Dopaminergic Suppression of Synaptic Transmission in the Lateral Entorhinal Cortex

Dopaminergic projections to the superficial layers of the lateral entorhinal cortex can modulate the strength of olfactory inputs to the region. We have found that low concentrations of dopamine facilitate field EPSPs in the entorhinal cortex, and that higher concentrations of dopamine suppress synaptic responses. Here, we have used whole-cell current clamp recordings from layer II neurons to determine the mechanisms of the suppression. Dopamine (10 to 50 μM) hyperpolarized membrane potential and reversibly suppressed the amplitude of EPSPs evoked by layer I stimulation. Both AMPA- and NMDA-mediated components were suppressed, and paired-pulse facilitation was also enhanced indicating that the suppression is mediated largely by reduced glutamate release. Blockade of D2-like receptors greatly reduced the suppression of EPSPs. Dopamine also lowered input resistance, and reduced the number of action potentials evoked by depolarizing current steps. The drop in input resistance was mediated by activation of D1-like receptors, and was prevented by blocking K+ channels with TEA. The dopaminergic suppression of synaptic transmission is therefore mediated by a D2 receptor-dependent reduction in transmitter release, and a D1 receptor-dependent increase in a K+ conductance. This suppression of EPSPs may dampen the strength of sensory inputs during periods of elevated mesocortical dopamine activity.


INTRODUCTION
The entorhinal cortex is animportant interface that links primary sensory and association cortices to the hippocampal formation, and it is critical for the sensory and mnemonic functions of the medial temporal lobe [1][2][3][4]. In the rat, the lateral division of the entorhinal cortex receives most of its cortical inputs from the olfactory cortex and perirhinal cortex, and the medial entorhinal cortex receives visual and multimodal inputs mainly via the postrhinal cortex [5][6][7]. This pattern of cortical input to the medial and lateral divisions of the entorhinal cortex contributes to their different roles in sensory and cognitive processing [8][9][10]. In addition, neuromodulatory transmitters innervate both the medial and lateral entorhinal cortices and can have powerful effects on sensory and mnemonic function in these regions. Specifically, acetylcholine and serotonin both modulate synaptic transmission and rhythmic EEG activities in the medial entorhinal cortex [11][12][13][14][15]. Further, midbrain dopamine neurons send one of their largest cortical projections to the superficial layers of the lateral entorhinal cortex where they target principal cell islands [16][17][18]. Relatively little is known, however, regarding the neuromodulatory effects of dopamine in the lateral entorhinal cortex.
The large dopaminergic projection to the prefrontal cortex is known to regulate cellular processes related to working memory [19][20][21], and dopaminergic inputs to the lateral entorhinal cortex are also likely to affect mechanisms of sensory and mnemonic function. In the prefrontal cortex, activation of D 1 receptors can suppress glutamate release in layer V [22][23][24] but can enhance glutamatergic transmission in layer III [25,26]. Further, the positive effect of D 1 receptor activation on working memory follows an inverted Ushaped function [27], and strong or weak stimulation of D 1 receptors can also have opposite effects on NMDA receptormediated synaptic currents [20,28]. We have also found that dopamine has dose-dependent bidirectional effects in layer II of the lateral entorhinal cortex. In awake animals, increasing levels of dopamine with a selective reuptake inhibitor facilitates synaptic responses evoked by stimulation of the piriform cortex, and field excitatory postsynaptic potentials (fEPSPs) are also facilitated by a low concentration of dopamine in vitro [29]. Higher concentrations of dopamine, however, suppress fEPSPs, and similar suppression effects have been observed by others in medial entorhinal cortex layer II [30] and layer III [31]. Dopamine can also reduce the input resistance of layer IIneurons in the medial entorhinal cortex [30] and reduce temporal summation in layer V neurons of the lateral division through an increase in the I h current [32]. Dopamine may therefore modulate synaptic function in the lateral entorhinal cortex through multiple mechanisms.
We have used whole-cell current clamp recordings to investigate the mechanisms of the suppression of EPSPs by dopamine in electrophysiologically identified "fan" cells in layer II of the lateral entorhinal cortex. Receptor blockers were used to determine the dopamine receptors that mediate the suppression of EPSPs, and paired-pulse tests were used to assess whether the suppression is expressed pre-or postsynaptically. Changes in the intrinsic excitability of fan cells were also monitored using responses to hyperpolarizing and depolarizing current steps. In addition to a D 2like receptor-mediated suppression of transmitter release, we show evidence that EPSPs are also reduced by an increased K + conductance dependent on activation of D 1 receptors.

Stimulation and recording
Patch recording pipettes for whole cell recordings were prepared from borosilicate glass (1.0 mm OD, 4 to 8 MΩ) using a horizontal puller (P-97, Sutter Instr., Calif, USA) and were filled with a solution containing (in mM) 140 K-gluconate, 5 NaCl, 2 MgCl 2 , 10 HEPES, 0.5 EGTA, 2 ATP-Tris, and 0.4 GTP-Tris (pH adjusted to 7.24-7.32 with KOH; 270-280 mOsm). Pipettes were placed in contact with somata of layer II neurons, and gentle suction was applied under voltage clamp to form a tight seal (1-3 GΩ). Whole cell configuration was achieved by increased suction, and experiments began after cells stabilized (typically within 3 to 5 minutes after break-in). Current clamp recordings were obtained using an Axopatch 200B amplifier (Axon Instr., Calif, USA) and displayed on a digital oscilloscope (Gould 1604). Recordings were filtered at 10 kHz and digitized at 20 kHz (Axon Instr., Digidata 1322A) for storage on computer hard disk. Recordings were accepted if the series resistance was ≤25 MΩ (mean = 16.9 ± 0.9 MΩ) and if input resistance and resting potential were stable. A bipolar stimulating electrode made from two tungsten electrodes (FHC, 1.0 MΩ) was positioned to span layer I near the border with layer II approximately 0.2 to 0.6 mm rostral to the recording electrode. Synaptic responses were evoked with 0.1 millisecond constant current pulses delivered using a stimulus timer and isolation unit (WPI, Mass, USA, models A300 and A360). Stimulation intensity was adjusted to evoke responses approximately 75% of maximal (75 to 300 μA).
All neurons (n = 118) included for analyses were identified as "fan" cells based on electrophysiological characteristics described previously [40,41]. In comparison tostellate cells of the medial entorhinal cortex, fan cells show modest inward rectification during hyperpolarizing current steps, a small depolarizing afterpotential following single spikes, and do not show prominent theta-frequency membrane potential oscillations at subthreshold voltages [40][41][42].

Dopaminergic modulation of synaptic responses
The effects of dopamine on glutamate-mediated synaptic transmission in the lateral entorhinal cortex are largely uncharacterized. We therefore recorded both mixed and isolated components of excitatory postsynaptic potentials (EPSPs) evoked by stimulation of layer I before and after 5-minute bath-application of 1, 10, or 50 μM dopamine. Results obtained using high concentrations of dopamine must be interpreted cautiously because of the possibility of nonspecific effects. However, dopamine degrades through oxidization within the slice preparation, and similar concentrations of dopamine have been used previously, and interpreted in light of the effects of specific antagonists, in reports examining the effects of dopamine on synaptic transmission in both the entorhinal [29][30][31][32] and prefrontal [23,43] cortices. Responses were evoked once every 20 seconds, and the mean of 10 responses was obtained for analysis. Baseline responses were obtained at resting potential and, because dopamine usually hyperpolarizes fan cells, constant current was often required to return cells to the original membrane potential for recordings in the presence of dopamine. Sodium metabisulfite (50 μM) was coapplied to slow the oxidation of dopamine [29,31,43], and ambient lighting was also reduced. Possible effects of sodium metabisulfite were assessed with a vehicle control group. Drugs were routinely stored at −20 • C as concentrated stock solutions until needed, but dopamine HCl was dissolved just prior to bath application.
Paired-pulse tests were used to determine whether dopamine modulates EPSPs through a pre-or postsynaptic mechanism [13]. Pairs of stimulation pulses separated by an interval of 30 milliseconds were delivered before and after 5-minute bath-application of 1, 10, or 50 μM dopamine. Stimulation intensity was reduced to evoke EPSPs approximately 50% of maximal, and ten responses were averaged for analyses. Paired-pulse facilitation was quantified by expressing the amplitude of the second response as a percentage of the first response.
Mechanisms mediating the suppression of EPSPs by high concentrations of dopamine were investigated by assessing the effects of 50 μM dopamine on isolated components of synaptic responses. After baseline recordings in normal ACSF, AMPA receptor-mediated responses were isolated with bath application of 50 μM 2-amino-5-phosphonovaleric acid (APV) and 25 μM bicuculline methiodide, or NMDA receptor-mediated responses were isolated with 20 μM 7-nitro-2,3-dioxo-1,4-dihydroquinoxaline-6-carbonitrile (CNQX) and 25 μM bicuculline. GABA-mediated IPSPs were isolated with either 1 mM kynurenic acid or 20 μM CNQX with 50 μM APV. Isolated synaptic responses were recorded before and after 5-minute application of 50 μM dopamine. Isolated AMPA receptor-mediated responses were also used to determine if dopamine suppresses EPSPs primarily throughD 1 -or D 2 -like receptors. Baseline responses were recorded in the presence of either the D 1 receptor antagonist SCH23390 (50 μM) or the D 2 receptor antagonist sulpiride (50 μM) [29][30][31], and 50 μM dopamine was then applied for 5 minutes. Sulpiride was prepared daily in a stock solution of 6% DMSO in ACSF titrated with 0.1 N HCl, and there was a final concentration of 0.1% DMSO with sulpiride.
The effects of dopamine on the intrinsic excitability of fan cells were assessed by monitoring responses to hyperpolarizing and depolarizing current steps. Changes in action potentials, afterhyperpolarizations, input resistance and inward rectification were examined before and after 5-minute bath application of 1, 10, or 50 μM dopamine. The number of action potentials elicited in response to suprathreshold current injection can be used to characterize neuronal excitability [32], and we therefore determined the number of spikes fired in response to a single 500 millisecond-duration depolarizing current pulse from a constant holding potential (typically rest) using a pulse amplitude that elicited 3 to 5 action potentials [32]. Receptors that mediate the dopamine-induced change in input resistance were investigated using SCH23390 or sulpiride, and the ionic conductances involved were assessed using 0.5 μM tetrodotoxin (TTX) or 30 mM tetraethylammonium (TEA). Blockers were preapplied for 5-10 minutes prior to coapplication of dopamine for 5 minutes.

Data analysis
Electrophysiological characteristics of fan cells and changes in synaptic responses were analyzed using the software program Clampfit 8.2 (Axon Instr., Calif, USA). The amplitudes of averaged EPSPs were measured relative to the prestimulus baseline, and paired-pulse facilitation was determined by expressing the amplitude of the second response as a proportion of the amplitude of the first response. Action potential amplitude was measured from resting potential, and action potential width and fast and medium afterhyperpolarizations were measured from threshold. Input resistance was calculated by measuring peak and steady-state voltage responses to −200 pA current steps (500 milliseconds), and inward rectification was quantified by expressing the peak input resistance as a proportion of the steady-state resistance (rectification ratio). All data were expressed as the mean ±SEM for plotting, and changes in response properties were assessed using paired samples ttests or mixed design ANOVAs.

Electroresponsiveness of layer II fan cells
A total of 118 fan cells in layer II of the lateral entorhinal cortex were identified electrophysiologically and included for analysis, and the characteristics of these cells were similar to those reported previously [40,41]. Fan cells had a mean resting membrane potential of −58.8 ± 0.6 mV, and a peak input resistance of 99.1 ± 2.1 MΩ. Most cells (108 of 118) demonstrated a small delayed inward rectification in response to hyperpolarizing current steps (rectification ratio: 1.10 ± 0.01). Action potentials (amplitude: 128.8 ± 0.7 mV, width: 4.1 ± 0.1 milliseconds, threshold: −44.1 ± 0.8 mV) were typically followed by fast and medium afterhyperpolarizations (−3.3 ± 0.3 mV and −5.8 ± 0.3 mV) with a small depolarizing afterpotential. Averaged EPSPs evoked by stimulation of layer I had a mean amplitude of 4.4 ± 0.2 mV. Continuous recordings of membrane potential were obtained in a subset of 28 cells to assess subthreshold membrane potential oscillations and, similar to findings of Tahvildari and Alonso [40], fan cells did not display prominent oscillations (data not shown).

Dopaminergic modulation of EPSPs
We previously found concentration-dependent effects of dopamine on field EPSPs in layer II in vitro, in which 10 μM dopamine facilitated fEPSPs and 50 to 100 μM dopamine suppressed fEPSPs [29]. We obtained similar concentrationdependent effects in whole cell EPSPs recorded here before and after 5-minute bath application of dopamine. Application of 50 μM dopamine resulted in a strong suppression of synaptic response to 38.5 ± 5.8% of baseline levels (see Figure 1(a); t 8 = 7.75, P < .001; n = 9) that could be reversed by 15 minutes washout in normal ACSF (3 cells). We initially expected 10 μM dopamine to facilitate EPSPs [29], but foundthat10 μM dopamine instead caused a small synaptic suppression (to 87.0 ± 5.8% of baseline; see Figure 1 = 2.31, P < .05; n = 18). However, a lower concentration of 1 μM dopamine significantly enhanced responses to 132.7 ± 4.4% of baseline levels (see Figure 1(c); t 6 = 5.04, P < .01; n = 7). In our previous study using a gas-fluid interface chamber, a larger bath volume and slower flow-rate may have increased dopamine oxidation and reduced the effective concentration of dopamine at the slice, and this may account for why a higher applied concentration facilitated responses in that study [29]. Bath application of the antioxidant sodium metabisulfite alone had no significant effect on the amplitude of whole cell EPSPs (see Figure 1(d); n = 8). Paired-pulse tests were used to determine if synaptic suppression and facilitation effects were likely expressed preor postsynaptically. Pairs of pulses were delivered before and after 5-minute dopamine application, and a 30-millisecond interpulse interval was used that results in optimal pairedpulse facilitation [13,[44][45][46]. If EPSPs are reduced through a reduction in transmitter release, then a greater amount of transmitter should be available for release in response to the second stimulation pulse and paired-pulse facilitation should be enhanced [47][48][49]. Changes in EPSPs mediated by alterations in postsynaptic receptors, however, should not be associated with changes in paired-pulse ratio. High concentrations of dopamine that reduced EPSP amplitude were also found to enhance paired-pulse facilitation (see Figures 2(a), 2(b); t 13 = 2.78, P < .05 for 10 μM; t 8 = 2.97, P < .05 for 50 μM), suggesting that dopamine reduced EPSPs by suppressing glutamate release. In contrast, the low concentration of 1 μM dopamine that facilitated EPSPs had no significant effect on paired pulse facilitation (see Figure 2(c)), suggesting that the facilitation of EPSPs was mediated primarily by an increased postsynaptic response to glutamate. The dopaminergic facilitation of the conditioning response was smaller during paired-pulse tests in which stimulus intensity was reduced to avoid spiking (see Figures 1(c) versus 2(c)) but a similar dopaminergic facilitation of fEPSPs with no effect on paired-pulse ratio has been observed in the entorhinal cortex in vivo [29].

Isolated synaptic responses
The suppression of EPSPs by high concentrations of dopamine was examined more closely using pharmacologically isolated synaptic responses. Consistent with a suppression of glutamate release from presynaptic terminals, bath application of 50 μM dopamine significantly attenuated both the isolated AMPA-and NMDA-mediated responses. The NMDA component was reduced to 26.0 ± 7.5% of baseline (see Figure 3(b); t 7 = 3.32, P < .05; n = 8) and the AMPA component was reduced to 41.7 ± 5.6% of baseline (see Figure 3(a); t 5 = 3.50, P < .05; n = 6). Dopamine receptor subtypes underlying the suppression of AMPA-mediated synaptic responses were investigated by applying 50 μM dopamine in the presence of either the D 1 receptor antagonist SCH23390 (50 μM) or the D 2 receptor antagonist sulpiride (50 μM). Similar to previous reports that have used selective agonists in the medial [30,31] and lateral [29] entorhinal cortex, application of either the D 1 agonist SKF38393 (25 to 50 μM; n = 9) or the D 2 agonist quinpirole (20 to 40 μM; n = 10) had no effect on EPSPs (data not shown), and we therefore used receptor blockers known to affect synaptic responses in the lateral entorhinal cortex [29]. Application of antagonists alone had no effect on EPSPs, and the D 1 antagonist SCH23390 did not block the suppression of AMPA-mediated EPSPs (see Figure 4(a); t 4 = 3.0, P < .05; n = 5), suggesting that D 1 receptors do not mediate the suppression. However, blockade of D 2 receptors with sulpiride significantly reduced the effects of dopamine on AMPA-mediated EPSPs. Coapplication of dopamine with sulpiride (n = 5) resulted in a nonsignificant suppression of synaptic responses, and the size of the suppression was significantly smaller than that observed with dopamine alone (79.8 ± 7.2% versus 41.7 ± 5.6% of baseline; F 1,9 = 18.10, P < .001; see Figure 4(b1)). Sulpiride also prevented the enhancement of paired-pulse facilitation induced by 50 μM dopamine (see Figure 4(b2)). Although this indicates that the dopaminergic suppression of EPSPs is largely dependent upon activation of D 2 -like receptors, the suppression of responses in the presence of sulpiride was close to statistical significance (t 4 = 2.65, P = .06), suggesting that a non-D 2 receptor-mediated mechanism mediates the residual suppression.

Dopaminergic suppression of IPSPs
Biphasic IPSPs were recorded from fan cells held near action potential threshold (−51 to −48 mV) and exposed to either 1 mM kynurenic acid or a combination of 50 μM APV and 20 μM CNQX to block ionotropic glutamate transmission. A concentration of 50 μM dopamine suppressed both the early GABA A -and late GABA B -mediated components of the IPSP.
The early IPSP was reduced to 84.5 ± 8.7% of baseline levels, and the late IPSP was reduced to 62.3 ± 11.1% of baseline levels (see Figure 5(b);early, t 8 = 2.41, P < .05, n = 9; late, t 7 = 2.46, P < .05, n = 8). The dopaminergic suppression of GABA synapses indicates that the reduction of EPSPs by dopamine is unlikely to be due to increased GABAergic inhibition of fan cells.

Modulation of intrinsic excitability
Bath application of dopamine also hyperpolarized resting membrane potential and reduced the input resistance of fan cells. Membrane potential was increased from −56.1 ± 2.0 to −59.7 ± 1.4 mV (see Figure 6(a); t 8 = 4.73, P < .001; n = 9), and peak input resistance was reduced from 90.3 ± 7.6 to 68.9 ± 3.1 MΩ by 50 μM dopamine (see Figure 6(b); t 7 = 4.27, P < .01; n = 8). Similar changes in membrane potential and input resistance were observed for 10 μM dopamine (not shown) and have also been reported following application of high concentrations of dopamine in whole-cell recordings from medial entorhinal cortex stellate cells [30]. Changes were not due to the vehicle, because control cells and cells exposed to 1 μM dopamine did not show a drop in input resistance or hyperpolarization of membrane potential. In layer V entorhinal cortex cells dopamine causes a reduction in excitability and a drop in input resistance through an increase in the hyperpolarization-activated current I h [32], and changes in I h were therefore assessed in layer II fan cells. However, dopamine did not significantly affect the amount of inward rectification, and the rectification ratio remained stable (see Figure 6(d); 1.09 ± 0.02 in ACSF and in 50 μM dopamine, t 7 = 0.00, P = 1.00).
Dopamine suppressed the excitability of fan cells, and application of 10 and 50 μM dopamine reduced the number of action potentials evoked by brief 500 milliseconds depolarizing current pulses (see Figure 7). The number of spikes was reduced from 4.1 ± 0.1 to 2.8 ± 0.5 spikes by 10 μM dopamine (see Figure 7(b); t 17 = 2.54, P < .05; n = 18). A higher 50 μM concentration of dopamine caused a similar reduction in the number of spikes (from 3.9 ± 0.2 to 2.8 ± 0.6) that was not statistically significant (t 8 = 1.82, P = .11; n = 9). The reduction in spiking could result in part from reduced input resistance, but it was not due to membrane hyperpolarization because cells were tested at the same membrane potential both before and after dopamine application.
The drop in input resistance induced by 50 μM dopamine was blocked by coapplication of the D 1 receptor antagonist SCH23390 (and there was actually a very small but reliable increase in R in in 4 of 5 cells; t 4 = 2.60, P = .06; see Figure 8(a)). The drop in input resistance was not affected by coapplication of the D 2 receptor antagonist sulpiride (t 4 = 9.71, P < .001; n = 5; Figure 8(b)). The reduction in input resistance induced by dopamine is therefore dependent on activation of D 1 , but not D 2 , receptors.
The conductances that mediated the reduced input resistance were investigated using blockers of Na + and K + channels. The Na + channel blocker TTX was used to verify that reductions in input resistance were not due to an increase in action potential-dependent synaptic inputs to fan cells, or due to an altered Na + conductance. Blockade of Na + channels with TTX did not prevent the drop in input resistance induced by dopamine (see Figure 9(a); peak, t 4 = 6.02, P < .01; steady-state, t 4 = 8.21, P < .01; n = 5). It has been suggested that the reduced input resistance induced by dopamine in medial entorhinal cortex stellate cells might be mediated by an increased K + conductance [30], and we therefore assessed the effects of dopamine on input resistance in the presence of the K + channel blocker TEA (30 mM; n = 5). Coapplication of TEA blocked the reduction in input resistance induced by dopamine (see Figure 9(b)), indicating that the D 1 receptor-dependent reduction in input resistance involves an increased K + conductance. The increased K + conductance is likely to contribute to the hyperpolarization of membrane potential induced by dopamine, and may also account for the reduced excitability of fan cells (see Figure 7). The reduced input resistance may also contribute to the dopamine-induced suppression of EPSPs; the D 2 receptor blocker sulpiride did not fully prevent the suppression of AMPA-mediated EPSPs (see Figure 4(b1)), and the D 1 receptor-mediated reduction in input resistance could contribute to part of the EPSP suppression.

DISCUSSION
We show here that dopamine has powerful suppressive effects on glutamate-mediated synaptic transmission in layer II fan cells ofthe lateral entorhinal cortex. Our findings suggest that the suppression of EPSPs involves the combined actions of a D 2 receptor-mediated reduction in neurotransmitter release and a D 1 receptor-mediated increase in a K + conductance that reduces cellular input resistance. Previously, we found that field EPSPs were enhanced by low concentrations of dopamine in vitro, and by blocking dopamine reuptake in awake animals [29]. This suggested that moderate increases in dopamine release might facilitate synaptic responses in the entorhinal cortex, and enhance transmission of sensory information to the rest of the hippocampal formation.
Here, we have replicated the synaptic facilitation with a low 1 μM concentration of dopamine and have also shown that high concentrations of dopamine induce a strong and reversible suppression of intracellular EPSPs. Similar suppression effects have been observed in the medial entorhinal cortex [30,31] and prefrontal cortex [22,23,50,51] using comparable doses of dopamine.

Suppression of glutamate release
The suppression of EPSPs by high concentrations of dopamine was found to be largely dependent on D 2 receptors since coapplication of the D 2 receptor antagonist sulpiride blocked most of the reduction. Dopamine also enhanced paired-pulse facilitation which suggests that the suppression of EPSPs resulted from a reduction in presynaptic glutamate release [47,49]. The suppression of both AMPA-and NMDA-mediated components of the synaptic response is also consistent with reduced transmitter release. Although similar reductions in EPSPs have been shown in stellate cells of the medial entorhinal cortex, the suppression was dependent on D 1 , and not D 2 , receptor activation [30]. However, Stenkamp et al. (1998) showed a reduction in synaptic responses in layer III of the medial entorhinal cortex through activation of both D 1 and D 2 receptors, and results of paired-pulse tests in their study suggested that the suppression was also mediated by reduced glutamate release. Dopamine has been shown to suppress AMPA-mediated synaptic responses in the prefrontal cortex through a D 1 receptor-mediated suppression of transmitter release [22][23][24]. Strong activation of D 1 receptors can also suppress synaptic responses through a retrograde signaling cascade. Weak D 1 receptor activation can enhance NMDA responses, but stronger D 1 receptor activation can lead to more intense NMDA receptor activation and the release of adenosine that suppresses transmitter release by acting on presynaptic A 1 receptors that suppress voltage-gated Ca 2+ channels [28,52,53]. In the striatum, activation of presynaptic D 2 receptors suppresses N-type Ca 2+ currents and inhibits acetylcholine release from striatal cholinergic interneurons [54]. D 2 receptors have also been linked to a suppression of responses in the parabrachial nucleus [55], ventral tegmental area [56], and striatum [57,58] via a D 2 -mediated reduction in glutamate release. A similar D 2 -mediated mechanism underlies the suppression of GABA release from striatal inhibitory cells onto cholinergic interneurons [59]. Similar mechanisms may mediate the dopaminergic suppression of glutamate release in the entorhinal cortex.
The dopaminergic suppression of EPSPs observed here cannot be explained by increased transmission at GABA synapses because we found that dopamine reduced monosynaptic GABA A and GABA B IPSPs. The suppression is also unlikely to be due to increased activation of feedback inhibition [60] because dopamine reduced both glutamatergic transmission and the number of spikes in fan cells (see Figure 7). The suppression of monosynaptic IPSPs that we observed may have resulted from a D 2 -mediated reduction in GABA release [59,61] and reduced input resistance in fan cells could also have contributed. These possibilities are consistent with the parallel reductions observed in GABA A and GABA B IPSPs. Recordings of spontaneous and/or miniature IPSCs would be useful to determine the mechanisms of the reduced IPSPs.

Modulation of intrinsic excitability
In addition to the D 2 -mediated suppression of transmitter release, high concentrations of dopamine also appear to sup-press synaptic transmission through a D 1 -receptor dependent mechanism. Sulpiride did not completely block the suppression of EPSPs (see Figure 4(b1)), and a D 1 receptordependent activation of a TEA-sensitive K + conductance appears to mediate the residual suppression via a reduction in input resistance. Blockade of synaptic transmission and voltage-gated Na + channels with TTX did not prevent the drop in input resistance induced by dopamine indicating that it is not due to increased spontaneous synaptic drive or to an increased Na + conductance. However, the broadly acting K + channel blocker TEA prevented the drop in input resistance, indicating that dopamine activates a K + conductance. The drop in input resistance was also prevented by blockade of D 1 , but not D 2 , receptors, indicating that dopamine activates K + channels via D 1 receptors. High concentrations of dopamine also hyperpolarize membrane potential and reduce input resistance in stellate cells of the medial entorhinal cortex, and it was also suggested that these changes might be mediated by an increased K + conductance [30]. A large number of K + conductancesare affected by TEA, and it is therefore not clear which type(s) may be responsible for the drop in input resistance observed here. Background leak channels are insensitive to TEA [62] and are therefore not likely to contribute. Voltage-gated K + currents are blocked by TEA, but dopamine in the prefrontal cortex tends to enhance neuronal excitability by suppressing these currents (see also [43,63]). Several reports in CA1 pyramidal cells have found that dopamine hyperpolarizes membrane potential, reduces input resistance, and increases afterhyperpolarizations through a D 1 -receptor mediated increase in Ca 2+ -activated K + currents ( [64,65], see also [66]), but others have found an increase in the excitability of CA1 neurons due to a suppression of Ca 2+ -activated K + currents (see also [32,67,68]). Here, there was no clear increase in afterhyperpolarizations, suggesting that Ca 2+dependent K + currents do not mediate the change in input resistance. Activation of D 1 receptors can also have dosedependent effects on activation of inward rectifying K + currents (IRKCs). In the prefrontal cortex, D 1 receptor activation typically inhibits IRKC by direct effects of cAMP on IRK channels, but strong activation can increase IRKC via phosphorylation of the channels through elevated levels of PKA [69]. This could explain why a significant reduction in input resistance was observed here only at the higher concentrations of dopamine. Clearly, however, further experiments will be required to determine the nature of the D 1 receptordependent K + conductance in fan cells.
We observed a decrease in fan cell firing during depolarizing current steps after dopamine, and the reduced spiking may reflect the drop in cellular input resistance. A surprising finding was that while the D 1 receptor antagonist SCH23390 prevented the dopamine-induced reduction in input resistance it did not completely eliminate the reduction in the number of spikes, suggesting that reduced input resistance cannot entirely account for the reduction in spiking, and that other mechanisms may also contribute. D 1 receptor activation can increase spiking in prefrontal neurons by enhancing the persistent Na + current (I NaP ) and suppressing a slowly-inactivating K + conductance [43,70], but a suppression of spiking via a reduction in I NaP has also been observed [71]. In layer V entorhinal cortex neurons, dopamine reduces input resistance and leads to a reduction of spiking though an increase in I h [32].
Here, there was no apparent change in I h in fan cells, and action potential threshold and afterhyperpolarizations were not affected, suggesting that the underlying currents were not modified. Dopaminergic effects on I NaP were not directly assessed in the present study, and the drop in input resistance could mask possible reductions in depolarizing responses to current injection related to I NaP . However, in tests in which SCH23390 prevented a change in input resistance, we found no reduction in the response to +20 pA pulses. This argues against a D 1 -mediated reduction in I NaP , but it is still possible that dopamine may reduce spiking via a D 2 receptor-mediated reduction in I NaP [71].

CONCLUSIONS
We have shown here that dopamine has concentrationdependent, bidirectional effects on glutamate-mediated synaptic transmission in principal cells of layer II of the lateral entorhinal cortex. The lateral entorhinal cortex receives a major input from the piriform cortex [5][6][7], and dopaminergic innervation of the superficial layers is likely to have a strong modulatory effect on olfactory processing. In the prefrontal cortex, moderate activation of dopaminergic inputs promotes workingmemory function, but excessive dopamine activation leads to a decrement in performance [20,27]. In the entorhinal cortex, moderate increases in dopamine concentration may enhance the salience of olfactory representations carried to the lateral entorhinal cortex (see Figure 1(c); see also 29), but large increases in dopamine associated with drug effects or acute stress [27] may dampen synaptic inputs to the superficial layers and suppress working memory function [72][73][74] or induction of lasting synaptic plasticity [75]. The dopaminergic suppression of synaptic transmission in layer II is also likely to inhibit the propagation of sensory information to the rest of the hippocampal formation such that only strong and synchronous inputs to the entorhinal region may be sufficient to activate entorhinal projection neurons.