Introduction

Parkinson’s disease (PD) is characterized by the degeneration of several neuronal ensembles, including dopaminergic neurons in the substantia nigra pars compacta (SNc), and by the occurrence of cytoplasmic inclusions containing misfolded alpha-synuclein among other aggregated components. In addition to the cardinal motor symptoms of the disease, several non-motor symptoms that are often underdiagnosed are frequently observed at disease onset. These include apathy, anhedonia and subtle executive dysfunction, as evidenced by attentional deficits, memory complaints and decreased cognitive flexibility1,2,3. Some of these cognitive impairments have been associated with reduced [18F] fluorodopa uptake in the caudate nucleus and frontal cortex4, reflecting the ongoing loss of dopaminergic innervation of the medial prefrontal and orbitofrontal cortices arising from the ventral tegmental area (VTA) and to a lesser extent from the SNc5,6.

The consequences of dopaminergic denervation on the functioning and information integration in basal ganglia motor circuits have been largely investigated in animal models of PD and have revealed that loss of dopaminergic innervation leads to altered activity in both the motor cortex and dorsolateral striatum. Loss of midbrain dopaminergic neurons leads to decreased excitability7 as well as reduced firing rate of pyramidal neurons in the primary motor cortex in rats8 and non-human primates9, although such effects seem to be influenced by the type of dopamine depletion (acute vs chronic) and species used to model PD10. At the striatal level, dopamine depletion leads to synaptic remodeling, together with altered morphology and activity of striatal output neurons11. Furthermore, cortico-striatal transmission is also markedly altered following nigrostriatal dopamine depletion, as evidenced by impairments of Hebbian plasticity (long-term potentiation and long-term depression) (for review, see ref. 12). Whether these cortical and striatal functional alterations demonstrated in motor circuits similarly operate in associative domains of the basal ganglia remain currently poorly known.

The introduction of dopamine replacement therapy (DRT) alleviates most of the motor deficits and normalizes to some extent the abnormal activity of cortical and subcortical nuclei but its long-term use leads to the development of maladaptive forms of synaptic plasticity that underlie the expression of motor side effects such as dyskinesia or non-motor side effects, such as impulse control disorders13,14,15. Based on the functional organization of the basal ganglia, both motor and non-motor side effects of dopaminergic treatments have been theorized to represent a pathological continuum of abnormal dopaminergic stimulation, with possible shared underlying molecular and cellular mechanisms15.

In this study, we aimed to determine if circuit dysfunctions described in motor cortico-striatal loops following nigrostriatal degeneration and DRT find their counterparts in associative frontostriatal loops. To this end, we used a rat model of PD based on viral-mediated expression of alpha-synuclein in the SNc to investigate the consequences of the degeneration of nigral dopaminergic neurons and of a 10-day treatment with the D3/D2 agonist pramipexole (PPX) on neuronal activity and plasticity in frontostriatal circuits. To this end, we recorded the spontaneous activity of orbitofrontal cortex (OFC) neurons and evaluated frontostriatal plasticity, as assessed by recording the response of neurons in the dorsomedial striatum (DMS) to OFC stimulations, before and after a high-frequency stimulation protocol known to induce a long-term potentiation16.

Results

Effect of alpha-synuclein and PPX treatment on nigrostriatal dopaminergic neurons and dopamine receptors expression

We used TH immunohistochemistry and Nissl staining coupled to stereological quantification to assess the level of dopaminergic cell loss in each experimental group (Fig. 1A). Stereological counts of TH-positive neurons showed that viral-mediated expression of alpha-synuclein induced a loss of neurons in the SNc (F1, 72 = 296.1, P < 0.0001). Post hoc analysis showed a significant loss of dopaminergic neurons both in saline (P < 0.0001) and PPX (P < 0.0001) treated alpha-SYN rats compared to saline or PPX-treated GFP rats (Fig. 1B). Total counts in the SNc (Nissl and TH-stained neurons, Fig. 1C) confirmed that the reduced number of TH-positive neurons in the SNc is due to neuronal loss and not to a loss of TH expression (F1, 72 = 203.6, P < 0.0001). This result is in agreement with recent studies using this AAV model17,18. Quantification of striatal TH immunofluorescence (Fig. 1D, E) demonstrated that the loss of TH neurons was associated with a significant reduction of striatal dopaminergic innervation (F1, 49 = 63.52, P < 0.0001) both in saline and PPX-treated alpha-SYN rats (P < 0.0001 for both groups).

Fig. 1: Nigrostriatal degeneration following viral-mediated expression of alpha-synuclein in the SNc.
figure 1

A Representative images of the SNc in sham (top) and lesioned rats (bottom). B Stereological counts of tyrosine hydroxylase (TH)-positive neurons in the SNc. C Stereological counts of tyrosine TH- positive plus Nissl-stained neurons in the SNc. GFP SALINE (N = 17); GFP PPX (N = 20); alpha-SYN SALINE (N = 19); alpha-SYN PPX (N = 20). D Representative images of striatal TH immunofluorescence in sham and lesioned rats. Scale bar = 1 mm. E Quantification of fluorescence intensity in rats used for electrophysiology experiments. GFP SALINE (N = 12); GFP PPX (N = 14); alpha-SYN SALINE (N = 14); alpha-SYN PPX (N = 13). ***P < 0.001 and ****P < 0.0001; two-way ANOVA followed by Tukey’s post hoc.

Full size image

Since impairment of dopamine homeostasis is known to impact dopamine receptors expression in the striatum, we assessed D1, D2, and D3 mRNA expression following surgery and PPX treatment using RNAscope in situ hybridization (Fig. 2A). PPX treatment nor DA neurons degeneration affected dopamine receptors mRNA levels in the DMS (D1 lesion: F1, 12 = 0.13, P = 0.72; PPX: F1, 12 = 0.42, P = 0.52; D2 lesion: F1, 12 = 0.36, P = 0.55; PPX: F1, 12 = 0.23, P = 0.64; D3: lesion: F1, 14 = 0.53, P = 0.48; PPX: F1, 14 = 0.59, P = 0.45) (Fig. 2B). These results indicate that dopaminergic denervation was successfully induced in the SNc by local viral-mediated expression of alpha-synuclein and that this dopamine cell loss and PPX treatment had no impact on striatal D1, D2, and D3 mRNA expression.

Fig. 2: Effect of nigrostriatal degeneration and PPX on D1, D2 and D3 receptors expression in the DMS.
figure 2

A RNAscope in situ hybridization for D1 (top panel), D2 (middle panel) and D3 mRNA (bottom panel) at ×10 and ×40 magnifications. Scale bar = 20 µm. B Fluorescence intensity of D1 (left panel), D2 (middle panel), and D3 mRNA (right panel) in the DMS. GFP SALINE (N = 4); GFP PPX (N = 4); alpha-SYN SALINE (D1, D2: N = 4, D3: N = 5); alpha-SYN PPX (D1, D2: N = 4, D3: N = 5).

Full size image

OFC neuronal activity was impacted by dopaminergic lesion and PPX treatment

We investigated the effects of both dopamine denervation and PPX treatment on OFC function by recording OFC neurons using juxtacellular electrophysiology in isoflurane-anesthetized rats (Fig. 3A). To assess the density of spontaneously active neurons, we performed a 3D screen of the OFC (1 mm × 1 mm × 1 mm). As this strategy does not allow for a systematic identification by neurobiotin injection, we thus defined inclusion criteria for neurons as putative pyramidal neurons based on spike duration (Fig. 3B, C). Out of 737 recorded neurons, 668 neurons whose spike duration was comprised between 0.5 and 1.4 ms were included in the study as putative pyramidal neurons. In average, one spontaneously firing neuron was recorded per mm in all groups (Fig. 3D), indicating that the overall neuronal activity was not impaired by alpha-synuclein-induced dopaminergic loss nor by 10-day PPX treatment (lesion: F1, 25 = 0.036, P = 0.85; PPX: F1, 25 = 0.002, P = 0.96).

Fig. 3: Impact of nigrostriatal degeneration and PPX on neuronal activity in the OFC.
figure 3

A Schematic representation of OFC recording site and example of a recording track (arrowhead). B Representative electrophysiological traces of OFC neurons in the four experimental groups. Insets represent the action potential shape (averaged over 2 min recording), the action potential duration is measured from rising up to the hyperpolarization peak (gray window). C Distribution of spike durations (ms), the dashed lines indicate the limits of inclusion criteria for putative pyramidal neurons. D Density of spontaneously active putative pyramidal neurons per mm. E Spike frequency (Hz). F Proportion of neurons exhibiting burst firing activity (%). G Bursting frequency (Bursts per minute). H Burst duration (seconds). I Percentage of spikes in burst. GFP SALINE (n = 157 neurons; N = 6 rats); GFP PPX (n = 162; N = 8); alpha-SYN SALINE (n = 191; N = 8); alpha-SYN PPX (n = 158; N = 7). *P < 0.05; **P < 0.01; two-way ANOVA followed by Tukey’s post hoc.

Full size image

We then analyzed in more details spike firing patterns by measuring average firing frequency (Fig. 3E) and burst firing properties (Fig. 3F–I). Spike firing frequency was increased by PPX treatment (F1, 25 = 5.09, P < 0.05), but no effect of the lesion nor interaction between lesion and treatment was found, indicating that the dopaminergic denervation had no effect on OFC firing frequency. However, there was a significant effect of alpha-synuclein-induced dopaminergic lesion on the proportion of neurons exhibiting burst firing (F1, 25 = 6.39, P < 0.05), with a decreased proportion of bursty OFC neurons in saline-treated alpha-SYN compared to GFP rats (P < 0.01) and this effect was rescued by 10-day PPX treatment (Fig. 3F). Mean frequency in bursts, burst duration and percentage of spikes in burst were not altered by the lesion nor by PPX treatment (Fig. 3G–I). These results indicate that alpha-synuclein-induced nigral degeneration had no effect on the overall spontaneous activity of OFC neurons, but impacted their firing pattern, as shown by a decreased bursting activity. Interestingly, this effect was rescued by 10-day PPX treatment.

Frontostriatal plasticity is dramatically impacted by nigrostriatal dopaminergic neurodegeneration and PPX treatment

HFS was applied in the OFC to induce plastic changes in the connection strength between OFC and DMS neurons (Fig. 4A, B). Juxtacellular recordings from striatal neurons were performed before and up to 45 min after HFS. Three-way ANOVA indicated a main effect of time (F4.783, 95.66 = 22.35, P < 0.0001), treatment (F1, 20 = 23.55, P < 0.0001), lesion (F1, 20 = 23.38, P < 0.0001) and time × treatment × lesion interaction (F11, 220 = 8.563, P < 0.0001) (Fig. 4C). As expected16, HFS induced an increase in spike probability in saline-treated GFP rats (GFP SALINE pre-HFS vs GFP SALINE post-HFS, P < 0.05) while a decrease of DMS response to OFC stimulation was observed following dopaminergic lesion (alpha-SYN SALINE pre-HFS vs alpha-SYN SALINE post-HFS, P < 0.0005) and this was not rescued by PPX treatment (alpha-SYN PPX pre-HFS vs alpha-SYN PPX post-HFS, P < 0.001). PPX treatment alone also resulted in a long-term decrease of spike probability following HFS (GFP PPX pre-HFS vs GFP PPX post-HFS, P < 0.0001). The OFC-DMS response probability was increased only in control rats, no difference was observed between lesioned, PPX-treated or lesioned and PPX-treated rats. These results indicate that both dopaminergic lesion and 10-day PPX treatment induced a reversal of plastic changes measured following HFS compared to control rats and that PPX treatment failed to rescue the effect of dopamine denervation on this parameter.

Fig. 4: Nigrostriatal degeneragtion and PPX impair frontostriatal plasticity.
figure 4

A Schematic representation of OFC stimulation and DMS recording sites with examples of stimulation and recording tracks (arrowhead). B Representative electrophysiological traces of spike responses evoked in DMS neurons by OFC stimulation in the different groups. Black traces represent trials that do not exhibit spike responses to OFC stimulations and colored traces represent trials exhibiting spike response to OFC stimulation both before (baseline, blue) and after (post-HFS, green) high-frequency stimulation of the OFC-DMS pathway. Spike probability was measured as the number of responding trials over 5 min. C Spike probability (% of baseline) 15 before and 30–45 min after HFS. Delta change of response probability after HFS compared to baseline (%). GFP SALINE (N = 6); GFP PPX (N = 6); alpha-SYN SALINE (N = 6); alpha-SYN PPX (N = 6); ****P < 0.0001; two-way ANOVA followed by Tukey’s post hoc.

Full size image

PPX treatment leads to altered striatal BDNF/TrkB protein levels

To probe potential mechanisms involved in the alteration of OFC-DMS plasticity following alpha-synuclein-induced dopaminergic neurodegeneration or PPX treatment, we assessed the protein levels of BDNF and TrkB in the DMS by western blot (Fig. 5A). Nigrostriatal lesion induced a trend for a decrease of striatal BDNF levels (F1, 17 = 3.64, P = 0.07). This decrease was further exacerbated in all groups treated with PPX (F1, 17 = 5.62, P < 0.05). Similarly, there was a trend for reduced levels of full-length TrkB following nigrostriatal degeneration (F1, 17 = 4.35, P = 0.052), which were further decreased in all animals exposed to PPX (F1, 17 = 8.53, P < 0.01) (Fig. 5B). Protein levels of truncated TrkB were unaffected by the lesion or PPX (data not shown).

Fig. 5: Effects of nigrostriatal degeneration and PPX on BDNF and TRkB in the DMS.
figure 5

A BDNF immunoblot levels and B TrkB immunoblot levels in the DMS. GFP SALINE (N = 5); GFP PPX (N = 5); alpha-SYN SALINE (N = 5); alpha-SYN PPX (N = 6). *P < 0.05; **P < 0.01; two-way ANOVA followed by Tukey’s post hoc.

Full size image

Discussion

Motor basal ganglia loops have been extensively studied following dopamine denervation and DRT with L-dopa14. However, little is known regarding associative circuits and the effects of PPX. We hypothesized that dopamine denervation and dopamine agonist administration could induce OFC-DMS dysfunction paralleling those described in motor loops (Fig. 6A). To this end, we investigated the respective contributions of nigrostriatal dopaminergic neurodegeneration and treatment with the D2/D3 agonist PPX on the OFC-DMS pathway in a viral-mediated rat model of PD. This AAV-A53T model has been shown to reproduce some key features of the disease, such as the progressive loss of nigral dopaminergic neurons associated with a progressive motor phenotype19. This model is also amenable to investigate cognitive aspects of the disease and has been shown to induce non-motor deficits such as impulsivity20, altered behavioral flexibility21, depressive-like symptoms22, emotional memory impairment (when targeting the ventral tegmental area)23, as well as short-term and spatial memory impairment (when targeting the substantia nigra and dentate gyrus)24.

Fig. 6: Basal ganglia loops and their associated behaviors.
figure 6

A Overview of motor, associative and limbic basal ganglia loops with their associated types of behaviors (top) and corresponding anatomy of the associative loop (blue) in human (left) and rat (right). B Disruption of cortico-striatal plasticity in motor and associative loops following degeneration of the substantia nigra pars compacta. AIC agranular insular cortex, BG basal ganglia, CMA cingulate motor area, DLS dorsolateral striatum, DMS dorsomedial striatum, IL infralimbic cortex, lOFC lateral orbitofrontal cortex, M1 primary motor cortex, PMC premotor cortex, PrL prelimbic cortex, SMA supplementary motor area, vmOFC ventromedial orbitofrontal cortex.

Full size image

Here, we highlight an impairment of associative basal ganglia loops in a viral-mediated model of PD. Indeed, we demonstrated that OFC putative pyramidal neurons exhibited a decrease in burst firing following dopamine denervation in the SNc. However, it has to be noticed that this impairment of OFC activity remains restricted to this particular firing pattern and that the mean firing activity did not seem to be impacted. This effect of DA lesion on OFC neurons seem to differ from that reported in the motor cortex25, but studies conducted in the motor cortex from rats showed diverging results. In freely moving rats8 as well as in urethane anesthetized rats26, 6-hydroxydopamine (6-OHDA)-induced nigrostriatal dopamine lesion resulted in decreased pyramidal neurons firing rate while in isoflurane-anesthetized rats 6-OHDA treatment induced a marked increase in firing rate and bursting pattern of motor cortical pyramidal neurons27. Such discrepancies may be related to differences in the types of DA denervation (medial forebrain bundle vs intranigral 6-OHDA injections) and anesthetics. Overall, our results indicate that progressive bilateral nigrostriatal DA denervation had no effect on the mean firing rate from OFC neurons but decreased their bursting activity. This latest consequence of DA cell loss was rescued by the D2/D3 agonist, PPX.

While the role of dopamine receptors has been well studied in the medial prefrontal cortex (mPFC)28, there is little information on how dopamine modulates pyramidal neurons in the OFC. Our results showing enhanced putative pyramidal neuron activity after D2 receptor activation are consistent with the finding that quinpirole acting on D2 receptors increases the excitability of mPFC pyramidal neurons in adult mice29. Only a few studies performed ex vivo on mice OFC neurons indicate that DA bath application decreases pyramidal neurons excitability30 through D2 receptor activation31. Our results, obtained from in vivo recordings, indicate however that in control animals D2/D3 receptor activation by PPX induces an increase in putative pyramidal neurons firing both in lesioned and in non-lesioned rats. The discrepancy between our data and the previously reported effect of D2 receptor activation on pyramidal OFC neurons can be due to the differences in species and preparation. Indeed, in anesthetized animals the network is intact resulting in preserved local and distal inputs. Thus D2/D3 receptor effect in our conditions could also be indirect through local GABAergic interneurons inhibition29,32.

The modest effect observed in this PD model on OFC function was expected since OFC dopamine innervation mostly comes from VTA dopamine neurons5,30 that are preserved in this model20. Although discrete with regards to the overall neuronal activity, a change in burst firing might account for meaningful alteration of information processing.

The efficacy of OFC to DMS connection was assessed by measuring the efficacy of OFC electrical stimulation to evoke action potential responses in DMS neurons. Striatal projection neurons are medium-sized spiny neurons (MSNs) that represent 95% of the striatal neurons and are classified according to their projection target. Striatonigral neurons project directly to the substantia nigra pars reticulata and express D1 dopamine receptors while striatopallidal neurons form the indirect pathway and express D2 receptors33. Mallet et al.34 have shown in an elegant study the differential modulation of both MSN populations in the dorsolateral striatum following unilateral 6-OHDA lesion34. Dopamine cell loss inhibited the direct pathway, whereas indirect pathway neurons were activated. Even though we were not able to identify their D1 or D2 phenotype, all recorded DMS neurons were similarly affected by the progressive DA loss and by PPX treatment.

Our results demonstrate that the efficacy of this pathway is increased following high-frequency stimulation in control rats as expected16, while the same stimulating protocol induced a decrease in the OFC-DMS efficiency following dopamine denervation, PPX treatment and the combination of both DA loss and PPX. These results are echoing those reported by the extensive work conducted in the motor cortico-striatal loop to decipher the synaptic mechanisms of L-dopa-induced dyskinesias following dopamine cell loss. Synaptic plasticity is differentially altered following loss of DA homeostasis. Partial DA denervation can impact either LTD or LTP while extensive denervation impairs both forms of plasticity35. Accordingly, the progressive and partial nigrostriatal DA lesion in our study resulted in a loss of LTP, and a switch to LTD (Fig. 4). Moreover, while L-dopa treatment restores the loss of LTP in the motor cortico-striatal pathway36, our data show no such recovery following PPX treatment in the associative loop. Eventually, PPX itself induced a switch from LTP to LTD in control rats, similar to that observed after DA lesion. This effect of the D2/D3 agonist is coherent with the promotion of LTP by D2 antagonist haloperidol37. It should however be noted that high-frequency stimulation of the OFC may not only recruit cortico-striatal projections but may also involve direct and antidromic activation of thalamic nuclei38 as well as corticocortical connections such as to the anterior cingulate cortex and the medial prefrontal cortex39. Although thalamic nuclei innervated by the OFC are not those projecting to the striatum38,40, we cannot rule out that direct or indirect activation of cortical territories beyond the OFC may contribute to the observed effect of nigrostriatal degeneration and PPX on frontostriatal plasticity.

In addition to its potent effects on neuronal survival, function and dendritic spine morphology, BDNF plays a key role in synaptic plasticity and LTP41,42. To probe for potential mechanisms associated with the alterations of frontostriatal plasticity observed following alpha-synuclein-induced nigral neurodegeneration and PPX treatment, we assessed BDNF and TrkB protein levels in the DMS. In PD, BDNF levels are reduced in the substantia nigra, caudate and putamen43. In lesioned rats, there was only a trend for decreased BDNF levels in the DMS. Consistent with a predominant cortical, and to a lesser extent nigral origin of striatal BDNF44, this marginal effect on BDNF levels may be attributable to the fact that the loss of nigral dopaminergic neurons occurring in our model is only partial. Interestingly, 10-day treatment with PPX significantly decreased DMS BDNF protein levels in both sham and lesioned rats with no additive effect of the lesion, thus indicating a main contribution of PPX itself. These results extend previous findings showing that a high dose of PPX (1 mg/kg twice a day) can reduce BDNF mRNA levels in the habenula in normal rats and in the amygdala and nucleus accumbens in 6-OHDA lesioned rats45.

BDNF binding with high affinity to TrkB results in downstream activation of the mitogen-activated protein kinases and extracellular signal-regulated kinases signaling pathways that are key effectors of postsynaptic LTP42. Similar to BDNF, nigral neurodegeneration led to a trend for decreased TrkB levels in the DMS. PPX treatment had a more pronounced effect, leading to significantly reduced protein levels of TrkB in both sham and lesioned rats. Similar effects of PPX on TrkB mRNA levels were found in the striatum of control and 6-OHDA-lesioned rats45. Given the demonstrated role for BDNF and TrkB signaling in mediating LTP41,42, our results suggest that alterations in BDNF and TrkB may contribute to the alterations of frontostriatal plasticity observed in our study.

Our results suggest that nigral dopaminergic loss and PPX lead to both pre- and postsynaptic alterations in the OFC-DMS pathway. Altered OFC activity may affect the encoding of stimulus reinforcement association learning46 and impaired frontostriatal plasticity may impact the changes in synaptic efficacy that are necessary for optimal responses and behavioral adaptations (Fig. 6B). Such frontostriatal dysfunction precluding adequate information processing in associative basal ganglia loops may contribute to the expression of non-motor symptoms and create a vulnerability state for the emergence of cognitive symptoms following DRT in susceptible individuals. Future studies using electrophysiological recordings during the completion of behavioral tasks should help to identify the relationships between altered neuronal activity in associative basal ganglia loops and non-motor symptoms.

Methods

Animals

All experiments were approved by the local ethical committee (Comethea Poitou-Charentes C2EA-84, study approval #24114-2020021310245838) and performed under the European Directive (2010/63/EU) on the protection of animals used for scientific purposes. Male Sprague Dawley rats (N = 76, 175 g, Janvier, Le Genest St Isle, France) were housed on a reversed 12 h cycle. Food and water were available ad libitum. Rats were segregated into four groups: sham treated with saline (GFP SALINE, N = 17), sham treated with PPX (GFP PPX, N = 20), lesioned treated with saline (alpha-SYN SALINE, N = 19) and lesioned treated with PPX (alpha-SYN PPX, N = 20).

AAV-alpha-synuclein-mediated lesion

Rats were anesthetized with isoflurane (2% at 1.5 L/min), vitamin A was applied to the rat’s eyes to prevent corneal dryness and 1 mL NaCL 0.9% was injected subcutaneously to prevent dehydration. Pre-surgical analgesia (ketoprofen, 10 mg/kg, i.p.) and local analgesia (xylocaine gel 2%) were performed following vetidine disinfection. Rats were placed in a stereotaxic frame (Kopf Instruments®, Tujunga, CA, USA) and after skin incision, the skull was drilled to perform two bilateral injections in the SNc (in mm from bregma and dura, AP: −5.1 and −5.6; ML: +/−2.2 and +/−2; DV: −8) with 1 µL of AAV2 expressing GFP (green fluorescent protein, CMVie/SynP-GFPdegron-WPRE, 3.7 × 1013 gcp/mL) or AAV2 expressing human A53T alpha-synuclein (CMVie/SynP-synA53T-WPRE, 5.2 × 1013 gcp/mL) at 0.2 µL/min as previously described20,21,47. Post-surgical analgesia was performed with ketoprofen (10 mg/kg/day, i.p.) the day after surgery and during three days if necessary.

Treatment

Eight weeks after stereotaxic surgery, rats received either saline (NaCl 0.9%) or PPX 0.3 mg/kg/day (Sequoia Research Products, Berkshire, UK) prepared daily in sterile saline solution. Both were administrated subcutaneously for 10 days47.

Electrophysiological experiments

Experiments were performed after 10 days of treatment with PPX or saline. Rats were anesthetized with isoflurane (2% at 1.2 L/min) and placed in a stereotaxic frame (LPC, France) 30 min following saline or PPX injection. Analgesia was performed with ketoprofen (10 mg/kg/day, i.p.) and locally with 2% xylocaine gel.

Orbitofrontal activity monitoring was performed using juxtacellular in vivo recordings with a glass microelectrode (PG150-T, Harvard Apparatus, Holliston, MA, USA) filled with 0.4 M NaCl solution containing 2% Chicago Sky Blue dye (2610-05-01, Sigma-Aldrich, St Louis, MO, USA). The signal was amplified (AxoClamp 900 A, Molecular Devices), filtered (low pass: 300 Hz; high pass: 10 kHz) and digitized (Micro 1401 mk II, Cambridge Electronics Design, England). The recording electrode was lowered into the OFC to screen the spontaneous activity. For each rat, from 5 to 9 descents of 3 mm were done, starting at the coordinates: AP: +3.4; ML: +1.8: DV: −4 to −7 (in mm from bregma and dura). Following descents were then carried out to go around the starting point, by moving the electrode 0.2 mm on the mediolateral or anterio-posterior axis. Each neuron was recorded for at least 120 s, using Spike2 software (Cambridge Electronic Design Limited, Milton, Cambridge, UK).

The overall activity of the neuronal population was assessed by measuring the number of spontaneously firing neurons per mm for each rat. Spike and firing patterns were analyzed offline, using Spike2 and NeuroExplorer softwares. Spike duration was measured between the beginning of the depolarization and the end of hyperpolarization. Spike firing patterns were analyzed using NeuroExplorer burst analysis (maximum interval to start a burst = 0.17 s, maximum interval to end a burst = 0.3 s, minimum interval between bursts = 0.2 s, minimum duration of a burst = 0.01 s and minimum number of spikes in a burst = 3) as previously described48. Neurons that did not present any burst (0% spikes in burst) were considered as non-burtsy neurons, other neurons were considered as bursty neurons.

Frontostriatal plasticity recordings were acquired by combining cortical stimulation to striatal recordings. A bipolar concentric stimulating electrode (SNEX-100, Rhodes Medical Instruments, USA) was placed in the OFC (AP: +3; ML: +2; DV: −5, in mm from bregma and dura), and a glass microelectrode (PG150-T, Harvard Apparatus, Holliston, MA, USA) filled with 0.4 M NaCl solution was lowered in the DMS (AP: +0.5; ML: +2.75; DV: −4.5 to −6, in mm from bregma and dura). OFC simulations (1 pulse, 600 µs, 1 mA) were applied every 3 s to evoke a spike in striatal neurons. DMS neurons response to OFC stimulation was recorded using Signal6 software (Cambridge Electronic Design Limited, Milton, Cambridge, UK) during baseline and after High-Frequency Stimulation (HFS) of the cortico-striatal pathway. Spike probability was measured as the number of spikes evoked in response to OFC stimulation over 5 min (100 repetitions) and expressed in percent. OFC stimulation intensity was adjusted to induce 50% spike probability during baseline. HFS (2 trains of 100 pulses at 50 Hz, inter-train interval: 10 s) was performed as previously described16 using Spike2 software (Cambridge Electronic Design Limited).

Electrophysiological recordings were analyzed offline, using Spike2 and Signal6 softwares. As for OFC recordings, spike duration was measured between the beginning of the depolarization and the end of hyperpolarization.

Tissue processing and histopathological analysis

After electrophysiological recording, electrode placement was verified via iontophoretic ejection of Chicago Sky Blue dye at the recording site. Sodium pentobarbital (120 mg/kg, i.p.) was injected prior to intracardiac perfusion with 200 mL NaCl (0.9%) followed by 200 mL ice-cold paraformaldehyde (4%). Brains were collected and postfixed overnight in paraformaldehyde (4%), cryoprotected in sucrose (20% in H2O) at 4 °C and frozen in isopentane at −40 °C before storage at −80 °C. Serial 50 µm coronal free-floating sections were collected and stored at −20 °C in a cryoprotectant solution.

For Tyrosine Hydroxylase immunochemistry, sections of the mesencephalon were washed three times in Tris Buffer Saline 1× (TBS, 15 min each), sections were incubated 10 min in H2O2 solution (S202386-2, Agilent Technologies, Santa Clara, CA, USA) to quench endogenous peroxidases. After three washes, an incubation of 90 min in a blocking solution containing 3% BSA (bovine serum albumin) and 0.3% Triton in TBS 1× was performed. Sections were then incubated for 18 h at 4 °C with mouse anti-TH (1/5000, MAB318, Sigma-Aldrich) in a blocking solution. Sections were washed again three times in TBS 1×, and then incubated 1 h at room temperature with EnVision HRP system anti-mouse (K400111-2, Agilent Technologies) in blocking solution. Then, sections were washed again three times in TBS 1× and immunoreactions were revealed with DAB peroxidase substrate (K346811-2, Agilent Technologies). Nissl counterstaining was performed with 0.1% cresyl violet. Finally, sections were mounted on gelatin-coated slides and coverslipped with DePeX. TH-positive neurons were counted using the optical fractionator method on every 6th section of the SNc as previously described20. Stereological cell counts were performed with the Mercator Pro V6.5 (Explora Nova, La Rochelle, France) software coupled with a Leica 5500B microscope. Systematic random sampling was performed with counting probes of 80 × 60 µm applied on a sampling grid of 150 × 150 µm. Each TH-positive or Nissl-stained neuron with its nucleus included in the probe or intersecting any of the acceptance lines were counted. Guard zones of 2 µm ensured the exclusion of lost profiles at the top and bottom of the sections sampled. Following the delineation of the SNc with x5 objective, counting was done with ×40 objective.

For striatal TH immunofluorescence, sections were incubated for 90 min in a blocking solution containing 3% BSA (bovine serum albumin) and 0.3% Triton in TBS 1×, then incubated 18 h at 4 °C with mouse anti-TH (1/5000, MAB318, Sigma-Aldrich) in blocking solution. Sections were washed three times in TBS 1×, and then incubated 1 h at room temperature with goat anti-mouse Alexa fluor 647 secondary antibody (1/1000, A21235, ThermoFisher). Finally, sections were rinsed three times in TBS, mounted on slides, and coverslipped with DePeX. Images were acquired using an Olympus VS 120 slide scanner at ×10 magnification and analyzed using QuPath. To allow for quantitative and qualitative comparisons, standardized settings remained constant.

For recording site visualisation, frontal sections were stained with neutral red and Cresyl violet to highlight the recording site. Slices were incubated 4 min in a solution of Triton 0.3% before two successive water washes. Slices were placed in a bath of alcohol 50° before incubation in the dye solution and successive incubations with alcohol. Finally, slices were coverslipped with DePeX (Sigma-Aldrich) mounting medium.

Multiplex fluorescence RNAscope in situ hybridization was performed on 50 µm thick paraformaldehyde-fixed rat coronal brain sections (Bregma +1.8 mm). RNAscope Multiplex Fluorescent Reagent Kit v.2 (323100, Advanced Cell Diagnostics, Newark, CA, United States) was used according to the manufacturer’s instructions. Briefly, following three consecutive washes in TBS, free-floating sections were treated with hydrogen peroxide for 10 min at room temperature and mounted on Superfrost Plus and were dried completely. A hydrophobic barrier was created around sections. After hydrophobic boundaries had dried overnight, protease digestion by the protease plus was carried out for 30 min. Following tissue pretreatment, samples were hybridized with a probe specific to rat dopamine receptors D1, D2, or D3 (D1: probe Rn-DRD1-C2 #895071; D2: probe Rn-DRD2-C1 #315642; D3: probe Rn-DRD3-C3 #449961, ACDBio) mRNA at 40 °C for 2 h in a HybEZ II oven (ACDBio). A three-step amplification process was performed at 40 °C for 30, 30, and 15 min, respectively, followed by an HRP development of channel-specific signal at 40 °C for 15 min. Final labeling was realized using 520 (D2), 570 (D1) or 690 (D3) Opal fluorophores (Akoya Biosciences). Nuclei were counterstained with DAPI and sections were mounted using Mowiol (Sigma- Aldrich). Sections were washed 3 × 5 min in wash buffer at room temperature between each step of the protocol. In each experiment, a slide was used for negative control incubated with a universal negative control probe targeting the dapB gene (320871, ACDBio) and for positive control using the 3-plex positive Rn control probe (320891, ACDBio). For quantification of the intensity of fluorescence, images were acquired with a Zeiss Axio Imager.M2 Apotome microscope (Zeiss, Oberkochen, Germany) at ×10 and ×40 magnifications with ZEN imaging software (Zeiss). To allow for quantitative and qualitative comparisons, standardized settings remained constant.

For western blot experiments, rats received an injection of sodium pentobarbital (120 mg/kg, i.p.) prior to an intracardiac perfusion with 200 mL 0.9% NaCl. Brains were retrieved, frozen in isopentane at −40 °C, and stored at −80 °C until use. Regions of interest (DMS and OFC) were collected using a tissue puncher on 150-µm cryostat sections. Proteins were extracted using 1% sodium dodecyl sulfate (SDS, Sigma-Aldrich) solution in Tris HCl 0.1 M with ethylenediaminetetraacetyl (EDTA, Sigma-Aldrich) 0.01 M and phenylmethylsulfonyl fluoride (PMSF, Sigma-Aldrich), protease inhibitor, and phosphatase inhibitor cocktails at 1% (Sigma-Aldrich). Following centrifugation at 13,000× g for 10 min, the supernatants were collected, and the protein concentrations were determined by Pierce BCA protein assay kit (17888228, Thermo Scientific, Rockford, USA). Equal amounts of proteins (30 µg) were separated by 7.5% SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad, Marnes-la-Coquette, France). After blocking membranes in TBS with Tween-20 0.1 M (TBST) and 5% non-fat milk for 1 h30 at room temperature, they were incubated overnight at 4°C with the following diluted primary antibodies: 1/1500 to BDNF (ab108319, Abcam, Cambridge, UK), 1/800 to TrkB (4603, Cell Signaling Technology, Leiden, The Netherlands) and 1/8000 to TUBULIN (Sigma-Aldrich). After washing membranes in TBST buffer, they were incubated with HRP-conjugated secondary antibodies (Cell Signaling) for 1h at room temperature. Following three washes in TBST, a chemiluminescence signal was produced with Immobilon solution (Merck Millipore, Bedford, MA, USA) and measured and visualized on a PXi image system (Syngene, Cambridge, UK). Bands were quantified by densitometry using GeneTools software (Syngene). Protein levels were determined after normalizing with alpha-tubulin.

Statistical analysis

Data are expressed as mean ± standard error of the mean and analyzed using GraphPad Prism 10 software (Boston, MA, USA). Normality was tested with the Kolmogorov–Smirnov test. Data having a Gaussian distribution were analyzed using two-way and three-way analysis of variance (ANOVAs) followed by Tukey’s post hoc multiple comparisons test.