The spatiotemporal characteristics of the dopaminergic transmission mediated by D1 receptors were investigated in vivo. For this purpose dopamine (DA) release was evoked in the striatum of anesthetized rats by train electrical stimulations of the medial forebrain bundle (one to four pulses at 15 Hz), which mimicked the spontaneous activity of dopaminergic neurons. The resulting dopamine overflow was electrochemically monitored in real time in the extracellular space. This evoked DA release induced a delayed increase in discharge activity in a subpopulation of single striatal neurons. This excitation was attributable to stimulation of D1 receptors by released DA because it was abolished by acute 6-hydroxydopamine lesion and strongly reduced by the D1 antagonist SCH 23390. Striatal neurons exhibiting this delayed response were also strongly excited by intravenous administration of the D1 agonist SKF 82958. Whereas the DA overflow was closely time-correlated with stimulation, the excitatory response mediated by DA started 200 msec after release and lasted for up to 1 sec. Moreover, functional evidence presented here combined with previous morphological data show that D1 receptors are stimulated by DA diffusing up to 12 μm away from release sites in the extrasynaptic extracellular space. In conclusion, DA released by bursts of action potentials exerts, via D1 receptors, a delayed and prolonged excitatory influence on target neurons. This phasic transmission occurs outside synaptic clefts but still exhibits a high degree of spatial specificity.
- D1 receptor
- G-protein-coupled receptor
- in vivo electrochemistry
- synaptic cleft
- diffusion in the extracellular space
Fast neurotransmission is achieved within a few milliseconds by the release of a neurotransmitter and its action inside synaptic clefts on receptors coupled to ligand-gated ion channels. In vitro studies in mammalian brain show that neurotransmission involving G-protein-coupled receptors is much slower (Cole and Nicoll, 1984; Surprenant and Williams, 1987; Hille, 1992;Isaacson et al., 1993; Batchelor et al., 1994) and might occur outside synaptic clefts (Hille, 1992; Isaacson et al., 1993; Batchelor et al., 1994). To my knowledge, these temporal characteristics have never been investigated in vivo and have not yet been described regarding dopaminergic transmission. Dopaminergic receptors of the D1 type are coupled with G-proteins (Hille, 1992) and are involved in the excitatory influence of endogenous dopamine (DA) on the in vivo activity of target neurons in the rat striatal complex (Young et al., 1991; Haracz et al., 1993; Chergui et al., 1996; Gonon and Sundström, 1996; Moratalla et al., 1996). Here I describe a new experimental model that shows in real time the excitatory effect of DA released by impulse flow on the discharge activity of target neurons. This model was used to describe the spatiotemporal characteristics of the dopaminergic transmission mediated by D1 receptors.
Dopaminergic neuronal cell bodies exhibit two kinds of discharge activity: single spikes and bursts of two to six action potentials (mean, 2.9) (Grace and Bunney, 1984; Freeman and Bunney, 1987). Individual dopaminergic neurons can switch from one pattern to another, and sensory stimuli favor the bursting pattern in unrestrained rats (Freeman and Bunney, 1987). In monkeys dopaminergic neurons respond to an appetitive stimulus by a burst (Mirenowicz and Schultz, 1996). In a previous in vivo study we recorded in target areas the DA overflow (i.e., the changes in the extracellular DA concentration) evoked by electrical stimulation mimicking both discharge patterns. We showed that bursting stimulations (four pulses at 15 Hz) induced a larger DA overflow than stimulations with the same number of pulses at a lower frequency (Chergui et al., 1994). Therefore, these train stimulations (one to four pulses at 15 Hz) were used here to investigate the effect of synaptic DA release on the activity of target neurons in the striatum.
To delineate in vivo the temporal characteristics of the dopaminergic transmission, its pre- and postsynaptic sides, i.e., the DA release and the discharge response of target neurons, were recorded sequentially in every experiment with a millisecond time resolution. The electrically evoked DA overflow in the extracellular space was electrochemically monitored with a carbon fiber microelectrode, and the discharge activity of single striatal neurons was extracellularly monitored. Regarding the spatial characteristics of this transmission, morphological studies have shown that most D1 receptors are located outside synaptic contact formed by dopaminergic terminal fibers on target neurons (Levey et al., 1993; Caillé et al., 1996). In vivo experiments were conducted here to determine whether these extrasynaptic receptors play a functional role.
MATERIALS AND METHODS
Male Wistar rats (300–350 gm) were anesthetized with urethane (1.15 gm/kg) and fixed on a stereotaxic frame. Electrical stimulations of the medial forebrain bundle (MFB) were applied as described (Chergui et al., 1994; Dugast et al., 1994). The depth of the stimulating electrode was adjusted for every experiment to maximize the DA overflow (Dugast et al., 1994); the highest and lowest effective placements were found at 8.1 and 8.9 mm, respectively, below the cortical surface. This overflow was monitored continuously with amperometry by a carbon fiber electrode, the active part of which is the surface of one carbon fiber 8 μm in diameter and 250 μm long (Chergui et al., 1994; Dugast et al., 1994). Despite the poor selectivity of continuous amperometry to DA versus other oxidizable compounds, we showed that the rapid changes in the oxidation signal evoked by MFB electrical stimulation were caused entirely by the evoked DA overflow (Chergui et al., 1994; Dugast et al., 1994). This electrode was implanted in the striatum with the stereotaxic coordinates 2.7 to 3.1 mm lateral to medial line, 0.9 to 1.3 mm rostral to bregma, and 3.5 to 5.5 mm below the cortical surface. Electrical artifacts evoked by the stimulation pulses were subtracted (Chergui et al., 1994; Dugast et al., 1994). To improve the signal-to-noise ratio, a series of 16 or 20 individual DA overflows were averaged, because when they were evoked every 15 or 20 sec their amplitude remained stable (Chergui et al., 1994; Dugast et al., 1994). DA overflow was estimated as changes in DA concentration on the basis of in vitro calibration of the carbon fiber electrode performed after in vivo measurement (Dugast et al., 1994).
The discharge activity of single striatal neurons was recorded extracellularly as described (Gonon and Sundström, 1996) with carbon fiber electrodes, the active surface of which was one carbon fiber with a conic shape 8 μm in diameter and 20 μm long (impedance, 6–10 MΩ). The recording electrode was implanted in the striatum at the same coordinates as above. Voltage signals were amplified, bandpass-filtered (300 Hz to 10 kHz), and digitized at 20 kHz as shown in Figure 1. Single spike discharges were isolated by a window discriminator (World Precision Instruments, Sarasota, FL) and recorded using a data acquisition system (CED) running “Spike2” software. Whereas most striatal neurons exhibited a low and very irregular basal activity, some others were tonically active and more regular; these were not considered in this study. Striatal neurons were identified as excited by DA overflow on the basis of their peristimulus time histograms (bin width, 0.1 sec) constructed from their spike responses to 20 consecutive train stimulations (four pulses at 15 Hz) applied every 20 sec (Fig.2 a). These histograms must fulfill the following criteria: the largest bin must be observed between +0.2 and +0.7 sec after the first stimulus, and its amplitude must be at least twofold higher than that of the largest bin recorded during the 10 sec before stimulation and at least five times larger than the averaged bin size recorded during the same control period.
To lesion the dopaminergic axons, a glass pipette (external tip diameter, 35 μm) filled with 6-hydroxydopamine (6-OHDA) (Sigma, St. Louis, MO) (2 μg/μl in a saline solution containing ascorbic acid 1 mm) was implanted in the vicinity of the stimulating electrode. Three injections (0.5 μl, 90 sec each) were made 0.3 mm above, at the level of, and 0.3 mm below the tip of the stimulating electrode. The effectiveness of this lesion was assessed in each experiment by monitoring every 5 min the DA overflow evoked by a strong stimulation (10 pulses at 40 Hz) before and for 1 hr after injections.
The full D1 dopaminergic agonist SKF 82958 (Andersen and Jansen, 1990) (RBI, Natick, MA) can be detected electrochemically by a carbon fiber electrode because it oxidizes at +100 mV versus the Ag/AgCl reference electrode. Following the same procedure as described previously regarding apomorphine (Marcenac and Gonon, 1985), I monitored the extracellular SKF 82958 concentration resulting from its intravenous administration in the frontal cortex by means of an electrochemically treated carbon fiber electrode combined with differential normal pulse voltametry. This SKF 82958 oxidation peak cannot be recorded in the striatum because it is obscured by the large endogenous oxidation peak corresponding to the main DA metabolite 3,4-dihydroxyphenyl-acetic acid (Marcenac and Gonon, 1985). The D1 antagonist SCH 23390 (RBI) was dissolved as described previously (Gonon and Sundström, 1996) and administered subcutaneously, whereas SKF 82958 was dissolved in physiological saline solution and administered via the jugular vein.
Striatal neurons considered in this study exhibited a low and very irregular discharge activity. Electrical stimulation of the MFB with four pulses at 15 Hz evoked, in distinct neuronal populations, two types of excitatory effects: either a delayed excitation (Fig.1 a) or an early response consisting of single spikes closely time-correlated with every stimulation pulse (Fig. 1 b). The delayed excitation started after the last stimulation pulse and lasted for up to 1 sec (Fig. 1 a). Similar reproducible delayed excitations that fulfilled criteria defined in Materials and Methods were observed in 125 neurons. Their basal activity was low and irregular (mean, 0.90 Hz; range, from 0 to 6.8 Hz; n = 125). During vertical penetrations in the striatum by steps of 15 μm for 2 mm, the number of neurons exhibiting this delayed response was found to be 2.86 ± 0.90 (mean ± SD; 24 determinations in 12 animals) (Fig. 2 a); however, when the DA overflow was almost completely abolished (>90% decrease) by acute injection of 6-OHDA in the MFB (Fig. 2 b), only 0.27 ± 0.20 (mean ± SD; 14 determinations in seven animals) neurons per track were found. The difference between these numbers is statistically significant (unpairedt test; p < 0.0001).
Striatal neurons exhibiting this delayed excitation were also strongly excited by intravenous administration of the full D1 agonist SKF 82958 (50 μg/kg) (Fig. 3). This delayed excitation was strongly reduced when the D1 dopaminergic antagonist SCH 23390 (0.5 mg/kg, s.c.) was administered to the animal (Fig. 4), whereas this treatment did not affect the DA overflow evoked by four pulse stimulation (three experiments). Accordingly, with use of another electrochemical approach we showed that SCH 23390 administration does not affect the evoked DA overflow (Suaud-Chagny et al., 1991).
In contrast, the early excitatory response (Fig. 1 b) was still often observed after 6-OHDA lesion and was not affected by SCH 23390 administration (four neurons tested). Moreover, neurons exhibiting this early excitation were not excited by intravenous administration of the D1 agonist SKF 82958 at the same dose (five neurons tested). Apart from spike responses evoked by MFB stimulation, these neurons were usually completely silent. Tonically active neurons exhibiting a regular discharge activity were not considered in this study because MFB stimulation evoked in these neurons a complex response (excitation followed by inhibition) that was not altered by 6-OHDA lesion or by SCH 23390 administration.
Train MFB stimulations consisting of one to four pulses at 15 Hz evoked a DA overflow that was closely time-correlated with the stimulation because DA is rapidly cleared with a half-life of 74 msec (Figs. 4 and 5). In contrast the response of target neurons mediated by D1 receptors was delayed and prolonged. It reached its maximum 394 ± 85 msec (mean ± SD; 125 neurons) after the first stimulus and lasted for up to 1 sec (see Figs. 1, 2, 4, and5). Both DA overflow and delayed excitation were enhanced by increasing the number of stimulation pulses from one to four (Fig. 5). More precisely, the amplitude of the DA excitatory response and that of the DA overflow, when it is estimated in terms of area (concentration multiplied by time), were closely correlated and grew linearly with the number of pulses (Fig. 5); however, the maximal amplitude of the DA overflow (i.e., maximal extracellular DA concentration) was not linearly related to the number of pulses (Fig. 5).
The DA overflow was estimated as changes in DA concentration on the basis of in vitro calibration of the carbon fiber electrode after in vivo measurement. In the experiments illustrated in Figures 4 and 5, the DA overflow evoked by four pulses at 15 Hz corresponded to a maximal increase in DA concentration of 0.57 ± 0.21 μm (mean ± SD; 21 determinations). The extracellular concentration of SKF 82958 resulting from its intravenous administration was electrochemically monitored in the frontal cortex. Doses of 1 and 2 mg/kg were necessary to observe a detectable oxidation peak caused by SKF 82958, the amplitude of which was linearly dose dependent (data not shown). In vitro calibration of the electrode after in vivo measurement for increasing SKF 82958 concentrations was used to estimate the in vivoconcentration. The maximal concentration observed after an intravenous injection of 1 mg/kg was found to be 56 ± 13 nm(mean ± SD; four experiments). Assuming that the extracellular concentration of SKF 82958 was also linearly related to the dose administered from 0 to 1 mg/kg, it can be calculated that a dose of 50 μg/kg induced a maximal SKF 82958 concentration of 2.8 nmin the brain extracellular fluid.
The DA overflow evoked by a single pulse stimulation is dramatically prolonged by pharmacological inhibition of the DA reuptake (Fig. 6 a). The DA half-life, i.e., the time for 50% decrease from the maximum, was 55 ± 28 msec (mean ± SD; five experiments) before nomifensine administration; however, 20 min after this drug treatment, the DA half-life was maximally enhanced up to 192 ± 62 msec (mean ± SD; n = 5). The differences in half-life were statistically significant (pairedt test; p < 0.003). The maximal amplitude of the DA overflow was enhanced to a smaller extent by nomifensine (+92 ± 62%; mean ± SD; n = 5).
At a distance r from a release site, the DA concentration resulting from DA diffusion in the extracellular space after exocytosis of a single vesicle has been calculated (Fig.6 b,c) according to equation 13 of Nicholson’s study (Nicholson, 1985), with the additional assumption that the vesicle content corresponds to 1800 DA molecules (Pothos et al., 1996). Values concerning DA diffusion coefficient, extracellular volume fraction, and tortuosity were as reported (Nicholson, 1995). This concentration reaches a maximum at a time t m (Fig.6 c), which depends on the square of the distancer (Nicholson, 1985). According to this estimate, DA diffusion at 4 and 12 μm lasts 9 and 74 msec, respectively (Fig.6 c).
Excitatory effects induced by released DA via D1 receptors
The effect of DA on target neurons is still a matter of debate. Several recent studies using various in vivo approaches suggest that the release of endogenous DA exerts, via D1 dopaminergic receptors, an excitatory influence on target neurons in the striatal complex (Young et al., 1991; Haracz et al., 1993; Chergui et al., 1996;Gonon and Sundström, 1996; Moratalla et al., 1996). According to studies using application of exogenous DA and D1 agonists, however, stimulation of D1 receptors either increases (Pierce and Rebec, 1995;Levine et al., 1996a,b; Wickens et al., 1996) or inhibits (Calabresi et al., 1987; Xu et al., 1994; Surmeier et al., 1995; Nicola et al., 1996) the excitability of target neurons. Opposite effects have also been reported, depending on the DA concentration (Akaike et al., 1987; Hu and Wang, 1988). These contrasting results might be reconciled when the DA concentration and procedure of application are reconsidered (Williams and Millar, 1991; Gonon and Sundström, 1996; Wickens et al., 1996).
The present study further supports the view that stimulation of D1 receptors by released DA enhances the discharge probability of target neurons. In fact, three observations strongly suggest that the delayed excitatory response observed here was actually caused by the evoked DA overflow via D1 receptors. First, dopaminergic fibers were involved in this delayed response because it was almost totally abolished when these fibers were specifically lesioned; second, this delayed response was strongly reduced by administration of the D1 antagonist SCH 23390; and third, striatal neurons exhibiting this delayed response were also strongly excited by administration of the full D1 agonist SKF 82958. The specificity of these approaches toward the dopaminergic transmission mediated by D1 receptors is further documented here. In fact, the early excitatory response evoked in another population of striatal neurons by MFB stimulation was not altered by 6-OHDA lesions and administration of the D1 antagonist. Moreover, these neurons were not excited by administration of the D1 agonist. The present conclusion that stimulation of D1 receptors by released DA excites target neurons does not imply that DA is capable of triggering discharge activity on its own. As suggested by several in vivo and in vitro studies, DA instead may facilitate the activity of striatal neurons receiving other excitatory inputs (Haracz et al., 1993; Pierce and Rebec, 1995; Gonon and Sundström, 1996; Levine et al., 1996a,b).
Kinetics of the dopaminergic transmission mediated by D1 receptors
Since 1967, many in vivo electrophysiological studies have investigated the effects of electrical stimulation of the dopaminergic nigrostriatal pathway on striatal neurons (Frigyesi and Purpura, 1967; Siggins, 1978; Wilson et al., 1982). None of the observed effects, however, have been unequivocally attributed to the evoked DA release (Siggins, 1978; Wilson et al., 1982). In 1978 Siggins pointed out the main difficulty concerning these studies: “selective activation of the DA system should reveal long response latencies.” These expected latencies should be attributable to the slow conduction velocity of the dopaminergic axons and to “the additional time required to activate adenylate cyclase” after stimulation of the dopaminergic receptors (Siggins, 1978; Hille, 1992).
Here the latency of the postsynaptic response to DA was shown for the first time. In fact, the DA overflow evoked by MFB stimulation was closely time-correlated with the stimulation, because released DA is rapidly cleared from the extracellular fluid by neuronal reuptake (Suaud-Chagny et al., 1995). In contrast, the response of target neurons to this DA overflow was delayed and prolonged. It could be hypothesized that this latency might be the consequence of intercellular delays. Released DA might activate presynaptic D1 heteroreceptors located on terminals using others neurotransmitters. Morphological studies, however, rule out this presynaptic hypothesis. In the striatum, “D1 immunoreactive axon terminals were exceedingly rare” (Hersch et al., 1995; Caillé et al., 1996). Moreover, more complex intercellular delays involving tonically active striatal neurons cannot be responsible for this latency because MFB stimulation evoked in these neurons a complex response that was not altered by blockage of D1 receptors.
The kinetics of the delayed excitation observed here is very similar to that observed in vitro concerning effects mediated by GABAB, metabotropic glutamatergic, muscarinic, and adrenergic receptors, which are also coupled with G-proteins (Cole and Nicoll, 1984; Surprenant and Williams, 1987; Isaacson et al., 1993;Batchelor et al., 1994). Altogether it is highly likely that the kinetics of the delayed excitation observed here is governed mainly by the kinetics of the intracellular G-protein-mediated messenger systems into target neurons resulting from stimulation of D1 receptors.
Importance of bursts of action potentials
Dopaminergic neurons exhibit two kinds of discharge activity: single spikes and bursts of action potentials. Electrical stimulations of the MFB consisting of one to four pulses at 15 Hz were used here to mimic both discharge patterns. It was observed that train pulse stimulations evoked larger DA overflow and larger DA excitatory responses than single pulse stimulations. Similar observations concerning neurotransmissions involving other G-protein-coupled receptors have already been reported (Cole and Nicoll, 1984; Surprenant and Williams, 1987; Hille, 1992; Isaacson et al., 1993; Batchelor et al., 1994).
More precisely, the amplitude of the DA excitatory response was more closely correlated with the amplitude of the DA overflow when it is estimated in terms of area (concentration multiplied by time) than when the maximal DA overflow (i.e., maximal extracellular DA concentration) was considered. This suggests that a full response develops when D1 receptors are stimulated by DA for a sufficient duration (i.e., for up to 300 msec, as observed with the longest stimulation). This observation supports Hille’s hypothesis of a “temporal summation” at G-protein-coupled receptors (Hille, 1992).
At rest, dopaminergic neurons discharge at a low frequency and mainly with the single spike mode, but sensory stimuli, and especially appetitive ones, often evoke a burst of two to six action potentials (Freeman and Bunney, 1987; Mirenowicz and Schultz, 1996). Rewarding stimuli induce sharp changes in extracellular DA (Richardson and Gratton, 1996). At rest the DA extracellular concentration is in the 10 nm range, but it is transiently enhanced by one order of magnitude when dopaminergic neurons are forced to discharge in bursts by chemical stimulation (Suaud-Chagny et al., 1992). In agreement with a previous study using another approach (Chergui et al., 1996), I show here that electrical stimulations mimicking bursts of action potentials are much more potent than single pulse stimulations in triggering DA overflow and evoking the D1-mediated response.
Geometry of the dopaminergic transmission mediated by D1 receptors
Most D1 dopaminergic receptors are located on striatal dendrites outside symmetric contacts formed by dopaminergic terminal fibers (Levey et al., 1993 Hersh et al., 1995; Caillé et al., 1996). These receptors are often located in the vicinity of asymmetric synapses formed by glutamatergic terminals on the head of dendritic spines (Levey et al., 1993; Caillé et al., 1996), whereas most dopaminergic terminal fibers form symmetric synapses on the neck of these spines (Groves et al., 1994). Three lines of evidence presented here show that these extrasynaptic receptors actually play a functional role.
First, considering the size of the carbon fiber electrode, it is obvious that it monitored the DA overflow in the extrasynaptic extracellular space. Therefore, the extrasynaptic hypothesis predicts that an excellent correlation between the measured DA overflow and the amplitude of the excitatory response must be observed. This expected correlation is shown here.
Second, this extrasynaptic hypothesis implies that D1 receptors can be stimulated by DA at the concentration range actually measured by the carbon fiber electrode during evoked DA overflow: 0.2 to 1 μm. This estimate is in excellent agreement with in vitro observations concerning the affinity of D1 receptors for DA (Andersen and Jansen, 1990) and the efficacy of released DA for stimulating adenylate cyclase activity (Kelly and Nahorski, 1987). Moreover, excitatory effects were also induced in the same neurons by intravenous administration of the D1 agonist SKF 82958. Measurement of the SKF 82958 extracellular concentration shows that this drug exhibits a nanomolar in vivo efficacy. This is consistent with the fact that in vitro it exhibits a much better affinity than DA for D1 receptors (Andersen and Jansen, 1990).
Third, this hypothesis implies that DA can diffuse a few micrometers away from its release sites before elimination by reuptake. Morphological studies show that most DA reuptake sites are located on dopaminergic fibers outside synaptic contacts formed by dopaminergic terminals (Nirenberg et al., 1996). Functional evidence reported here and in a previous study (Garris et al., 1994a) actually shows that DA can diffuse outside synaptic cleft. In fact, inhibition of DA reuptake dramatically slowed down DA elimination but only moderately enhanced the maximal amplitude of the DA overflow evoked in the striatum by a single pulse stimulation (present study) or in the nucleus accumbens by a “pseudo-one-pulse” stimulation (Garris et al., 1994a). These observations show that released DA can invade the whole extracellular space before significant elimination. This is attributable to the fact that DA diffusion for 4 μm in the extracellular fluid is much faster than DA clearance. Because dopaminergic terminal fibers densely innervate the striatum, the average distance between release sites is ∼4 μm (Doucet et al., 1986; Groves et al., 1994). Moreover, at DA release sites the probability of release per action potential seems very high (Garris et al., 1994a). Therefore, this diffusion step, leading to homogenous DA concentration in the extrasynaptic extracellular space, might be achieved within 9 msec.
Functional evidence and morphological data consistently show that the dopaminergic transmission mediated by D1 receptors occurs outside synaptic clefts formed by dopaminergic terminals; however, this transmission still exhibits a high degree of spatial specificity. In fact, in normal conditions, the time for DA diffusion is limited by reuptake. From the measured DA half-life (74 msec) it can be calculated that the maximal distance for DA diffusion is ∼12 μm. Therefore, regional variations regarding the dopaminergic innervation (e.g., a lower probability of exocytosis per action potential) would result in heterogeneous DA overflow with well defined regional boundaries. Sharp regional variations in DA overflow have actually been observed (Garris et al., 1994b).
The present study underlines the importance of bursting impulse flow regarding the phasic dopaminergic transmission. In fact, train stimulations mimicking bursts of action potentials evoked in dopaminergic neurons by rewarding stimuli are more potent than single pulse in triggering DA overflow and DA excitatory responses mediated by D1 receptors. This phasic transmission exerts a delayed and prolonged excitatory influence on target neurons and occurs outside synaptic clefts. It is restricted, however, inside well defined spatiotemporal domains: it lasts for up to 1 sec, and the distance for DA diffusion is <12 μm.
This work was supported by funds from “La région Aquitaine.” I thank B. Bloch, K. Chergui, D. Choquet, M. Jaber, C. Mulle, and P. Svenningsson for helpful discussions.
Correspondence should be addressed to Dr. F. Gonon, Centre National de la Recherche Scientifique UMR 5541, Laboratoire d’Histologie-Embryologie, UFR 2, Université Bordeaux II, 33076 Bordeaux, France.