Address correspondence and reprint requests to Philippe Brun, Institut Fédératif des Neurosciences de Lyon; UCBL, Faculté de Pharmacie, Laboratoire de Neuropharmacologie et Neurochimie, 8 avenue Rockefeller, 69373 Lyon Cedex 08, France. E-mail: email@example.com
Neuroleptics are thought to exert their anti-psychotic effects by counteracting a hyper-dopaminergic transmission. Here, we have examined the dopaminergic status of STOP (stable tubule only polypeptide) null mice, which lack a microtubule-stabilizing protein and which display neuroleptic-sensitive behavioural disorders. Dopamine transmission was investigated using both behavioural analysis and measurements of dopamine efflux in different conditions. Compared to wild-type mice in basal conditions or following mild stress, STOP null mice showed a hyper-locomotor activity, which was erased by neuroleptic treatment, and an increased locomotor reactivity to amphetamine. Such a behavioural profile is indicative of an increased dopaminergic transmission. In STOP null mice, the basal dopamine concentrations, measured by quantitative microdialysis, were normal in both the nucleus accumbens and the striatum. When measured by electrochemical techniques, the dopamine efflux evoked by electrical stimulations mimicking physiological stimuli was dramatically increased in the nucleus accumbens of STOP null mice, apparently due to an increased dopamine release, whereas dopaminergic uptake and auto-inhibition mechanisms were normal. In contrast, dopamine effluxes were slightly diminished in the striatum. Together with previous results, the present study indicates the association in STOP null mice of hippocampal hypo-glutamatergy and of limbic hyper-dopaminergy. Such neurotransmission defects are thought to be central to mental diseases such as schizophrenia.
Neuroleptics are major anti-psychotic agents, widely used in the treatment of schizophrenia symptoms. Because neuroleptics are dopamine (DA) receptor antagonists, their beneficial effects have been attributed to their ability to block dopaminergic transmission. Thus, it is postulated that enhanced DA levels occur in psychotic illnesses, including schizophrenia. Due to technical limitations, this hypothesis has been difficult to test in patients. Therefore animal models exhibiting behavioural defects alleviated by neuroleptics are of definite interest to investigate the relationship between DA status and sensitivity to neuroleptics. To this day, the most widely used model consists of rats with ventral hippocampus excitotoxic lesions (VHL rats: Lipska and Weinberger 1994; Bardgett et al. 1998; Le Pen and Moreau 2002). VHL rats exhibit an enhanced locomotor reaction to acutely stressful situations, such as a single saline injection or exposure to novelty, indicative of an increased dopaminergic transmission in the mesolimbic system. Additionally, some of these behavioural alterations are alleviated by chronic neuroleptic treatment. However, the mechanisms responsible for these changes have been unclear, since no clear increase in DA levels could be detected in neurochemical studies (Lipska and Weinberger 2000).
Recently, severe behavioural disorders, specifically sensitive to chronic neuroleptic treatment, have been observed in mice deficient for a microtubule stabilizing protein called STOP (Stable Tubule Only Polypeptide) (Bosc et al. 1996; Guillaud et al. 1998; Andrieux et al. 2002). Based on these observations STOP null mice have been proposed as a potential novel model relevant for study of neuroleptics in schizophrenia related disorders.
Here, we have examined the DA status of STOP null mice. We find that, in addition to previously described complex behavioural defects (Andrieux et al. 2002), STOP null mice exhibit a hyper-locomotor activity that is amended by neuroleptic treatment, and are hyper-responsive to amphetamine, as expected in the case of hyper-DA. We have determined the DA efflux status of STOP null mice using neurochemical approaches. We find that, compared to wild-type mice, STOP null mice display hyper-reactivity of the DA efflux linked to an exacerbated DA release, in the limbic system.
Materials and methods
STOP null mice males (STOP–/–) and their wild-type (WT) littermates (STOP+/+) were generated as previously described (Andrieux et al. 2002). Mice were housed eight per cage in a temperature-controlled (22°C) environment under a 12:12 light/dark cycle (light from 6.00 am to 6.00 pm), with ad libitum access to food and water. Mice were allowed to habituate at least 1 week to the animal holding room prior to use. In all tests, mice (11–16 weeks old and 13–18 weeks old for drug free and long-term neuroleptic treated animals, respectively) were randomly allocated to treatment groups within each genotype and investigated in a random order for comparisons both between and within genotypes.
Locomotor activity battery tests
Mice were tested for locomotor activity using an automated digiscan apparatus linked to a PC computer (Imetronic, Pessac, France). Locomotor activity was monitored in a photocell testing cage equipped with an array of four parallel horizontal infrared beams (two at the front and two at the back) positioned 0.7 cm above the floor to measure horizontal activity. The number of beam breaks was recorded automatically. Horizontal activity was expressed in term of cage crossovers (i.e. consecutive breaks on either side of the cage). The number of cage crossovers was continuously recorded and cumulated over 10 or 60-min intervals.
Locomotor activity was assessed after exposure to a novel environment, after saline and after amphetamine injection during inactive phase (light period), and for a circadian period. For the novelty test, mice were removed from their home cage and placed into an individual photocell cage and the locomotor activity was measured for 1 h. For the saline injection test, mice were removed from their home cage and placed into an individual photocell cage for 1 h (habituation period). They received then a saline injection (10 mL/kg, i.p.) and their activity was monitored for one additional hour. In the amphetamine reactivity test, d-amphetamine (sulphate 1, 3 or 5 mg/kg, i.p.) was administered after a 2-h habituation period and activity was recorded for two additional hours. In the circadian locomotor activity test, each mouse was placed in a photocell cage under the same light/dark cycle as in its holding room. Food pellets (13 g) were scattered over the cage floor and water was available ad libitum from a drinking bottle that did not interfere with the activity measurement system. The locomotor activity was then measured for a 24-h period after a 2-h habituation period.
Dopamine overflow was investigated in vivo, both in the nucleus accumbens (Nacc) and the striatum. Basal DA concentrations were assayed using microdialysis no-net-flux method (Smith and Justice 1994). For assessment of evoked DA efflux we used in vivo electrochemical measurements. Such measurements do not yield absolute DA concentrations but have a much better temporal resolution than microdialysis (Wightman and Robinson 2002) and are required for determination of dynamic responses (Suaud-Chagny 2004).
Animals for dopamine analysis
Mice were anaesthetized with urethane (1.8 g/kg, i.p.) and immobilized in a stereotaxic apparatus. Experiments were performed in accordance with French and European Economic Community guidelines for the care and use of laboratory animals.
Basal dopamine levels
A microdialysis probe (CMA/Microdialysis, Stockholm, Sweden) was implanted either in the Nacc (membrane 1 mm long; CMA-7) [coordinates adapted from the atlas of Franklin and Paxinos (1997): 1.7 mm anterior to bregma, 0.8 mm lateral and 4.1 mm from dura] or in the striatum (membrane 2 mm long; CMA-11) [coordinates adapted from the atlas of Franklin and Paxinos (1997): 0.2 mm anterior to bregma, 2.0 mm lateral and 4.0 mm from dura] of each mouse. The dialysis probe was perfused at 1 µL/min with artificial cerebral spinal fluid [145 mm NaCl, 2.7 mm KCl, 1.0 mm MgCl2, 1.2 mm CaCl2, 2.33 mm NaH2PO4, 0.3 mm ascorbate (pH 7.4)]. The outlet of the probe was connected to an on-line derivatization system allowing continuous derivatization of nanoliter volumes containing DA at a subnanomolar concentration as previously described (Bert et al. 1996). Microdialysis sampling started 120 min after the probe placement. After basal DA levels were established, different DA concentrations (0, 3, 30 and 100 nm) were successively applied through the probe in order to measure for each concentration the in vivo loss or gain from dialysis efferent samples. For each DA concentration perfused, three 3 min-perfusate samples, derivatized online, were collected after a 30-min stabilization. The true DA concentrations applied were measured from each solution perfused. For this, while the perfusion solutions were changed, a 100-µL sample from the solution just finished from being perfused was taken, manually derivatized and three aliquots were measured for DA as described below. All samples (final volume 4.2 µL) were analysed for DA by capillary zone electrophoresis with laser-induced fluorescence detection at 442/490 nm as previously described (Bert et al. 1996). The difference between the measured DA concentrations applied (Cin) and in vivo collected (Cout) was plotted as a function of Cin for each animal. The point of zero flux (no-net-flux: the Cin at which Cin − Cout = 0), extrapolated by linear regression, represents the basal DA extracellular concentration (DAext). The slope of each linear regression yields an apparent DA recovery index (Ra) which has been used to assess DA uptake (Smith and Justice 1994).
Evoked dopamine efflux
Electrochemical techniques. The electrically evoked variations in extracellular DA concentration were monitored with a carbon fibre electrode combined either with differential pulse amperometry (DPA) or continuous amperometry. Measuring electrodes were made as previously described (Gonon et al. 1984). Their active part was the surface of one pure pyrolytic carbon fibre 250-µm long and 8 µm in diameter. For continuous amperometry, the carbon fibre is used in its raw state, i.e. without any further treatment, to catalyse the oxidative reaction. In vitro calibration of such electrodes was performed using a flow injection system (40 µL/s) (Dugast et al. 1994). Electrodes yielded a linear response over a wide range of DA concentrations (50 nm to 5 µm) with a slope of 3.81 ± 0.33 nm/pA (n = 6). For DPA, electrodes are electrochemically treated just before in vivo use. A step of quality control is then added to the above procedure. In this step, the treatment current as well as sensitivity is systematically checked before use. When successively immersed in DA solutions of varying concentrations, electrochemically treated electrodes yielded a linear response over a wide range of DA concentrations (5 nm−5 µm) with a slope of 0.40 ± 0.03 nm/pA (n = 11).
Throughout the present study, electrical responses have been converted to DA concentrations, based on the electrode calibration curves. The conversion of electrical signals into DA concentrations relies on hypothesis that calibration curves do not change dramatically in in vivo conditions (Suaud-Chagny 2004).
Electrodes were implanted either into the Nacc [coordinates adapted from the atlas of Franklin and Paxinos (1997): 1.9 mm anterior to bregma, 0.8 mm lateral and 4.0–4.2 mm from dura] or the dorso-lateral striatum [coordinates adapted from the atlas of Franklin and Paxinos (1997): 0.9 mm anterior to bregma, 1.8 mm lateral and 2.6–2.8 mm from dura] of each mouse. Stereotaxic coordinates for the striatum are slightly different from that used for microdialysis experiments to give enough space to combine the implantation of a stimulating electrode. In a first set of experiments, the amplitude of evoked DA overflow was measured by means of DPA, a technique with a high time resolution (1 Hz) and sensitivity (Suaud-Chagny 2004). Electrochemically treated (Suaud-Chagny 2004) carbon fibre electrodes were connected to a pulse voltammetric recorder (Biopulse, Radiometer Analytical, Villeurbanne, France). The Ag/AgCl reference and stainless steel auxiliary electrodes were maintained in contact with the skull by means of a semi-liquid junction. The final potential was adjusted to +80 mV (vs. the Ag/AgCl reference electrode) because DA oxidized at this potential at the surface of these treated electrodes (Suaud-Chagny 2004). In a second set of experiments, the variations of extracellular DA concentration were monitored by means of continuous amperometry, which allows the study of the kinetic parameters of DA overflow thanks to a 1 ms time constant (Suaud-Chagny 2004). A two-electrode potentiostat (AMU 110, Radiometer Analytical) was used to apply +0.4 V to an untreated carbon fibre electrode vs. the reference electrode and to record the current passing through it. The amplified signal was digitized by a PowerLab/5.0.2 system coupled to a PC computer running the ‘Scope’ program (ADInstruments, Castle Hill, Australia).
Medial forebrain bundle electrical stimulation. A bipolar stimulating electrode (SNEX-200; Rhodes Medical Instruments) was implanted in the medial forebrain bundle of each mouse [coordinates adapted from the atlas of Franklin and Paxinos (1997): 2.1 mm posterior to bregma, 1.0 mm lateral]. The depth was adjusted for each experiment so that the DA efflux was maximal. Series of cathodal monophasic current pulses (300 µA, 0.5 ms duration) were delivered to the stimulating electrode via an isolated stimulator (DS3; Digitimer, Welwyn Garden City, Hertfordshire, UK). For DPA experiments, the DS3 stimulator was triggered by a pulse generator (Electronique Lyonnaise, Lyon, France). In a first experimental paradigm, stimulations consisted of 200 pulses applied consecutively (3 min apart) at 4, 6, 10, 15, 20, 30, 40, 50 and 100 Hz. In a second experimental paradigm, regularly spaced (4 Hz, 20 s) and bursting (20 trains of 6 pulses at 15 Hz, train period: 1 s) stimulations were alternately applied 3 min apart. The 3-min delay between stimulations has been chosen so that there was no decline in the DA signal evoked by two identical stimulations applied consecutively. For continuous amperometry experiments, the DS3 stimulator was triggered by a PowerLab/5.0.2 system. In a first experimental paradigm, train stimuli consisted of 1 or 2, 3, 4 or 50 pulses at 100 Hz. In a second experimental paradigm, Paired Stimuli Depression (PSD), two stimulations consisting of 3 pulses at 100 Hz were separated by various time intervals (300, 600 and 1200 ms). In the PSD study, the amplitude of DA overflow evoked by the second stimulation was expressed as a percentage of that evoked by the first. In all continuous amperometry experimental paradigms, except 50-pulse stimulation experiments, each sequence of stimulations consisted of train stimuli applied 10 times every 15 s and recorded and averaged online. Fifty-pulse stimuli were applied once. Time interval between consecutive sequences was 5 min.
In the behavioural study, d-amphetamine sulphate (Sigma, St Louis, MO, USA) was dissolved in saline solution (NaCl 0.9%) and injected i.p. (10 mL/kg) at the dose of 1, 3 or 5 mg/kg. In the PSD study, haloperidol (Haldol™, 0.5 mg/kg, 10 mL/kg, s.c., Janssen-Pharmaceutica, Beerse, Belgium) was acutely administered immediately after the last sequence of control stimulations. The first sequence of test stimulations started 30 min after the injection.
Long-term neuroleptic treatment
Mice received haloperidol [Haldol™, 0.5 mg/(kg day), Janssen-Cilag] and chlorpromazine [Largactil™, 5 mg/(kg day), Aventis Pharma, Waterford, Ireland] dissolved in the drinking water from weaning to the day of experiment, as previously described (Andrieux et al. 2002).
In electrochemical assays, electrode placement was verified histologically as described in (Suaud-Chagny 2004). The recording sites were determined according to the atlas of Franklin and Paxinos 1997). In microdialysis experiments, the location of the probe was determined from the probe track using digital photomicrograph of coronal sections of the mouse brain (Bert et al. 2004). Briefly, at the end of each experiment, animals were killed by decapitation. The brain was removed, frozen in isopentane (−30°C) and then subjected to coronal sections 25-µm thick. Sections were removed until the probe track was visible on the exposed section of the frozen brain. Photomicrographs of the track were taken directly in the cryostat using the macro function of a digital camera [model Coolpix 950 (i.e. 2 million pixels), Nikon, France]. The digital photomicrographs were transferred from the camera to a PC computer and the probe track location was determined according to the atlas of Franklin and Paxinos 1997). Animals for which electrode or probe placement was not correct were eliminated from the present study.
STOP null mice were tested for locomotor activity, both in basal conditions (spontaneous circadian activity) and following a mild stress (exposure to novelty or saline injection). In a circadian locomotor activity test, in both free-drug genotypes, the total activity was increased during the dark phase compared to the light phase (Fig. 1a). However, the horizontal activity of STOP null mice was significantly greater than that of WT mice during the dark phase (Fig. 1a). Compared to WT mice, STOP null mice also displayed an enhanced locomotor reaction to both exposure to novelty (Fig. 2a) and saline injection (Fig. 2c). Interestingly, in the same tests, a long-term neuroleptic treatment erased the hyperlocomotor activity of STOP null mice, both in basal conditions (Fig. 1b) and after a mild stress (novelty Fig. 2b and saline injection Fig. 2d). The total locomotor activity of two treated genotypes remained increased during the dark phase compared to the light phase (Fig. 1b).
We then tested the effect of amphetamine in STOP null mice. Amphetamine is an indirect dopaminergic agonist. A supersensitive locomotor reaction to amphetamine is thought to be a signature of hyper-DA transmission (Lipska and Weinberger 2000; Gainetdinov et al. 2001; Miyakawa et al. 2003). STOP null mice or control mice were injected with amphetamine doses ranging from 1 to 5 mg/kg. At all tested doses, the response to amphetamine was significantly higher in STOP null mice compared to WT mice. Both the responses to amphetamine and the differences between genotypes increased with increasing amphetamine doses (Fig. 3).
All the locomotor alterations in STOP null mice can be considered as specific to environmental and to pharmacological challenge, as they were observed during the light phase where basal locomotor activity is normal. These results are consistent with the possibility of hyper-DA transmission in STOP null mice. To assess the dopaminergic efflux status of these mice directly, a detailed neurochemical analysis was performed in vivo by checking the basal DA levels and the DA efflux evoked by electrical stimulation.
Basal dopamine levels
The basal DA extracellular concentration was measured by quantitative microdialysis combined with capillary zone electrophoresis and laser-induced fluorescence detection (Table 1). Striatal and accumbal basal DA extracellular concentrations were similar in WT and STOP null mice, indicating that basal DA levels are unmodified in STOP null mice. Ra was also similar between genotypes, indicating normal DA uptake in STOP null mice, although there are some questions concerning the validity of Ra as an index of physiological DA uptake (Peters and Michael 1998).
Table 1. Basal dopamine extracellular concentrations (DAext) and apparent DA recovery (Ra) measured by quantitative microdialysis (no-net-flux) combined with capillary zone electrophoresis and laser-induced fluorescence detection
Data are given as the mean ± SEM (n in each group).
ns, non-significant compared to corresponding WT mice, Student's t-test.
DAext: 11.1 ± 3.5 nm (6)
DAext: 17.4 ± 3.6 nm (6)
Ra: 0.28 ± 0.08 (6)
Ra: 0.51 ± 0.1 (6)
STOP null mice
DAext: 9.3 ± 2.8 nmns (5)
DAext: 21.7 ± 5.5 nmns (6)
Ra: 0.35 ± 0.13ns (5)
Ra: 0.29 ± 0.04ns (6)
Evoked dopamine efflux
Dopaminergic transmission is governed by both tonic and phasic signalling modes. The tonic signalling corresponds to a discharge of dopaminergic neurones in a low regular spike mode (mean firing rate about 4–6 Hz) (Grace and Bunney 1984b; Sanghera et al. 1984) establishing a steady state level of DA efflux responsible for the basal DA level. The phasic signalling, notably conveying motivationally relevant information, corresponds to a discharge in a high frequency bursting pattern (Grace and Bunney 1984a; Sanghera et al. 1984; Freeman and Bunney 1987) eliciting transient DA effluxes on top of dopaminergic tone. Usually, bursts, favoured by physiological stimulations, consist of two to six spikes at 15 Hz but can reach 100 Hz (Freeman and Bunney 1987; Kiyatkin and Rebec 1998) and a single neurone can switch from one mode to the other. In summary, DA effluxes depend of both the frequency and the pattern of stimulation. The influence of both parameters was examined in the following experiments. These experiments were run on anaesthetized animals. However, it is thought that extracellular DA dynamics determined using electrical stimulations that mimic physiological stimuli in anaesthetized animals correlate tightly with responses elicited by behaviourally relevant events (Garris and Rebec 2002; Sabeti et al. 2003).
The accumbal and striatal DA effluxes evoked by electrical stimulations, in WT and STOP null mice, were measured using carbon fibre electrode combined with DPA. Ascending fibers from dopaminergic mesencephalic neurones were stimulated at various frequencies and plots of dopaminergic response vs. frequency were determined (Fig. 4). At low stimulation frequencies, there was no significant difference in the evoked accumbal DA effluxes (Fig. 4), between WT and STOP null mice, consistently with dialysis results (see above). In contrast, we observed a large potentiation of the DA efflux evoked by 15–50 Hz frequencies in the Nacc of STOP null mice compared to WT mice (Fig. 4). In the striatum of STOP null mice, the evoked DA effluxes were globally lower in STOP null mice compared to WT mice, the difference between genotypes mainly affecting responses in the high frequency range.
To investigate the phasic vs. tonic dopaminergic response directly, in conditions close to physiology, we used stimulation protocols mimicking the two natural discharge patterns of dopaminergic neurones, tonic (4 Hz, regular) or phasic (6 Hz in mean, in bursts of 15 Hz) (Grace and Bunney 1984a, b; Sanghera et al. 1984; Freeman and Bunney 1987). The ratios of the DA effluxes evoked by phasic vs. tonic stimulation were enhanced in the Nacc of STOP null compared to WT mice (Fig. 4b, insert). Strikingly, phasic vs. tonic ratios were not significantly different between genotypes, in the striatum (Fig. 4c, insert).
Our results indicate a normal response to tonic stimulation and a dopaminergic hyper-reactivity in the phasic signalling mode, in the Nacc of STOP null mice.
Frequency dependent changes in DA effluxes during long-lasting stimulations depend on DA release, clearance, and autoregulation (Wightman and Zimmerman 1990). In the following sections, we have used continuous amperometry and rapid kinetics to investigate these various factors in STOP null mice.
Dopamine release and re-uptake
Autoregulation is maximum 300–600 ms after the onset of stimulation (Benoit-Marand et al. 2001; Schmitz et al. 2002). As a result, the DA efflux evoked by one stimulation pulse or a very short train stimulation is not affected by autoregulation and depends principally on DA release and uptake (Garris et al. 1994). Therefore, to investigate DA release in the absence of autoregulation, we used single pulse electrical stimuli (0.5 ms, 300 µA). In STOP null mice the DA overflow in response to a single electrical stimulus reached 155% of the amplitude of the DA overflow recorded in WT mice (22.7 ± 1.8 nm, n = 16 and 14.4 ± 1.2 nm, n = 16, respectively, p < 0.001 Student's t-test), indicating that DA hyper-reactivity involves an increased DA release in STOP null mice.
However, the DA-hyper-reactivity and the increased overflow observed in STOP null mice could also involve a decrease in DA clearance. The DA clearance is usually expressed as T1/2, which corresponds to the time required for the DA concentration to drop to half its initial value. In a Michaelis–Menten model of the DA transport, T1/2 depends on both the intrinsic parameters of the transport molecules (Km and Vmax) and on the total DA concentration. To avoid DA concentration influence, T1/2 corresponding to fixed initial DA concentrations were compared between genotypes, over a wide range of evoked DA overflows. Results showed absence of significant differences between genotypes for all fixed initial DA concentrations (Fig. 5), indicating a normal DA clearance in STOP null mice. In this study, the clearance phase was analysed starting at a fixed DA concentration over the baseline, as measured by continuous amperometry, not at a fixed total DA concentration. However, no-net-flux experiments indicate that there is no difference between genotypes with respect to basal DA levels and, additionally, T1/2 measurements were performed over a range of DA concentrations including values at which basal DA levels are small compared to the DA overflow (Fig. 5). Therefore the comparison of T1/2 between genotypes is most probably not dependent on putative variations in basal levels, and the absence of difference between T1/2 suggest a normal DA uptake, meaning unmodified Km and Vmax, in STOP null mice.
In support to this conclusion, we used the curves observed at 50 stimulation pulses, which evoked saturating DA overflows, to estimate Vmax. Vmax corresponds to the initial slope of the DA over time plot, during the DA clearance phase. The estimated Vmax for WT and STOP null mice (564 ± 91 nm/ms, n = 6 and 747 ± 191 nm/ms, n = 8, respectively) did not differ significantly between genotypes (p > 0.05, Student's t-test).
The apparent absence of difference in both Km and Vmax, which are directly related to the affinity and number of transport molecules, respectively, indicates that both the affinity and the abundance of DA transporters are unmodified in STOP null mice compared to WT mice. Thus, excess DA response in STOP null mice apparently involves excess release, in the absence of decrease in DA uptake.
Dopamine release autoinhibition
DA release is normally negatively modulated by D2 autoreceptors (autoinhibition). We wondered whether the increased DA release observed in STOP null mice during long-lasting stimulations could reflect an impairment of autoinhibition. To investigate autoregulation, we have tested the effects of axon-terminal D2 autoreceptors on DA overflow using various stimulating protocols combined with continuous amperometry recordings. Autoinhibition was assessed by measuring the ratio of the maximal response amplitudes for the second and first stimulus (Paired Stimuli Depression, PSD). When occurring, autoinhibition induces a PSD drop inhibited by the D2 blocker haloperidol. We first used brief stimulations (3 pulses) at very high frequency (100 Hz) to minimize DA-re-uptake between stimuli (Suaud-Chagny et al. 1995). We did not observe autoregulation between stimulating pulses in such high frequency conditions, the amplitude of the effect evoked by 3 pulses at 100 Hz corresponding to three times the effect evoked by one stimulus pulse (Fig. 6). This result suggests that autoregulation does not operate in this time range (30 ms), in our mouse model, as previously observed in other mouse strains (Benoit-Marand et al. 2001; Schmitz et al. 2002). To explore a larger time range, paired stimuli (3 pulses, 100 Hz) were applied at intervals of 300, 600 and 1200 ms. In this time range, autoinhibition was manifest in both phenotypes as shown by drops in PSD (Fig. 6b) and the reversing effect of acute haloperidol (0.5 mg/kg, s.c., Fig. 6c). Consistent with previous estimates of the duration of DA release autoinhibition (Benoit-Marand et al. 2001; Schmitz et al. 2002), the maximal autoinhibition was observed for stimulation intervals of 300–600 ms, for both genotypes (Fig. 6b). Interestingly, at all stimulation intervals, the paired stimulation ratio was smaller (i.e. the PSD was larger) in STOP null mice compared to WT mice and the difference reached the level of statistical significance for a stimulus interval of 300 ms (Fig. 6b), the difference between genotypes being erased by acute haloperidol (0.5 mg/kg, s.c., Fig. 6c).
These results suggest a stronger autoinhibition in STOP null mice compared with WT mice. Since autoinhibition is stimulated by extracellular DA, the increased autoinhibition in STOP null mice could simply reflect the observed hyper-DA release in these mice. Alternatively, autoregulation may be reinforced in STOP null mice through an adaptive process. In any case, our results indicate that hyper-DA in STOP null mice is not due to a decreased activity of the D2 autoreceptors.
STOP null mice have been proposed as a model for studying neuroleptics in schizophrenia-related diseases because they exhibited neuroleptic sensitive behavioural deficits in the absence of detectable anatomical anomaly and of neurological defects (Andrieux et al. 2002). Establishing a mechanistic basis for the neuroleptic sensitivity displayed by STOP null mice was obviously a crucial step for a rational use of these mice. Here, we have tested whether the neuroleptic sensitivity observed in STOP null mice was associated with hyper-DA as expected in current models of schizophrenia. We demonstrate that STOP null mice exhibit a locomotor reaction to environmental and to pharmacological challenge typically linked to an increased dopaminergic transmission. Additionally, STOP null mice exhibit a DA hyper-reactivity in the limbic system, with a DA overflow linked to an exacerbated DA release.
Our behavioural study added to the results previously reported (Andrieux et al. 2002) allows a comprehensive behavioural analysis of STOP null mice to be given. These mice exhibit increased basal locomotor activity during the dark phase of the diurnal cycle (this study) associated with purposeless and disorganized activity, with an increased shifting between behavioural states and breaking in characteristic sequences of activity (Andrieux et al. 2002). Additionally, STOP null mice display a supersensitivity locomotor reaction to acutely stressful situation, such as a single saline injection and exposure to novelty and an increased locomotor stimulatory effect of amphetamine. STOP null mice also display severe perturbations of complex behaviours including impaired nesting and severe nurturing defects, anxiety-related behaviour, social withdrawal and inability to perform object recognition tests (Andrieux et al. 2002). Finally, a chronic treatment with neuroleptics alleviates a number of these behavioural disorders, erasing the hyperlocomotor activity of STOP null mice in addition to improving nurturing. Such a combination of phenotypes and drug sensitivity corresponds to what is currently expected for an animal model of schizophrenia (Lipska and Weinberger 2000; Gainetdinov et al. 2001; Miyakawa et al. 2003).
At the neurochemical level, previous studies have yielded strong indications for glutamatergic deficits in STOP null mice, with synaptic abnormalities in the hippocampus, including depleted synaptic vesicle pools in glutamatergic synapses associated with defects in both short- and long-term synaptic plasticity (Andrieux et al. 2002). Glutamatergic deficits are thought to be an important causal mechanism in schizophrenia related syndromes (Goff and Coyle 2001). However, the beneficial effect of neuroleptics in schizophrenia cannot be explained only by their modulating effect on glutamatergic transmission (Konradi and Heckers 2003) and hyper-DA-ergy is thought to be an another crucial landmark of schizophrenia-related syndromes. In the present study we demonstrate that STOP null mice display a dramatic dopaminergic hyper-reactivity, with an exacerbated DA efflux in response to electrical stimulations mimicking physiological stimuli, in the absence of dopaminergic alterations in basal conditions. Remarkably, dopaminergic hyper-reactivity is restricted to the Nacc, a brain system involved in reward and motivational processes that is important for behaviour. The supersensitivity of the mesolimbic system seems to be due to an increased capacity to release DA, whereas DA re-uptake and autoregulatory mechanisms seem normal or reinforced. These changes in DA regulation processes could explain the DA overflow phenotype seen in STOP null mice. In tonic conditions, an increased autoinhibition could mask the increase in DA release because autoinhibition requires, for activation following initial DA release, a finite time in the time scale of the tonic firing, this combination resulting in unchanged DA overflow between WT and STOP null mice. In contrast, in phasic conditions the increase in DA release escapes, at least in part, to autoinhibition and thus could trigger an increased DA overflow. The slight hypo-DA reactivity observed in the striatum of STOP null mice is not associated with obvious motor defects (Andrieux et al. 2002).
We do not know which molecular mechanisms are involved in the neurochemical alterations observed in STOP null mice. Many cellular processes regulating neurochemical transmission, such as cellular morphology and connectivity, long-range transport of vesicles and organelles and fusion of these vesicles with their target membranes are dependent on microtubule functioning and could be affected in these mice. Impaired glutamatergic transmission in the hippocampus of STOP null mice is associated with a decrease in synaptic vesicle density (Andrieux et al. 2002). Alterations of the synaptic vesicle density could also be involved in the enhanced DA release observed in STOP null mice. However, one would then expect an increase in vesicle density in DA neurone terminals, as opposed to a decrease, as in glutamatergic neurones. In this view, STOP gene deletion would affect in an opposite manner systems using different neurotransmitters. These heterogeneous consequences of the STOP gene deletion suggest that this deletion affects complex regulations with different consequences according to structural or functional specificities of synapses. For instance neurotransmitter exocytosis can occur both through ‘kiss-and-run’ events, where vesicle releases only a part of its content, the vesicle being reused, and through a ‘full fusion’ where the vesicle releases all its content and is then temporarily incorporated into the plasma membrane (Wightman and Haynes 2004). These two mechanisms could be involved to varying extent among synapses and be affected differently by a change in STOP or in microtubule function. The effect of the STOP gene deletion may also vary within a given neurotransmission system. For example, glutamatergic synapses differ in bouton size, vesicle size and density and size of the postsynaptic structures targeted (French and Totterdell 2004). Alterations in microtubule dynamics could have different effects on glutamatergic transmission according to these variable synaptic features, with resulting heterogeneous effects on DA transmission. It is also possible that the heterogeneous neurochemical alterations observed in STOP null mice depend on STOP protein distribution. STOP proteins are present in all the brain areas but are more strongly expressed in the olfactory bulb glomeri, hippocampus, and cerebellum (Andrieux et al. 2002).
In our study, we have concentrated our efforts on the glutamatergic and dopaminergic systems whose perturbations are central in current models of mental diseases. These perturbations may not be, however, exclusive. Perturbations in other neurotransmission systems, may participate to the apparent complexity of the neurotransmitter anomalies observe in STOP null mice, and suspected in mental diseases (Carlsson et al. 2001).
Models of schizophrenia (Grace 1991) as well as human brain imaging data (Breier et al. 1997; Abi-Dargham et al. 1998; Martinez et al. 2003) suggest that schizophrenia involves an exacerbated dopaminergic response to stimulation, with no increase in basal dopaminergic impregnation, in the mesolimbic system. The DA efflux anomalies displayed by STOP null mice fit with such models and are compatible with the observed neuroleptic sensitivity of behavioural disorders in these mice. Altogether, STOP null mice exhibit neurotransmitter anomalies, involving an association of reduced glutamatergic transmission and of hyper-DA-ergy, that are currently emerging as landmarks of schizophrenia (Carlsson et al. 2000; Jablensky 2004).
Thus, cytoskeletal anomalies can cause neurotransmitter perturbations that have, logically, long been expected to originate from defects affecting neurotransmission molecules such as neuroreceptors. Recently, evidence has accumulated that schizophrenia and other psychoses such as autism can arise from primary synaptic defects involving structural proteins (Harrison and Eastwood 2001; Mirnics et al. 2001; Jamain et al. 2003; Stefansson et al. 2004). Interestingly, in a recent, study, the product of a gene disrupted in human schizophrenia (DISC1) has been identified as a protein associated with microtubule organizing centres and with microtubule-associated proteins (Morris et al. 2003). STOP null mice provide an animal model that directly demonstrates that cytoskeletal alterations can indeed lead to behavioural abnormalities mimicking certain aspect of schizophrenia and neurotransmitter deficits thought to be involve in mental diseases. Additionally and most importantly STOP null mice react to anti-psychotic drugs used in human mental diseases.
In conclusion, we have assembled several lines of evidence indicating that the suppression of microtubule stabilization in STOP null mice induces an hyper-dopaminergic state that accounts for the sensitivity of these mice to neuroleptics. These and previous observations demonstrate that primary synaptic defects affecting cytoskeletal proteins can cause a combination of neurotransmission disorders that have been expected to occur in human psychosis. Based on these results, we propose that STOP null mice represent a meaningful model for the investigation of the origin of schizophrenia and the evaluation of new anti-psychotic agents.
This work was supported in part by grants from the French National League against Cancer to DJ, AA and AS. We thank Dr Lionel Bert for helpful expertise in quantitative microdialysis experiments.