Diffusion Tensor Imaging Marks Dopaminergic ... - Wiley Online Library

3 downloads 0 Views 675KB Size Report
entire monkey brain. Preprocessing. Analyses of diffusion tensor data were performed using the Oxford Centre for Functional Magnetic Res- onance Imaging of ...
RESEARCH

ARTICLE

Diffusion Tensor Imaging Marks Dopaminergic and Serotonergic Lesions in the Parkinsonian Monkey  te reau, PhD,1 Maude Beaudoin-Gobert, PhD,1 Sandra Duperrier, BTEC, HND,1 Ste phane Thobois, MD, PhD,1,2 Elise Me on Tremblay, PhD,1 and Ve ronique Sgambato-Faure, PhD 1* Le 1

 de Lyon, Centre National de la Recherche Scientifique, Institut des Sciences Cognitives Marc Jeannerod, Bron, France Universite 2 ^ pital Neurologique Pierre Wertheimer, Lyon, France Hospices Civils de Lyon, Ho

A B S T R A C T : Background: Diffusion tensor imaging has received major interest to highlight markers of neurodegeneration in Parkinson’s disease. Whether the alteration of diffusion parameters mostly depicts dopaminergic lesions or can also reveal serotonergic denervation remains a question. Objectives: The aim of this study was to determine the best diffusion tensor imaging markers of 1-methyl-4phenyl-1,2,3,6-tetrahydropyridine (MPTP) and 3,4-methylene-dioxy-methamphetamine (MDMA; also known as ecstasy) lesions in the nonhuman primate. Methods: We acquired measures of mean diffusivity and fractional anisotropy longitudinally (before and after MPTP and MDMA) and correlated them with severity of parkinsonism, PET imaging, and postmortem fiber quantification.

Both dopaminergic (DA) and serotonergic (5-HT) lesions are found in Parkinson’s disease (PD).1 By the time a PD clinical diagnosis is made, up to 60% of dopaminergic cells are already lost in the substantia

------------------------------------------------------------

ronique Sgambato, Univ Lyon, CNRS, *Corresponding author: Dr. Ve Institut des Sciences Cognitives Marc Jeannerod, UMR5229, 67 Boulevard Pinel, F-69 675 Bron, France; [email protected]

Funding agencies: This work was supported by Fondation de France grants (201234497 and 00016818), a Fondation pour la Recherche dicale grant (DEQ20110421326), a Agence Nationale de la Recherche Me al grant (For Women grant (ANR-09-MNPS-018), and a Fondation L’Ore in Science). This work was performed within the framework of the LABEX  de Lyon, within the program CORTEX (ANR-11-LABX-0042) of Universite “Investissements d’Avenir” (ANR-11-IDEX-0007) operated by the French National Research Agency. V.S-F. is supported by the Institut National  et de la Recherche Me dicale. de la Sante Relevant conflicts of interests/financial disclosures: Nothing to report. Received: 2 June 2017; Revised: 24 August 2017; Accepted: 27 August 2017 Published online 00 Month 2017 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/mds.27201

Results: MPTP-induced lesions were associated with increases of mean diffusivity within both the caudate nucleus and the anterior cingulate cortex, whereas MDMA-induced lesions caused an increase of fractional anisotropy within the caudate nucleus. These variations of diffusion tensor imaging correlated with the motor score. Conclusion: Taken together, these results demonstrate that diffusion measures within specific brain regions can mark severity of dopaminergic and serotonergic induced lesions in a neurotoxic nonhuman primate C 2017 International Parmodel of Parkinson’s disease. V kinson and Movement Disorder Society

K e y W o r d s : Parkinson’s disease; diffusion tensor imaging; monkey; dopamine; serotonin

nigra (SN).2,3 For raphe serotonergic neurons, the percentage of loss is unknown at the time of the diagnosis, but it is clearly linked to disease progression.3-5 Imaging techniques based on specific radiotracers have revealed the presence of DA and 5-HT dysfunction in PD patients.1,6-8 Recent advances in magnetic resonance imaging (MRI) have allowed the visualization of structural abnormalities in the brain of PD patients.9-11 Diffusion tensor imaging (DTI) holds promise for studying abnormalities in both white matter tracts12 and gray matter areas.13 Brain diffusion parameters are altered within both the SN (where dopaminergic cell body loss takes place) and projecting structures in early14-16 and advanced11,14,17-23 PD patients. DTI studies in rodent24,25 and monkey26,27 models of PD have found alterations of diffusion parameters in the SN. However, none of these DTI studies could address whether those diffusion alterations mostly depict DA lesions or also show 5-HT denervation.

Movement Disorders, Vol. 00, No. 00, 2017

1

T E R E A U M E

E T

A L

The aim of this study was to test the reliability of DTI to specifically detect 1-methyl-4-phenyl-1,2,3,6tetrahydropyridine (MPTP) and 3,4-methylene-dioxymethamphetamine (MDMA) lesions induced in the monkey. We used our double intoxicated monkey model of PD, the MPTP- and MDMA-treated macaque,28 and correlated diffusion measures with severity of parkinsonism, PET imaging, and postmortem quantification.

(post-MDMA state; Supplementary Fig. 1 and Supplementary Tables 1 and 2). The mean time between last MPTP dose and DTI scans was (mean 6 SD) 16 6 6 days for the peak-MPTP state and 60 6 23 days for the post-MPTP state. MDMA injection occurred an average of 202 6 56 days after the last MPTP injection. The pre-MDMA scan was acquired 13 6 5 days before the first MDMA injection and the post-MDMA scan 17 6 8 days after.

Methods Ethical Statement All studies were carried out in accordance with European Communities Council Directive of 2010 (2010/63/UE) and the recommendations of the French National Committee (2013/113).

Animals and Treatments A total of 11 male cynomolgus monkeys (Macaca Fascicularis, MF; Supplementary Fig. 1) between 3 and 5 years old and weighing 4 to 6 kg were used in this study. Taking into consideration the 3 Rs (Reduction, Refinement, and Replacement) for animal experimentation, we have been using some of the behavioral and anatomical data from a previous study.28 All monkeys were kept under standard conditions (12hour light cycles, 23 8C and 50% humidity). A double DA/5-HT lesion was induced by MPTP followed by MDMA injections (Sigma-Aldrich, SaintQuentin-Fallavier, France, Erlangen, Germany) as described previously.28 The animals were given intramuscular MPTP injections (mean total dose of 1.55 mg per kg) under light anesthesia (ketamine 0.5 mg/kg, atropine 0.05 mg/kg) to limit the acute effects driven by MPTP. As expected, the monkeys that received progressive MPTP intoxication (3 to 5 injections at 0.3-0.6 mg/kg every 4 to 5 days) recovered from their symptoms (MPTPrec n55) while monkeys that received acute MPTP intoxication (3 to 5 injections at 0.3-0.4 mg/kg on consecutive days) remained symptomatic (MPTPsym, n 5 6; see detailed protocol in ref. 28. Around 6 months after the beginning of MPTP intoxication, 8 of the MPTPintoxicated monkeys (all MPTPrec [n 5 5] and half of MPTPsym [n 5 3]) received MDMA injections twice daily for 4 consecutive days (5 mg/kg, s.c.). Among the 6 MPTPsym monkeys, only 3 received MDMA because the 3 others were used in another study.28

Experimental Schedule DTI measures were acquired at the following 5 different times: (1) baseline, (2) symptom expression (peak-MPTP), (3) symptom recovery for MPTPrec or stabilization for MPTPsym (post-MPTP), (4) before MDMA (pre-MDMA state), and (5) after MDMA

2

Movement Disorders, Vol. 00, No. 00, 2017

DTI Imaging DTI Acquisition On the day of each DTI scan, the monkeys were pretreated with atropine (0.05 mg/kg) and 15 minutes later anesthetized by an intramuscular dose of zoletil (15 mg/kg). The monkeys were then transported to the Imaging Centre (CERMEP, Lyon, France) and kept anesthetized during the acquisition time. The respiratory frequency, pO2, and heart rate were monitored throughout the experiment. During scanning, monkeys were fixed in a sphinx position using an MRI-compatible stereotaxic frame to secure the head. MRI data were acquired using a horizontal 1.5 Tesla magnetic resonance scanner (Sonata, Siemens, Erlangen, Germany) with a radial receive-only surface coil (10-cm diameter) positioned above the monkey’s head. Anatomical MRI acquisition consisted of T1-weighted magnetization-prepared 180 degrees radio-frequency pulses and rapid gradient-echo (MP RAGE) 3D sequences (repetition time (RT) 5 2160 ms, time echo (TE) 5 2.89 ms, flip angle 5 15; Field of View (FoV) 5 154 mm; matrix size 5 256 3 256; voxel size 5 0.6 3 0.6), yielding 176 sagittal 0.6-mm thick slices, repeated 2 times, and averaged to increase signal-to-noise ratio. The acquisition of our DTI sequences was based on the study performed by Hannoun and collaborators.29 Diffusion weighted images were acquired using a 2dimensional spin echo-planar imaging DTI sequence (RT 5 7800 ms; TE 5 103 ms; flip angle 5 90; FoV 5 168 mm; matrix size 5 96 3 96; voxel size 5 1.75 3 1.75) with diffusion-sensitizing gradients in 24 directions and b-factors of 0 and 1000 s/mm2, yielding 51 transversal 2.5-mm thick volumes through the entire monkey brain. Preprocessing Analyses of diffusion tensor data were performed using the Oxford Centre for Functional Magnetic Resonance Imaging of the Brain (FMRIB) Software Library (FSL).30 First, the volumes were eddy-current corrected using FMRIB Diffusion Toolbox (Fdt). The brain tissue was carefully extracted on a nondiffusion-weighted image using the Brain Extraction Tool.31 A binary mask was then created from the

D T I

M A R K S

D A

A N D

processed non-diffusion-weighted image and used to mask all other images in the diffusion dataset. Tensor Estimation DTIFIT (FSL) was used to analyze the diffusionweighted data and provided maps of the tensor diffusivities (k1, k2, k3). Mean diffusivity (MD), axial diffusivity, and radial diffusivity maps were calculated as defined previously.31 Fractional anisotropy (FA) is a measure of the directionality of water diffusion within a given voxel, scaled from 0 (isotropic) to 1 (anisotropic).32 Registration Procedures DTI maps were transformed into the Macaca fascicularis template33 standard space using the following 4-step realignment procedure: (1) FA maps (Supplementary Fig. 2) were registered and resampled into the corresponding individual anatomical MRI using the FSL linear transform algorithm (flirt) with the normalized mutual information cost function, (2) between-FA intramonkey linear registration was performed in this space on a reference FA map (baseline scan FA map) with the correlation ratio cost function, (3) the individual anatomical MRIs were registered on the Macaca fascicularis template using the FSL nonlinear transform algorithm (FNIRT), (4) the transformation matrices calculated in steps 1 to 3 were concatenated to provide transformation from DTI native spaces to the template standard space and applied to the MD, FA, axial diffusivity, and radial diffusivity maps. Region of Interest Delineation Regions of interest (ROIs) were defined based on the Macaca fascicularis maximum probability atlas.33 ROIs were divided into brain stem areas (SN and raphe), basal ganglia regions (anterior and posterior caudate, anterior and posterior putamen, ventral striatum (VS), and pallidum [both external and internal parts]), other subcortical regions (the thalamus, hippocampus, amygdala, and insula) and cortical areas: the sensorimotor cortex (SMC, resulting from the combination of the primary motor cortex, primary sensory cortex, and premotor cortex), anterior cingulate cortex (ACC, encompassing Brodmann areas 24, 25, and 32), and frontal cortex (PFC, resulting from the combination of the dorsolateral frontal cortex, the medial frontal cortex, the ventral frontal cortex, and the orbitofrontal cortex). Because raphe nuclei are not defined in the atlas, the raphe region was delineated in the template space as previously.28 Regional FA and MD were calculated by averaging the values in all the voxels constituting each of these ROIs in the standard space. The numbers of voxels in each ROI were as follows: SN 226, raphe 350,

5 - H T

L E S I O N S

I N

T H E

P R I M A T E

B R A I N

anterior caudate 1361, posterior caudate 860, anterior putamen 1299, posterior putamen 1617, VS 564, pallidum 406, thalamus 2526, hippocampus 1919, amygdala 1547, insula 1734, PFC 2892, ACC 1213, SMC 21416. Whole-Brain Analysis For illustrative purpose only, SPM analyses were performed on the normalized parametric images of FA and MD values using SPM8 software (Wellcome Trust Centre for NeuroImaging, University College London, UK) in Matlab (version 8.1, R2013a, The Mathworks Inc., Natick, Massachusetts). Initially, a Gaussiansmoothing kernel of 4 mm was applied to increase signal to noise ratio. Behavioral, PET Imaging, and Immunohistochemical Data The severity of parkinsonism was evaluated using the rating scale proposed by Schneider and Kovelowski.34 As expected, when compared with MPTPrecovered monkeys, MPTP-symptomatic monkeys exhibited more severe parkinsonism associated with a more severe dopamine lesion.28 Moderately lesioned monkeys totally recovered from their motor symptoms (motor score 5 0 before and after MDMA), whereas severely-lesioned monkeys remained symptomatic (with a stable motor score of around 10 before and after MDMA without any medication). MPTP and MDMA lesions were assessed by PET imaging and immunohistochemistry for each monkey. These data were described previously,28,35 and some were used in this study to define their relationship with DTI measures. For PET imaging, the radiotracers used were 18 F-fluoro- L -DOPA ([18F]DOPA) for amino acid decarboxylase activity and 11C-N,N-dimethyl-2-(-2amino-4-cyanophenylthio)benzylamine ([11C]DASB) for the serotonergic transporter. We acquired [18F]DOPA PET scans at baseline versus peak-MPTP states and [11C]DASB PET scans at pre-MDMA versus post-MDMA states.28 Using the cerebellum (excluding the vermis) as the reference area, the uptake rate (Ki; for [18F]DOPA) and the nondisplaceable binding potential (BPND; for [11C]DASB) were calculated in the ROIs for each state. For postmortem studies, the animals were euthanized at the end of the experiments, their brains removed, fixed, and sliced, and anti-tyrosine hydroxylase (TH), anti-SERT (serotonin transporter), and anti-tryptophan hydroxylase 2 (TPH2) markers were used on the postmortem tissues. These studies were performed under blinded conditions relative to the clinical rating and the imaging measures and were compared to immunohistochemical data obtained in control monkeys (3 male cynomolgus and 1 female rhesus aged 3-5 years). For each monkey, TH 1 cells were counted in the SN on 9 regularly

Movement Disorders, Vol. 00, No. 00, 2017

3

T E R E A U M E

E T

A L

spaced sections encompassing the A8 (peri- and retrorubral area), A9 (pars compacta), and A10 (ventral tegmental area) dopamine regions as reported previously.36 For each monkey, TPH2 1 cells were counted in the raphe nucleus on 5 regularly spaced sections encompassing the rostrocaudal extent of the region. The total number of TH 1 and TPH2 1 cells was estimated after correction by the Abercrombie method.37 TH 1 and SERT 1 fibers were quantified in the ACC, anterior caudate, pallidum, and thalamus at different distances relative to the anterior commissure to sample ROIs (as previously).28 For the anterior caudate and the ACC, the fibers were counted on 2 coronal sections (AC16 and AC10). For the pallidum, the fibers were counted on 3 coronal sections (AC-2, AC-5, and AC-7). For the thalamus, the fibers were counted on 2 coronal sections (AC-5, AC-7). Briefly, for each animal, stained fibers were quantified in the ROIs by counting the number crossing the perimeter of circles (diameter 100 mm) randomly distributed in each region drawn by the computer (Mercator, ExploraNova, La Rochelle, France). As previously,38 we were unable to count TH 1 fibers in the control group (excessive staining). Optical density measurement of TH immunostaining was therefore used to assess the loss of TH in MPTP-treated monkeys when compared with controls. Briefly, 15 circles were randomly distributed by the computer (Mercator) in ROIs. Optical density (expressed in arbitrary units) was obtained by the difference of luminosity between the ROI and the reference region (corpus callosum). Variations of immunohistochemical quantification were expressed relatively to the control count.28

Statistical Analysis ROIs Statistical analyses were conducted using statistical software package Stata 10.1 (StataCorp., College Station, Texas, USA). The differences between baseline and peak-MPTP and between pre-MDMA and postMDMA were assessed using the nonparametric Wilcoxon signed-rank test (on 9 and 7 monkeys, respectively) on FA and MD measures. We then tested for group effect (MPTPrec vs MPTPsym) on the differences of peak-MPTP 2 baseline, post-MPTP 2 baseline, and post-MDMA 2 pre-MDMA using nonparametric Mann–Whitney U tests. To control for multiple comparisons (n 5 14 regions), we used a P value of < .0036 based on the Bonferroni method. However, for completeness, less-confident results with an uncorrected threshold of P < .05 were also reported as trends. Furthermore, we investigated the relationships of DTI variations with lesion and symptom severities. This was done using nonparametric Spearman rank correlation statistics in the ROIs that showed significant effects (after correction for multiple

4

Movement Disorders, Vol. 00, No. 00, 2017

comparison) between baseline versus peak-MPTP (10 regions for MD and 4 for FA) or pre-MDMA versus post-MDMA (2 regions for MD and 3 for FA), with a threshold corrected for multiple comparisons, P < (.05/ number of regions). Finally, analyses of the immunohistochemistry data were conducted as described in our previous study28 using nonparametric Mann– Whitney U tests with P < .05. The SERT 1 fiber count in the anterior caudate nucleus and the ACC were compared for: (1) control versus MPTPrec, (2) control versus MPTPsym, and (3) MPTPrec versus MPTPsym monkeys. The TH 1 fiber counts were compared in the same regions for the MPTPrec versus MPTPsym monkeys. Finally, SERT 1 fiber counts and DTI variations in the pallidum and the thalamus were compared for MPTPrec versus MPTPrec-MDMA monkeys with a P < .05. Whole-Brain Analysis Paired t tests were used to compare FA and MD values at peak-MPTP versus baseline states (9 animals) and at post-MDMA versus pre-MDMA states (7 animals). Subsequently, a multiple regression analysis was used to look for correlation between the maximal motor score and the MD increase after MPTP. We used ImCalc as implemented in SPM8 to create parametric images of the MD increase after MPTP by subtracting MD values at the baseline state from MD values at the peak-MPTP state. Whole-brain analyses were performed within an explicit mask composed of the ROIs (see Region of Interest Delineation section) using only the gray matter part for the cortical ROIs. The aim of these analyses was only to illustrate the main results from the ROI analyses, therefore the threshold level was set to P < .001 for the MPTP effect and P < .005 for the MDMA effect, without correction for multiple comparisons. For visualization, voxel-wise levels of significance (t-maps) were projected onto the Macaca fascicularis MRI template.33

Results Impact of MPTP Lesion on Diffusion Parameters At baseline, FA ranged from 0.15 in prefrontal and sensorimotor cortices to 0.36 in the SN, whereas the MD ranged from 7.06 3 1024 mm2/s in the PFC to 9.88 3 1024 mm2/s in the ACC (Supplementary Table 1). When comparing DTI measures at baseline and peak-MPTP states (Fig. 1A-B, Table 1 and Supplementary Tables 1 and 2), we found significant FA decreases within the posterior caudate, ACC, sensorimotor cortex, and hippocampus (P < .0036, corrected for multiple comparison). We also found significant MD increases within the anterior and posterior

D T I

M A R K S

D A

A N D

5 - H T

L E S I O N S

I N

T H E

P R I M A T E

B R A I N

FIG. 1. Impact of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) on diffusion tensor imaging measurements. (A) Fractional anisotropy (FA) and mean diffusivity (MD) changes at the peak-MPTP state relative to baseline in the regions of interest. A comparison of peak-MPTP versus baseline DTI values in each region; **P