The pharmacological blockade of medial forebrain bundle induces an acute pathological synchronization of the cortico–subthalamic nucleus–globus pallidus pathway

Abstract

Pathological oscillations characterize the firing discharge of different basal ganglia (BG) stations in rat models of Parkinson's disease. Most recent literature focused on the prominence of the beta frequency band in awake rats. Yet, in 6-hydroxydopamine-lesioned animals, the firing discharge of the globus pallidus (GP) and the substantia nigra reticulata are in phase with urethane-induced slow wave cortical activity. The neuronal basis of this pathological synergy at low frequency is widely debated. In order to understand the role of substantia nigra pars compacta (SNc) signalling in the development of pathological synchronization, we performed a pharmacological inactivation of the medial forebrain bundle (MFB) through tetrodotoxin (TTX), which led to a dramatic, but reversible, reduction of the dopamine content in the striatum. This procedure caused a significant contralateral akinesia, detectable as soon as anaesthesia vanished, and lasting about 3–4 h. We sought to determine the electrophysiological counterpart of this transient Parkinsonian-like hypokinetic syndrome. Hence, we obtained the electrocorticogram (ECoG) and single unit recordings from GP and subthalamic nucleus (STN) in normal rats before and after the TTX injection in MFB. Intriguingly, the TTX-mediated inactivation of MFB induced a fast developing coherence between cortex and GP and a significant increase of the cortex/STN synchronization. The intra-GP iontophoretic delivery of haloperidol or the GABA_A receptor antagonist bicuculline induced a short term cortex/GP synchronization. Strikingly, STN inactivation by local muscimol reversed both haloperidol- and TTX-mediated coherence between cortex and GP. Our data show that an abnormal cortical/BG synchronization, at low frequency, can be reproduced also without SNc neuronal loss and striatal cytoarchitectonic alterations. In addition, our results, which represent an acute and reversible Parkinsonism based upon impaired cable properties, seem compatible with the interpretation of acute changes of the functional interplay between cortex and the STN–GP pathway as a key factor mechanism underlying the fast deep brain stimulation-induced acute Off–On transitions.

Parkinson's disease (PD) is a disabling movement disorder centred on the degeneration of substantia nigra pars compacta (SNc) dopamine (DA) neurons that provide a massive innervation of the striatum. The latter, the main input structure of the basal ganglia (BG), is linked through direct and indirect projections to the output BG stations, whose functional activity strongly influences motor performance (Parent & Hazrati, 1995a,b; Hammond et al. 2007). Recordings in PD patients have shown that an abnormally synchronized oscillatory discharge at multiple structures of the BG–cortical loop, and not mere changes in excitability in single nuclei, dictate the clinical transitions (Bergman et al. 1998; Boraud et al. 2005; Weinberger et al. 2006). For instance, during dopaminergic therapies or deep brain stimulation (DBS) in awake patients under stereotactic neurosurgery, the pattern of endogenous frequency bands rapidly changes even without documented alterations of striatal functional activity (Brown et al. 2001; Levy et al. 2002; Kühn et al. 2008). On the other hand, in rats with 6-hydroxydopamine (6-OHDA)-induced Parkinsonism, one of the best characterized and consistent features consists of the powerful expression of cortical slow wave activity (SWA) of about 1 Hz, generated under urethane anaesthesia, in both globus pallidus (GP) and substantia nigra pars reticulata (SNr) (Magill et al. 2001; Tseng et al. 2001).

Although changes in the BG oscillatory activity reflect different states of DA activation (Vila et al. 2000; Brown, 2003), the precise correlation amongst neuronal BG oscillations, DA release and motor impairment is not fully understood. The systemic administration of DA receptor blockers in freely moving rodents caused clear-cut synchrony in the theta range (Burkhardt et al. 2007) but had no effect on the prominence of the beta band (Mallet et al. 2008b). Others reports suggested that alterations of subthalamus (STN) intrinsic voltage-gated conductance (Nambu & Llinás, 1994; Beurrier et al. 1999) or STN–GP network (Plenz & Kitai, 1999) determine the oscillatory activity. Despite such controversies, several groups would agree on a common background: that altered oscillations and pronounced synchrony between cortex and BG are attributable to the co-occurrence of SNc neuronal death and adaptive changes developing weeks after 6-OHDA (Mallet et al. 2008b).

The clinically effective DBS treatment for PD patients raised new fundamental questions and challenged some a priori assumptions. Which mechanism explains the efficacy of DBS when benefits on akinesia are perceived in a few seconds or minutes? Should we consider the possibility that these fast responses occur because of abrupt changes to cable properties in the nigrofugal pathways?

Thus, in the present report we describe to what extent the acute nigrostriatal pathway blockade by intra-MFB TTX (Nieuwenhuys et al. 1982) affects cortex/BG coherence. The effects of TTX (and riluzole) injections were first examined with behavioural tests, which demonstrated a clear but reversible asymmetric motor impairment reminiscent of a classical ‘Parkinsonism’. Hence, we detailed the biochemical (intra-striatal microdialysis), histological and electrophysiological hallmarks of our approach. Our goal is to understand if pathological cortex/BG synchronization – apart from expression of a chronic DA depletion and related compensatory mechanisms, already well known (Magill et al. 2000, 2001; Tseng et al. 2001) – may also represent an acute product of an altered information flow between SNc, BG and cortex.

Further, we investigated if ‘extra-striatal’ mechanisms govern the TTX-mediated changes, limited to the GP. It was verified that intra-GP (versus systemic) administration of haloperidol may produce the pathological synchronization. In addition, to test whether the ECoG/GP synchronization is mediated by glutamatergic STN drive and/or by GABAergic striatal inputs, we recorded GP either during the inactivation of the STN with the GABA agonist muscimol or following the intra-GP iontophoresis of the GABA_A receptor antagonist bicuculline.

Preliminary excerpts were presented as abstracts at the 2008 Federation of European Neuroscience Societies (FENS) meeting.

Methods

Ethical approval

Experimental procedures (electrophysiology, microiontophoresis, histology, microdialysis and behaviour) were carried out on 95 adult male Wistar rats weighing 250–300 g, in compliance with Italian laws on animal experimentation (D.L. 116/1992) and with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (for synopsis see Table 1). In addition, experiments fully comply, in all aspects, with UK regulations and policies for animal experimentation, as clearly stated in a recent article (Drummond, 2009).

Table 1.

Groups of animals utilized in this study

ECoG/GP and /STN recordings before and after TTX injection into MFB (n= 14)	⇒	ECoG/GP recordings (n= 7)
ECoG/STN recordings (n= 7)
ECoG/GP recordings before and after saline and riluzole injection into MFB (n= 2)
ECoG/GP recordings during systemic DA receptor blockade (n= 4)
ECoG/GP recordings coupled with micro-iontophoresis intra-GP drug delivery (n= 20)	⇒	Iontophoresis of haloperidol (n= 10)
		Iontophoresis of SCH23390 (n= 6)
		Iontophoresis of bicuculline (n= 4)
ECoG/GP recordings after muscimol injection under TTX-mediated MFB blockade (n= 4)
ECoG/GP LFP after muscimol injection under the effect of haloperidol intra-GP delivered (n= 4)
Long-lasting ECoG/GP-STN LFP after TTX-mediated MFB blockade (n= 4)
Histological data (SNc counting and fibre density measurements) after TTX injection into MFB (n= 15)	⇒	1 h after TTX injection (n= 5)
		2 weeks after TTX injection (n= 5)
		5 weeks after TTX injection (n= 5)
Microdialysis from striatum (n= 12)	⇒	Before and after TTX (n= 8)
		Before and after vehicle (n= 4)
Behavioural data (n= 16)	⇒	Step test (n= 6)
		Elevated body swing test (EBST, n= 10)

Pre-recording surgery

Rats were anaesthetized with urethane (1.4 g kg⁻¹, i.p.) (Sigma Chemical Co., St Louis, MO, USA) and mounted on a stereotaxic instrument (Stoelting Co., Wheat Lane, Wood Dale, IL, USA). Body temperature was maintained at 37–38°C with a heating pad placed beneath the animal. A midline scalp incision was made and the skull was almost completely drilled on the left side. The dura was then spread out to expose the cortical surface. All wound margins were infiltrated with a local anaesthetic (bupivicaine). Overall, 52 rats (out of 95) were submitted to BG recordings (Table 1).

Electrophysiology and intra-GP micro-iontophoresis

Electrocorticogram (ECoG) recordings coupled with GP or STN extracellular activity were performed (Fig. 1A). The ECoG was recorded via silver chloride ball electrodes placed on the cortical surface above the ipsilateral frontal cortex (3.0 mm anterior of bregma and 2.0 mm lateral to the midline) and referenced against an indifferent electrode. In additional experiments, we acquired the local field potential (LFP) from GP or STN coupled with ECoG for a long-lasting evaluation of the TTX-mediated effects, using a macro-electrode (NEX-100, Kopf). Raw ECoG and LFP were band-pass filtered (0.1–300 Hz), amplified (×2000; model 12A5 amplifier, Grass Instrument Company, Quincy, MA, USA), sampled (1000 Hz) on-line and stored on a computer connected to an analog/digital interface (Micro1401 mk II, Cambridge Electronic Design, Cambridge, UK). During ECoG recording, extracellular action potentials of GP or STN neurons were acquired using 3–5 MΩ glass electrodes (tip diameter ∼1.5 μm) containing 2 m NaCl saturated with 2% pontamine sky blue dye. Another set of experiments was conducted using microelectrodes for micro-iontophoresis and extracellular spike recording (with a carbon fibre, 0.4–0.8 MΩ; Carbostar-4, Kation Scientific, Minneapolis, MN, USA). As shown in Fig. 1B the three pipettes for micro-iontophoresis drug delivery were filled with haloperidol (6.0 mm, pH 4.2; Sigma), SCH23390 hydrochloride (10 mm, pH 4.0; Tocris), and bicuculline (20 mm, pH 4.0; Sigma). Retaining currents (negative for all solutions) of 10 nA were applied to the drug pipettes between ejection periods. In order to exclude direct effects of the pH of drug solutions on neuronal activity, preliminary tests were performed passing currents of the same intensity and polarity as for drugs ejection through barrels containing buffered saline solutions, with pH and concentrations identical to drug solutions (data not shown). The recording electrode signal was amplified (×10 000; ISO-DAM8; World Precision Instruments, Hertfordshire, UK), band-pass filtered (300–1000 Hz), sampled (60 kHz) on-line and stored on a computer connected to the CED 1401 interface (Galati et al. 2006, 2008).

Figure 1. Schematic representation of the utilized electrophysiological procedures and the impact of the pharmacological blockade of medial forebrain bundle (MFB) on motor performance assessed by the step test (Olsson et al. 1995) and by the elevated body swing test (EBST) (Borlongan & Sanberg, 1995).

A, scheme showing electrophysiological setting on a parasagittal section of rat brain. Extracellular action potentials from globus pallidus (GP) or substantia nigra pars reticulata (SNr) or subthalamic nucleus (STN) were simultaneously acquired with ECoG before and after TTX injection into the MFB. B, diagram depicting the micro-iontophoretic experiments by which a GP neuron is recorded during local (or intra-STN) drug delivery. C, micrograph of three antero-posterior sections showing the needle track targeting MFB. The inset depicts a magnified view of the needle target in the MFB area as shown in the joined atlas section (coronal section −2.80 mm from the bregma). Note that the MFB target was reached through a trajectory angled 1 deg in the anterior–posterior plane. Protocols: D, dramatic increase of the initiation step time in the right body forelimb (controlled by the left injected hemisphere) 1 h after MFB block. Deficit is still significant after 3 h. E, increase of the right forelimb stepping time evaluated 1 h after the TTX injection into the left MFB. F, clear-cut decrease of the right forelimb step length following the pharmacological blockade of the left MFB. G, following pharmacological blockade of the left MFB, rats tend to move toward right.

Pharmacological blockade of the MFB

During ECoG and single unit extracellular acquisition as well as LFP, a functional TTX (5 μm dissolved in saline) blockade of the nigrostriatal pathway was performed at the MFB (Nieuwenhuys et al. 1982; stereotaxic coordinates: 2.56 mm posterior to the bregma, 2 mm lateral to the midline, and 8.6 mm below the cortical surface; Paxinos & Watson, 1986) by means of a 30 gauge stainless steel tube (o.d. 0.2 mm) connected via tubing to a 25 μl pump-driven syringe (CMA 400 syringe pump) at an infusion rate of 1 μl min⁻¹ for 2–5 min (Fig. 1A and C).

In a subset of experiments, Evans blue (EB) was injected into the same target (0.1 μl, n= 4). These rats were transcardially perfused, under deep anaesthesia, with 60 ml of saline solution containing 0.05 ml heparin, followed by 200 ml of 4% paraformaldehyde in saline solution. The brains were removed 2 days after, and fixed overnight at 4°C, cryoprotected in 0.1 m phosphate buffer containing 10% sucrose, 20% glycerol and 0.02% sodium azide for 48 h at 4°C. Brains were sectioned frozen on a sliding microtome at 50 μm thickness. The fluorescent tracer, originally utilized for its retrograde transport, leaked around the target region but did not label STN, raphe nuclei, or locus coeruleus (data not shown), supporting the lack of involvement of extra-MFB areas.

Pharmacological inactivation of STN

Muscimol, a GABA agonist, was injected into the STN during ECoG and GP recordings (Table 1). A 30 gauge stainless steel cannula was stereotactically implanted into the STN (3.8 mm posterior to the bregma, 2.4 mm lateral to the midline, and 8.0 mm below the cortical surface) and left in place. The STN cannula was connected through a tube to a 50 μl syringe driven by a micro-infusion pump (CMA 400 syringe pump). After a baseline recording of at least 5 min, the TTX pipette was implanted into the MFB and injected as described above. Ten minutes after the MFB toxin block either muscimol (0.1 μg in 0.5 μl of saline) or saline (0.5 μl) was injected (0.25 μl min⁻¹), and recordings were continued for up to 30 min.

Electrophysiological data analysis

Single unit activity and ECoG were analysed off-line by Spike 2 software (Cambridge Electronic Design, Cambridge, UK). As described elsewhere, under urethane-induced anaesthesia, frontal ECoG is characterized by regularly occurring slow waves of large amplitude resembling physiological sleep. As a result of spontaneous fluctuation, or from sensory stimulation, the ECoG is also characterized by brief periods of cortical activation (Steriade et al. 1993; Magill et al. 2000). We focused our investigation on the cortical SWA that was visually selected from the whole ECoG recording. During robust SWA, a portion of the coincident 500 spikes (GP or STN) was previously transformed into a series of events (sampled at 1000 Hz) and then converted into a continuous function by weighting each event with a raised cosine (1 ms smoothing period; Levy et al. 2000; Galati et al. 2006). The resulting waveform was used in a subsequent coherence analysis with a size of 4096 points, yielding a frequency resolution of 0.25 Hz. The power spectral analysis as well as the measure of dominant frequency within the SWA of the ECoG recording before and after TTX treatment was performed using the fast Fourier transform function of Spike 2. The same spikes were utilized for inter-spike interval (ISI) analysis including firing frequency (as the reciprocal of the ISI mean) and autocorrelation histograms (AutoCrls). The coefficient of variation (CV) of the inter-spike intervals was calculated; this value is widely used as an indicator of regularity in point processes. Spike-triggered waveform averaging (AvWv) of the coincident ECoG was performed and used to estimate ECoG/spikes phase relationships. AutoCrls were constructed for each unit in order to test for long-term oscillations using a 10 ms bin resolution for a period of 5 s.

Moreover, the LFPs from GP or STN were cross-correlated with ECoG using a dedicated script of Spike 2 software. The value of the correlation (from +1 to −1) was calculated for epochs of 15 s and then plotted in a time course scheme.

Behavioural data

Motor assessment, conducted without pharmacological challenge (such as apomorphine or amphetamines; Hudson et al. 1993) was performed with two different procedures: the step test (Olsson et al. 1995) and the elevated body swing test (EBST; Borlongan & Sanberg, 1995; Table 1). The stereotactic pipette insertion was carried out under halothane and behavioural data were obtained 1 h after wakefulness. The first test was used to monitor initiation time, stepping time and step length using a 1 m ramp connected to the rats’ home cage. During each session, the rat was held by the experimenter with one hand fixing the hindlimbs and slightly raising the hind part above the surface. The initiation time was calculated waiting for the rat-initiated movement with the forelimb not fixed by the experimenter, using 180 s as break-off point. The stepping time was measured from initiation of movement until the rat reached the home cage; the step length was calculated by dividing the length of the ramp by the number of steps required for the rat to run up the ramp. The sequence of testing was right paw testing followed by left paw testing, repeated twice before and after TTX injection. For objective monitoring of step time and length, experimental sessions were filmed. The EBST was carried out in an additional series of animals. This simple and easy behavioural test measures the swings made by the animal handled by its tail during 30 s.

Microdialysis data

An additional series of experiments were conducted on 12 rats (Table 1) with standard microdialysis procedures (Keefe et al. 1992; Galati et al. 2006; Marte et al. 2008). Microdialysis was performed using concentric microdialysis probes (MAB2, 3 mm PES membrane, outside diameter 0.6 mm, cut-off 35 kDa, Microbiotech, Stockholm, Sweden) which were positioned into the left striatum according to the following coordinates: 0.8 mm posterior to the bregma, 4.0 mm lateral to the midline, and 6.0 mm below the cortical surface. Probes were then connected to a high precision micro-injection pump (CMA/100, CMA Microdialysis, Stockholm, Sweden) and infused at a flow rate of 4 μl min⁻¹ with artificial cerebrospinal fluid containing (in mm): NaCl 145, KCl 3, MgCl₂ 1, CaCl₂ 1.26, buffered at pH 7.4 with 2 mm phosphate buffer. After 4 h of stabilization, consecutive samples were collected every 20 min; TTX (n= 8) or vehicle (n= 4) was applied at the end of the 5th sample. All samples were immediately frozen at −80°C until analysis. At the end of the experiment, rats were killed and the correct position of the probe was histologically verified. The concentrations of DA and homovanillic acid (HVA) in samples were determined by HPLC analysis coupled with electrochemical detection (Model 5100A with a 5014B analytical cell, ESA, USA). Analysis was performed using a mobile phase consisting of 75 mm sodium phosphate, 50 mm EDTA, 0.6 mg l⁻¹ sodium dodecyl sulphate (final pH 3.0) and 5% CH₃CN at a flow rate of 1.0 ml min⁻¹. Chromatographic separation was obtained using a column for reverse-phase chromatography (Symmetry C18 3.5 μm 4.6 mm × 75 mm, Waters, Milford, USA). Applied potentials were +0.30 V and −0.30 V at the first and second electrode, respectively. Signal was obtained from detector II output. Detector gain was adjusted to obtain a detection limit of 25 pg ml⁻¹. Each sample was assayed in duplicate. A first series of experiments collected samples up to 3 h post-treatment (demonstrating the abrupt fall of the amine but a complete DA or HVA level recovery was not reached, data not shown). In order to assess a reliable recovery of DA/HVA, a second series of microdialysis experiments was performed (n= 8 post-TTX, 4 post-saline), with samples collected up to 260 min after the treatment (see Fig. 1).

Histological data

Tyrosine hydroxylase immunostaining

After each experiment, the recording site was marked by the ejection of pontamine sky blue dye from the electrode using a 40 μA current for 20 min. Brains were removed and placed in 10% buffered formalin for 2 days before histological examination. Frozen sections were cut at 40 μm intervals and stained. Microscopic examination of the sections was carried out to verify that the electrode tip was within the recorded BG structures. Immunostaining of tyrosine hydroxylase (TH) showed the degree of damage of dopaminergic neurons and nerve fibres in the substantia nigra and the ipsilateral striatum, respectively (Sancesario et al. 2004). TH staining required the following steps: free-floating sections were washed three times with Tris-buffered saline, pH 7.4, and endogenous peroxidase activity was inactivated by 5 min incubation in Tris-buffered saline containing 2% H₂O₂. Sections were rinsed with Tris-buffered saline containing 0.1% Triton X-100 and incubated with 2% normal goat serum followed by an overnight incubation at 4°C with the primary antibodies. Mouse anti-tyrosine hydroxylase (1:1000; Immunostar Inc., Hudson, WI, USA) primary antiserum was used for immunohistochemical identification of specific neuronal markers in separate series of sections from each brain. Primary antibodies were detected using a biotinylated secondary antibody (Vector Laboratories, Vectastain ABC Kit, Burlingame, CA, USA) and an avidin horseradish peroxidase–diaminobenzidine–H₂O₂ chromogen system (Sigma Fast; Sigma-Aldrich). After the diaminobenzidine reaction, sections were rinsed with Tris-buffered saline and mounted on gelatin-coated slides, dehydrated and coverslipped with Permount for light microscope examination.

SNc cell counting

After TH staining, the counting of the dopaminergic cell bodies in the SNc was carried out on six sections per animal, matched to comparable anatomical structures such as the accessory optic tract level. Neurons were counted if they were intact and had a clear nucleus and/or cytoplasm. The number of neurons was expressed as the average of the counts obtained from the six sections either in the TTX-injected (left) hemisphere or the control (right) side (Sancesario et al. 2004).

Image analysis for striatal fibre density measurement

After TH staining, the images of striatal sections were acquired with a camera (Sony DXC, Sony, Belgium), connected to the microscope and a computer. Density was measured using dedicated software (IAS 2000; Delta Sistemi, Milan, Italy) operated by an expert operator blind to the treatment. Five separate sections in each hemisphere, at the same level for each animal, were selected for density measurement. A cortex measurement was chosen as reference and subtracted from each measurement in the striatum. The observed density in the intact right striatum was considered as 100% and was compared to the left (injected) side (Sancesario et al. 2004).

Statistical analysis

Statistics were calculated using statistical software (SigmaStat 3.1). Statistical comparisons of firing rates and ISI parameters (mean, median and CV) were conducted using the non-parametric Mann–Whitney U test. The comparison within the power spectra of coherence (17 frequencies, from 0 to 3.90 Hz) was performed using the non-parametric Friedman ANOVA followed by Tukey's multiple comparison test. The comparison between each power at the different frequencies of coherence of pre-and post-TTX, or before and after the iontophoretic treatment, was performed by the Mann–Whitney U test. Bonferroni correction was applied for the multiple (the 17 frequencies) comparison setting the P value threshold up to 0.003. Microdialysis data were expressed as percentages of the mean basal value (defined as 100%) that was determined by averaging the content of the first five samples collected before TTX treatment. Differences were analysed by Friedman's ANOVA followed by Tukey's test and were considered significant at the level of P < 0.05. Likewise, behavioural data were submitted to one-way ANOVA followed by Tukey's multiple comparison test and were considered significant at the level of P < 0.05. All data are expressed as mean ± standard error of the mean (s.e.m.).

Results

Behavioural data

All behavioural tests indicated, as a consequence of TTX injection, the development of reversible contralateral akinesia, as might be expected. The following specific tests highlighted this issue.

Initiation time

The post-TTX test revealed a remarkable delay (n= 6; H= 63.967; P < 0.001; Fig. 1D) in the paw contralateral to the injected side, reaching significance already in the first test at 1 h (P < 0.05; Fig. 1D); as might be expected, this effect began to recover after 3 h and fully recovered at 4 h (Fig. 1D). No changes were observed ipsilateral to TTX.

Stepping time

No significant differences were detected pre-TTX between the two sides (Fig. 1E). Following TTX, a longer time was needed only when the rats were forced to use the right paw (n= 6; H= 33.127; P < 0.001; Fig. 1E). The effect was evident after 1 h and fully recovered at 4 h (Fig. 1E).

Step length

Step length was reduced contralateral to the TTX injection (n= 6; H= 38.338; P < 0.001; Fig. 1F); step length was significantly reduced in the 3 h post-TTX (P < 0.05). The score at the 4th hour was not different from basal.

EBST

The EBST clearly showed that right swings were preferred, being significantly higher if compared to control rats (n= 10; H= 49.234; P < 0.001; Fig. 1G) 1 h after TTX injection into the MFB. This effect was already evident at 1 h following TTX and recovered after 4 h (Fig. 1G).

Microdialysis data

The TTX injection induced a time-lock decrease of DA content in the striatum that reached significance at the first fraction following TTX injection (n= 8, H= 35.078, P < 0.001; Fig. 2A). The last five fractions were not significantly different from basal values (P > 0.05; Fig. 2A). It is important to note that the main DA metabolite, HVA, also decreased although in a different time course (n= 8, H= 38.746, P < 0.001; Fig. 2B). Sham microdialysis experiments (n= 4, determination of DA and HVA concentration following vehicle introduction alone) resulted in no significant alteration (Fig. 2A and B).

Figure 2. Microdialytic measurement of dopamine (DA) and homovanillic acid (HVA) striatal levels before and after TTX infusion and assessment of dopaminergic nigrostriatal damage after TTX-mediated MFB blockade.

A, effect of infusion of 2 μl of 10 μm TTX (•) or saline (○) into the MFB on the DA levels in striatal dialysate. Numbers on the abscissa represent successive 20 min dialysate samples after 4 h of stabilization. TTX or saline was applied at the end of the 5th sample. B, effect of infusion of 2 μl of 10 μm TTX (•) or saline (○) into the MFB on the HVA levels in striatal dialysate. Numbers on the abscissa represent successive 20 min dialysate samples after 4 h of stabilization. TTX or saline was applied at the end of the 5th sample. *P < 0.05 versus basal values. C, example of a histological section at the level of the left substantia nigra pars compacta (SNc) (TTX-injected side) compared to the right SNc (control side) performed 5 weeks after the pharmacological blockade of the MFB: no substantial differences are evidenced. D, quantitative analysis of neuronal loss within SNc carried out at 1, 2 and 5 weeks following the TTX injection of MFB compared to control side. E, histological section of the left dorsal striatum (injected side) substantially similar to the right control side at 5 weeks from the TTX-mediated blockade of the MFB. F, quantitative analysis of dopaminergic fibre content within the striatum performed 1, 2 and 5 weeks after the TTX injection of MFB, compared to control side.

Histological data

Cell counting in SNc after TH staining

No significant decrease in the amount of TH-immunoreactive (TH-IR) cells was detected in the SNc of the injected (left) side in comparison to the control (right) side (Fig. 2C). As shown in Fig. 2D the number of SNc neurons was unchanged immediately after, as well as 2 weeks after, TTX injection (P > 0.05). The result was confirmed 5 weeks post-injection (Fig. 2D).

Density measurement in the striatum

In agreement with cell counting analysis we did not observe any significant effect of the TTX injection on the percentage density of TH-positive fibres in the striatum between the two sides (P > 0.05; Fig. 2E and F). These data were obtained either acutely or 2 and 5 weeks after surgery (Fig. 2F).

Electrophysiological data

ECoG analysis

As previously described (Steriade et al. 1993; Steriade, 1999; Magill et al. 2000, 2001) the urethane anaesthesia caused SWA of ∼1 Hz with large amplitude (> 500 μV) in which a smaller and faster component (< 200 μV) overlays specific portions (Fig. 3A). The latter activity is attributed to the discharge of cortical projection neurons and is thus considered as the active component of the cortical SWA (Steriade et al. 1993). During SWA the TTX-mediated blockade of MFB promotes a clear increase in the power of the higher frequencies, as exemplified in Fig. 3B and C. Despite these relevant changes in the frequency band distribution, the dominant frequency of SWA (0.96 ± 0.03 Hz) in the frontal ECoG was not affected by TTX treatment (Fig. 3D). In our preliminary data, ECoG activity was also assessed during spontaneous or induced awake/arousal (cortical activation) epochs. In agreement with previous findings (Sharott et al. 2005, Mallet et al. 2008a,b;), showing a clear increase of beta bands in 6-OHDA rats’ expression of chronic DA denervation, the acute TTX-mediated blockade of MFB failed to induce a similar ECoG pattern (data not shown). Thus, this paper focuses on the key findings related to TTX-mediated impairment of MFB and consequent acute developing synchronization in the dominant low frequency band, between cortex and GP as well as between cortex and STN.

Figure 3. Example ECoG traces during urethane-induced slow wave cortical activity (SWA) before and after TTX-induced MFB blockade.

A, note the active component superimposed on the 1 Hz SWA. B and C, after TTX injection the active component of SWA expressed a higher frequency as highlighted by the grey boxes. D, in the course of MFB blockade no significant differences of the dominant frequency and its power were observed.

Single unit analysis of GP firing and its coherence with the simultaneously acquired ECoG before and after TTX blockade of MFB

A total of 133 neurons were recorded in the GP before and after TTX injection into MFB. Of these 104 were considered eligible for analysis (stable recordings over 3 min and over 2 Hz firing discharge). Firing features of neurons recorded are shown in Table 2. Before TTX injection into the MFB, GP neurons (n= 45) showed a tonic and regular firing discharge (Table 2). Firing rate distribution showed the presence of two major subpopulations, reminiscent of type I and type II cells (Kelland et al. 1995). Nevertheless, GP firing properties, not being the aim of the study, were not further analysed. We performed a population study by analysing the firing properties of GP neurons before and after the treatment. Thus, we collected more than one neuron before and after the TTX injection from the same animal. However, in a subset of these neurons (as depicted in Fig. 4), it was possible to maintain stable recordings before, during and after the TTX pipette insertion and the subsequent injection of TTX (n= 7).

Table 2.

Electrophysiological properties of extracellular recorded neurons before and after TTX-mediated blockade of MFB

	Basal				Post-TTX
	Firing rate	ISI mean	CV		Firing rate	ISI mean	CV
GP (n= 45)	17.6 ± 1.42	0.05 ± 0.01	0.74 ± 0.04	GP (n= 59)	23.5 ± 1.66*	0.04 ± 0.01*	1.41 ± 0.08*
STN (n= 26)	3.75 ± 0.58	0.26 ± 0.08	1.81 ± 0.15	STN (n= 46)	7.38 ± 0.69*	0.13 ± 0.01*	2.17 ± 0.07*

Data are mean ±s.e.m.

^*P < 0.001, Mann–Whitney rank sum test

Figure 4. The impact of TTX-induced MFB blockade on GP activity and ECoG/GP coherence.

A and B, an example change of synchronization revealed by the simultaneous acquisition of ECoG and GP single unit before (A) and after (B) TTX infusion. The lower insets represent the GP spike-triggered waveform average (AvWv) and the autocorrelogram histograms (AutoCrl). Before the pharmacological MFB manipulation (A) the AvWv's do not present any phase between cortex and GP. Accordingly, the AutoCrl is lacking a clear oscillatory activity. In the post-TTX period (B), the GP firing pattern changed (AutoCrl) demonstrating an undoubted coherence with the cortex as shown in AvWv. C, the corresponding contour plot of the neuron in A and B describing the mean power changes time course following the TTX injection into the MFB. D is the mean of ECoG/GP coherence power before (○, n= 45) and following TTX (•, n= 59). Dashed line in D is P= 0.05; in the same figure * is P < 0.003 within analysis; § is P < 0.003 between analysis.

The TTX injection into the MFB caused a significant increase in GP neuron firing rate (n= 59, T= 314.000, P < 0.005; Table 2). In agreement, we observed a significant reduction of the ISI mean (T= 1046.000, P < 0.001; Table 2). Changes in firing rate were parallel with firing pattern variations as demonstrated by the significant increase in CV (T= 363.000, P < 0.001; Table 2).

In normal conditions, GP neurons’ (n= 45) activity was quite regular during cortical SWA. An example of the recording in Fig. 4A shows the absence of coherence between the cortical SWA and GP spikes. The corresponding AvWv and AutoCrl did not display any significant phase and oscillatory activity, respectively (Fig. 4A). On the other hand, after TTX injection into the MFB, GP neurons showed a significant oscillation in their activity coincident with the cortical SWA (as exemplified by Fig. 4B). Consistently, the phase relationship between the two structures was expressed in the AvWv plots (Fig. 4B) and oscillation was shown in AutoCrl (Fig. 4B). The increased coherence was clearly time-locked with the TTX injection into the MFB (Fig. 4C).

This reinforced the correlation with behavioural data. As shown in the coherence average in Fig. 4D, TTX-mediated synchronization of GP/cortex produced a peak ranging between 0.48 and 1.22 Hz (n= 59; Z_0.24–0.48= 6.13; Z_0.97–1.22= 3.36; P < 0.003) at the same SWA frequencies range. The comparison between the coherence before and after TTX showed a significant difference from 0.2 to 2.19 Hz (P < 0.003; Fig. 4D). Since this experimental approach promoted relevant changes in GP firing pattern we designed two series of observations in order to validate its specificity (see Table 1 and Fig. 5A). First, control experiments were performed by recording GP neurons (n= 15) with the pipette insertion alone or with the pipette insertion plus saline, utilizing the same TTX infusion parameters in order to exclude any mechanical component (Fig. 5A and B). Second, another set of experiments was conducted recording 10 GP units during the MFB injection of another sodium voltage channel blocker, riluzole (1–5 μm in NaCl), in order to confirm the pharmacological specificity (Fig. 5C).

Figure 5. Pipette insertion as well as vehicle infusion into MFB did not cause any synchronization between cortex and GP.

A, in basal condition the GP firing is regular and not correlated with the cortex as demonstrated by AvWv and coherence histograms. B, further, the simple pipette insertion as well as the vehicle infusion into the MFB field did not cause any GP firing pattern change. C, however, riluzole (injected into MFB) mimicked the TTX-mediated behaviour of GP neurons producing good synchronization between the two structures with a clear phase in AvWv and in the coherence histogram.

Further experiments were conducted to assess the time-span effect of the TTX-mediated blockade of MFB on ECoG/GP activity. For a long-lasting evaluation up to 4 h we cross-correlated the ECoG with GP-LFP in two animals. As expected, we found in both a negative correlation between the two activities following the TTX injection that reversed after about 3.5 h (Supplemental Fig. 1A, available online only).

Single unit analysis of STN firing and its coherence with the simultaneously acquired ECoG before and after TTX blockade of MFB

Seventy-two units were recorded from the STN before and after TTX injection into the MFB (Table 2). However, only in 10 animals (and thus 10 cells) we were able to record the same neuron before, during and after the TTX pipette insertion and the subsequent injection of TTX. In pre-TTX recordings (n= 26), STN neurons showed a mean firing rate of 3.75 ± 0.58 Hz that increased to 7.38 ± 0.69 Hz after TTX injection (n= 46) into the MFB (T= 633.500, P < 0.001, Table 2), followed by a significant reduction in ISI mean (T= 1264.500, P < 0.001; Table 2). Although STN neurons showed some dispersion of ISI mean already before TTX, the CV of ISIs was significantly augmented in STN units recorded after TTX injection (T= 589.000, P < 0.001; Table 2).

Cortical oscillations were reflected in the spontaneous firing pattern of the STN already in pre-TTX control rats (n= 26; Fig. 6A) leading to a clear phase in the AvWv and oscillatory activity in the AutoCrl (Fig. 6A). However, after TTX injection we observed an increased phase correlation (Fig. 6B and C). The average ECoG/STN coherence histogram before TTX injection showed a peak between 0.48 and 1.46 Hz (Z_0.48–0.73= 4.45; Z_1.22–1.46= 3.31; P < 0.003; Fig. 6D). In the post-TTX period, STN units (n= 46) became more intensely synchronized with the cortex, with a more pronounced peak in the coherence histogram from 0.24 to 1.46 Hz (Z_0.24–0.48= 5.90; Z_1.22–1.46= 5.07; P < 0.003; Fig. 6D). The comparison between pre- and post-TTX coherence frequencies showed a significant difference from 0.48 to 1.22 Hz (P < 0.003; Fig. 6D).

Figure 6. The impact of TTX blockade on STN activity and ECoG/STN coherence.

A and B, following TTX, STN units already synchronized with the cortex in normal conditions (A) displayed a more profound coherence (B). In the pre- (A) and post-TTX (B) period the AvWv's are suggestive of a synchronized activity between cortex and STN and the AutoCrls display an oscillatory behaviour of the discharge. C, the corresponding contour plot of the STN activity described above (A and B) in which the mean power is represented during the time course following the TTX injection into the MFB. D is the mean of ECoG/STN coherence power before (○, n= 26) and following TTX (•, n= 46). Dashed line in D is P= 0.05; in the same figure * is P < 0.003 within analysis; § is P < 0.003 between analysis.

As for GP activity, we performed further experiments in order to assess the time-span effect of the TTX-mediated blockade of MFB on ECoG/STN activity. For a long-lasting evaluation up to 4 h we cross-correlated the ECoG with STN-LFP in two animals. In agreement with the single unit data we found a negative correlation in both animals between the two activities augmented by the TTX injection. About 3 h after the treatment we observed a slow but significant decrease of this effect that reached basal levels (Supplemental Fig. 1B).

Cortical/GP synchronization is caused by an imbalance between STN and striatal inputs to GP

In order to investigate whether changes of STN firing activity affected the observed ECoG/GP TTX-mediated synchronization, two orders of experiments were designed: the intra-STN injection of the GABA_A agonist muscimol and the iontophoretic injection of bicuculline into GP. The ECoG/GP synchronization induced by TTX-mediated MFB blockade (Fig. 7A and B) was reversed after the STN inhibition by muscimol (n= 4; Fig. 7C and D). Thus, no peak was detected in the average coherence histograms after muscimol inactivation of STN (P > 0.003) and no significant differences were observed in the frequencies between the basal and post-muscimol period (n= 4; P > 0.003, Fig. 7E). In another group of animals (n= 4, Table 1) the GABA_A antagonist bicuculline was micro-iontophoretically in loco injected during concurrent GP single unit recordings (n= 19, Table 3). Similar to what was observed after TTX-mediated MFB blockade, bicuculline induced a clear phase between cortex and GP (Fig. 8A, B and C). The average ECoG/GP coherence histogram after bicuculline delivery showed a peak between 0.73 and 1.70 Hz significantly different from the basal coherence (Z_0.73-0.97=−3.69; Z_1.70-1.92=−3.36; P < 0.003; Fig. 8D).

Figure 7. Pharmacological inactivation of STN abolished the TTX-induced ECoG/GP synchronization.

A–C, the impact of pharmacological blockade of MFB and the subsequent inactivation of activity of STN is illustrated by means of traces and the AvWv and AutoCrl series (see Results). The GP firing does not show any synchronization with the ECoG in basal condition (A) while it is in phase after the TTX protocol (B). However, as a consequence of the STN inactivation by micro-iontophoresis of muscimol, the GP reacquired the pre-TTX firing pattern (C). D is the corresponding contour plot of A–C, showing the mean power in the course of the TTX injection into the MFB and the subsequent inactivation of the STN by muscimol. E, mean of ECoG/GP coherence power before (○, n= 4), following TTX (•, n= 4) and post-muscimol (▴, n= 4). Dashed line in E is P= 0.05; in the same figure * is P < 0.003 within analysis; § is P < 0.003 between analysis.

Table 3.

Electrophysiological properties of extracellular GP neurons during micro-iontophoretic and systemic drug treatments

	Firing rate	ISI mean	CV
Basal (n= 18)	19.9 ± 1.52	0.05 ± 0.01	0.76 ± 0.06
Haloperidola (n= 16)	14.2 ± 1.78*	0.07 ± 0.02*	1.09 ± 0.09*
SCH23390a (n= 10)	13.2 ± 1.42	0.07 ± 0.01	0.86 ± 0.07
Haloperidol + SCH23390b (n= 10)	14.7 ± 1.56*	0.07 ± 0.00*	0.25 ± 0.07*
Bicucullinea (n= 19)	39.7 ± 2.54*	0.03 ± 0.00*	0.41 ± 0.04*

Data are mean ±s.e.m.

^aMicro-iontophoretically administered.

^bSystemically administered.

^*P < 0.05, Mann–Whitney rank sum test.

Figure 8. Bicuculline micro-iontophoretically delivered within GP induced an ECoG/GP peak of coherence.

A and B, the GABA_A antagonist, bicuculline caused the shift of the GP firing pattern from regular (A) to clustered (B). Changes of discharge are highlighted by the AvWv and AutoCrl showing a clear oscillation (A and B). C, the corresponding contour plot of the GP activity described in A and B showing the mean power in the course of bicuculline micro-iontophoretically injected at 50 nA within GP. D is the mean of ECoG/GP coherence power before (○, n= 18) and post-bicuculline (•, n= 19). Dashed line in D is P= 0.05; in the same figure * is P < 0.003 within analysis; § is P < 0.003 between analysis.

TTX-mediated changes of ECoG/GP coherence are mimicked by haloperidol only when locally applied

In another subset of animals (n= 4, Table 1), GP firing units (n= 10, Table 3) were recorded before and following the i.p. administration of the DA D1 antagonist SCH23390 (0.5 mg kg⁻¹) plus the D2 antagonist haloperidol (1 mg kg⁻¹). This acute treatment failed to affect consistently the ECoG/GP coherence (H= 14.73, P > 0.003, Fig. 9A). Surprisingly, the ECoG/GP synchronization was obtained when haloperidol was intra-GP delivered by iontophoresis in normal rats (n= 10, Table 1). GP firing changed its firing rate (Table 3) becoming in phase with the cortex as shown by representative recording in Fig. 9C and E. In contrast, the selective D1 receptor antagonist SCH23390 did not induce any oscillatory behaviour in GP units (P > 0.003, Fig. 9D). Haloperidol caused a clear coherence peak ranging from 0.48 and 1.22 (Z_0.48-0.73= 2.29; Z_0.96-1.22=−2.44; P < 0.003; Fig. 9F).

Figure 9. Changes of ECoG/GP coherence induced by D2 antagonist haloperidol only when locally applied.

A, the systemic treatment with haloperidol and SCH23390 caused a significant firing rate reduction not associated with changes in firing pattern as evidenced in the corresponding insets (on the right). B–D, when haloperidol was micro-iontophoretically applied, basal GP pattern (B) was shifted towards ‘clustered’ discharge (C); the D1 antagonist SCH23390 failed to cause clear pattern changing (D). E, the corresponding contour plot of the GP activity shown in B and C, picturing the mean power in the course of haloperidol micro-iontophoretically injected at 50 nA within GP. F, mean of ECoG/GP coherence power in control (○, n= 18), post-haloperidol (•, n= 16), post-SCH23390 (▴, n= 10) and post-systemic treatment of both drugs (▪, n= 10). Dashed line in F is P= 0.05; in the same figure * is P < 0.003 within analysis; § is P < 0.003 between analysis.

We also evaluated if the cortical/GP synchronization induced by the intra-GP delivery of haloperidol was susceptible to the STN inhibition by muscimol. In order to address this point we performed additional experiments in two animals by recording the ECoG and the GP-LFP. Soon after the synchronization induced by the intra-GP delivery of haloperidol, the subsequent muscimol injection into the STN caused a clear drop of the correlation between the two activities. A representative ECoG/GP-LFP cross-correlation plot is shown in Supplemental Fig. 1C (available online only).

Discussion

This paper has shown the electrophysiological changes acutely detected in different BG stations following the TTX-mediated blockade of the nigrostriatal pathway at the level of MFB. This procedure, although not associated with SNc neuronal damage, caused a reversible akinesia contralateral to the toxin injection and increased the firing frequency in GP and STN. More importantly, TTX installed a pathological coherence between the cortex and the GP. Abnormal GP synchronization was also mimicked by intra-GP haloperidol or bicuculline. TTX-mediated changes of GP firing pattern were abolished by intra-STN muscimol.

These findings do not provide a contradictory view of the 6-OHDA model, which is tailored to detect the progressively developing BG changes (Mallet et al. 2008a,b;). However, our data raise the possibility that some facets of Parkinsonism may be promoted by an acute DA deprivation through low frequency synchronization in extra-striatal pathways.

The fast TTX-mediated block of MFB efficiency reflects an acute DA deficiency as demonstrated by microdialysis and behavioural correlates (not an equivalent of catalepsy). As shown, all the examined parameters (akinesia, intrastriatal DA level, changes of pallidal LFP) tend to occur in a similar time-frame (about 4 h). MFB transient inactivation, even in the absence of cytoarchitectonic damage (e.g. spine loss which requires several days to build up, Solis et al. 2007) could represent a reliable model to identify early/acute changes in the cortical–BG interplay without chronic/compensatory changes in mammalian BG circuitry. In other words, extended SNc neuronal death is not a sine qua non for the occurrence of the whole spectrum of abnormal motor PD-like signs (Blandini et al. 2007); the key factor is the impairment of MFB transmission (which, in turn, would be an obvious result of SNc death). These results, though expected, retain an important clue: that abnormal synchrony in the low frequency range occurs even without an advanced stage of degeneration.

So far, the DA dependence of BG synchronization was studied not by inactivating MFB, but instead, by utilizing DA antagonists. On the one hand, the acute antagonism of D1/D2 receptors (Mallet et al. 2008b) did not have a ‘substantial impact either on synergy between cortex and BG or on oscillations’. However, Burkhardt and coauthors showed that the chronic administration of haloperidol in freely moving rats was sufficient to produce a synchronous oscillatory activity (in the theta range) across BG neuron populations (Burkhardt et al. 2007). In our hands, the local, intra-GP, delivery of haloperidol introduced a significant coherence (which was partially reversed by intra-STN muscimol), hence supporting the hypothesis that DA deprivation following TTX treatment affects the D2 receptor sensitivity of GP neurons. In contrast, by combining systemic haloperidol and SCH23390, the firing pattern changes in GP were negligible. This negative finding is not, per se, a counterargument. DA antagonists, if systemically administered, should not interfere with the MFB-dependent phasic release of the endogenous amine, whilst they might affect multiple targets such as different ion channels.

Functional mechanisms

The altered synchronization between cortex and GP, following TTX, could involve at least two different territories: extra- and intra-striatal areas.

First, the sudden TTX-mediated DA deprivation might have direct effects on target BG nuclei, without, or in combination with, striatal disregulation.

Baufreton and Bevan recently documented that DA, via D2 pre-synaptic receptors on GP terminals into STN, influences STN firing patterning. One of their key findings was that DA limits ‘the propensity for GABAergic transmission to generate correlated, bursting activity in STN neurons’ (Baufreton & Bevan, 2008). In agreement, we observed a twofold increase of STN excitability after TTX injection.

In STN slices of control rats, the prevalent response to DA and D2 agonists was membrane depolarization and firing excitation (Zhu et al. 2002) but in the STN of 6-OHDA-lesioned rats, l-DOPA had reduced STN firing activity. The modulation was mimicked by quinpirole plus the D1 agonist SKF (and not by the D2-selective agonist), thus challenging the notion that the GP–STN network is mainly dependent upon the tonic D2-mediated control of striatal medium spiny neurons (MSNs) governing the indirect pathway (Kreiss et al. 1997).

Systemic quinpirole reduced GP inverse-phase coupling with slow waves, whilst striato-pallidal hyperactivity was only slightly affected (Zold et al. 2007). Also, a robust D2-mediated modulation of N-type calcium conductance was observed in dissociated GP cells (Stefani et al. 2002), in indirect support of extra-striatal mechanisms affecting GP oscillations. Overall, these results present the possibility that some effects of DA deprivation rely on direct influence on the STN–GP projection. Interestingly, in parasagittal slices that preserved the connectivity between the GP and STN, no evidence of a dynamic relevant network emerged, at least in control rodents (Loucif et al. 2005); again, this may reinforce the possibility that peculiar to the Parkinsonian state are changes in the cortico–striato–pallidal and cortico–subthalamo–pallidal circuitry.

Recent data (Aron et al. 2007; Dejean et al. 2008) on the so-called ‘hyperdirect’ pathway indicate as clinically relevant the fast disruption of the cortico–STN–fugal projections (Nambu et al. 2002). Consistently, our muscimol experiment confirms the key role played by STN on abnormal low frequency GP oscillations (Ni et al. 2000; Magill et al. 2001).

Our results do not exclude a priori, as a simultaneous consequence of the acute blockade of MFB, acutely developing alterations of the intra-striatal DA-dependent plasticity, affecting, in turn, the firing pattern in target stations. TTX injected in MFB might interfere with physiological long term potentiation (LTP) and/or long term depression (LTD) (Charpier & Deniau, 1997; Kreitzer & Malenka, 2007; Picconi et al. 2008) and/or facilitate pathological LTP. Yet, recent studies show how LTP and LTD even persist in the DA-depleted striatum (Schen et al. 2008). In addition, arguing against this interpretation is the fact that under urethane anaesthesia, neither tetanic nor low frequency stimulation (5 Hz, Charpier et al. 1999) was driven and that TTX-mediated changes occurred in a few minutes (probably insufficient to recruit and maintain plastic mechanisms).

That said, we are aware that this study has not addressed the possible concomitant changes of striatal firing discharge (in both the main MSN subpopulations, Calabresi et al. 1993). Although the pathological synchrony which follows the intra-GP micro-iontophoretic delivery of bicuculline or haloperidol may indicate that extra-striatal mechanisms shape STN and GP synchrony, a concomitant modulation of presynaptic binding sites on striato-GP terminals cannot be ruled out and will be the subject of future investigations.

Clinical implications

In the clinical practice of movement disorders, it is evident that the quality of life of many PD patients is hampered by dramatic changes of motor performance, which may occur unpredictably with respect to pharmacokinetic/pharmacodynamic properties of prescribed drugs. Some PD patients, for instance, experience not simply predictable ‘wearing-off’ but also sudden ‘On-freezing’ even in early disease stages. Conversely, decades of experience with the STN-DBS-mediated response (i.e. switching on DBS following overnight drug withdrawal) provides unique evidence: the almost sudden reduction of akinesia, which is difficult to attribute to an endogenous remodelling of striatal plasticity.

One of the key messages implicit in the experimental data discussed here, is that acute and reversible unmasking of pathological synchrony, indeed triggered by DA depletion, reverberate swiftly in the cortex–STN–GP pathway. Transient alterations of the functional activity of the so-called hyperdirect pathway should, thus, play a relevant role in influencing fast occurrence of (or rapid relief from) akinesia.

Conversely, long-lasting adaptive processes are known to tune patients’ rigidity and to shape involuntary movement intensity. Consider, for instance, the slowly developing changes of dendritic morphology of MSNs or the adaptive down- and up-regulation of striatal peptide mRNAs, or the variability in terms of DA terminal sprouting and sub-regional altered release (in association with the variable degree of failed storage capacity as SNc neuronal loss becomes > 60%, Lee et al. 2008), or even the loss of LTD depotentiation (Picconi et al. 2008): all factors which develop within days or weeks after 6-OHDA and increase the chance of developing dyskinesias (Guigoni et al. 2005).

Although speculatively, our hypothesis may contribute to explaining the notorious paradoxical question posed by Marsden and Obeso in the nineties (Marsden & Obeso, 1994) on how stereotactic lesions of the motor thalamus apparently do not worsen Parkinsonian bradykinesia nor regularly cause dyskinesias. One of their provocative conclusions was that ‘the motor circuits of the basal ganglia are part of a distributed motor system which can operate, albeit imperfectly, in the absence of striato–pallido–thalamo–cortical feedback’ (Marsden & Obeso, 1994). As a matter of fact, at least in TTX-treated rats, akinesia may transiently peak and reverse through a straightforward alteration of cortico–STN synchrony.

Glossary

Abbreviations

ANOVA

analysis of variance

AutoCrl

autocorrelogram

AvWv

spike-triggered waveform averaging

BG

basal ganglia

CV

coefficient of variation

DA

dopamine

DBS

deep brain stimulation

EBST

elevated body swing test

ECoG

electrocorticogram

GP

globus pallidus

HVA

homovanillic acid

ISI

inter-spike interval

LFP

local field potential

MSN

medium spiny neuron

MFB

medial forebrain bundle

6-OHDA

6-hydroxydopamine

PD

Parkinson's disease

s.e.m.

standard error of the mean

SNc

substantia nigra pars compacta

SNr

substantia nigra pars reticulata

STN

subthalamic nucleus

SWA

slow wave cortical activity

TH

tyrosine hydroxylase

TTX

tetrodotoxin

Author contributions

Study concept and design: Stefani, Galati, Stanzione. Acquisition of data: Galati, D’Angelo, Fedele, Procopio, Marzetti. Analysis and interpretation of data: Galati, D’Angelo, Marzetti, Stefani. Drafting of the manuscript: Galati, Stefani. Critical revision of the manuscript for important intellectual content: Stefani, Fedele, Stanzione. Study supervision and final approval: Stefani. All experiments were performed in the Electrophysiological In Vivo Unit of the Department of Neuroscience, University Tor Vergata, Rome, Italy.