Multimodal imaging study of the 5-HT1A receptor biased agonist, NLX-112, in a model of L-DOPA-induced dyskinesia

Graphical abstract


Introduction
Parkinson's disease (PD) is the second most common neurodegenerative disorder after Alzheimer's disease (de Lau and Breteler, 2006); (Lees et al., 2009).The gold standard treatment for PD is the oral administration of the dopamine (DA) precursor, L-3,4-dihydroxyphenylalanine (L-DOPA), which compensates for the progressive reduction of DA release in the projection areas (especially the striatum) (Poewe et al., 2017) due to loss of DAergic neurons in the substantia nigra pars compacta (SNc).
Although L-DOPA is an effective treatment to alleviate motor symptoms, its therapeutic efficacy is gradually lost and its long-term use often leads to development of troublesome abnormal involuntary movements (AIMs) also called L-DOPA-induced dyskinesia (LID) (Lees et al., 1977;Marsden and Parkes, 1977).Around 80% of patients will be affected by LID after 10 years of treatment (Grandas et al., 1999).The management of these side-effects related to L-DOPA protracted therapy is problematic and there is no fully efficacious and well tolerated treatment.
While the pathophysiology of LID is still incompletely understood, altered signaling by DA D 1 receptors (D 1 R) (Bastide et al., 2015), NMDA and AMPA (Bastide et al., 2015) receptors, and striatal serotonergic 5-HT 1A receptors (5-HT 1A R) (Bishop et al., 2009) are probably involved.In particular, the DAergic nigrostriatal denervation seen in PD results in striatal serotonergic neurons taking up and converting L-DOPA into DA through their enzymatic machinery (L-amino acid decarboxylase).Consequently, DA is aberrantly released as a 'false neurotransmitter' by serotonergic neurons without the autoregulatory feedback mechanisms specific to DAergic neurons (Carta and Bezard, 2011;Cenci, 2014).In this context, targeting inhibitory serotonergic autoreceptors, such as 5-HT 1A R located on raphe nuclei, appears to be a promising strategy to reduce swings in synaptic levels of DA, and consequently decrease LID (Huot et al., 2017;Ohno et al., 2013).
Initial [ 18 F]FDG PET imaging studies in naïve rats showed that NLX-112 modifies glucose metabolism in the cerebellum, striatum, brainstem and raphe nuclei (Levigoureux et al., 2019).In fMRI experiments, NLX-112 also elicited functional connectivity effects in regions expressing pre-synaptic 5-HT 1A R (Vidal et al., 2020).Changes in the functional state of 5-HT 1A R were also observed in HPK rats, especially when treated repeatedly with L-DOPA to elicit dyskinesia (Vidal et al., 2021).
The aim of this study was to define for the first time the brain activation and functional connectivity profiles in rats presenting LID, based respectively on glucose metabolism changes ([ 18 F]FDG PET) and temporal correlations of BOLD signal fluctuations (fMRI imaging), and to visualize the different neural circuits affected by NLX-112 in HPK rats with LID (HPK-LID) or without LID (HPK-non-LID rats).

Animals
Seventy-nine adult male Sprague-Dawley rats (Charles River laboratories, France) weighing 200-250 g were used in the present study.Animals were initially housed six per cage in standard temperature (22 ± 2 • C) and humidity (50%) conditions with a 12 h/12 h light/dark cycle, light on at 8.00 am; food and water were provided ad libitum.All experiments were performed in accordance with the European guidelines for care of laboratory animals (directive 2010/63/EU) and approved by the University of Lyon review board (CELYNE).

Unilateral 6-OHDA lesions in the medial forebrain bundle
Prior to surgery, buprenorphine (0.05 mg/kg, s.c.) and lidocaine cream were used for analgesia.Anesthesia induction was performed with a mixture of 4% isoflurane and air (1 L/min), lowered to 2% isoflurane during the stereotaxic surgical procedure.Rats were lesioned in the right medial forebrain bundle (MFB) by infusing 6-OHDA (14 µg/4 µL, 1 µL/min) dissolved in saline containing ascorbic acid (0.02%), using the following coordinates from the Paxinos and Watson's (Iderberg et al., 2015;Paxinos and Watson, 2014) atlas: AP = − 4.0 mm and ML = − 1.3 mm, relative to bregma, and DV = -7.6 mm from the dural surface.Following the procedure, rats were single-housed for 24 h, before being housed 6 per cage again for the rest of the experiment.After 3 weeks, rats were challenged with apomorphine (0.05 mg/kg, s.c.) and turning behavior assessed over 30 min post-injection, in order to verify a striatal DAergic depletion of at least 90% (Hudson et al., 1993).Thirtyone rats were selected and used in subsequent neuroimaging experiments and AIMS scoring for 15 of them (see supplementary figure for details).

Autoradiography
After selection with the apomorphine test, an additional cohort of HPK rats (n = 7) was used for validation of the unilateral dopaminergic lesion by autoradiography of the dopamine transporter (DAT).After a short inhalation of isoflurane, rats were decapitated and brains were carefully removed and immediately frozen in 2-methylbutane cooled with dry ice (− 29 • C).Coronal sections (30 μm thick) were cut using a − 20 • C cryostat, thaw-mounted on glass slides, and allowed to air dry before storage at − 80 • C until used.The day of [ 11 C]PE2I synthesis, the slides were incubated for 20 min in Tris phosphate-buffered saline buffer, pH adjusted to 7.5, containing 37 kBq/ml of [ 11 C]PE2I.Nonspecific binding was determined in duplicate serial sections coincubated with GBR12909 10 µM (Sigma-Aldrich).After incubation, the slides were dipped in cold buffer and distilled water (4 • C) and then dried and juxtaposed to a phosphor imaging plate for 120 min (BAS-1800 II; Fujifilm).The striatum was drawn manually using Multigauge software (Fujifilm).The specific binding was determined from the subtraction of the non-specific binding from the total binding in the duplicate sections, and results were expressed in optical densities (PSL/ mm 2 ).

L-DOPA-induced dyskinesia
Rats of the HPK-LID cohort received chronic daily treatment with a combination of L-DOPA and benserazide (6/12 mg/kg s.c., administered as a single injection) for 3 weeks to induce stable reproducible AIMs.

PET imaging procedures and data analyses 2.6.1. PET/CT protocol and image acquisition
All 15 rats that were tested in the AIMs scoring experiment were scanned 3 weeks after the apomorphine-induced rotational test.Each animal underwent 4 different PET scans with [ 18 F]FDG in a randomized order i.e., one scan for each of the 4 treatments mentioned above for the AIMs scoring (on Day 43, 45, 50 & 52 post-6-OHDA lesion).Rats were fasted for 4 h before each scan.Fifteen minutes after i.p. injection of L-DOPA,NLX-112, L-DOPA + NLX-112 or saline, [ 18 F]FDG (36.4 ± 3.1 kBq/g) was injected under awake conditions.Thirty min later, rats were anesthetized with constant insufflation of isoflurane (4% for induction then 2% during acquisition, 1 L/min) and placed in the PET-CT imaging scanner (INVEON®, Siemens).PET scans were acquired in list mode during 30 min, completed by a computed tomography (CT) scan.Static images of the 30 min acquisitions were reconstructed with attenuation and scatter corrections.

PET data analyses
Data acquisition and processing was carried out using INVEON Research Workplace (IRW®, Siemens) software and the statistical parametric mapping software SPM12®.Images were analyzed in two steps, i.e., a voxel-based analysis and a regions of interest (ROIs) analysis.

Voxel-based analyses (VBA)
Individual PET images were realigned and spatially normalized based on our own previously built CT template, co-registered on an anatomical MRI template.Each PET volume with voxel size 0.1 × 0.1x 0.2 mm 3 was then spatially normalized and smoothed using an isotropic Gaussian filter [1 × 1 × 1 mm 3 ].Local voxel activities were normalized to the mean uptake of the whole brain taken individually as a volume of interest (VOI), in order to minimize potential confounding factors such as changes in peripheral glucose metabolism.Indeed, the dyskinetic behavior can indirectly decrease the global brain uptake by increasing the uptake in the muscles and confirmed by SUV measurements (see supplementary Figs. 2 and 3).For each group (HPK-non-LID and HPK-LID separately), a voxel-to-voxel statistical analysis was performed, using a variance analysis in a generalized linear model, to compare mean uptakes between the different conditions.A significant threshold was set up at p < 0.001 uncorrected, at the voxel level.A second analysis was performed in order to compare the metabolic profiles of HPK-LID and HPK-non-LID rats under saline condition by using an unpaired Student's t-test, in order to assess the impact of dyskinesia (chronic administration of L-DOPA) on the absorption ratios of [ 18 F]FDG.

Regions of interest (ROIs) analyses
Brain ROIs were already defined in the Lancelot rat brain atlas on the corresponding MRI template (Lancelot et al., 2014).Mean uptake ratios were extracted in the different ROIs, and a two-way mixed ANOVA for repeated measures (factors: treatment and ROI) with Tukey's post-hoc tests (corrected for multiple comparisons) were performed using GraphPad Prism 8.0 software to compare the [ 18 F]FDG uptake ratios between the different conditions.

fMRI protocol and image acquisition
Out of the 31 rats that were retained following the apomorphine challenge (see above), the remaining 16 (8 HPK-non-LID and 8 HPK-LID) rats were subjected to an fMRI study, 3 weeks after the apomorphine-induced rotational test.Anesthesia induction was performed with 4% isoflurane (1 L/min), then lowered to 2% isoflurane during the catheterization procedure and fMRI acquisition.Saline, NLX-112, L-DOPA or L-DOPA + NLX-112 (same doses as for the PET/scan experiments) were injected i.p. over one minute during the fMRI scan.The fMRI protocol was carried out on a 7-Tesla Bruker Biospec MR system (Bruker Bio-spinGbmH, Germany).A 2D anatomical T2-RARE (Rapid Acquisition with Relaxation Enhancement) image was obtained with 14 contiguous slices of 1.5-mm thickness covering the whole rat brain.To measure BOLD variations, a T2* (Echo Planar Imaging) sequence was used and fourteen slices acquired with the identical geometry of the anatomical T2-RARE scan.Each fMRI session lasted 1 h, corresponding to repetitions.Drugs were injected 15 min after the beginning of the session, resulting in 45 min of post-injection period (900 repetitions).

fMRI data analyses
Data were preprocessed using the SPM12® software.Images were realigned and spatial normalization was performed using the standardized MRI template (Lancelot et al., 2014).For each acquisition, the time courses of the BOLD signal were extracted from different regions of interest (ROIs) on the 900 post-injection images.These regions were automatically delineated using the Lancelot rat brain atlas (Lancelot et al., 2014) defined in the MRI template space (Lancelot et al., 2014).Temporal correlations were evaluated by the Pearson correlation coefficient between the BOLD time-courses of each pair of ROIs over the min after injection.The correlation matrix displaying the correlation coefficients for every pair of regions was generated using the Graphpad Prism 8.0 software.Matrices were averaged to obtain mean correlation matrices for each group.For statistical analysis, correlation coefficients of each acquisition were converted into Z-scores by Fisher transformation.For HPK-non-LID and HPK-LID rats, each drug condition was compared to the saline condition using a two-way ANOVA followed by Dunnett's multiple comparisons post-hoc tests using the Graphpad Prism 8.0 software.

Results
[ 18 F]FDG uptake ratios were compared between the different groups and drug treatments.They are illustrated in Fig. 1 that shows a comparison of the averaged images of [ 18 F]FDG uptake ratio after administration of saline, L-DOPA or L-DOPA + NLX-112 (n = 8 per condition).

Effects of acute challenge with L-DOPA and NLX-112 on AIMs in HPK-LID and HPK-non-LID rats
Upon administration of an acute i.p. challenge with L-DOPA/benserazide, HPK-LID rats develop high AIMs scores (average ± SEM: 2.71 ± 0.18), i.e. presented orolingual and moderate to strong limb & axial dyskinesia.By contrast, HPK-non-LID rats exhibit low AIMs score (1.33 ± 0.21), i.e., presented mainly a normal behavior or some orolingual dyskinesia.Co-administration of NLX-112 (at 0.16 mg/kg i.p., a dose that produces no dyskinesia by itself) fully reversed the AIMs elicited by L-DOPA in both cohorts of rats.

PET imaging
Images of voxel-based analyses of [ 18 F]FDG uptake ratios in HPKnon-LID and HPK-LID rats following an acute saline injection are shown in Fig. 2. The metabolic profiles of HPK-non-LID and HPK-LID rats are identical in terms of [ 18 F]FDG uptake ratios, except for hypermetabolism in the mesencephalic locomotor region on the contralateral lesion side of HPK-LID rats.

Effects of L-DOPA on imaging profiles of HPK-non-LID and HPK-LID
Fig. 3A represents the voxel-to-voxel statistical comparisons of [ 18 F] FDG uptake ratios between L-DOPA/benserazide and saline injections in HPK-non-LID and HPK-LID rats.The metabolic pattern of activation in HPK-LID rats is more pronounced than that of HPK-non-LID rats after L-DOPA acute administration.Indeed, significant activations were detected in the cerebellum, brainstem, raphe area and thalamus, on the unilateral side of the lesion in HPK-LID rats.Significant clusters of hypometabolism were found in the hippocampus, the ipsilateral striatum and in the cortex, especially the prefrontal cortex.
Concerning [ 18 F]FDG uptake ratio in HPK-LID rats, the results are also representative of the images of voxel-based analysis.A decrease was detected in the right (-6.6%) and in some cortical regions compared to saline injection in HPK-LID rats.The results were similar concerning HPK-non-LID rats.A decrease was observed in the left striatum (-5.3%) and in some cortical regions: − 8.6% and − 6.0% for right and left cingulate cortex and − 5.4% and − 8.7% for right and left temporal cortex.

fMRI imaging functional connectivity
The development of LID produced changes in functional connectivity following acute administration of saline (Fig. 4A) and acute administration of L-DOPA (Fig. 4B and Fig. 4C).There was a statistically significant effect of the pairs of regions and the interaction between the pairs of regions and the groups on the Z-score.Post-hoc tests revealed increases of connectivity in HPK-LID rats compared to HPK-non-LID rats between the brainstem and the parietal cortex and between the parietal cortex of the right and left hemispheres.In contrast, a decrease between the left cingulate cortex (non-lesioned side) and the right cerebellum (lesioned side) was observed.
Functional connectivity was also compared between the L-DOPA and the saline injections in each cohort of rats using two-way ANOVAs.In HPK-non-LID rats, there was a statistically significant effect of the pairs of regions, the drug conditions and their interaction on the Z-scores.Post-hoc tests revealed that following acute administration of L-DOPA in HPK-non-LID rats (Fig. 4B), a strong and specific loss of connectivity was found between the left occipital cortex (non-lesioned side) and many other brain structures (thalamus, hippocampus, temporal cortex, left amygdala, brainstem, and raphe).When compared with saline treatment, there was also a disconnection between the lesioned and non-    lesioned side in the occipital cortex.Concurrently, other significant decreases of connectivity were found between the left cingulate cortex and the right cerebellum.A connectivity increase between the right frontal cortex and left cingulate cortex was also observed.
In HPK-LID rats, there was a statistically significant effect of the pairs of regions, the drug conditions but not their interaction on the Z-scores.The post-hoc tests showed that when comparing L-DOPA and saline (Fig. 4C), the alteration of the connectivity of the left occipital cortex was no longer observed in that group.On the other hand, increases of connectivity were found between both hemispheres of the frontal cortex and the left striatum and between the right striatum and the left frontal cortex, as well as between the left parietal cortex and the right striatum (lesioned side).A significant increase of connectivity between the right and left sides of the parietal and the frontal cortices was also observed.

PET imaging
Images of voxel-based analysis of the effects of NLX-112 on glucose metabolism in HPK-LID animals are shown in Fig. 5A.By itself, NLX-112 produced an increase in glucose metabolism in cortical areas and a decrease in motor regions (thalamus, cerebellum, brainstem) and raphe nuclei.Co-administration of L-DOPA and NLX-112 partially erased the cerebral effects of L-DOPA in some, but not all, brain regions (Fig. 5A).In particular, voxel-to-voxel comparisons suggested that NLX-112 reverses the L-DOPA-induced hypometabolism observed in cortical regions and reduces the L-DOPA-induced hypermetabolism seen in brainstem and cerebellum (Fig. 5A).These effects were echoed in the uptake ratios (Fig. 5B) which also found a robust normalization by NLX-112 of [ 18 F]FDG uptake in raphe nuclei, a main target region underlying NLX-112′s therapeutic activity.

Functional connectivity
NLX-112 at a therapeutic dose (0.16 mg/kg) produced numerous changes in functional connectivity in HPK-LID rats (Fig. 6).Post-hoc tests following two-way ANOVAs in HPK-LID rats showed that NLX-112 alone (Fig. 6A) produced connectivity increases between the right amygdala and the right hippocampus), between both sides of the cingulate cortex and between the left and right striatum and the right frontal cortex.The co-administration of NLX-112 with L-DOPA (Fig. 6B) prevented the functional hyper-connectivity at the corticostriatal level that was observed with L-DOPA alone.Significant increases of connectivity were found between the right amygdala and brainstem, right cingulate cortex and left and right hippocampus.

Effects of NLX-112 on L-DOPA-induced activity in non-LID rats
The metabolic and functional connectivity changes occurring after NLX-112 injection in non-LID animals were also evaluated (Supplementary Fig. 4).NLX-112 alone produced few effects on [ 18 F]FDG uptake, but prevented the frontal decrease and attenuated the cerebellar increase induced by L-DOPA.NLX-112 alone also produced few effects on the functional connectivity in non-LID rats, but co-administration with L-DOPA was followed by numerous decreases of connectivity between various subcortical regions and the occipital cortex, as observed with L-DOPA alone, accompanied by a bilateral increase of connectivity between the frontal cortex, the cingulate cortex and the hippocampus.

Validity of the rat HPK-LID model
The present neuroimaging study describes, for the first-time, specific brain metabolic patterns of LID produced by an acute injection of L-DOPA and/or the 5-HT 1A R biased agonist, NLX-112, in vigil HPK rats with unilateral 6-OHDA lesions of the substantia nigra.This animal model of PD has the benefit of being robust, reproducible and widely described, with a lesion success rate of at least 50% (Hudson et al., 1993) and a good tolerance to the unilateral lesion (Schwarting and Huston, 1996).The HPK-LID rat model presents similar characteristics to "peak dose dyskinesia" (good face validity) and responds to antidyskinetic treatments (good predictive validity).About 80% of rodents with a DAergic lesion greater than 90% rapidly develop dyskinesia (Cenci and Lundblad, 2007).Indeed, with this methodology we found a massive unilateral decrease of the dopamine transporter binding in an additional cohort of rats (Supplementary Fig. 1).

[ 18 F]FDG PET imaging in HPK-LID rats
Changes in glucose uptake ratios revealed by [ 18 F]FDG PET imaging reflect cerebral activity dynamics (Wehrl et al., 2013).Thus, in the basal state (i.e., after saline injection), HPK-non-LID and HPK-LID rats had generally similar metabolic profiles (Fig. 2), but clear differences in brain activity were observed after acute L-DOPA challenge (Fig. 3A), which confirms the "peak-dose" aspect of the dyskinesia obtained in HPK-LID rats (Cenci and Crossman, 2018;de la Fuente-Fernández et al., 2004).The metabolic pattern elicited by L-DOPA in HPK-LID rats was characterized by decreased glucose metabolism in the cingulate and temporal cortex and the ipsilateral striatum.In contrast, glucose hypermetabolism was found in the dorsal raphe nuclei, cerebellum, brainstem, regions that are involved in motor control and influence basal ganglia activity (Bezard et al., 2001).
Interestingly, the regions mentioned above are similar to those affected in [ 18 F]FDG PET studies of PD patients and are referred to as the "PD-related pattern" (PDRP), with a hypermetabolism in the cerebellum, pons, thalamus and lentiform nucleus and a hypometabolism in premotor and parietal cortex (Eidelberg et al., 1994;Ma et al., 2007).In contrast, parkinsonian tremor seems to have its own pattern in humans, with an increase in metabolic activity in the cerebellum, pons, primary motor cortex, caudate nucleus and putamen (Mure et al., 2011).
In our HPK-LID rat model, we only found hypoactivity in the prefrontal cortex and the ipsilateral striatum.The striatal hypoactivity may be due to the effect of endogenous DA on inhibitory D 2 receptors, since it possesses higher affinity for them than for activatory D 1 receptors (Marcellino et al., 2012).Furthermore, the activating effects of L-DOPA in the raphe nuclei point to the involvement of the serotonergic system in LID.This is consistent with the role of serotonergic neurons in converting L-DOPA to DA and releasing it as a "false neurotransmitter" (Carta and Bezard, 2011;Cenci, 2014).

Functional connectivity in HPK-LID rats
The functional connectivity experiments provide further evidence of the alteration of basal ganglia circuitry in PD (Perlbarg et al., 2018) and in the development of LID (Wichmann and Delong, 2006) and are consistent with functional connectivity studies carried out in humans.In particular, aberrant cortico-striatal changes in connectivity are found between cortical areas and the putamen (Herz et al., 2016;Tessitore et al., 2019).In addition, a correlation is observed between the connectivity changes in the prefrontal cortex and the occurrence of AIMs (Cerasa et al., 2016(Cerasa et al., , 2013)).A major involvement of the cerebellum is also found in human with an increase in connectivity between it and other regions (Mueller et al., 2019;Tessitore et al., 2019).
Interestingly, a marked decrease of connectivity between other cerebral structures and the left occipital cortex was found in HPK-non-LID rats with acute L-DOPA administration.This disconnection was not found in dyskinetic rats and may be correlated with the known occipital hypoperfusion in treated parkinsonian patients (Abe et al., 2003).Another study reported an alteration of fMRI resting-state measurements in the occipital cortex which was correlated with decreased DAuptake by the DA transporter in the caudate, showing the link between this region and the striatal dopaminergic system (Oh et al., 2020).

NLX-112 as a promising anti-dyskinetic drug for LID in Parkinson's disease?
Beyond the imaging phenotyping of the HPK-LID model, the second objective of this study was to investigate the central effects of NLX-112, a selective 5-HT 1A R biased agonist.Administration of NLX-112 completely prevented L-DOPA-induced AIMs in HPK-LID rats, confirming the antidyskinetic activity of the molecule (Depoortere et al., 2020;Fisher et al., 2020;Iderberg et al., 2015;McCreary et al., 2016).The behavioral effects of NLX-112 were accompanied by a partial reversal of the metabolic changes induced by L-DOPA: thus, NLX-112 restored cortical metabolism and completely reversed the hypermetabolism caused by L-DOPA in the raphe nuclei, reducing it in the mesencephalic motor region, brainstem, and cerebellum.The metabolic profile observed upon combined NLX-112 and L-DOPA administration did not, however, result in a complete return to a basal state and the remaining changes may be due to the intrinsic 5-HT 1A agonist activity of NLX-112, as characterized previously in non-lesioned rats (Vidal et al., 2020).
Furthermore, administration of NLX-112 corrected the functional connectivity profile in HPK-LID rats treated with L-DOPA.Taking these observations together with previous behavioral data in 6-OHDAlesioned rats (Iderberg et al., 2015), it would appear that the main mechanism for the antidyskinetic effects of NLX-112 is a preferential stimulation of somatodendritic 5-HT 1A autoreceptors in the raphe nuclei.This activation decreases the activity of the ascending serotonergic system and inhibits the excessive release of DA by 5-HT terminals (Iderberg et al., 2015;Lladó-Pelfort et al., 2012;Muñoz et al., 2008;Ohno et al., 2013).

Conclusion
This study defined for the first time the brain activation and inhibition patterns in functional [ 18 F]FDG PET imaging and the functional connectivity profiles of HPK rats presenting dyskinesia after chronic administration of L-DOPA.These data enrich the limited knowledge available concerning the functional imaging of LID.Moreover, the present observations further reinforce the involvement of the serotonin system in LID (Huot et al., 2017) and support targeting 5-HT 1A R as a therapeutic strategy for LID treatment (Huot, 2018;Muñoz et al., 2008;Ohno et al., 2013).Finally, these preclinical results obtained by translational imaging approaches are directly transferable to human studies and inform ongoing clinical development of NLX-112 in Parkinson's disease patients suffering from LID.

Authors' Roles
Sarah Chaib carried out the experiments with PET imaging, performed data analysis, participated in the study design, and wrote the manuscript.
Benjamin Vidal carried out the experiments with fMRI imaging and autoradiography, performed data analysis, participated in the study design and reviewed the manuscript.
Caroline Bouillot carried out the experiments with PET imaging.Ronan Depoortere participated in the behavioral analysis of the study and reviewed the manuscript.
Adrian Newman-Tancredi provided NLX-112 and reviewed the manuscript.
Luc Zimmer initiated the study, participated in its design and reviewed the manuscript.
Elise Levigoureux coordinated the study, carried out the experiments, participated in its design and reviewed the manuscript.
All authors have read and approved the final manuscript.

Funding sources for study
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Financial disclosures
Dr. Adrian Newman-Tancredi and Dr Ronan Depoortere are employees and stockholders of Neurolixis.The other authors report no conflict of interest and have nothing to disclose.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 2 .
Fig. 2. Comparison of metabolic profiles using [ 18 F]FDG PET scans between HPK-non-LID and HPK-LID rats following an acute i.p. saline injection.Voxel-to-voxel statistical comparisons of [ 18 F]FDG uptake ratio between saline injections in HPK-non-LID (n = 7) and HPK-LID (n = 8).T scores are represented in color scales (significant increases from red to light yellow, significant decreases from dark to light blue; p < 0.01).Left panel: significant increase of glucose metabolism in HPK-LID rats compared to HPK-non-LID rats; Right panel: decrease of glucose metabolism in HPK-LID rats compared to HPK-non-LID rats.HPK: hemi-parkinsonian, LID: L-DOPA-induced dyskinesia, NL = non-lesioned side, L = lesioned side.Coronal sections are from + 4 to − 13 mm with respect to Bregma.

Fig. 4 .
Fig.4.Functional connectivity analysis using fMRI between HPK-non-LID (n = 7) and HPK-LID rats (n = 8) following a saline i.p. catheter infusion.A-Analysis following injection of saline between HPK-LID rats and HPK-non-LID rats.B-Analysis between L-DOPA and saline injections in HPK-non-LID rats.C-Analysis between L-DOPA and saline injections in HPK-LID rats.Schematic representations (on the left) of the significant changes in functional connectivity (p < 0.001) as measured by comparing the correlation coefficient after Fisher transformation of r values into z values.The dashed arrows represent the significant decreases of functional connectivity between two regions and the solid arrows represent the significant increases of functional connectivity between two regions, as compared to the saline.The correlation matrices (on the right) are expressed as mean values of Pearson correlation coefficient.Positive values of correlation coefficient are shown in red and negative values of correlation coefficient are shown in blue.For statistical comparisons, correlation coefficients were transformed into Z-scores using Fisher transformation before two-way ANOVA, followed by Sidak (in A) or Dunnett (in B and C) multiple comparisons post-hoc test (*p < 0.05, ** p < 0.01, **** p < 0.0001, HPK-LID vs HPK-non-LID (in A), or L-DOPA versus saline in HPK-non-LID (in B) or in HPK-LID (in C)).

Fig. 6 .
Fig. 6.Effects of administration of NLX-112 at 0.16 mg/kg on functional connectivity using fMRI in HPK-LID rats.A -Functional connectivity analysis between NLX-112 alone and saline i.p. catheter infusions in HPK-LID rats (n = 8) and B -between co-administration of L-DOPA + NLX-112 injections and saline in HPK-LID rats (n = 8).Schematic representations (on the left) of the significant changes in functional connectivity (p < 0.001) as measured by comparing the correlation coefficient after Fisher transformation of r values into z values.The dashed arrows represent the significant decreases of functional connectivity between two regions and the solid arrows represent the significant increases of functional connectivity between two regions, as compared to the saline.The correlation matrices (on the right) are expressed as mean values of Pearson correlation coefficient.Positive values of correlation coefficient are shown in red and negative values of correlation coefficient are shown in blue.For statistical comparisons, correlation coefficients were transformed into Z-scores using Fisher transformation before two-way ANOVA followed by Dunnett's multiple comparisons test (*p < 0.05, ** p < 0.01, **** p < 0.0001).