Altered levels of tonic/background dopamine in prefrontal cortex (PFC) may underlie modifications of executive cognitive function. We showed previously in rat PFC slices that exogenously supplied background dopamine facilitates induction of long-term potentiation (LTP), a possible cellular substrate for the long-term component of executive cognitive function. In the present study, we characterized cellular and molecular mechanisms underlying this modulatory dopamine effect. We show first that the LTP-facilitating effect of tonic/background dopamine follows an inverted-U shape concentration curve and that the effective level of background dopamine slowly activates postsynaptic extracellular signal-regulated kinases (ERKs) to facilitate LTP. Furthermore, we show the evidence that LTP-inducing high-frequency stimulation evokes endogenous release of dopamine in PFC slices. This fast dopamine serves as a trigger for LTP in the presence of the background dopamine. In its absence, the endogenous dopamine triggered, instead, long-term depression. These results indicate that appropriate levels of tonic/background dopamine serve to activate critical molecular factors in PFC neurons and thereby facilitate induction of synaptic potentiation.
Modification of executive cognitive function as seen in schizophrenia may be related to alterations in the level of dopaminergic activity in prefrontal cortex (PFC; Okubo et al. 1997; Seamans and Yang 2004; Lewis and Gonzalez-Burgos 2006). Particularly, it is thought that a deviation of tonic/background level of dopamine in the PFC may underlie the deficits of cognitive function (e.g., Schultz 2002). Thus, a working hypothesis particularly states that the relation between dopamine receptor activation in the PFC and the performance of executive cognitive function follows an inverted-U shape relation where only a narrow range of the dopamine level benefits the PFC function (Seamans and Yang 2004; Williams and Castner 2006; Goto et al. 2007). Indeed, manipulation of dopamine receptor activation alters PFC neural activity and spatial working memory performance in monkeys (Sawaguchi and Goldman-Rakic 1994; Williams and Goldman-Rakic 1995). In rats, the application of dopamine agonists modifies response properties of PFC pyramidal neurons in vitro (Yang and Seamans 1996; Gulledge and Jaffe 1998; Tseng and O'Donnell 2004) and can also disrupt working memory performance (Zahrt et al. 1997). Further studies are needed to more precisely determine dopamine roles in physiological and pathophysiological modifications of PFC function.
Recent progress in PFC research rekindles the classical view (e.g., Fuster 1995) that the executive cognitive function may involve long-term cognitive components, particularly long-term memory for strategic organization of behavior (Runyan et al. 2004; Takehara-Nishiuchi et al. 2006; Touzani et al. 2007). This view appears to be substantiated by the abundant evidence that PFC synapses support long-term augmentation/diminution of synaptic efficacy (Hirsch and Crepel 1990; Vickery et al. 1997; Otani et al. 1998, 1999; Gurden et al. 1999, 2000; Huang et al. 2004; Matsuda et al. 2006). It is thus thought that executive cognitive function may be at least in part supported by stable synaptic changes formed within the PFC. Abnormal induction of synaptic plasticity in the PFC would contribute to certain pathological states of executive function (Goto and Otani 2007). Consistent with this view, recent studies demonstrated impaired synaptic/neuronal plasticity induction in the brain of schizophrenic patients (Daskalakis et al. 2008; Frantseva et al. 2008). Furthermore, many schizophrenia susceptibility genes encode proteins that are thought to regulate synaptic plasticity (Harrison and Weinberger 2005). In rats, equally, it was observed that chronic stress that is known to impair executive cognitive function and reduce PFC dopamine levels (Mizoguchi et al. 2000) disrupts synaptic plasticity induction in the hippocampus–PFC pathway (Goto et al. 2007).
In the present study, we therefore explored the relation between tonic/background dopamine levels and the induction of synaptic plasticity, particularly long-term potentiation (LTP), a leading candidate for cellular mechanism of long-term memory (Whitlock et al. 2006). We showed previously in rat PFC slices that a transient background stimulation of D1 and D2 receptors by a relatively high concentration of dopamine (100 μM, 12.5 min), or a continuous “tonic” application of low concentration of dopamine (3 μM, 40 min), acts to restore induction of LTP (Matsuda et al. 2006). Here we adopted this latter, more physiological protocol and revealed some critical cellular and molecular mechanisms underlying the modulatory action of background dopamine on PFC synaptic plasticity.
Materials and Methods
All experiments were conducted in accordance with Regional Committee for Ethics of animal experimentation. Male Sprague-Dawley rats (23–30 days; Hirsch and Crepel 1990; Otani et al. 1998) were decapitated and their brains rapidly removed from their skull. Coronal slices containing the prelimbic area (300 μm; 2.2–3.7 mm from the bregma) were sectioned by the use of a Campden vibratome (Campden Instruments, Loughborough, UK) in chilled (≈0–4 °C) oxygenated (95% O2/5% CO2) artificial cerebrospinal fluid (ACSF) of the following composition (mM): NaCl 124, KCl 2, NaHCO3 26, KH2PO4 1.15, MgCl2 1, CaCl2 2, and D-glucose 11. The slices were then allowed to recover for at least 3 h at room temperature (≈20 °C) in oxygenated ACSF. Experiments were performed in a submersion-type recording chamber with continuous perfusion (1 mL/min) of warmed ACSF (28 °C).
By the use of sharp glass micropipettes (GC120F-10; Harvard Apparatus, Holliston, MA) filled with 3 M K-acetate (80–120 MΩ tip resistance), the soma of 184 layer V pyramidal neurons in the prelimbic area was penetrated. Negative currents were initially injected with an Axoclamp 2A (Molecular Devices, Union City, CA) or a BVC-700A amplifier (Dagan Corporation, Minneapolis, MN) to help stabilize the cells. After stabilization of the cells, all or most currents were removed, and thus, experiments were performed at or near resting membrane potentials of the cells. On average, resting membrane potential was −70 ± 0.5 mV, and the potential maintained during experiments was −73 ± 0.3 mV. In addition, initial and final spike heights of these cells were 76 ± 0.5 and 73 ± 0.6 mV, respectively. Initial and final input resistances were 64 ± 1.4 and 65 ± 1.5 MΩ, respectively. Prior to the experiments, the mode of spike discharge was routinely examined by applying a depolarizing current step (500-ms duration) from resting membrane potential. Sixty-eight percent of the cells were classified as regular spiking, 14% as bursting, and 18% as “adaptation cells” in which initial repetitive spikes ceased abruptly with a strong adaptation. There were no correlations between a discharge mode and the other measurements including the degree of synaptic plasticity as verified in our previous studies (Otani et al. 1998, 1999; Matsuda et al. 2006).
Stimulation, Recording, and Data Analysis
A bipolar Teflon-coated tungsten stimulating electrode (external diameter 125 μm, A-M Systems, Carlsborg, WA) was placed on layer I–II of the prelimbic area (Fig. 1A). The excitatory postsynaptic potential (EPSP) of approximately 10 mV in amplitude was evoked at 0.033 Hz by the application of monophasic constant current square pulses (100-μs duration, stimulus intensity range 20–60 μA; A360 stimulus isolator, WPI, Sarasota, FL). All drugs were included in perfusing medium except the extracellular signal-regulated kinase (ERK) substrate peptide, which was loaded in the recording electrodes. In a given set of experimental conditions, the experiments were performed in the interleaved manner. To induce plasticity, a train of tetanic stimuli (100 pulses at 50 Hz, i.e., 2 s) was delivered 4 or 6 times at 0.1 Hz. All evoked responses were fed to an amplifier in current clamp mode and digitized at 10 kHz through a Digidata 1322A interface (Molecular Devices) by the use of Elphy data acquisition analysis program (G. Sadoc, Institut Alfred Fessard, CNRS, Gif-sur-Yvette, France).
For the analysis, the initial rising slope of the EPSP (1-ms period from its onset, mV/ms), which contains only monosynaptic component (Hirsch and Crepel 1990), was calculated for each individual EPSP (Law-Tho et al. 1993). Under our experimental conditions, the slope measurement is traditionally used to express changes in the monosynaptic EPSP (Law-Tho et al. 1993; Otani et al. 1998, 1999; Matsuda et al. 2006), whose amplitude is often difficult to accurately assess due to the successively occurring polysynaptic components. Changes of the EPSP slope after conditioning (40- to 45-min period after conditioning, denoted as “40 min after conditioning”) were expressed as a percent increase or decrease from the preconditioning baseline level (the 10-min period just before conditioning stimulation or drug application). To reduce variability, the percent values were grouped for each successive 2-min period (i.e., 4 responses). The values were then compared between different experimental groups. For statistical analysis, the analysis of variance (ANOVA) was used to compare increases/decreases obtained in the entire postconditioning or post-drug application periods. To compare LTP/long-term depression (LTD) magnitudes expressed at 40 min after conditioning, 2-tailed student t-test was used.
All experiments were performed in the presence of γ-aminobutyric acid-A (GABA-A) receptor antagonist bicuculline methiodide (1 μM; with the exception of the experiments depicted in Fig. 4C,D, where drug concentration was 2 μM). The presence of bicuculline did not disrupt the EPSP. Percent changes of the EPSP slope 60 and 120 min after the beginning of bicuculline perfusion were 4.6 ± 7.3% and −2.1 ± 6.9%, respectively (n = 4).
Morphological identification of the pyramidal neurons was routinely performed using biocytin (1.5% in recording electrodes) as previously described (Otani et al. 1999; Matsuda et al. 2006). All identified neurons were classified as layer V pyramidal neurons, based on the shape and location of their cell body and on the long apical dendrite bearing a tuft, which occurs in the superficial layer. In a subpopulation of cells, an immunofluorescent method was used to allow a better visualization of cells (Fig. 1A).
The drugs used were as follows: bicuculline methiodide (Sigma Aldrich, Saint Quentin Fallavier, France), biocytin (Sigma Aldrich), 6-cyano-7-nitroquinoxaline-2,3-dion (CNQX) (Tocris, Bristol, UK), DL-2-amino-5-phosphonopentanoic acid (DL-AP5) (Tocris), dopamine (Sigma Aldrich), mitogen-activated protein (MAP) kinase substrate peptide Ala-Pro-Arg-Thr-Pro-Gly-Gly-Arg-Arg (Alexis Biochemicals, Paris, France), SCH23390 (Tocris), sulpiride (Tocris), SKF38393 (Tocris), quinpirole (Tocris), and PD98059 (Tocris).
Brain slices were identically prepared as in the electrophysiological experiments. Layer I–II fibers were identically stimulated with stimulus intensity set at the mean intensity calculated from last 20 electrophysiological experiments (∼40 μA). Drugs and conditioning stimulation were identically applied. At the end of experiments, brain slices were placed on ice, and a small prelimbic area at the level of stimulation was rapidly dissected and transferred to ice-cold lysis buffer ([10 mM Tris–HCl, 50 mM NaCl, 1% Triton X-100, 30 mM sodium pyrophosphate, 50 mM NaF, and 5 μM ZnCl2] and a mix of protein protease and phosphatase inhibitors [100 mM Na3VO4, 0.5 mM dithiothreitol, 100 nM okadaic acid, 2.5 μg/mL aprotinin, 2.5 μg/mL pepstatin, 0.5 mM phenylmethylsulfonyl fluoride, 0.5 mM benzamidine, and 2.5 μg/mL leupeptin]). Tissues were then homogenized and centrifuged (13 000 × g; 20 min; 4 °C) to pellet insoluble materials. Lysates (10–30 μg) were boiled for 5 min and analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis on 10% gels. After blocking of nonspecific sites with 5% nonfat dry milk, blots were incubated overnight at 4 °C with antiphospho-(Thr202–Tyr204)-ERK1/2 or antiphospho-(Thr180–Tyr182)-p38 primary antibodies (Cell Signalling Technology, France) diluted to 1:5000 and 1:200, respectively, in blocking solution with 0.1% Tween-20. Blots were then extensively washed before being incubated for 2 h at room temperature with anti-rabbit horseradish peroxidase (HRP)-conjugated antibodies (Amersham Pharmacia Biotech, Orsay, France) diluted to 1:3000 in the blocking solution with 0.1% Tween-20. Proteins were detected by chemiluminescence using the ECL kit (Amersham Pharmacia Biotech). As a loading control, blots were subsequently incubated with an ERK (Santa Cruz, Le Perray en Yvelines, France) or β-tubulin antibody (1:5000, Sigma Aldrich) and an anti-mouse HRP-conjugated secondary antibody. Films were scanned and analyzed using Scion Image software. Relative kinase phosphorylation levels were measured by normalization of the optical density obtained from the protein of interest with that of ERK or β-tubulin.
Exogenously Applied Background Dopamine Facilitates LTP in Concentration-Dependent Manner
We stimulated layer I–II afferent fibers at 0.033 Hz and recorded evoked monosynaptic EPSP (Hirsch and Crepel 1990; Law-Tho et al. 1993) from the soma of postsynaptic layer V pyramidal neurons in rat PFC slices (Fig. 1A; see Materials and methods for the details of EPSP analysis). In the first condition, high-frequency afferent stimulation (100 pulses at 50 Hz, i.e., 2-s duration, repeated 4 times at 0.1 Hz), which is within the frequency range of cortical gamma band activity (Keil et al. 1999), was found to induce no lasting plasticity (Fig. 1B; 0.1 ± 1.2% 40 min after conditioning, n = 12). However, when the identical tetanic stimulation was delivered in the presence of tonic/background dopamine at 3 μM concentration (40-min perfusion prior to the tetani), it induced LTP (Fig. 1D black triangles, 32 ± 14% 40 min after conditioning, n = 7). Dopamine application itself had no significant effects on the EPSP slope during the entire course of response recording (Fig. 1D gray squares; F6,46 = 0.569, P > 0.9 ANOVA repeated measures). Two-way ANOVA revealed a significant postconditioning group difference between these 2 groups depicted in Figure 1D (F1,12 = 6.906, P < 0.025), and the EPSP increase measured at 40 min after conditioning in the LTP group was significantly different from both the above nondopamine tetani group (P < 0.01, 2-tailed t-test) and the nontetani dopamine control group (Fig. 1D gray squares, P < 0.05). Importantly, in the dopamine LTP group (Fig. 1D), tetani were not delivered to a more depolarized membrane potential than in the nondopamine tetani group (Fig. 1B): the 3 μM dopamine did not change the membrane potential significantly during the 40-min preconditioning period (predopamine level −74.0 ± 1.7 mV and postdopamine level −74.1 ± 1.8 mV, n = 7), and the mean membrane potential at the time of tetani in this group was not different from that in the nondopamine tetani group (−74.3 ± 0.9 mV, n = 12).
The above results suggest that the supply of tonic/background dopamine facilitates LTP, which is hampered in the PFC slices due to the lack of sufficient levels of ambient dopamine (Matsuda et al. 2006). It is thought that the level of dopamine receptor activation to optimize PFC-mediated executive cognitive function follows the so-called inverted-U shape curve (Seamans and Yang 2004; Williams and Castner 2006; Goto et al. 2007). Therefore, we next tested whether the concentration of the tonic/background dopamine serves as a critical factor for the LTP facilitation. Thus, in 2 separate groups of neurons, we applied dopamine either at 1 or 10 μM for 40 min before tetanic stimulation (Fig. 1E,F). In these cases, the same 50-Hz tetani failed to induce LTP (1 μM condition, 1.9 ± 14% increase 40 min after tetani, n = 7, P > 0.7 compared with nontetani control group by both 2-way ANOVA and t-test; 10 μM condition, 5.6 ± 17%, P > 0.6). The relation between the tonic/background dopamine concentration and the magnitude of LTP expressed at 40 min postconditioning was found to follow an apparent inverted-U shape curve (Fig. 1C1), where a significant difference from nondopamine condition was obtained only for the 3 μM dopamine group (P < 0.01, 2-tailed t-test; for the 1 and 10 μM groups, P > 0.6).
Our previous study showed that the background application of dopamine (100 μM, 12.5 min in this case) acts on both D1 and D2 receptors to facilitate the later induction of LTP (Matsuda et al. 2006). We therefore tested whether the present, continuously applied low concentration of dopamine acts on D1 and D2 receptors. Because we found (see the later section) that endogenous fast release of dopamine evoked by tetanic stimulation is also necessary for the LTP, dopamine receptor antagonists could not be used to block dopamine receptors only during the pre-tetani dopamine perfusion phase due to slow drug washout from the bath. We therefore tried to mimic the dopamine effect by the use of receptor agonists. First, D1-like receptor agonist SKF38393, at concentrations 2–3 μM which approximately mimic the action of 3 μM dopamine on D1-like receptors (Seeman and Van Tol 1994), was applied alone in the bath. The SKF38393 induced no significant changes in the baseline EPSP during 40-min perfusion (1.5 ± 7.9% from baseline at 40 min, n = 6). Along with the no effect of 3 μM dopamine on the baseline EPSP (Fig. 1D gray squares), this shows that weak stimulation of D1 receptors does not affect basal synaptic responses. Under this condition, delivery of tetani (40 min after the beginning of the SKF38393 perfusion) induced no lasting changes in the EPSP slope (−3.1 ± 5.7% from the baseline 40 min after tetani, n = 6, Fig. 1C2, annotated as “D1 agonist,” P > 0.4 compared with the group depicted in Fig. 1B). Second, D2-like agonist quinpirole at 2–3 μM, which approximately mimics dopamine action on at least D2 and D4 receptors (Seeman and Van Tol 1994), was perfused for 40 min before tetani. Quinpirole also did not change the baseline synaptic responses (−1.9 ± 2.6% at 40 min, n = 7). Delivery of 50-Hz tetani again induced no lasting changes in the EPSP (1.5 ± 6.5% at 40 min, n = 7, Fig. 1C2, “D2 agonist,” P > 0.3). Third, in contrast, when SKF38393 and quinpirole were applied together for 40 min (−2.9 ± 13% change from the baseline during 40 min), a subsequent delivery of 50-Hz tetani induced a significant increase of the EPSP slope (20 ± 11% at 40 min, Fig. 1C2, “D1 + D2,” n = 6, P < 0.03 compared with the nonagonists control group). The somewhat smaller LTP compared with the LTP induced in the presence of 3 μM dopamine (32 ± 14%, Fig. 1D) may be due to the difference in the affinity of the synthetic agonists to dopamine receptor subtypes (Seeman and Van Tol 1994). Nevertheless, these results indicate that both D1 and D2 receptors have to be stimulated by tonic/background dopamine to obtain an LTP, consistent to our previous observation (Matsuda et al. 2006).
The Effect of Endogenously Released Dopamine Is Affected by Background Dopamine
Endogenous Dopamine Triggers Synaptic Plasticity in PFC Slices
It is thought that certain neuronal modifications seen under goal-directed behavior are triggered by event-related phasic release of dopamine at target sites (Reynolds et al. 2001; Schultz 2007). In the PFC, enhanced phasic release of dopamine indeed augments LTP (Gurden et al. 1999). In the slices of PFC, residual dopaminergic axon terminals still can release endogenous dopamine when stimulated at high frequency by extracellular electrodes (Young and Yang 2005). Such a “phasic” release of dopamine acts to trigger synaptic plasticity in PFC slices (Young and Yang 2005). These results prompted us to test whether the LTP-facilitating effect of tonic dopamine described above (Fig. 1D) is related to the action of fast endogenous release of dopamine. The action of phasic dopamine might indeed depend on background levels of dopamine (Grace 1991), and such an adaptation process might underlie the genesis of certain cognitive modifications (Grace 1991). We note that whereas the brief stimulus parameter used by Young and Yang (2005; 15 pulses delivered at 50 Hz) is within the range of a typical burst activity of single ventral tegmental area (VTA) neurons (<15 spikes at 15–200 Hz; Freeman and Bunney 1987; Kiyatkin and Rebec 1998, 2001), when a cluster of VTA neurons is taken into account, the duration of an afferent burst activity can range to ≥2 s (Kosobud et al. 1994). Indeed, 1.7- to 2-s trains of pulses (at 20–100 Hz) delivered to the VTA have been used to phasically release dopamine in the PFC (Gurden et al. 1999; Lavin et al. 2005). Therefore, we tested first whether the present 50-Hz stimulation (duration of 2 s) evokes endogenous release of dopamine in our slice preparation and triggers plasticity. As has been already shown in Figure 1B, a 4-time application of 50-Hz tetanus did not induce lasting synaptic plasticity. However, we found that the same tetanus, when repeated 6 times, induces LTD (Fig. 2A, −16 ± 5.8% 40 min after tetani, n = 7, P < 0.005 compared with Fig. 1B). Importantly, this LTD was completely blocked by combined application of D1 (SCH23390 2 μM) and D2 (sulpiride 20 μM) receptor antagonists (Fig. 2B), or by sole application of D1 or D2 antagonist (Fig. 2C), during the tetani (SCH23390 + sulpiride condition, 5.5 ± 9.3% 40 min after tetani, n = 7, P < 0.05 compared with Fig. 2A; SCH23390 condition, 41 ± 19%, n = 6, P < 0.05; sulpiride condition, 30 ± 18%, n = 6, P < 0.05). A significant group effect in the postconditioning period was also found by ANOVA analysis between the nonantagonist LTD group (Fig. 2A) and the SCH23390 or sulpiride group (F1,11 = 9.117, P < 0.02, and F1,11 = 17.722, P < 0.005, respectively). These results indicate that the 50-Hz (2 s) extracellular stimulation evokes endogenous release of dopamine that acts on D1 and D2 receptors to trigger LTD.
The Effect of Endogenous Dopamine Changes in the Presence of Background Dopamine
Our next question was whether this effect of endogenous dopamine changes in the presence of tonic/background dopamine. To test this, we repeated the aforementioned LTD-inducing 6-time tetanus protocol (Fig. 2A) in the presence of 3 μM background dopamine (40-min pre-tetani perfusion). Under this condition (Fig. 2D black squares), the 50-Hz tetani induced LTP, but not LTD (37 ± 17% 40 min after tetani, n = 7, P < 0.02 compared with nondopamine LTD group in Fig. 2A, P < 0.05 compared with nontetanus dopamine control group in Fig. 2D, 2-tailed t-test). The ANOVA analysis revealed a significant group difference in postconditioning period between the 2 groups depicted in Figure 2D (F1,12 = 5.630, P < 0.04). Importantly, this tonic dopamine–facilitated LTP was again completely blocked by the application of D1 or D2 receptor antagonist (3- to 5-min application just before and during tetani, Fig. 2E; SCH23390 group, −5.9 ± 6.6% 40 min after tetani, n = 6, P < 0.05 compared with LTP group in Fig. 2D and P > 0.9 compared with nontetani group in Fig. 2D; sulpiride group, −7.4 ± 7.1%, n = 6, P < 0.05 and P > 0.8, respectively). ANOVA revealed a strong trend for the group effect in the postconditioning period between the SCH23390 group (Fig. 2E) and the LTP group depicted in Figure 2D (F1,11 = 4.110, P = 0.068) and a significant group effect between the sulpiride group (Fig. 2E) and the LTP group (Fig. 2D; F1,11 = 6.585, P < 0.03). These results indicate that endogenous fast dopamine triggered LTD in the absence of background dopamine (Fig. 2A) but instead triggered LTP in its presence. In addition, we verified whether the LTP induced by 4-time application of tetanus in the presence of 3 μM dopamine (as depicted in Fig. 1D) also depends on endogenous dopamine released upon the delivery of tetani. As shown in Figure 2F, this LTP was also blocked by D1 or D2 antagonist applied during the tetani (SCH23390 condition, −10 ± 6.0%, n = 7, P < 0.02 compared with LTP group in Fig. 1D, P > 0.4 compared with nontetani group in Fig. 1D; sulpiride condition, 8.0 ± 8.8%, n = 7, P > 0.3 compared with nontetani group in Fig. 1D). The summary of these series of experiments is shown in Figure 3. Thus, the endogenous fast dopamine that triggers no plasticity or LTD when tonic/background dopamine is absent (Figs 1B and 2A) triggers LTP in the presence of tonic/background dopamine (Figs 1D and 2D).
To gain some insight into cellular mechanisms underlying this LTP trigger, we additionally performed the analysis on the postsynaptic depolarization evoked during tetanus delivery. It is thus possible that tonic/background dopamine enhances postsynaptic excitability (Yang and Seamans 1996; Tseng and O'Donnell 2004), which is manifested by endogenous dopaminergic input evoked by tetanus and triggers LTP. To express the degree of postsynaptic depolarization, we quantified the number of spikes evoked to a tetanus as well as the area of membrane depolarization above the basal membrane potential during the 2-s period of tetanus delivery [i.e., Σtf(t) dt, 0 < t < 2000 ms]. First, we found significant increases in the number of spikes evoked by single tetanus under 3 μM tonic/background dopamine (i.e., the LTP groups depicted in Figs 1D and 2D) compared with the nondopamine non-LTP groups (depicted in Figs 1B and 2A) in each of the 6 tetanus episodes. The range of spike number in dopamine condition was 2.7 ± 0.9 to 3.3 ± 0.4 spikes per tetanus, whereas that in nondopamine condition was 0.4 ± 0.3 to 1.2 ± 0.2 spikes per tetanus (P < 0.05 in each of 6 episodes, t-test). The increase in the area of membrane depolarization in the dopamine LTP condition showed a statistical significance for the first tetanus episode (P < 0.025) and a trend for the second episode (P = 0.08) compared with the nondopamine non-LTP condition. Notably, such enhancements of depolarization did not occur when N-methyl-D-aspartate (NMDA) receptor antagonist DL-AP5 (100 μM) was present in the bath during tetani (i.e., non-LTP condition, see Fig. 4A and the following section, P > 0.4 compared with the nondopamine non-LTP condition or AP5 control condition; the range of spike number in the AP5 condition 0.9 ± 0.6 to 1.1 ± 0.4 spikes per tetanus). However, importantly, unlike AP5, D1 antagonist SCH23390 (2 μM) or D2 antagonist sulpiride (20 μM), which were applied during tetani and completely blocked LTP (Fig. 2E,F), did not reduce at all the increases in these depolarization parameters achieved by 3 μM tonic dopamine. The range of spike number in SCH23390 group was 4.8 ± 0.7 to 5.4 ± 1.1 spikes per tetanus and that of sulpiride group was 3.0 ± 0.9 to 4.2 ± 1.0 spikes per tetanus (in both groups P < 0.05 compared with the nondopamine non-LTP condition for each of 6 tetanus episodes). The increase in the area of membrane depolarization in these groups also reached a statistical significance in the first and second tetanus episodes (SCH23390 group, P < 0.02) or in the first 4 tetanus episodes (sulpiride group, P < 0.05). Furthermore, interestingly, these depolarization enhancements were still seen under 1 or 10 μM dopamine perfusion (see Fig. 1E,F), which are other non-LTP conditions (P < 0.01 compared with the nondopamine, non-LTP group). Moreover, consistent increases in spike number and depolarization were in fact found in the other, subsequently tested drug/manipulation conditions where LTP was blocked (i.e., Figs 4B and 5C,D). These results taken together indicate that although the tonic dopamine indeed enhances postsynaptic depolarization in an NMDA receptor–dependent manner (cf., Seamans et al. 2001; Tseng and O'Donnell 2004), such an enhancement is insufficient for LTP. Also, the endogenous fast dopamine does not appear to contribute to such a depolarization enhancement during tetanus.
Dopamine-Facilitated LTP Requires NMDA Receptor Activation
LTP Is Blocked by Inhibition or Reduction of NMDA Receptor Activation
Dysfunction of NMDA subtype of glutamate receptors in the PFC may be another underlying factor for cognitive deficits seen in schizophrenia (Lewis and Gonzalez-Burgos 2006). Such a dysfunction may be associated with impaired synaptic plasticity. It is therefore important to test whether the present dopamine-facilitated LTP requires the activation of NMDA receptors. In the initial pilot study, the NMDA receptor antagonist DL-AP5 (100 μM) was applied throughout the experiment and found to block LTP (n = 8, P < 0.05, not shown). We were however keen to reveal more temporally specific involvements of NMDA receptors in LTP induction, that is, tetani-associated induction phase versus the pre-tetani tonic dopamine perfusion phase. First, to test NMDA receptor involvement in the induction phase, we started DL-AP5 perfusion just 5 min prior to the delivery of 50-Hz tetani following a 40-min tonic perfusion of 3 μM dopamine (Fig. 4A). LTP was blocked under this condition (Fig. 4A black triangles, −6.3 ± 10% 40 min after tetani, n = 7, P > 0.7 compared with nontetani control by t-test or no group effect detected during the postconditioning period, F1,10 = 0.532, P > 0.4 by ANOVA). This result indicates that NMDA receptor activation by 50-Hz tetani is necessary for the induction of LTP. Second, we tested the involvement of NMDA receptors during pre-tetani tonic dopamine perfusion phase. Because it is difficult to pharmacologically block NMDA receptors only during the pre-tetani phase due to slow drug washout, we reduced synaptic activation of NMDA receptors during pre-tetani phase by stopping 0.033-Hz test pulses. Test pulses were stopped for 27.5 min from the start of the pre-tetani dopamine perfusion (Fig. 4B), and in order to reestablish the baseline prior to tetani, the test pulses were resumed 12.5 min before tetani delivery (Fig. 4B). Our previous result (Matsuda et al. 2006) showed that this 12.5-min perfusion of 3 μM dopamine with test pulses applied is itself insufficient for LTP. Under this condition (Fig. 4B), LTP was completely absent (−2.0 ± 8.5% 40 min after tetani, n = 7, P < 0.05 compared with LTP group in Fig. 1D by t-test; a group effect by ANOVA, F1,12 = 6.608, P < 0.05). This result supports the possibility that repeated synaptic activation of NMDA receptors by test pulses (>12.5 but <40 min) during tonic dopamine perfusion phase is necessary for the LTP.
NMDA Receptor–Mediated Synaptic Transmission Is Unaffected by Background Dopamine
We further tested whether the tonic/background perfusion of dopamine potentiates the basal NMDA receptor–mediated synaptic transmission because NMDA-induced currents or NMDA receptor–mediated synaptic responses were reported to potentiate after D1 receptor stimulation in the PFC (Zheng et al. 1999; Seamans et al. 2001; Chen et al. 2004). Such a D1 receptor–mediated potentiation may contribute to LTP induction (Gurden et al. 2000). We thus isolated NMDA receptor–mediated synaptic responses by adding CNQX (10 μM) and bicuculline (2 μM) in the bath and by increasing the stimulus intensity from 31 to 76 μA on average. Under this condition, a 40-min perfusion of 3 μM dopamine induced no potentiation of NMDA receptor–mediated synaptic responses (n = 6). Instead, the NMDA receptor–mediated responses nonsignificantly decreased after dopamine perfusion (Fig. 4C,D; from 0.28 to 0.22 mV/ms on average at 40 min after dopamine perfusion, P = 0.080, paired t-test). One study reported that dopamine at concentrations of 10 μM or higher potentiates NMDA-induced membrane currents (Zheng et al. 1999). We therefore used dopamine at 10 μM and tested whether it increases the isolated NMDA receptor–mediated evoked synaptic potential. Again, dopamine failed to induce potentiation of NMDA receptor–mediated responses (Fig. 4C, n = 5; from 0.19 to 0.14 mV/ms on average, n = 5, P > 0.3 paired t-test). The differences from the previous studies may be derived from different origins of stimulating fibers (layer I–II vs. layer V; Seamans et al. 2001) or the method used to evoke NMDA receptor–mediated responses (synaptic stimulation vs. direct application of agonist NMDA; Zheng et al. 1999; Chen et al. 2004). Our recent observation found indeed a large difference of synaptic efficacy modification in response to another catecholamine noradrenaline between layer I–II and layer V–VI glutamatergic synapses (A Marzo, J Bai, J Caboche, P Vanhoutte, and S Otani, unpublished data). Whatever the case, we conclude that in the present condition, the basal NMDA receptor–mediated synaptic response does not undergo potentiation by dopamine.
The Effective Concentration of Background Dopamine Activates ERKs to Facilitate LTP
What is an underlying molecular mechanism for the LTP-facilitating effect of tonic/background dopamine? We focused on ERK1 and ERK2 MAP kinases because these molecular messengers are implicated in synaptic plasticity in various brain regions (Sweatt 2004) and because in the PFC ERK activation acts to consolidate long-term memory of behavioral organization and adaptation (Runyan et al. 2004; Hugues et al. 2006). Furthermore, we have previously shown that dopamine (100 μM) activates ERKs through D1 and D2 receptors in the PFC (Otani et al. 1999). By the use of western blot analysis with phospho-specific antibodies, we measured the phosphorylated form of ERK1 and ERK2 in the PFC (the prelimbic area, see Fig. 1A) after tonic application of 3 μM dopamine. All experiments were performed identically to the electrophysiological experiments described above except the intracellular recording, including the delivery of 0.033-Hz single test pulses (stimulus intensity was set at the mean intensity calculated from the last 20 electrophysiological experiments; typically ∼40 μA). First, as shown in Figure 5A, phosphorylation of the ERKs was significantly enhanced after 40-min perfusion of 3 μM dopamine (P < 0.0001, 2-tailed t-test), but not a 12.5-min perfusion (P > 0.1) which is a non-LTP condition (Matsuda et al. 2006). These slow ERK activations are selective because p38, another MAP kinase subtype known to be involved in synaptic plasticity (Butler et al. 2004), showed no such increases of phosphorylation in response to dopamine (Fig. 5A inset, P > 0.1). Furthermore, a 40-min perfusion of 1 or 10 μM dopamine, the other non-LTP conditions (Fig. 1E,F), failed to induce significant activations of the ERKs (Fig. 5B; #P < 0.05 and ##P < 0.01 compared with 3 μM dopamine group). These ERK activations occur through both D1 and D2 receptor stimulation by dopamine because 40-min perfusion of D1 agonist SKF38393 (2–3 μM) or D2 agonist quinpirole (2–3 μM) increased the ERK activity compared with the control (33 ± 7.4% ERK2 increase, n = 4, P < 0.01, and 28 ± 14% ERK2 increase, n = 4, P < 0.05, respectively). The combination of the 2 induced on average a larger activity increase (48 ± 20%, n = 4, P < 0.05 compared with control), but it is statistically not different from the SKF38393 or quinpirole alone group (P > 0.1). These results indicate that the increase of the ERK activity by 3 μM dopamine is mediated by both D1-like and D2-like receptors.
We further showed that the delivery of 50-Hz tetani following 40-min perfusion of 3 μM dopamine evoked no further increases of ERK activity (Fig. 5A, P > 0.1), indicating that the tetani and associated endogenous dopamine activate other factors to trigger LTP. However, the accumulated increases of the ERK activity by the tonic/background dopamine are causally related to LTP because the specific ERK inhibitor PD98059 (20 μM), applied just for 5 min, reduced the ERK phosphorylation to the control level (Fig. 5A, P < 0.0002) and blocked the LTP (Fig. 5C; −3.6 ± 8.6% 40 min after tetani, n = 7, P < 0.05 compared with Fig. 1D by t-test; a group effect by ANOVA F1,11 = 6.450, P < 0.05). Importantly, however, unlike PD98059, 3- to 5-min application of SCH23390 (2 μM) and sulpiride (20 μM) during tetani, which also blocked LTP (Fig. 2E,F), did not reduce the ERK activation achieved by 40-min prior perfusion of 3 μM dopamine (n = 3, data not shown), suggesting that LTP block by these dopamine antagonists is not due to a cancellation of increased ERK activation. Finally, critical ERK activations occur in the postsynaptic sites because postsynaptic infusion of specific ERK substrate synthetic peptide (Otani et al. 1999) blocked LTP (Fig. 5D black triangles; −1.2 ± 14% 40 min after tetani, n = 7, P > 0.7 compared with nontetani control, gray squares). In these experiments, a sufficient duration of time (>45 min) was allowed to pass between the penetration of a neuron and the application of tetani in order to guarantee postsynaptic diffusion of the peptide (cf., Otani et al. 1999; see Fig. 5 legend for details). These results together indicate that slowly occurring postsynaptic activations of the ERKs by an appropriate level of tonic/background dopamine, acting through D1 and D2 receptors, set the condition for LTP induction.
Clinical and behavioral pharmacological studies indicate critical roles of tonic or background dopamine for the regulation of PFC-mediated cognitive function (Schultz 2002; Seamans and Yang 2004; Lewis and Gonzalez-Burgos 2006). To date, however, it is not yet entirely clear how changes in the level of tonic/background dopamine modify the function of the PFC. Using an in vitro model system, we showed that an appropriate concentration of exogenously applied tonic dopamine facilitates the induction of LTP in PFC glutamatergic synapses. Without this dopamine, the conditioning stimuli failed to trigger LTP or triggered rather LTD. Our molecular analysis indicated that the effective level of tonic/background dopamine slowly activates the postsynaptic ERK pathway, a critical intracellular cascade implicated in long-term memory formation in the PFC (Runyan et al. 2004; Hugues et al. 2006; Nagai et al. 2007).
The importance of neuroplasticity in physiological and pathophysiological adaptations of brain networks receives strong current attention (Lewis and Gonzalez-Burgos 2008). For example, by the use of transcranial magnetic stimulation method, an impaired LTP has been detected in the brain of schizophrenia patients (Frantseva et al. 2008), and dopaminergic manipulation was shown to enhance LTP in the brain of healthy human subjects (Kuo et al. 2008). In rodents, on the other hand, chronic restraint stress that reduces the basal dopamine level in the PFC and impairs working memory (Mizoguchi et al. 2000) was shown to block LTP and instead facilitate LTD in the cognitively important hippocampus–PFC pathway (Goto et al. 2007). The present study reinforces these previous studies by providing the evidence that, by the use of in vitro model system, a deviated level of tonic/background dopamine may indeed impair normal induction of LTP in the PFC.
A possible criticism, which may be addressed to the present study, is that all experiments were performed in the slices prepared from 23- to 30-day-old juvenile rats. The utilization of juvenile rats may limit the generalization of our results to the model of cellular mechanisms underlying schizophrenia cognitive symptoms, which manifest in the majority of cases during adolescence or early adulthood. However, schizophrenia is likely to be a developmental disorder (Lewis and Levitt 2002; Kellendonk et al. 2006) where their cognitive abnormalities can be detected earlier in life, well before clinical onset of the symptoms (Muratori et al. 2005; Lewis and Gonzalez-Burgos 2006). Moreover, we have recently verified that the 23- to 30-day-old rats perform the PFC-dependent attentional-set shifting task (Birrell and Brown 2000) as well as adult rats do (J Bai, A Marzo, B Giros, M Nosten-Bertrand, E Tzavara, and S Otani, unpublished data), indicating that the PFC of these young rats is matured enough to support the higher cognitive functions. We suggest therefore that our results still provide useful information about the physiology and pathophysiology of executive cognitive function.
Another point that may require attention is the fact that the experiments were conducted in the presence of GABA-A receptor antagonist bicuculline (1 μM). The addition of GABA-A receptor antagonist has been generally practiced in plasticity research in order to reduce synaptic inhibition and effectively isolate excitatory transmission (Wigström and Gustafsson 1983; Artola and Singer 1987; Hirsch and Crepel 1990). This was practiced in the present study also to directly compare results with those in our preceding experiments (Otani et al. 1998, 1999; Matsuda et al. 2006). Whereas we do not know to what extent the presence of bicuculline affected plasticity induction in these studies, our preliminary data using a modified stimulus protocol (6 trains of 5 burst pulses at 50 Hz) indicate that LTP is indeed facilitated by bicuculline (B Kolomiets and S Otani, unpublished data). Whether it is the case for the present stimulus protocols remains an important future investigation.
A main finding of the present study is that the ERK activation and the LTP facilitation by tonic/background dopamine followed an inverted-U shape curve. The relatively narrow concentration window for the effective dopamine (>1 but <10 μM) provides a cellular and molecular basis to at least partly explain the subtle regulation of PFC function by this neuromodulator (Seamans and Yang 2004; Williams and Castner 2006; Goto et al. 2007). Particularly, ERK activation plays a key role in memory formation in the PFC (Runyan et al. 2004; Hugues et al. 2006; Nagai et al. 2007), which can be indeed mediated by (D1) dopaminergic inputs to the PFC (Nagai et al. 2007). But we note that ERK activation by dopamine is also involved in LTD induction in the PFC (Otani et al. 1999), similar to the hippocampal case where both LTP and LTD require the ERKs (English and Sweatt 1997; Thiels et al. 2002). We thus by no means suggest that the ERK activation is sufficient for LTP: instead, the ERKs appear to gate the induction of plasticity by playing a key role in various molecular networks that involve diverse different factors (Girault et al. 2006). Indeed, although the present ERK activation can be mostly achieved by either D1 or D2 receptors (see Results), these receptors must be concurrently activated to facilitate LTP (Fig. 1C2). D1 receptor–mediated cellular changes are well described in the PFC such as NMDA receptor upregulation (Seamans et al. 2001; Chen et al. 2004; Tseng and O'Donnell 2004) and increase of the insertion and surface expression of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (Sun et al. 2005), but D2 receptors may also be coupled to critical cellular effects (Vial and Piomelli. 1995; Nilsson et al. 1998). Importantly, however, in frontal neurons, a link between concurrent activation of D1 and D2 receptors and phospholipase C–mediated postsynaptic [Ca2+] rises appears to exist (Lee et al. 2004). A synergistic production of arachidonic acid by D1 and D2 receptor stimulation also has been described in Chinese hamster ovary cells (Piomelli et al. 1991). Behaviorally, requirement of D1 and D2 receptors in the PFC was found using a memory-guided goal direction task (Goto and Grace 2008). Collectively, we suggest that concurrent stimulation of D1 and D2 receptors may potentiate phospholipase C/protein kinase C pathway in addition to the ERK activation and contribute to plasticity and memory-related behavior. We note further that our previous detection of ERK activation after high concentration of dopamine (100 μM, LTD condition) was rapid (<2 min; Otani et al. 1999) but that the low-concentration dopamine-induced increase of background ERK activity proceeds slowly (>12.5 min, Fig. 5A), perhaps depending on the 0.033-Hz test synaptic inputs acting on NMDA receptors (see Fig. 4B).
Young and Yang (2005) pharmacologically verified that 50-Hz stimuli evoke endogenous fast release of dopamine and trigger synaptic plasticity in PFC slices. Our results supported this finding and extended it by showing that the effect of this endogenous dopamine depends on the level of tonic/background dopamine (Fig. 3). It is important to note, however, that the 50-Hz conditioning stimuli did not enhance the ERK activity any further than the level already achieved by 3 μM background dopamine (Fig. 5A). Therefore, currently, cellular/molecular actions of the endogenous fast release of dopamine are also unknown. It may stimulate other molecular factors such as protein kinase A (Gurden et al. 2000).
Lines of evidence now reinforce the classically held view (Fuster 1995) that PFC-mediated cognitive function depends on certain types of long-term memory (Runyan et al. 2004; Takehara-Nishiuchi et al. 2006; Touzani et al. 2007). Although yet speculative, we have suggested that long-term neuronal traces are formed within the PFC through synaptic plasticity inductions and encode the cognitive aspects of the temporal organization of behavior, which enable one to perform context-dependent flexible adaptations in goal direction (Otani 2002; Goto and Otani 2007). The present study presented a novel piece of evidence that for a successful induction of use-dependent synaptic potentiation in the PFC, the tonic/background concentration of dopamine has to be tuned within an appropriate range so that it serves to upregulate critical postsynaptic factors to secure LTP induction.
Centre National de la Recherche Scientifique (CNRS); Institut National de la Sante et de la Recherche Medicale (INSERM); University of Paris VI; French Minister of Research (SO, AM); The Sophia University Open Research Center Grant (SO); International Brain Research Organization (IBRO) (BK).
We thank Drs Yukiori Goto, Cliff Abraham, Angelo Arleo, Susan Sara, Bruno Giros, Jean-Michel Deniau, Bail Lu, and Charles Yang for their helpful comments on the manuscript and Dr Therese Cronin for English correction. Conflict of Interest: None declared.