Spontaneous Partial Recovery of Striatal Dopaminergic Uptake Despite Nigral Cell Loss in Asymptomatic MPTP-Lesioned Minipigs

The gold standard animal model of Parkinson’s disease is the non-human primate rendered parkinsonian with the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) toxin. Low availability, ethical issues, and primate-specic biohazards make alternative large animal models necessary. Here, we investigate the temporal evolution of presynaptic dopaminergic function after MPTP in another large animal model, the Göttingen minipig. We subcutaneously injected seven sedated minipigs with 1–2 mg/kg of MPTP, and two minipigs with saline, three times a week over 4 weeks. We monitored behavioral decits using a validated motor scale and a Gait4Dog® walking mat. Minipig brains were imaged with (+)-(cid:0)-[ 11 C]-dihydrotetrabenazine ([ 11 C]-DTBZ) and [ 18 F]-uorodopa ([ 18 F]-FDOPA) PET at baseline and 1, 3, 9 and 12 months after the nal MPTP injection. Immunohistochemical tyrosine hydroxylase (TH) staining was used to assay nigral TH + area loss post-mortem. The minipigs showed only mild bradykinesia and impaired coordination at early timepoints after MPTP. PET revealed decreases of striatal [ 11 C]-DTBZ and [ 18 F]-FDOPA uptake post-MPTP with a partial spontaneous recovery of [ 18 F]-FDOPA after 9 months. Postmortem histological analysis showed a loss of 71% TH-immunopositive area in the substantia nigra. When testing the ecacy of putative neuroprotective agents, partial spontaneous recovery of dopamine terminal function must be taken into account in the MPTP minipig model of parkinsonism.


Introduction
Parkinson's disease (PD) is a chronic neurodegenerative disease characterized by degeneration of nigrostriatal dopaminergic neurons associated with motor symptoms including bradykinesia, rigidity and tremor. The 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) model of PD in non-human primates is a well-established animal lesion model of PD 1 . MPTP is metabolised to a toxic by-product, MPP+, which enters dopaminergic neurons and induces neuronal degeneration by blocking complex 1 activity of the mitochondrial electron transport chain 2 . The model presents with parkinsonian motor symptoms, shows loss of nigral dopamine neurons and striatal levels of dopamine, and responds to levodopa therapy, making it an ideal model system to evaluate the e cacy of neuroprotective agents [3][4][5][6] . MPTP also induces dopamine de ciency in mice [7][8][9] and a number of other small species, including the gold sh 10 .
Although it is regarded as the 'gold standard' model of PD, there are major limitations inherent to the model, including variable dose response, poorly understood compensatory mechanisms, and spontaneous recovery from motor symptoms.
Non-human primates are often the preferred species for the MPTP model, but low availability, ethical, and cost issues limit their use. The Göttingen minipig is a strain speci cally designed for research as their limited growth permits long-term longitudinal follow-up with non-invasive neuroimaging while their brain size allows identi cation of the sub-regional striatal distribution of pre-and post-synaptic dopaminergic tracers [11][12][13][14] . Our team has previously developed and validated MPTP-lesion pig models and studied the e cacy of stem cell transplantation and deep brain stimulation of the subthalamic nucleus 15-20 . In the current report, we aimed to study the longitudinal effects of MPTP administration to minipigs on behavior and uptake of presynaptic markers of the dopamine system using serial positron emission tomography (PET) over one year. Despite large doses of MPTP, the minipigs became only mildly symptomatic at early stages after MPTP injections. Exogenous dopa decarboxylation in dopaminergic terminals was serially studied with uoro-3,4-dihydroxyphenylalnine ([ 18 F]-FDOPA) PET 21

Behavior
During the rst weeks after MPTP injections, all seven pigs had mildly reduced motility, three had abnormal standing/leg positions, one displayed freezing of gait, one had head tremor, and three had abnormal swallowing. All of these symptoms, however, were mild and the total score for each animal ranged from 0-2, with only one animal with a score of 3 out of 12 in the rst weeks after the MPTP treatment. The animals in this study were, therefore, classi ed as asymptomatic or mildly symptomatic during the most affected period, and did not need treatment of symptoms. No score above 0 was observed in the two saline-injected animals at any time during the study.
Using the GAIT4Dog® walking mat, we measured gait symmetry of the minipigs by assessing the Total Pressure Index (TPI) for all limbs, which is expected to be 30% for each front limb (LF and RF) and 20% for each hind limb (LH and RH) under healthy conditions. Both saline (n = 2) and MPTP injected minipigs (n = 7) had TPIs close to the expected 30% for front limbs (range: 28.0-32.5%) and 20% for hind limbs (range: 17.5-20.9%) at all measured timepoints accounting for small individual variances (Table 1). We also used the GAIT4Dog® walking mat to study gait velocity and found a small trend towards a decrease (12%, p = 0.23) in the 7 pigs injected with MPTP one month after the nal MPTP injection. By the next testing on the gait mat at 4 months after the nal MPTP injection, the gait velocity in all pigs had returned to baseline values, and by the nal timepoint 15 months after the last MPTP injection, the values had signi cantly increased by 29% from baseline values (p < 0.016) (Fig. 1).

Longitudinal effects of MPTP neurotransmission
Average parametric Ki and BP ND maps at baseline and 1, 3, 9 and 12 months after MPTP are shown in Fig. 2a  compared to baseline (one-way ANOVA, Bonferroni correction, p < 0.0001) while saline-injected animals showed no signi cant changes from baseline (Fig. 3b).
In order to test the impression that the binding values recovered between the time of initial intoxication and the end of the study, we performed a two-tailed, paired t-test on the data from the 6 minipigs injected with MPTP that were scanned at both the 1-and 9-month timepoints. We found that striatal [ 18 F]-FDOPA Ki values recovered signi cantly (38.6%, p < 0.05) (Fig. 3c) from 1 to 9 months but with no signi cant recovery in striatal [ 11 C]-DTBZ BP ND that only increased by 9.7% between the 1-and 9-month timepoints

Histology
The SN anti-TH staining was semi-quanti ed by measuring the nigral TH-immunopositive area on both hemispheres on ve corresponding sections per animal. In Fig. 4 the mean immunopositive area (µm 2 ) is shown for each animal (top) and each group (bottom). The TH-immunopositive area was signi cantly reduced by 71% (95% CI: 1,353,966 to 546,152, p < 0.001) in the MPTP-injected pigs (n = 7) compared to the saline-injected pigs (n = 2), con rming the presence of cell death in the model.

Discussion
We followed longitudinally the behavior and dopaminergic function of Göttingen minipigs exposed repetitively to MPTP with multi-tracer PET imaging. The MPTP paradigm was similar to that previously used in a large animal model which successfully induced parkinsonism [15][16][17][18][19][20] . Our data show that minipigs dosed with MPTP over a one month period demonstrate signi cant reductions in dopaminergic presynaptic function, observed up to one year later, along with cell death, even in the absence of sustained behavioral symptoms. Despite the signi cant reductions in tracer binding at all timepoints post-MPTP, we found spontaneous partial recovery of striatal [ 18 F]-FDOPA Ki uptake when comparing data 1 month vs 9 months after the nal MPTP injection. Even though the minipigs showed only mild and reversible behavioral changes after MPTP in the most aggressive phase, the administration protocol was effective and su cient to induce in vivo dopamine storage and VMAT2 binding changes along with pathological nigral dopamine cell loss.
The MPTP model in non-human primates is still the gold-standard in PD animal lesion models, especially for trialing the e cacy of novel symptomatic medications 1 . However, there are well known limitations to MPTP administration, including spontaneous recovery from motor symptoms 23,24 which is not observed in idiopathic PD patients. This limits its potential to investigate long-term e cacy of putative neuroprotective medications on motor symptoms. Studies have found that cell loss, however, is not statistically different between non-human primates that recover from MPTP exposure compared to those with stable motor de cits 25 suggesting variable degrees of dopamine terminal adaptation to the nigral lesions are occurring. Indeed, it has been found that remaining TH + bers can sprout and increase branching after partial lesion with MPTP in non-human primates 26 and our data suggest an increase in aromatic L-amino acid decarboxylase (AADC) activity important for the intracellular retention of [ 18 F]-FDOPA in remaining terminals at both 9 and 12 months post-MPTP. It is also known that the same MPTP dose can lead to mild to severe motor de cits in the same species. Previous studies of Göttingen minipigs administered MPTP have shown behavioral sensitivity and moderate parkinsonism, though they reported a non-signi cant tendency towards improved scores throughout the 6-month study period 17,18 . In contrast to their low 0.7-1 mg/kg daily SC injection (cumulative dose of 6.6 ± 1.4 mg/kg over a 22-day injection period), our initial dose of 1 mg/kg three times a week, was doubled in the nal 2 weeks of injections due to the apparent lack of motor symptoms, resulting in a cumulative dose of 18 mg/kg.
Previous studies in non-human primates administered cumulative doses of 1.6 mg/kg 27 and 1.8 mg/kg as a progressive intoxication dose, and 3.2 mg/kg as an acute intoxication dose 25 . It is unclear why the MPTP dose in the current cohort of minipigs, which far exceeded the dose administered in previous large animal studies, was not su cient to elicit moderate parkinsonian symptoms, or any stable behavioral impairment.
However, we speculate that our use of midazolam and (S)-ketamine anesthesia prior to MPTP-injections to increase safety of the staff handling MPTP might have protected the minipigs against extensive ber loss. Midazolam and MPTP are both competitive inhibitors of various cytochrome P450 isozymes, that are involved in the neurotoxicity of MPTP 28, 29 . Furthermore, a recent study found that (R)-ketamine attenuated MPTP-induced reduction in striatal dopamine transporter (DAT) and TH, while (S)-ketamine only had a small effect on DAT 30 . Therefore, the combination of midazolam (S)-ketamine may have reduced the dopamine terminal loss. Nevertheless, the dose was su cient to induce 71% TH + area loss in the SN and signi cant reductions in PET markers of striatal presynaptic dopamine neurotransmission.
[ 18 F]-FDOPA PET has previously been used to assess decline in striatal dopamine terminal function in MPTP-induced parkinsonism and its response to fetal mesencephalic dopaminergic allografts in pigs 17 and humans 31 . The previous transplant study reported motor symptoms and a stable 60% loss of dopa decarboxylase activity 3-and 6-months after MPTP in the non-grafted minipigs. We show here that asymptomatic minipigs demonstrate a clear and persistent reduction in [ 18 F]-FDOPA uptake even when behaviorally intact but, for the rst time, we also demonstrate a signi cant recovery at 9 months after MPTP intoxication, which was also sustained at 12 months post-MPTP. The observed decline in striatal [ 18 F]-FDOPA uptake of MPTP-lesioned pigs is consistent with changes reported in MPTP-exposed nonhuman primates 32 and both idiopathic and MPTP-exposed PD patients 33 . Franke et al. also suggested that mild pathological changes were reversible after cessation of MPTP administration in marmosets 34 .
We have previously shown reductions in striatal [ 11 C]-DTBZ binding in minipig models of parkinsonism induced by acute and chronic proteasome inhibition 11,14 , and preformed a-synuclein brils 35 45 . This difference in uptake and binding was argued to re ect up-regulation of AADC activity in the parkinsonian striatum as a plastic mechanism to increase availability of dopamine to postsynaptic dopamine receptors 45  In conclusion, we show here for the rst time the temporal evolution of two markers of presynaptic dopamine function in a large animal model, the Göttingen minipig, as an alternate to the non-human primate, in response to MPTP. Our asymptomatic to mildly symptomatic animals showed similar radioligand changes observed in behaviorally impaired animal models as well as in human PD. The recovery of [ 18 F]-FDOPA binding 9 and 12 months after the nal dose of MPTP suggests that caution should be taken when planning investigations of therapeutic interventions, which may be confounded by spontaneous recovery of the model. for several weeks prior to the experiment. We fed minipigs a restricted pellet diet (SDS Diet, Witham, UK) and fasted them overnight, with free access to tap water, prior to the day of the experiment. We grouphoused pigs in a 4.6 m 2 enclosure with fence-line visual contact to one another under environmental conditions of 20°C and 50-55 % relative humidity, with air changed 8 times every hour and 12 hour light/dark cycles. We sedated minipigs during subcutaneous MPTP-or saline-injections and anesthetized them for the scanning procedures. We followed humane endpoints of weight loss of over 10% body weight in combination with other observations suggestive of poor well-being, bleeding complications or signs of hypoxia under anaesthesia. All minipigs were clinically healthy prior to the study, and they were randomized to either MPTP or saline treatment groups.

Study overview
Minipigs were imaged at baseline with [ 18 F]-FDOPA and [ 11 C]-DTBZ PET and then injected 3 times per week with 1-2 mg/kg MPTP over a 4-week period. Pigs were scanned four additional times with the same two radioligands at 1, 3, 9 and 12 months after their last MPTP injection. Two animals were injected with saline 3 times per week over a 4-week period and were scanned at baseline and 1, 6 and 9 months after the nal saline injection. Due to the radiochemistry schedule, it was not possible to conduct PET with both radioligands in each animal at each timepoint and one [ 18 F]-FDOPA scan was excluded due to an acquisition error. This led to a varying number of scans across the animals. The MPTP group PET data was obtained with group sizes of at least 5 minipigs (5-6 for baseline and 6-7 for post MPTP scans). Neurological scoring and gait were assessed throughout the study. Histology for tyrosine hydroxylase (TH) immunopositive neurons was performed to assess nigral cell death.
The MPTP-injected minipigs were part of a larger study of the effects of human engineered stem cells as therapeutic agents in the MPTP pig model of PD. Due to lack of behavioral symptoms, spontaneous partial recovery of striatal dopaminergic neurotransmission, and insu cient immunosuppression (> 2 µg/l tacrolimus blood level), there were no differences between the minipigs that received stem cells and sham treatment in any of the studied parameters. We, therefore, pooled the data for the purpose of this study to investigate the long-term effects of MPTP on [ 18 F]-FDOPA and [ 11 C]-DTBZ binding, and potential recovery of PET measures in near asymptomatic MPTP-injected minipigs.

MPTP treatments
Three times a week over a 4-week period, we sedated the minipigs with approximately 0.8 mg/kg of midazolam (Hameln) and 5 mg/kg of S-ketamine (P zer) or 1.2-2.8 mg/kg midazolam alone. We normal, 1: slightly compromised feeding function, 2: rigid jaws and reduced feeding function (total maximum score = 12).
We also assessed walking speed and total pressure index (TPI) of the gait using a GAIT4Dog® walking mat prior to any treatments and at 1 week, 4 months and 15 months after the last MPTP or saline treatment. Minipigs walked at their preferred velocity on the pressure sensing walkway system. Three consecutive gait cycles (12 steps) or more accounted for one quality reading and the nal measures were an average of at least 3 quality readings per animal.

PET Imaging
As previously described 12 , minipigs were anesthetized with a mixture of 1.25 mg/kg midazolam (Hameln) and 6.25 mg/kg S-ketamine (P zer) IM, intubated for mechanical ventilation and prepared for PET imaging. We monitored physiological parameters which all remained at normal porcine values in line with   48 .
We placed minipigs supine in a human PET/CT scanner (Siemens Biograph 64 Truepoint PET) and immobilised their head and body. We positioned the animal within the eld of view of the PET camera and performed a low-dose CT scan prior to each PET recording for anatomical de nition and attenuation correction of PET emission data.
We intravenously administered an average of 369 ± 26 MBq (range 315-415 MBq) (+)-a-[ 11 C] labeled DTBZ (saline with 10% ethanol) via an ear vein catheter in 10 mL saline, during the initial 60 seconds of a 90-minute scan. After the DTBZ scan, we intravenously administered carbidopa ( 50 , we created custom-made masks of the cerebellum to obtain the cerebellar tissue radioactivity as a region of non-displaceable binding and extracted time activity curves for the cerebellum and striatum.

Histology
Deeply anesthetized (Zoletil® 50 Vet.) minipigs were euthanised 15 months post MPTP with an intracardial injection of 20 mL sodium pentobarbital 400 mg/mL (Exagon Vet, Richter Pharma, Austria) followed by transcardial perfusion with 5 L of 4% formaldehyde prior to brain removal as previously described 51,52 and stored in 4% formaldehyde at 4°C. Brains were then embedded in a sectioning chamber (Quick Slicer 100 mm, HistOtech, Aarhus, Denmark) lled with HistOmer mixed with water. Following polymerization of the HistOmer, 2 cm thick sections were cut and transferred to a 30% sucrose in isotonic buffer solution for 1-2 weeks. Brain slabs were then frozen in isopentane cooled to − 40°C for 1-2 min before being cut into eight series of coronal sections (50 µm) on a CryoStar NX70 (Thermo Scienti c) and stored free-oating in DeOlmos at -20°C. One series was mounted directly onto 0.5% gelatin-coated slides and stained with 0.1% toluidine blue in citrate buffer (pH 4) for 9 min followed by a rinse in distilled water, dehydration in 99% alcohol (2 x 30 s) and clearing with xylene before being coverslipped with Pertex®.

Statistics
Statistical tests were performed using Prism8 (GraphPad Software, California, USA). To compare PET and behavioral data in animals injected with MPTP at each timepoint, a one-way ANOVA was used with a Bonferroni correction for multiple comparisons where each post-MPTP timepoint was compared to baseline. The statistical signi cance level was set at p < 0.05 and normal distribution of the data was checked. In order to test the hypothesis that the animals recovered signi cantly from MPTP treatment, a two-tailed paired t-test was performed on the data acquired 1 month vs 9 months after the last MPTP treatment. Furthermore, an unpaired two-tailed t-test was performed comparing nigral TH-immunopositive area of saline-and MPTP-treated animals. Pearson correlation was performed to determine the relationship between FDOPA uptake and DTBZ binding at 1, 3, 9 and 12 months after MPTP and between the measures of each of the two PET tracers at 12 months and the TH-immunopositive area.