Accepted Article
Article Type: Original Article
Progressive brain metabolic changes under deep brain stimulation of subthalamic nucleus in parkinsonian rats Running head: Combined in vivo 1H-MRS and DBS in rodent Christophe Melon BS1*, Carine Chassain PhD2*, Guy Bielicki BS3, Jean-Pierre Renou PhD3, Lydia Kerkerian-Le Goff PhD1, Pascal Salin PhD1#, Franck Durif MD, PhD2# 1
Aix Marseille Université, CNRS, IBDM UMR 7288, 13288, Marseille, France
2
CHU Clermont-Ferrand and Université d’Auvergne, 63001, Clermont-Ferrand, France
3
NMR plateform, INRA Clermont-Ferrand, 63122, Saint Genes Champanelle, France
*Co-first authors #Co-last authors
Corresponding author’s contact information: Pr Franck Durif, EA7280, Université d’Auvergne, Faculté de Médecine et de Pharmacie, 28 place Henri Dunant, 63001 Clermont-Ferrand, France. E-mail: [email protected] Tel: + 33 4 73 754 790 Fax: + 33 4 73 752 129 Keywords: DBS, Parkinson's disease, MRS, microdialysis, striatum, substantia nigra. Abbreviations: 1H-MRS, proton magnetic resonance spectroscopy; 6-OHDA, 6-hydroxydopamine; BG, basal ganglia; CRLB, Cramer–Rao lower bounds; DBS, Deep brain stimulation; Gln, glutamine; Glu, glutamate; Lac, lactate; MM, macromolecules and lipids; MRI, magnetic resonance imaging; Myo-Ins, myo-inositol; NMR, nuclear magnetic resonance; NAA, N-acetylaspartate; PD, Parkinson’s disease; SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata; STN, subthalamic nucleus; Tau, taurine; tCho, total choline (glycerophosphocholine and phosphocholine); tCr, total creatine (creatine and phosphocreatine); VOI, voxel of interest.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as an 'Accepted Article', doi: 10.1111/jnc.13015 This article is protected by copyright. All rights reserved.
Accepted Article
Abstract
Deep brain stimulation (DBS) of the subthalamic nucleus (STN) is an efficient neurosurgical treatment for advanced Parkinson’s disease. Non-invasive metabolic neuroimaging during the course of DBS in animal models may contribute to our understanding of its action mechanisms. Here, DBS was adapted to in vivo proton magnetic resonance spectroscopy at 11.7T in the rat to follow metabolic changes in the main basal ganglia structures, the striatum and the substantia nigra pars reticulata. Measurements were repeated OFF and ON acute and subchronic (7 days) STN-DBS in control and parkinsonian (6-hydroxydopamine lesion) conditions. Acute DBS reversed the increases in glutamate, glutamine and GABA levels induced by the dopamine lesion in the striatum but not in the SNr. Subchronic DBS normalized GABA in both the striatum and SNr, and glutamate in the striatum. Taurine levels were markedly decreased under subchronic DBS in the striatum and SNr in both lesioned and unlesioned rats. Microdialysis in the striatum further showed that extracellular taurine was increased. These data reveal that STN-DBS has duration-dependent metabolic effects in the basal ganglia, consistent with development of adaptive mechanisms. In addition to counteracting defects induced by the dopamine lesion, prolonged DBS has proper effects independent of the pathological condition. Introduction Deep brain stimulation (DBS) at high frequency of the subthalamic nucleus (STN) is an effective symptomatic treatment of Parkinson’s disease (PD), which shows sustained long-term efficacy in relation to most motor symptoms (Krack et al. 2003; Rodriguez-Oroz et al. 2005; Kringelback et al. 2007; Benabid et al. 2009; Rizzone et al. 2014). Contrasting with its increasing use, little is known about the impact of prolonged application of this surgical treatment on brain function. Studies of DBS action mechanisms in animal models are mostly based on short duration stimulation (minutes to hours) (Gubellini et al. 2009). The few studies examining the effects of prolonged STN-DBS indicate duration-dependent impact on basal ganglia (BG) functioning and suggest that widespread adaptive changes within and outside the BG network may contribute to this treatment action (Bacci et al. 2004; Khaindrava et al. 2011; Lortet et al. 2013). Recent evidence has also been provided for neural plasticity in brain connectivity in a PD patient as a result of long-term DBS (van Hartevelt et al. 2014). The motor symptoms of PD are primarily attributed to dysfunction in the cortico-BG-thalamocortical loop circuits resulting from degeneration of the dopamine neurons in the substantia nigra
This article is protected by copyright. All rights reserved.
Accepted Article
pars compacta. The dopamine denervation-induced changes in glutamate (Glu) and GABA neurotransmission systems are key components of PD pathophysiology. Alterations in the glutamatergic drive include increased Glu transmission and maladaptive synaptic plasticity at the synapses formed by cortical afferents onto projection neurons of the striatum (Lindefors and Ungerstedt 1990; Calabresi et al. 1993; Centonze et al. 1999), the main BG input station, and pathological activation in bursts of STN (Hollerman et al. 1992; Bergman et al. 1994; Bevan et al. 2002). Dopamine denervation triggers opposite changes in the two populations of striatal GABA projection neurons, decreased activity of the direct pathway neurons and increased activity of the indirect pathway neurons. These changes contribute, together with pathological activation of STN, to an overactivity of the BG output structures, the internal globus pallidus and the substantia nigra pars reticulata (SNr) (Blandini et al. 2000; Obeso et al. 2008). Several studies in rodent models have investigated the impact of short duration (1-4 hours) STN-DBS on Glu and/or GABA tone in STN itself or in other BG components (Bruet et al. 2003; Windels et al. 2005; Boulet et al. 2006; Lee et al. 2007; Walker et al. 2009, 2012; Favier et al. 2013). The local impact of DBS has also been analyzed in human neocortical slices (Mantovani et al. 2006) or in the striatum of freely moving rats (Hiller et al. 2007). Despite some discrepancies, results suggest that the control of GABA and Glu transmission may be main components of DBS action mechanisms, at least in the short-term. However, whether and how prolonged STN-DBS impacts Glu and GABA neurotransmission systems remains unknown. In vivo proton (1H) magnetic resonance spectroscopy (MRS) is a powerful noninvasive means for evaluating brain regional levels of several metabolites. MRS studies in animal models of PD reported changes in striatal levels of GABA, Glu and Gln and their recovery following acute L-Dopa therapy (Chassain et al. 2008, 2010; Bagga et al. 2013; Coune et al. 2013). Here, we developed a stimulation procedure adapted for use in magnetic environment and coupled for the first time DBS with high field MRS in vivo in the rat to study the temporal evolution of STN-DBS effects between acute (1-3 hours) and subchronic (7 days) conditions in the striatum and the SNr in control and parkinsonian states. In addition, striatal extracellular levels of aminoacids were measured using in vivo microdialysis.
Materials and Methods Animal care and use All experiments were performed on young adult male Wistar HAN rats (Charles River, L'Arbresle, France) weighing 180–200g at first surgery time and were in accordance with the European
This article is protected by copyright. All rights reserved.
Accepted Article
Communities Council Directive (86/609/EEC). Protocols were approved by the local ethical committees (project number 4-03-03-2011 for DBS/microdialysis and CE07-11 for DBS/MRS). Animals were housed one per cage after DBS electrode implantation and maintained on a 12-h light/12-h dark cycle at a constant temperature (21±1°C) with free access to food and water. Surgery Stereotactic surgery was performed under Ketamine+Xylazine anaesthesia (100+10mg⁄kg, i.p.). 6-OHDA lesion. Animals received a unilateral injection inside the left SNc of 12µg of 6-OHDA (Sigma–Aldrich, St. Quentin-Fallavier, France) dissolved in 6µl of 0.9% sterile NaCl containing 0.1% ascorbic acid, at 1µl/min. Sham rats received the vehicle solution. The stereotaxic coordinates were: anteroposterior +2.2mm; lateral 2.0mm; dorsoventral +3.3mm; incisor bar +5.0mm above the interaural plane according to the rat stereotaxic atlas by De Groot (1959). Electrode implantation for DBS Fourteen days after the nigral injection, rats were unilaterally implanted with one electrode for DBS in the left STN. Home-made bipolar gold electrodes and polyethylene guides were used. These materials were selected based on pilot experiments showing compatibility with use in magnetic environment with no artifacts on NMR images and no tissue damage after several days of continuous DBS. The electrode design was as previously described (Bacci et al. 2004; Forni et al. 2012). In brief, the electrode is formed from two parallel gold wires (distance~400µm) insulated with Teflon and bared at the extremity for a length of 500µm (diameter of each wire 140µm insulated, 76µm bare). The electrode was implanted so that the two wires were placed in the anteroposterior axis and active zone covering the STN extent in depth. The stereotaxic coordinates were taken as the average of interaural (AP=+5.2mm; L=2.35mm and DV=+2.4mm) and bregma (AP=-3.8mm taken at equidistance of the two wires; L=2.35mm; DV=-8.1mm from dura) coordinates from the Paxinos and Watson atlas (Paxinos and Watson 1998). The electrode was fixed to the skull with dental cement and welded to a connector that allows connection with the stimulation device. In the same surgical session, animals for microdialysis were implanted on the dopamine lesion side with a cannula guide CMA⁄11 (Carnegie Medicine, Stockholm, Sweden) placed above the striatum (AP=+0.2mm and L=2.8mm from bregma; DV=-3.3mm from dura) (Paxinos and Watson 1998).
This article is protected by copyright. All rights reserved.
Accepted Article
Protocol design for DBS For MRS experiments, 3 successive acquisition sessions were performed: the first “OFF DBS” at five days post-implantation, the second at the start of DBS (between 1 to 3 hours of DBS) representing the “acute DBS” condition and the third after 7 days of DBS representing the “subchronic DBS” condition. For the 7 days DBS in between the MRS sessions, the connector was plugged in a portable microstimulator (Forni et al. 2012) placed in a backpack (Harvard Apparatus, MA, USA) that allows continuous stimulation in freely moving animals. For the two MRS sessions under DBS, we used a home-made external non-magnetic stimulator consisting of two main blocks: one pulse generator and one voltage to current converter. It can deliver pulses of adjustable intensity from 80 to 100µA at a frequency of 130Hz. Filter assemblies decoupled lead wires and/or electronic components of the device from undesirable electromagnetic interference (EMI) signals at the selected frequency of the MRI equipment. All components were non-magnetic. The entire assembly was packaged in a landscaped Faraday cage enclosure to prevent external electromagnetic disturbance. It was powered by a rechargeable lead gel battery. The electrode was connected to the stimulator via a twisted cable to reduce the effects of noisy electromagnetic inductions of environment. The DBS impact on NMR acquisitions was checked in vitro by implanting a gold electrode in a phantom solution. Signal to noise ratios from the spectra acquired at proximity of the electrode in stimulation OFF condition and after two hours of stimulation were estimated using the acquisition software by measuring the metabolites amplitude signal relatively to the signals amplitudes in the noise. They were identical within experimental errors (Fig. 1). Metabolites concentrations measured from these spectra were in accordance with the phantom preparation. For microdialysis experiments, OFF and ON acute conditions were successively assessed in the same first dialysis session at 5 days post-surgery and subchronic condition was assessed in a second session after 7 days of continuous DBS in freely moving rats. Electrical stimuli were delivered by an external pulse generator and a stimulus isolation unit (P2MP, Marseille, France). In all cases (MRS and microdialysis), monophasic rectangular current pulses were delivered. DBS was applied under conditions proven efficient for alleviating DA denervation-mediated motor deficits while avoiding tissue damage (Gubellini et al. 2006; Oueslati et al. 2007): frequency: 130Hz, pulse width: 80µs, intensity: 80µA (about two fold below the intensity evoking contralateral forelimb dyskinetic movements).
This article is protected by copyright. All rights reserved.
Accepted Article
Behavioral observations The cylinder test was performed twice for each animal under OFF and subchronic DBS conditions to assess the parkinsonian akinesia-like deficit induced by the 6-OHDA lesion and STN-DBS efficiency to relieve this deficit. Animals were placed in a Plexiglas® cylinder and video-recorded during 15min. The number of contacts made on the cylinder wall with the ipsilateral and the contralateral forepaws, either separately or together, were counted. Results were expressed as an asymmetry score, defined as the percent of contralateral contacts minus the percent of ipsilateral contacts, and were means±SD of the values determined from the n animals per group. MRS All MRS experiments were performed at 11.7T on a Bruker BioSpec 117/16 Ultra Shielded Refrigerated system. Animals under inhaled gas anesthesia (1-2.4% isoflurane and air, 300mL/min) were secured in a handling system with head centered in a circular polarized 1H rat brain radiofrequency coil used for excitation and signal reception. Respiratory rate was monitored all along the session to adjust anesthesia and corporal temperature was maintained at 37.8°C using a warm air system. 1H-MRS spectra were acquired from a voxel of interest (VOI) centered in the striatum and in the SNr (Fig. 3). Sizes of VOI are 8μL (2×2×2mm) and 2.9μL (1.8×1.0×1.6mm), respectively. The positions of VOI were based on T2 weighted images in accordance with the Paxinos and Watson atlas (Paxinos and Watson 1998). The middle of the STR voxel was positioned 0.36mm posterior and 3.0mm left from the bregma and 5.5mm from the skull surface. For the SNr voxel, it was positioned 4.92mm posterior and 2.5mm left from the bregma and 8.5mm from the skull surface. The 2.9-µl VOI localized on the SNr was selected as the smallest volume around the SNr and its location was positioned in the most reproducible way, using anatomical landmarks on the T2 weighted NMR images: the aspect of the corpus callosum and the volume of the ventricles. Furthermore, higher grey level allowed delineating the SNr from the SNc. To validate a reproducible positioning of the voxel in all animals and between the three MRS sessions, a percentage voxel coverage was estimated using Gimp2.6.11 (image registration toolbox). Automatic shimming of 1st and 2nd order were performed and then refined manually. For localized spectroscopy, a standard PRESS sequence was used with an echo time of 8.8ms, a repetition time of 4000ms and a spectral width of 5000Hz. The water signal from the VOIs was suppressed by variable-power RF pulses with optimized relaxation delays (VAPOR) (Tkac et al. 2004). Each spectrum corresponded to a mean of 512 scans for the dorsal striatum and a mean of 1664 scans for the SNr. A spectrum was acquired in the same conditions without water suppression for the absolute quantification of metabolites.
This article is protected by copyright. All rights reserved.
Accepted Article
Metabolite
concentrations
were
determined
with
jMRUI
software
(version
5.0,
http://www.mrui.uab.es/mrui). The quantification method used was the time-domain semiparametric algorithm QUEST, based on signals for a basis set of simulated metabolites comprising the following molecules: lactate (Lac), N-acetylaspartate (NAA), total creatine (creatine and phosphocreatine; tCr), total choline (glycerophosphocholine and phosphocholine; tCho), Glu, glutamine (Gln), GABA, taurine (Tau) and myo-inositol (Myo-Ins). The basis set also included a simulated signal of macromolecules and lipids (MM). The intensity of the water signal obtained from nonsuppressed water spectra was used as an internal reference. The reliability of metabolite quantification was assessed from the average Cramer–Rao lower bounds (CRLB) calculated by jMRUI. Only results with a CRLB ≤ 30% were included and were expressed in mM, as means±SEM. In vivo microdialysis and measurement of extracellular amino-acids levels The microdialysis probe (CMA/11, Carnegie Medicine, Stockholm, Sweden; microdialysis membrane length, 3mm; molecular weight cut-off, 20kDa; outer diameter, 0.5mm) was lowered through the guide cannula so that the tip of the membrane reached DV=-6.3mm from the dura surface. Probes were perfused with Ringer’s solution (147mM NaCl, 2.5mM CaCl2, 4mM KCl,) at a constant flow rate (1µL⁄min) using a CMA⁄102 microdialysis pump (Carnegie Medicine). The dialysates were collected every 20min in tubes containing 30µL of a 10µM ascorbic acid and 10µM homoserine (as internal sample control) solution, and were immediately frozen at -80°C until HPLC analysis. Sample collection started 120min after the perfusion onset to achieve stable levels of aminoacids. In the first dialysis session, 6 samples were collected for 120min as basal values OFF DBS then DBS was turned ON and 6 more samples were collected as acute DBS. In the second dialysis session, 6 samples were collected as chronic DBS. Analysis was performed on the 3 last samples out of the 6 for each condition (corresponding for the acute condition to the time window of 1 to 2 hours after DBS start). Measurement of aminoacids was performed using HPLC coupled with fluorimetric detection (Waters,
St.
Quentin,
France),
including
a
precolumn
derivatization
with
o-
phthaldialdehyde⁄mercaptoethanol and a C-18 (ODS2, 4.6·150mm) Spherisorb column. A non-linear gradient delivered through a Waters 600 pump was used to separate derivatives (solvent A: 0.1M potassium acetate with 25% methanol, pH 5.5; solvent B: 0.05M potassium acetate and 60% methanol, pH 5.5). Samples (15µL) were automatically injected (Waters 717 plus autosampler) and analysed using a Waters 474 fluorimetric detector. The limit of detection was 1.5pmol⁄sample. Data were computed with WATERS MILLENIUM software (via a Waters bus SAT⁄IN module), compound identification and peak quantification were achieved by comparison with standard aminoacid solution. The average of each aminoacid concentration from the three samples per condition was
This article is protected by copyright. All rights reserved.
Accepted Article
then determined for each animal. Extracellular GABA levels were close to the detection limit of the HPLC measurement, preventing sound assessment of the changes in the different experimental conditions. Data were means±SEM of the values obtained for the n animals of each experimental group. Histological verification of 6-OHDA lesion and electrode placement At the end of MRS or microdialysis experiments, animals were killed by decapitation and brains were directly frozen on dry ice and stored at -80°C. Coronal tissue sections (10µm thick) were cut at -20°C with a cryostat (Leica CM3050, Nussloch, Germany), mounted on SuperFrost Plus glass slides (Fisher Scientific, Elancourt, France) and stored at -80°C until specific treatment. The loss of DA terminals in the striatum was assessed as an index of the DA denervation extent by analysis of 3H-mazindol binding to DA uptake sites, as described previously (Salin et al. 2002; Bacci et al. 2004). The correct location of the stimulating electrode and of the dialysis probe was examined on Cresyl Violet-stained sections. Only animals showing extensive striatal DA denervation (>85%), location of the two wires of the DBS electrode within the STN and location of the dialysis probe within the striatum were selected for data analysis. Around 30% of animals which did not satisfy all the selection criteria have been discarded. The final cohort contained twenty-eight rats divided up into two groups (i) intranigral NaCl-injected rats (sham, n=6 for MRS and n=8 for microdialysis); (ii) intranigral 6-OHDAinjected rats (6-OHDA, n=6 for MRS and n=8 for microdialysis). Statistical analysis Data were analyzed statistically (Statistica, version 7.1; Stat Soft, Maisons-Alfort, France) using a two-way analysis of variance, ANOVA (group: sham, 6-OHDA and DBS conditions as repeated measures: OFF, acute (except cylinder test) and subchronic). When it was significant, ANOVA was followed by Newman-Keuls post-hoc test. Results Forelimb asymmetry In the cylinder test, sham rats OFF DBS equivalently used their right and left forepaws and this behavior was not affected under unilateral subchronic DBS (Fig. 2). 6-OHDA animals assessed OFF DBS showed marked asymmetry with negative score, the number of contralateral contacts being dramatically reduced due to akinesia-like deficit affecting the forepaw contralateral to the lesion. This asymmetry was completely reversed by subchronic DBS of the STN on the lesion side.
This article is protected by copyright. All rights reserved.
Accepted Article
MRS spectroscopy For each animal, we have checked that the stimulation did not induce artifact on NMR acquisitions in vivo and tissue damage in STN after histological control on brain sections (Fig. 3a). Typical in vivo 1H spectra obtained in the striatum are depicted in figure 3b. After shimming, a half-height linewidth of the water signal of 10-12Hz was achieved in the dorsal striatum. Measuring MR spectra in the SNr was challenging due to the small size of this structure, its location in the midbrain and its high iron content. However, despite these challenges in shimming and broad intrinsic linewidths relative to other brain area, spectral quality that allows assessment of the neurochemical profile of the SNr was achieved in 2.9µL volumes (Fig. 3c). The half-height linewidth of the water resonance was 13-15Hz. The majority of the metabolites had average CRLBs lower than 15% in the striatum, with the exception of GABA (18-22%) and Gln (15-18%), and lower than 25% in the SNr. Session-on-session percentage of voxel coverage was in mean 88.9±6.4% in the STR and 80.7±9.4% in the SNr. In the OFF DBS condition, 6-OHDA rats showed significant increases vs sham in striatal levels of Glu, Gln and GABA (Fig. 4). Levels of these metabolites were also increased in the SNr of 6-OHDA rats compared to sham (Fig. 4). DBS, either acute or chronic, did not affect the levels of Glu, Gln, or GABA in the striatum or the SNr of sham animals while having differential impact in 6-OHDA animals. In the striatum, acute DBS normalized Glu, Gln and GABA levels in 6-OHDA rats: the levels were significantly reduced versus 6-OHDA OFF DBS (Glu: -33±9%, p
Article Type: Original Article
Progressive brain metabolic changes under deep brain stimulation of subthalamic nucleus in parkinsonian rats Running head: Combined in vivo 1H-MRS and DBS in rodent Christophe Melon BS1*, Carine Chassain PhD2*, Guy Bielicki BS3, Jean-Pierre Renou PhD3, Lydia Kerkerian-Le Goff PhD1, Pascal Salin PhD1#, Franck Durif MD, PhD2# 1
Aix Marseille Université, CNRS, IBDM UMR 7288, 13288, Marseille, France
2
CHU Clermont-Ferrand and Université d’Auvergne, 63001, Clermont-Ferrand, France
3
NMR plateform, INRA Clermont-Ferrand, 63122, Saint Genes Champanelle, France
*Co-first authors #Co-last authors
Corresponding author’s contact information: Pr Franck Durif, EA7280, Université d’Auvergne, Faculté de Médecine et de Pharmacie, 28 place Henri Dunant, 63001 Clermont-Ferrand, France. E-mail: [email protected] Tel: + 33 4 73 754 790 Fax: + 33 4 73 752 129 Keywords: DBS, Parkinson's disease, MRS, microdialysis, striatum, substantia nigra. Abbreviations: 1H-MRS, proton magnetic resonance spectroscopy; 6-OHDA, 6-hydroxydopamine; BG, basal ganglia; CRLB, Cramer–Rao lower bounds; DBS, Deep brain stimulation; Gln, glutamine; Glu, glutamate; Lac, lactate; MM, macromolecules and lipids; MRI, magnetic resonance imaging; Myo-Ins, myo-inositol; NMR, nuclear magnetic resonance; NAA, N-acetylaspartate; PD, Parkinson’s disease; SNc, substantia nigra pars compacta; SNr, substantia nigra pars reticulata; STN, subthalamic nucleus; Tau, taurine; tCho, total choline (glycerophosphocholine and phosphocholine); tCr, total creatine (creatine and phosphocreatine); VOI, voxel of interest.
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process which may lead to differences between this version and the Version of Record. Please cite this article as an 'Accepted Article', doi: 10.1111/jnc.13015 This article is protected by copyright. All rights reserved.
Accepted Article
Abstract
Deep brain stimulation (DBS) of the subthalamic nucleus (STN) is an efficient neurosurgical treatment for advanced Parkinson’s disease. Non-invasive metabolic neuroimaging during the course of DBS in animal models may contribute to our understanding of its action mechanisms. Here, DBS was adapted to in vivo proton magnetic resonance spectroscopy at 11.7T in the rat to follow metabolic changes in the main basal ganglia structures, the striatum and the substantia nigra pars reticulata. Measurements were repeated OFF and ON acute and subchronic (7 days) STN-DBS in control and parkinsonian (6-hydroxydopamine lesion) conditions. Acute DBS reversed the increases in glutamate, glutamine and GABA levels induced by the dopamine lesion in the striatum but not in the SNr. Subchronic DBS normalized GABA in both the striatum and SNr, and glutamate in the striatum. Taurine levels were markedly decreased under subchronic DBS in the striatum and SNr in both lesioned and unlesioned rats. Microdialysis in the striatum further showed that extracellular taurine was increased. These data reveal that STN-DBS has duration-dependent metabolic effects in the basal ganglia, consistent with development of adaptive mechanisms. In addition to counteracting defects induced by the dopamine lesion, prolonged DBS has proper effects independent of the pathological condition. Introduction Deep brain stimulation (DBS) at high frequency of the subthalamic nucleus (STN) is an effective symptomatic treatment of Parkinson’s disease (PD), which shows sustained long-term efficacy in relation to most motor symptoms (Krack et al. 2003; Rodriguez-Oroz et al. 2005; Kringelback et al. 2007; Benabid et al. 2009; Rizzone et al. 2014). Contrasting with its increasing use, little is known about the impact of prolonged application of this surgical treatment on brain function. Studies of DBS action mechanisms in animal models are mostly based on short duration stimulation (minutes to hours) (Gubellini et al. 2009). The few studies examining the effects of prolonged STN-DBS indicate duration-dependent impact on basal ganglia (BG) functioning and suggest that widespread adaptive changes within and outside the BG network may contribute to this treatment action (Bacci et al. 2004; Khaindrava et al. 2011; Lortet et al. 2013). Recent evidence has also been provided for neural plasticity in brain connectivity in a PD patient as a result of long-term DBS (van Hartevelt et al. 2014). The motor symptoms of PD are primarily attributed to dysfunction in the cortico-BG-thalamocortical loop circuits resulting from degeneration of the dopamine neurons in the substantia nigra
This article is protected by copyright. All rights reserved.
Accepted Article
pars compacta. The dopamine denervation-induced changes in glutamate (Glu) and GABA neurotransmission systems are key components of PD pathophysiology. Alterations in the glutamatergic drive include increased Glu transmission and maladaptive synaptic plasticity at the synapses formed by cortical afferents onto projection neurons of the striatum (Lindefors and Ungerstedt 1990; Calabresi et al. 1993; Centonze et al. 1999), the main BG input station, and pathological activation in bursts of STN (Hollerman et al. 1992; Bergman et al. 1994; Bevan et al. 2002). Dopamine denervation triggers opposite changes in the two populations of striatal GABA projection neurons, decreased activity of the direct pathway neurons and increased activity of the indirect pathway neurons. These changes contribute, together with pathological activation of STN, to an overactivity of the BG output structures, the internal globus pallidus and the substantia nigra pars reticulata (SNr) (Blandini et al. 2000; Obeso et al. 2008). Several studies in rodent models have investigated the impact of short duration (1-4 hours) STN-DBS on Glu and/or GABA tone in STN itself or in other BG components (Bruet et al. 2003; Windels et al. 2005; Boulet et al. 2006; Lee et al. 2007; Walker et al. 2009, 2012; Favier et al. 2013). The local impact of DBS has also been analyzed in human neocortical slices (Mantovani et al. 2006) or in the striatum of freely moving rats (Hiller et al. 2007). Despite some discrepancies, results suggest that the control of GABA and Glu transmission may be main components of DBS action mechanisms, at least in the short-term. However, whether and how prolonged STN-DBS impacts Glu and GABA neurotransmission systems remains unknown. In vivo proton (1H) magnetic resonance spectroscopy (MRS) is a powerful noninvasive means for evaluating brain regional levels of several metabolites. MRS studies in animal models of PD reported changes in striatal levels of GABA, Glu and Gln and their recovery following acute L-Dopa therapy (Chassain et al. 2008, 2010; Bagga et al. 2013; Coune et al. 2013). Here, we developed a stimulation procedure adapted for use in magnetic environment and coupled for the first time DBS with high field MRS in vivo in the rat to study the temporal evolution of STN-DBS effects between acute (1-3 hours) and subchronic (7 days) conditions in the striatum and the SNr in control and parkinsonian states. In addition, striatal extracellular levels of aminoacids were measured using in vivo microdialysis.
Materials and Methods Animal care and use All experiments were performed on young adult male Wistar HAN rats (Charles River, L'Arbresle, France) weighing 180–200g at first surgery time and were in accordance with the European
This article is protected by copyright. All rights reserved.
Accepted Article
Communities Council Directive (86/609/EEC). Protocols were approved by the local ethical committees (project number 4-03-03-2011 for DBS/microdialysis and CE07-11 for DBS/MRS). Animals were housed one per cage after DBS electrode implantation and maintained on a 12-h light/12-h dark cycle at a constant temperature (21±1°C) with free access to food and water. Surgery Stereotactic surgery was performed under Ketamine+Xylazine anaesthesia (100+10mg⁄kg, i.p.). 6-OHDA lesion. Animals received a unilateral injection inside the left SNc of 12µg of 6-OHDA (Sigma–Aldrich, St. Quentin-Fallavier, France) dissolved in 6µl of 0.9% sterile NaCl containing 0.1% ascorbic acid, at 1µl/min. Sham rats received the vehicle solution. The stereotaxic coordinates were: anteroposterior +2.2mm; lateral 2.0mm; dorsoventral +3.3mm; incisor bar +5.0mm above the interaural plane according to the rat stereotaxic atlas by De Groot (1959). Electrode implantation for DBS Fourteen days after the nigral injection, rats were unilaterally implanted with one electrode for DBS in the left STN. Home-made bipolar gold electrodes and polyethylene guides were used. These materials were selected based on pilot experiments showing compatibility with use in magnetic environment with no artifacts on NMR images and no tissue damage after several days of continuous DBS. The electrode design was as previously described (Bacci et al. 2004; Forni et al. 2012). In brief, the electrode is formed from two parallel gold wires (distance~400µm) insulated with Teflon and bared at the extremity for a length of 500µm (diameter of each wire 140µm insulated, 76µm bare). The electrode was implanted so that the two wires were placed in the anteroposterior axis and active zone covering the STN extent in depth. The stereotaxic coordinates were taken as the average of interaural (AP=+5.2mm; L=2.35mm and DV=+2.4mm) and bregma (AP=-3.8mm taken at equidistance of the two wires; L=2.35mm; DV=-8.1mm from dura) coordinates from the Paxinos and Watson atlas (Paxinos and Watson 1998). The electrode was fixed to the skull with dental cement and welded to a connector that allows connection with the stimulation device. In the same surgical session, animals for microdialysis were implanted on the dopamine lesion side with a cannula guide CMA⁄11 (Carnegie Medicine, Stockholm, Sweden) placed above the striatum (AP=+0.2mm and L=2.8mm from bregma; DV=-3.3mm from dura) (Paxinos and Watson 1998).
This article is protected by copyright. All rights reserved.
Accepted Article
Protocol design for DBS For MRS experiments, 3 successive acquisition sessions were performed: the first “OFF DBS” at five days post-implantation, the second at the start of DBS (between 1 to 3 hours of DBS) representing the “acute DBS” condition and the third after 7 days of DBS representing the “subchronic DBS” condition. For the 7 days DBS in between the MRS sessions, the connector was plugged in a portable microstimulator (Forni et al. 2012) placed in a backpack (Harvard Apparatus, MA, USA) that allows continuous stimulation in freely moving animals. For the two MRS sessions under DBS, we used a home-made external non-magnetic stimulator consisting of two main blocks: one pulse generator and one voltage to current converter. It can deliver pulses of adjustable intensity from 80 to 100µA at a frequency of 130Hz. Filter assemblies decoupled lead wires and/or electronic components of the device from undesirable electromagnetic interference (EMI) signals at the selected frequency of the MRI equipment. All components were non-magnetic. The entire assembly was packaged in a landscaped Faraday cage enclosure to prevent external electromagnetic disturbance. It was powered by a rechargeable lead gel battery. The electrode was connected to the stimulator via a twisted cable to reduce the effects of noisy electromagnetic inductions of environment. The DBS impact on NMR acquisitions was checked in vitro by implanting a gold electrode in a phantom solution. Signal to noise ratios from the spectra acquired at proximity of the electrode in stimulation OFF condition and after two hours of stimulation were estimated using the acquisition software by measuring the metabolites amplitude signal relatively to the signals amplitudes in the noise. They were identical within experimental errors (Fig. 1). Metabolites concentrations measured from these spectra were in accordance with the phantom preparation. For microdialysis experiments, OFF and ON acute conditions were successively assessed in the same first dialysis session at 5 days post-surgery and subchronic condition was assessed in a second session after 7 days of continuous DBS in freely moving rats. Electrical stimuli were delivered by an external pulse generator and a stimulus isolation unit (P2MP, Marseille, France). In all cases (MRS and microdialysis), monophasic rectangular current pulses were delivered. DBS was applied under conditions proven efficient for alleviating DA denervation-mediated motor deficits while avoiding tissue damage (Gubellini et al. 2006; Oueslati et al. 2007): frequency: 130Hz, pulse width: 80µs, intensity: 80µA (about two fold below the intensity evoking contralateral forelimb dyskinetic movements).
This article is protected by copyright. All rights reserved.
Accepted Article
Behavioral observations The cylinder test was performed twice for each animal under OFF and subchronic DBS conditions to assess the parkinsonian akinesia-like deficit induced by the 6-OHDA lesion and STN-DBS efficiency to relieve this deficit. Animals were placed in a Plexiglas® cylinder and video-recorded during 15min. The number of contacts made on the cylinder wall with the ipsilateral and the contralateral forepaws, either separately or together, were counted. Results were expressed as an asymmetry score, defined as the percent of contralateral contacts minus the percent of ipsilateral contacts, and were means±SD of the values determined from the n animals per group. MRS All MRS experiments were performed at 11.7T on a Bruker BioSpec 117/16 Ultra Shielded Refrigerated system. Animals under inhaled gas anesthesia (1-2.4% isoflurane and air, 300mL/min) were secured in a handling system with head centered in a circular polarized 1H rat brain radiofrequency coil used for excitation and signal reception. Respiratory rate was monitored all along the session to adjust anesthesia and corporal temperature was maintained at 37.8°C using a warm air system. 1H-MRS spectra were acquired from a voxel of interest (VOI) centered in the striatum and in the SNr (Fig. 3). Sizes of VOI are 8μL (2×2×2mm) and 2.9μL (1.8×1.0×1.6mm), respectively. The positions of VOI were based on T2 weighted images in accordance with the Paxinos and Watson atlas (Paxinos and Watson 1998). The middle of the STR voxel was positioned 0.36mm posterior and 3.0mm left from the bregma and 5.5mm from the skull surface. For the SNr voxel, it was positioned 4.92mm posterior and 2.5mm left from the bregma and 8.5mm from the skull surface. The 2.9-µl VOI localized on the SNr was selected as the smallest volume around the SNr and its location was positioned in the most reproducible way, using anatomical landmarks on the T2 weighted NMR images: the aspect of the corpus callosum and the volume of the ventricles. Furthermore, higher grey level allowed delineating the SNr from the SNc. To validate a reproducible positioning of the voxel in all animals and between the three MRS sessions, a percentage voxel coverage was estimated using Gimp2.6.11 (image registration toolbox). Automatic shimming of 1st and 2nd order were performed and then refined manually. For localized spectroscopy, a standard PRESS sequence was used with an echo time of 8.8ms, a repetition time of 4000ms and a spectral width of 5000Hz. The water signal from the VOIs was suppressed by variable-power RF pulses with optimized relaxation delays (VAPOR) (Tkac et al. 2004). Each spectrum corresponded to a mean of 512 scans for the dorsal striatum and a mean of 1664 scans for the SNr. A spectrum was acquired in the same conditions without water suppression for the absolute quantification of metabolites.
This article is protected by copyright. All rights reserved.
Accepted Article
Metabolite
concentrations
were
determined
with
jMRUI
software
(version
5.0,
http://www.mrui.uab.es/mrui). The quantification method used was the time-domain semiparametric algorithm QUEST, based on signals for a basis set of simulated metabolites comprising the following molecules: lactate (Lac), N-acetylaspartate (NAA), total creatine (creatine and phosphocreatine; tCr), total choline (glycerophosphocholine and phosphocholine; tCho), Glu, glutamine (Gln), GABA, taurine (Tau) and myo-inositol (Myo-Ins). The basis set also included a simulated signal of macromolecules and lipids (MM). The intensity of the water signal obtained from nonsuppressed water spectra was used as an internal reference. The reliability of metabolite quantification was assessed from the average Cramer–Rao lower bounds (CRLB) calculated by jMRUI. Only results with a CRLB ≤ 30% were included and were expressed in mM, as means±SEM. In vivo microdialysis and measurement of extracellular amino-acids levels The microdialysis probe (CMA/11, Carnegie Medicine, Stockholm, Sweden; microdialysis membrane length, 3mm; molecular weight cut-off, 20kDa; outer diameter, 0.5mm) was lowered through the guide cannula so that the tip of the membrane reached DV=-6.3mm from the dura surface. Probes were perfused with Ringer’s solution (147mM NaCl, 2.5mM CaCl2, 4mM KCl,) at a constant flow rate (1µL⁄min) using a CMA⁄102 microdialysis pump (Carnegie Medicine). The dialysates were collected every 20min in tubes containing 30µL of a 10µM ascorbic acid and 10µM homoserine (as internal sample control) solution, and were immediately frozen at -80°C until HPLC analysis. Sample collection started 120min after the perfusion onset to achieve stable levels of aminoacids. In the first dialysis session, 6 samples were collected for 120min as basal values OFF DBS then DBS was turned ON and 6 more samples were collected as acute DBS. In the second dialysis session, 6 samples were collected as chronic DBS. Analysis was performed on the 3 last samples out of the 6 for each condition (corresponding for the acute condition to the time window of 1 to 2 hours after DBS start). Measurement of aminoacids was performed using HPLC coupled with fluorimetric detection (Waters,
St.
Quentin,
France),
including
a
precolumn
derivatization
with
o-
phthaldialdehyde⁄mercaptoethanol and a C-18 (ODS2, 4.6·150mm) Spherisorb column. A non-linear gradient delivered through a Waters 600 pump was used to separate derivatives (solvent A: 0.1M potassium acetate with 25% methanol, pH 5.5; solvent B: 0.05M potassium acetate and 60% methanol, pH 5.5). Samples (15µL) were automatically injected (Waters 717 plus autosampler) and analysed using a Waters 474 fluorimetric detector. The limit of detection was 1.5pmol⁄sample. Data were computed with WATERS MILLENIUM software (via a Waters bus SAT⁄IN module), compound identification and peak quantification were achieved by comparison with standard aminoacid solution. The average of each aminoacid concentration from the three samples per condition was
This article is protected by copyright. All rights reserved.
Accepted Article
then determined for each animal. Extracellular GABA levels were close to the detection limit of the HPLC measurement, preventing sound assessment of the changes in the different experimental conditions. Data were means±SEM of the values obtained for the n animals of each experimental group. Histological verification of 6-OHDA lesion and electrode placement At the end of MRS or microdialysis experiments, animals were killed by decapitation and brains were directly frozen on dry ice and stored at -80°C. Coronal tissue sections (10µm thick) were cut at -20°C with a cryostat (Leica CM3050, Nussloch, Germany), mounted on SuperFrost Plus glass slides (Fisher Scientific, Elancourt, France) and stored at -80°C until specific treatment. The loss of DA terminals in the striatum was assessed as an index of the DA denervation extent by analysis of 3H-mazindol binding to DA uptake sites, as described previously (Salin et al. 2002; Bacci et al. 2004). The correct location of the stimulating electrode and of the dialysis probe was examined on Cresyl Violet-stained sections. Only animals showing extensive striatal DA denervation (>85%), location of the two wires of the DBS electrode within the STN and location of the dialysis probe within the striatum were selected for data analysis. Around 30% of animals which did not satisfy all the selection criteria have been discarded. The final cohort contained twenty-eight rats divided up into two groups (i) intranigral NaCl-injected rats (sham, n=6 for MRS and n=8 for microdialysis); (ii) intranigral 6-OHDAinjected rats (6-OHDA, n=6 for MRS and n=8 for microdialysis). Statistical analysis Data were analyzed statistically (Statistica, version 7.1; Stat Soft, Maisons-Alfort, France) using a two-way analysis of variance, ANOVA (group: sham, 6-OHDA and DBS conditions as repeated measures: OFF, acute (except cylinder test) and subchronic). When it was significant, ANOVA was followed by Newman-Keuls post-hoc test. Results Forelimb asymmetry In the cylinder test, sham rats OFF DBS equivalently used their right and left forepaws and this behavior was not affected under unilateral subchronic DBS (Fig. 2). 6-OHDA animals assessed OFF DBS showed marked asymmetry with negative score, the number of contralateral contacts being dramatically reduced due to akinesia-like deficit affecting the forepaw contralateral to the lesion. This asymmetry was completely reversed by subchronic DBS of the STN on the lesion side.
This article is protected by copyright. All rights reserved.
Accepted Article
MRS spectroscopy For each animal, we have checked that the stimulation did not induce artifact on NMR acquisitions in vivo and tissue damage in STN after histological control on brain sections (Fig. 3a). Typical in vivo 1H spectra obtained in the striatum are depicted in figure 3b. After shimming, a half-height linewidth of the water signal of 10-12Hz was achieved in the dorsal striatum. Measuring MR spectra in the SNr was challenging due to the small size of this structure, its location in the midbrain and its high iron content. However, despite these challenges in shimming and broad intrinsic linewidths relative to other brain area, spectral quality that allows assessment of the neurochemical profile of the SNr was achieved in 2.9µL volumes (Fig. 3c). The half-height linewidth of the water resonance was 13-15Hz. The majority of the metabolites had average CRLBs lower than 15% in the striatum, with the exception of GABA (18-22%) and Gln (15-18%), and lower than 25% in the SNr. Session-on-session percentage of voxel coverage was in mean 88.9±6.4% in the STR and 80.7±9.4% in the SNr. In the OFF DBS condition, 6-OHDA rats showed significant increases vs sham in striatal levels of Glu, Gln and GABA (Fig. 4). Levels of these metabolites were also increased in the SNr of 6-OHDA rats compared to sham (Fig. 4). DBS, either acute or chronic, did not affect the levels of Glu, Gln, or GABA in the striatum or the SNr of sham animals while having differential impact in 6-OHDA animals. In the striatum, acute DBS normalized Glu, Gln and GABA levels in 6-OHDA rats: the levels were significantly reduced versus 6-OHDA OFF DBS (Glu: -33±9%, p
