti-inflammatory drugs (Mitsumoto et al., 1998; Kurkowska-Jastrzebska et al., 1999; Kowall et al., 2000; Barcia et al., 2004; Fornai et al., 2005; Meredith et al., 2005). In the present study, we found that striatal MPP+ levels were still elevated 6 h after the chronic MPTP treatment, suggesting a prolonged half-life as compared with acute MPTP regimens, in line with the clearance inhibitory effect provided by probenecid (Youdim & Arraf, 2004; Fornai et al., 2005).
All together the results of the present study provide conclusive evidence of neuroprotective effects of rosiglitazone in an experimental PD model that closely reproduces human PD features.
This work was supported by Fondazione Banco di Sardegna (grant number 953/2007.0399). The authors thank GlaxoSmithKline (GSK) for the supply of rosiglitazone; Dr Alessandra Silvagni and Dr Simona Vargiu for technical support in HPLC and HPLC/ESI analysis.
electrospray ionization source
glial fibrillary acidic protein
high-pressure liquid chromatography
1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine plus probenecid