The PPAR-γ agonist pioglitazone modulates inflammation and induces neuroprotection in parkinsonian monkeys

Adult rhesus monkeys (Macaca mulatta, 5-7 years old) were obtained from the Wisconsin National Primate Center (WNPRC) and singly housed with a 12-hr light/dark cycle at the WNPRC facility. Purina monkey chow and water was available ad libitum. The animals' diet was supplemented with fruit during the testing sessions and daily enrichment. All efforts were made to minimize the number of animals used and ameliorate their suffering. This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. Two sets of experiments were carried out: a neuroprotective and a CSF penetration analysis. The protocols were approved by the Institutional and Animal Care Committee at the University of Wisconsin-Madison (permits #: G00492 and G00569, respectively).

All behavioral training and evaluations used positive reinforcement to entice monkeys' cooperation. Monkeys were evaluated weekly using a clinical rating (CR) scale as previously described [23, 24]. The scale ranges from 0 to 32; a score of 0 corresponds to normal behavior and 32 to extreme severe parkinsonian symptoms. Fine motor skills (FMS) were tested using a movement assessment panel computerized system 3 days per week [24]. General activity was assessed pre-MPTP and once per month after MPTP using image digitization followed by computerized post-processing of animal movement (Viewpoint, Inc; [24]).

After baseline data collection was completed, 20 rhesus monkeys (male, 5-7 yrs, 4-8 kg) received a unilateral intracarotid artery injection of 3 mg of MPTP-HCl (Sigma) in 20 ml of saline (rate: 1.33 ml/min) as previously described [23]. The procedure was performed in a state-of-the-art surgical suite under isofluorane anesthesia (1-2%). Throughout the procedure vital signs were monitored and recorded. Each animal was given cefazolin (25 mg/kg i.m.) and buprenex (0.01 mg/kg i.m.) upon waking up response and 24 hours post surgery.

Twenty-four hours after MPTP administration, the 20 animals were behaviorally assessed with the CR scale and 16 monkeys were selected (CR score ≥ 9 points), matched according to PD signs, randomly assigned to one of three groups (see below) and their treatments started.

To assess for pioglitazone side effects in non-diabetic animals, one month before MPTP surgery a glucose tolerance test, general serum panel, glycosylated hemoglobin, and insulin levels were performed. Afterwards the animals received pioglitazone (5 mg/kg p.o.) once a day for 7 days and the tests repeated.

After 1 month of washout the animals were intoxicated with MPTP and 24 hours later animals were selected and treatments started. Once a day until the end of the study, the animals received oral dosing of vehicle, (n = 5), 2.5 (n = 6) or 5 mg/kg (n = 5) of pioglitazone [1]. During treatment animals were weighed weekly and serum samples were taken monthly. Before necropsy, a glucose tolerance test was performed. The animals were trained for blood sampling using positive reinforcement.

Three months post-MPTP the animals were anesthetized with sodium pentobarbital (25 mg/kg iv) and transcardially perfused with heparinized saline, followed by 4% paraformaldehyde (PFA; [23, 24]). Brains were post-fixated in 4% PFA for 12-24 hours and cryoprotected by immersion in a graded (10-40%) sucrose/0.1 M phosphate buffered saline (PBS, pH 7.2) solution. The tissue was cut frozen (40 μm sections) on a sliding knife microtome. All sections were stored in a cryoprotectant solution before processing.

All other major organs were examined and sampled for histology at the time of necropsy. Tissues were fixed in 10% neutral buffered formalin, and routinely processed for hematoxylin and eosin staining.

Brain coronal sections were stained with Nissl or by immunohistochemical methods according to our previously published protocols [23, 24]. Antibodies used include: tyrosine hydroxylase (TH; 1:20,000; Immunostar, Hudson, WI), vesicular monoamine transporter 2 (VMAT2; 1:1000; Phoenix Pharmaceuticals, Belmont, CA), glial fibrillary acidic protein (GFAP; 1:2,000; DakoCytomation, Glostrup, Denmark), heme-oxygenase-1 (HO-1; 1:1,000; Assay Designs "Stressgen", Ann Arbor, MI), nitrotyrosine (1:300; Millipore, Billerica, MA), and CD68 (1:3,000; DakoCytomation, Glostrup, Denmark).

The optical density (OD) of TH and VMAT2 immunoreactive (ir) striatal fibers was quantified within ventral, medial and dorsal sections of both the caudate and putamen using NIH ImageJ software. Images of nine coronal sections per monkey approximately 2 mm apart were captured using an Epson 1640XL-GA high-resolution digital scanner. ImageJ was calibrated using a step tablet, grey scale values were converted to OD units using the Rodbard function, and the mean OD for each area of interest was recorded.

The total number of TH-ir, VMAT2-ir, and Nissl neurons in the right and left substantia nigra (SN) was calculated using unbiased stereological cell-counting methods described previously [23, 25‐27]. The optical dissector system consisted of a computer assisted image analysis system including a Zeiss Axioplan 2 imaging photomicroscope (Carl Zeiss, Inc) hard-coupled to a MAC5000 high precision computer-controlled x-y-z motorized stage, and a MicroFire CX9000 camera (Optronics, Goleta, CA). Neuronal counts were performed using Stereo Investigator Version 7.5 (MicroBrightField, Williston, VT). The SN was outlined under a low magnification (2.5×). The total number of TH-ir and VMAT2-ir neurons within the counting frame was counted using a 100× oil immersion objective with a 1.4 numerical aperture. Six equally spaced sections from each subject containing the SN were used for analysis.

The amount of CD68, GFAP, HO-1, and nitrotyrosine immunoreactivity (ir) was quantified within the SN using NIH ImageJ software. Images from five coronal sections per monkey approximately 2 mm apart were captured using a Nikon E800 microscope equipped with a SPOT camera. CD68-ir was calculated using the particle count function of the ImageJ program, which refers to the quantification of objects of a certain size within a region of interest in a thresholded image. The object size was set between five and 75 square pixels. The particle number and their total area within each region of interest was then analyzed and recorded. For GFAP, nitrotyrosine and HO-1-ir, ImageJ was calibrated using a step tablet, grey scale values were converted to OD units using the Rodbard function, and the area in pixels above a threshold of 0.30 OD units was recorded.

To determine CSF penetration of pioglitazone, five rhesus monkeys (female, 6-7 yrs, 5-6 kg) received daily oral administration of placebo, 2.5, 5 or 7.5 mg/kg of pioglitazone. Each dosing was given for a period of 8 days. Plasma and immediately thereafter, CSF samples were obtained at baseline and approximately 5 hours after the last dose of pioglitazone. CSF and blood collections were performed under ketamine (7 mg/kg i.m.) and medetomidine (0.05 mg/kg i.m.) anesthesia. Animal vital signs were monitored until they completely recovered. Samples were stored at -80°C until analysis.

Drug levels were evaluated by HPLC methods as described by Sripalakit et al. [28] with modifications. Plasma and CSF samples were stored at -80°C until analysis. To purify the samples, they were thawed and processed through solid phase extraction (SPE) based on a previously described method with modifications [28]. One ml of sample was processed with 1 ml KH₂PO₄ (0.1 M) and 400 μl rosiglitazone as the internal standard (Rosiglitazone-maleate 2 mg tablets, Avandia^®, 10 μg/ml stock solution in acetonitrile (ACN)/buffer). The SPE columns (Strata C18-T, 100 mg/1 ml) were preactivated with 1 ml each of 100% ACN and then 1 ml KH₂PO₄ (0.1 M). Samples were added and then washed with 1 ml methanol-Kh₂PO₄ (30:70) and 1 ml K₂HPO₄. Columns were allowed to dry for 5 minutes and then were eluted by adding 500 μl ACN:H₂O (40:60) and 500 μl ACN:H₂O (50:50). For plasma samples, the elution was centrifuged at 3,000 for 5 minutes and transferred into HPLC vials. For CSF samples, the elution was dried under air at 60°C to concentrate the sample, resuspended in 250 μl of 50% CAN, and transferred into HPLC vials.

Pioglitazone standards were prepared as reported [28] with concentrations of 2,400, 1,200, 600, 300, 150, 75, 37.5, 18.75, 9.38, 4.69 ng diluted in 1 ml of plasma or CSF, 1.0 ml KH₂PO₄ with 400 μl internal standard. Standards were linear in plasma (r ² = 0.996) and in CSF (r ² = 0.997).

HPLC analysis used a Beckman HPLC system consisting of an automated sampler with a 100 μl loop, dual pumps, and a diode array analyzer for UV detection. Measurement was made at 269 nm wavelength. Chromatographic conditions and column were based on previously described methods and samples were injected as 100 μl. Retention times for rosiglitazone and pioglitazone were 4.3 and 10.1, respectively. Assay precision was 7.25% at 300 nag and 11.9% at 75 nag of pioglitazone (n = 9). Accuracy was 103.57% ± 2.53 (mean ± SEM).

Sample size was defined by a priori power analysis based on MPTP-treatment differences in motor function to achieve an alpha = 0.05 and beta < 0.2 (power > 80%). All data was collected and analyzed by investigators blind to the treatment groups. A P value of < 0.05 was considered statistically significant. All statistics were performed using SPSS version 17.0 software (SPSS, Chicago, IL).

Before MPTP treatment all monkeys showed no neurological impairments, scoring zero in the CR (Figure 1-A). Twenty-four hrs after intoxication, the animals presented a typical hemiparkinsonian syndrome, consisting of tremors in the side contralateral to MPTP infusion, slowness and decreased amount of movement, as well as balance and gait impairments (score ≥ 9 points). Over time the pioglitazone-treated monkeys showed a progressive improvement, in particular, bradykinesia, gross motor skills and gait that reached statistical significance for the 5 mg/kg treated monkeys compared to placebo at 9 (mean ± SEM, 5 mg/kg pioglitazone 5.60 ± 0.73, placebo 9.70 ± 0.46) and 11 weeks (5 mg/kg pioglitazone 5.60 ± 0.95; placebo 9.20 ± 0.34) (Kruskal-Wallis test; X ² = 8.500, df = 2, P = 0.014 and X ² = 7.833, df = 2, P = 0.02, respectively).

Pioglitazone's effect on FMS testing was more variable. Prior to MPTP administration, all subjects were able to complete the FMS task consistently and quickly with both hands (Figure 1-B). After MPTP administration, most monkeys had difficulty in completing the FMS task under 30 seconds using the hand contralateral to MPTP dosing (left hand), while exhibiting no deficits in task performance using the hand ipsilateral to MPTP administration (right hand). Two of the pioglitazone-treated monkeys in the 5 mg/kg group showed improvement in FMS. Due to individual variability, a statistically significant difference was found only in the third week post-treatment using the hand contralateral to MPTP dosing (repeated measures ANOVA; F [2,12] = 3.797, P = 0.04). Post hoc analysis further confirmed the observation that monkeys treated with 5 mg/kg of pioglitazone had significantly shorter latency times compared to placebo (Fisher's LSD P = 0.023). Performance with the hand ipsilateral to MPTP administration showed no significant differences between treatment groups over time (repeated measures ANOVA; F [2,12] = 0.771, P = 0.701).

The overall amount of activity of all monkeys was measured using a digitized monitoring system. No evidence of pioglitazone-induced hyperkinesia was observed (Table 1). The placebo monkeys showed an increase in activity in the first month after MPTP, probably due to a disbalance in brain DA that induced periods of spontaneous circling to the side of the brain lesion (observed as bursts of activity) but due to individual differences did not reach statistical significance. Repeated measures ANOVA failed to find in the activity data a significant effect of time (F [1,12] = 1.885, P = 0.195), treatment (F [2,12] = 1.767, P = 0.213) or an effect of time × treatment (F [2,12] = 1.462, P = 0.270).

To assess DA striatal terminal fiber condition after MPTP and pioglitazone treatments analysis of TH and VMAT2 immunostaining were performed. OD quantification of TH-ir striatal fibers showed a significant loss on the side ipsilateral to MPTP administration compared to the contralateral side (Wilcoxon Signed Rank test; z = -3.408, P = 0.001, two tailed). There was a significant main effect of treatment between groups (ANOVA; F [2,143] = 11.76, P = 0.012). Comparison between groups revealed that the ipsilateral putamen of the animals treated with 5 mg/kg of pioglitazone had a significant preservation of TH-ir fibers compared to placebo (ANOVA; Fisher's LSD post hoc P = 0.001; Figure 2). OD quantification of VMAT2-ir striatal fibers also showed significant loss on the side ipsilateral to MPTP compared to the contralateral side (Wilcoxon Signed Rank test; z = -3.408, P = 0.001). Although a slight preservation in the ipsilateral putamen for animals treated with pioglitazone was found, it did not reach statistical significance (ANOVA; F [2,12] = 0.555, P = 0.25).

DA SN cell survival was evaluated by analysis of TH and VMAT2 immunohistochemistry, as well as Nissl histochemistry. Qualitatively, all subjects had preservation of neurons in the side contralateral to MPTP administration as well as in the ventral tegmental area. In the ipsilateral SN the placebo-treated monkeys showed decreased numbers of neurons and the surviving neurons had a shrunken perikarya and diminished neuropil. The pioglitazone monkeys instead, displayed more nigral neurons with a large round perikarya and extensive multipolar neurites. Stereological cell counts showed a significant loss of TH-ir nigral neurons on the side ipsilateral to MPTP administration compared to the contralateral side (Wilcoxon Signed Rank test; z = -3.408, P = 0.001 two-tailed). There was a main effect of treatment between groups (ANOVA; F [2,13] = 3.493, P = 0.05). Monkeys treated with 5 mg/kg of pioglitazone had higher cell counts (53,7345 ± 8,715) in the ipsilateral SN compared to controls (29,175 ± 2,954) (ANOVA; Fisher's LSD post hoc P = 0.02; Figure 3). Interestingly, the contralateral SN of both pioglitazone-treated groups also had a significantly higher number of nigral TH-ir cells (2.5 mg/kg, 185,987 ± 6420; 5 mg/kg, 186,064 ± 3,991) compared to controls (159,44 ± 8234) (ANOVA; Fisher's LSD post hoc P = 0.014, P = 0.02, respectively).

VMAT2-ir nigral neuron stereological quantification showed a significant loss on the side ipsilateral to MPTP administration compared to the contralateral side (Wilcoxon Signed Rank test; z = -3.408, P = 0.01 two-tailed). There was a main effect of treatment between groups (ANOVA; F [2,13] = 5.958, P = 0.016). A significant preservation of VMAT2-ir neurons in the SN ipsilateral to MPTP administration was found in animals treated with 5 mg/kg of pioglitazone (49,701 ± 5,391) compared to placebo (29,911 ± 3,631) (ANOVA; Fisher's LSD post hoc P = 0.04; Figure 4). No significant differences between treatment groups were observed in the contralateral SN.

Nissl-stained nigral neuron stereological quantification confirmed a significant neuronal loss on the side ipsilateral to MPTP administration in comparison to the contralateral side (Wilcoxon Signed Rank test; z = -3.408, P < 0.001, two-tailed). There was a main effect of treatment between groups (ANOVA; F [2,12] = 5.414, P = 0.021). Monkeys treated with 5 mg/kg of pioglitazone had higher cell counts (57,297 ± 7,033) in the ipsilateral SN compared to controls (33,974 ± 3,275) (ANOVA; Fisher's LSD post hoc P = 0.017; Figure 5).

Measures of dopaminergic innervation correlated with behavioral performance (Figure 6). The animals that presented a lower score in the CR (indicating improvement in their PD signs) had more TH-ir OD in the putamen (Pearson's correlation; r ² = 0.320, P = 0.002), as well as more TH (r ² = 0.338, P = 0.024) and VMAT2 (r ² = 0.689, P = 0.003) positive neurons in the SN. Similar to the CR, better performance in the FMS test (observed as less time needed to complete the task) inversely correlated with TH-ir fibers in the putamen (r ² = 0.386, P = 0.004), as well as TH-ir (r ² = 0.654, P = 0.005) and VMAT2-ir (r ² = 0.879, P = 0.004) neurons in the SN indicating that better performance in the task was associated with more DA markers in the nigrostriatal system. Additionally, the number of TH-ir nigral cells was positively correlated with the amount of TH-ir OD in the putamen (r ² = 0.618, P = 0.001).

Reactive neuroinflammation was evaluated using CD68 (marker for microglia/macrophages) and GFAP (astrocyte marker) immunostainings. Qualitative observations of CD68 (Figure 7) immunostained brain coronal sections of placebo-treated subjects revealed the presence of numerous CD68 positive cells in the ventrolateral caudate, ventromedial putamen, and the ventral tier of the SN ipsilateral to MPTP administration. Most of these cells exhibited characteristic hyperramified morphology of activated microglia, and many clustered together. In contrast to placebo-treated animals, pioglitazone-treated monkeys displayed a dose-dependent CD68 immunoreactivity, characterized by mildly active microglia ipsilateral to MPTP administration. Quantification showed a significant increase in CD68-ir particle number in the SN ipsilateral to MPTP administration compared to the contralateral side (Wilcoxon Signed Rank test; z = -3.516, P < 0.01). A significant difference between treatment groups was found in the side ipsilateral to the MPTP lesion (ANOVA; F [2,13] = 3.778, P < 0.05). Multiple comparisons revealed that monkeys treated with 5 mg/kg of pioglitazone had significantly less CD68-ir particle counts in the ipsilateral SN (746 ± 142) compared to placebo (1,331 ± 152) animals (Figure 7; ANOVA Fisher's LSD P = 0.018). In addition, there was a significant difference in the total area covered by the particles between ipsi and contralateral sides (Wilcoxon Signed Ranks test; z = -3.516, P < 0.01, two-tailed) and a significant difference between treatment groups (ANOVA F [2.13] = 4.248, P = 0.038). Multiple comparisons analysis revealed that animals treated with 5 mg/kg of pioglitazone had a significantly smaller CD68 immunoreactive area compared to controls (ANOVA; Fisher's LSD P = 0.015).

GFAP-ir was significantly increased in the SN on the side ipsilateral to MPTP administration compared to the contralateral side (Wilcoxon Signed Ranks test; z = -2.840, P = 0.005 two-tailed), however no significant differences were found between treatment groups (ANOVA; F [2,11] = 0.495, P = 0.622).

Quantification of nitrotyrosine-ir (NO-dependent oxidative stress marker) showed significantly higher levels in the ipsilateral SN compared to the contralateral side (Wilcoxon Signed Rank test; P < 0.001). A trend was revealed toward less intensity in the pioglitazone-treated monkeys compared to placebo (ANOVA; F [2,13] = 3.079, Fisher's LSD P = 0.068).

Quantification of HO-1-ir (oxidative stress marker) did not show significant differences between ipsilateral vs. contralateral SN (Wilcoxon Signed Rank test; P = 0.569) or between treatment groups (ANOVA; F [2,13] = 0.555, P = 0.407).

The number of TH-ir nigral cells inversely correlated with CD68-ir (Pearson's correlation; r ² = 0.279, P = 0.014), and nitrotyrosine-ir (r ² = 0.387, P = 0.001) suggesting that the number of surviving DA cells are affected by inflammation and oxidative stress byproducts (Figure 6). Interestingly, CR and FMS test also correlated with CD68 staining intensity (r ² = 0.663, P = 0.007 and r ² = 0.697, P = 0.003, respectively).