The hippocampus and prefrontal cortex are two structures implicated in learning and memory and are related through a direct excitatory pathway. The characteristics of the synaptic influence of the hippocampus on pyramidal cells of the prefrontal cortex were determined using intracellular recordings in anesthetized rats. Single-pulse stimulation of the hippocampus induced an early EPSP of fixed latency in most of the recorded pyramidal cells (n = 106/116) thereby demonstrating a monosynaptic connection between hippocampal neurons and pyramidal cells of the prefrontal cortex. Furthermore, the EPSP was followed by a prolonged IPSP and suggests a simultaneous engagement of pyramidal and non-pyramidal neurons that may ultimately constrain the spread of excitation in response to hippocampal input. Paired-pulse stimulation induced short-term modifications in the synaptic responses and this short-term plasticity may contribute to the temporal filtering of information. Finally, tetanic stimulation of the hippocampus produced long-term potentiation of the monosynaptic EPSP with a concomitant potentiation of the IPSP, indicating that the hippocampo-prefrontal network can participate in the formation and consolidation of memories. In conclusion, the characteristics of the synaptic transmission in the hippocampo-prefrontal cortex pathway further supports the existence of a cooperative relationship between two structures known to be involved in higher cognitive processes.
Clinical and experimental studies indicate that both the hippocampus and the prefrontal cortex (PFC) are critically implicated in several aspects of learning and memory (Squire, 1992; Goldman-Rakic, 1994, 1995; Fuster, 1997). The PFC is involved in working memory and contributes to the temporal ordering of spatial and non-spatial events as well as to the organization and planning of responses. For example, increased cellular discharge occurs in the PFC during the delay period of delayed response tasks in the rat and the monkey (Sakurai and Sugimoto, 1986; Batuev et al., 1990; Goldman-Rakic, 1995). The hippocampus plays a crucial role in the encoding and formation of memories, its involvement being only temporary since memories are gradually translated to long-term neocortical sites (Squire, 1992).
A direct hippocampo-PFC pathway has been described in the monkey, the cat and the rat (Rosene and Van Hoesene, 1977; Swanson, 1981; Irle and Markowitsch, 1982; Cavada et al., 1983; Goldman-Rakic et al., 1984; Férino et al., 1987). A cooperative relationship between the hippocampus and the PFC in working memory processes has been suggested. In the rat, the role of the hippocampo-PFC pathway has been specifically addressed by studying performance in a delayed radial maze task, following unilateral lesion of the ventral subiculum in combination with a contralateral lesion of the PFC. These combined lesions produced a disruption of foraging only during the test phase of this delayed win-shift task, similar to that described after bilateral inactivation of the prelimbic (PL) area (Seamans et al., 1995), suggesting that transmission of information between the hippocampus and the PFC is required when trial-unique short-term memory is used to guide prospective search behavior (Floresco et al., 1997). More recently, using a similar asymmetric disconnection model, the involvement of the hippocampo-PFC pathway has also been reported in operant learning (Izaki et al., 2000). Finally, several behavioral studies in rats with restricted lesion of the PL area, a prefrontal region that receives direct hippocampal inputs, indicate that the PL area is involved in working memory and planning of motor response strategies (Kolb et al., 1974; Kesner and Gray, 1989; Dunnett, 1990; Seamans et al., 1995; Delatour and Gisquet-Verrier, 2000; Dias and Aggleton, 2000; Ragozzino and Kesner, 2001). Thus, the hippocampo-PFC pathway likely provides an essential circuit through which spatial information can be integrated into cognitive and motor planning processes mediated by the PFC (Goldman-Rakic, 1987; Fuster, 1991; Doyère et al., 1993). In the rat, the hippocampal innervation of the PFC originates from the temporal CA1/subiculum region and is mainly restricted to the prelimbic, medial orbital and infralimbic areas (Jay et al., 1989; Jay and Witter, 1991; Condé et al., 1995). Hippocampal nerve terminals form primarily asymmetric synapses on spiny pyramidal neurons (Carr and Sesack, 1996). Using retrograde transport of d-[3H]aspartate and extracellular unit recordings, it has been shown that the hippocampo-PFC pathway is glutamatergic and that excitatory responses evoked in prefrontal cells by single pulse stimulation of the hippocampus are primarily mediated by activation of AMPA receptors (Jay et al., 1992). In addition, paired-pulse stimulation can produce a short-term facilitation of these excitatory responses (Laroche et al., 1990). Finally, indicating that the hippocampo-PFC pathway supports long-term potentiation (LTP), the field potential evoked in the PFC by single pulse stimulation following tetanic stimulation of the hippocampus shows a rapid and sustained increase in amplitude (Laroche et al., 1990; Jay et al., 1995).
The present study was undertaken to further characterize the synaptic influence of hippocampal afferents in pyramidal cells of the PFC using intracellular recordings in anesthetized rats. Since three main classes of pyramidal cells, i.e. regular spiking (RS), inactivating bursting (IB) and non-inactivating bursting (NIB) have recently been distinguished in the rat PFC in vivo (Dégenètais et al., 2002), synaptic responses evoked by single pulse stimulation of CA1/subiculum in electrophysiologically and morphologically identified pyramidal cells were analyzed. Short-term and long-term synaptic modifications of these synaptic responses were also studied using paired-pulse or tetanic stimulation of the hippocampus, respectively.
Material and Methods
Experiments were conducted in 77 adult male Sprague–Dawley rats weighing 275–300 g. Animals were initially anesthetized with sodium pentobarbital (66 mg/kg, i.p.) and mounted in a stereotaxic apparatus (Horseley Clark, LPC, Asnières, France). Anesthesia was maintained throughout the experiments by additional doses of sodium pentobarbital (22 mg, i.p.) administered hourly. Level of anesthesia was assessed by testing the limb withdrawal reflex and additional doses of anesthetic were injected, if needed, to ensure areflexia. Wounds and pressure points were repeatedly infiltrated with lidocaine (xylocaïne 2%). Stability of recordings was ensured by cisternal drainage. Body temperature was maintained at 36.5°C with a homeothermic blanket.
Intracellular recordings of PFC neurons were made with glass micro-pipettes (50–80 MΩ) filled with 1 M K-acetate containing 1% neurobiotin to achieve labeling of recorded neurons. Ten neurons were recorded with microelectrodes filled with 1 M KCl, 1% neurobiotin instead of K-acetate in order to assess effects of chloride equilibrium potential modifications on inhibitory components of the responses. Recordings were performed in the prelimbic and medial orbital (PL/MO) areas of the PFC using the following stereotaxic coordinates: anterior: 3–4 mm from the bregma; lateral: 0.4–1 mm from the midline; and depth: 2–4 mm from the cortical surface (Paxinos and Watson, 1986). All recordings were obtained using an Axoclamp-2B amplifier (Axon Instruments, Foster City, CA) operated in the bridge mode. Impalements of neurons were considered acceptable when membrane potential was at least −60 mV and spike amplitude >50 mV. Signals were stored on digital audiotape (DTR-1800, Biologic, Claix, France) and subsequently digitized using a CED 1401 interface. Data were analyzed off-line using Spike2 software (Cambridge Electronic Design, Cambridge, UK).
Stimulation of the CA1/subiculum region of the hippocampus was made with a bipolar coaxial stainless steel electrode (300 μm diameter, 300 μm tip-barrel distance) stereotaxically positioned at the following coordinates: anterior: +2.5–3.5 mm from the interaural line; lateral: 5–5.5 mm from the midline; and depth: 4.5–5.5 mm from the cortical surface (Paxinos and Watson, 1986). Single pulse stimulation (300 μs duration, 0.1–0.5 mA intensity) was applied at the frequency of 0.5 Hz. Higher intensities (0.5–0.9 mA) were only used to study the influence of stimulation intensity on the synaptic responses.
Paired-pulse stimulation of the hippocampus, with increasing inter-stimulus intervals (ISI) from 40 to 800 ms, was applied at a frequency of 0.5 Hz. To analyze whether the net effect was facilitation or depression of the excitatory post-synaptic potential (EPSP), amplitudes of both the first and the second EPSP were measured with reference to the resting membrane potential and the absolute difference in amplitude between the first and the second EPSP was then calculated (Buonomano and Merzenich, 1998). To allow across-cells comparison, this difference was expressed as a percentage of the first EPSP amplitude. A similar analysis was performed for inhibitory post-synaptic potentials (IPSP). The data are expressed as mean ± SD. Student’s unpaired t-test was used to compare the mean amplitudes of the first and second post-synaptic potentials.
Tetanic stimulation of the hippocampus was performed to induce long-term modifications of synaptic transmission. High-frequency stimulation consisted of two series of 10 trains separated by 6 min. Trains were applied at a frequency of 0.1 Hz and consisted of 250 Hz, 300 μs pulses applied during 200 ms. The stimulation intensity of test pulses was adjusted before tetanic stimulation so that responses were 50% of the maximal response amplitude. The same intensity was used during tetanic stimulation.
Finally, recorded pyramidal cells were classified according to their firing patterns in response to intracellular application of prolonged (400 ms) depolarizing current pulses of increasing intensity, as previously described (Dégenètais et al., 2002). Resting membrane potential was calculated by subtraction of the tip potential occurring when the microelectrode was withdrawn from the neuron.
Following the electrophysiological characterization of the neuron, an intracellular injection of neurobiotin was performed by passing depolarizing pulses (100 ms, 0.3–0.6 nA, 1 Hz) for 5–15 min. Then, rats were deeply anesthetized with sodium pentobarbital (200 mg/kg, i.p.) and perfused intra-cardiacally with a 0.9% saline solution containing 1% sodium nitrite, followed by fixative solution (4% paraformaldehyde/0.1% glutaraldehyde in 0.1 M sodium phosphate buffer). Brains were removed and following 4 h post-fixation in a 4% paraformaldehyde phosphate-buffered solution, stored for 48 h in 20% phosphate-buffered sucrose. Frontal sections (50 μm) were cut on a freezing microtome and collected in 0.1 M potassium phosphate-buffered saline (pH 7.4). After three washes in 0.1 M sodium phosphate buffer (pH 7.4), slices were incubated 12 h in 1% avidin–biotin complex (ABC Kit Standard, Vector Laboratories, Burlingame, CA) in the presence of 0.5% Triton X-100. After three rinses in phosphate buffer, they were reacted in diaminobenzidine (1%) and cobalt chloride (1%)/nickel ammonium sulfate (1%)/H2O2 (0.01%) solution. The position of the stimulating electrode was marked by electrical deposit of iron (6 μA positive current, 20 s) and observed on histological sections following a ferri-ferrocyanide reaction.
Synaptic responses induced by stimulation of the ventral CA1/subiculum region of the hippocampus (Fig. 1) were studied in 116 neurons located in layers II to VI of PL/MO areas of the PFC. All recorded cells were electrophysiologically characterized. A subset of these cells (n = 41) was injected with neurobiotin for morphological characterization. All the labeled cells had morphological features of pyramidal neurons. As previously reported, three main classes of pyramidal cells were distinguished according to their firing pattern in response to the application of depolarizing intracellular current pulses: regular spiking (RS, n = 82), non-inactivating bursting (NIB, n = 26) and inactivating–bursting (IB, n = 8) cells (Dégenètais et al., 2002). Among the 30 labeled RS cells, five were located in layers II–III, 16 in layer V and nine in layer VI. The three labeled IB cells were all located in layer V. Among the eight labeled NIB cells 1 was located in layers II–III, six in layer V and one in layer VI. The synaptic responses induced by single pulse stimulation of hippocampus were analyzed in all the recorded cells. The effect of paired-pulse and tetanic stimulations of hippocampus were tested in 71 and 12 of these neurons, respectively.
Synaptic Responses Evoked by Single Pulse Stimulation of the Hippocampus
Single pulse stimulation of the hippocampus induced in 106 (91%) of the 116 tested cells, a complex synaptic response consisting of an early EPSP followed by a composite inhibitory response (Fig. 2). The early EPSP was also followed by a late EPSP in 58 cells (55%) (Fig. 2C). These complex synaptic responses were observed in 77 of the 82 RS cells (94%), in 21 of the 26 NIB cells (80%) and in the 8 IB cells (100%).
The mean latency of the early EPSP was 14.6 ± 4.0 ms. The latency of the early EPSP remained constant when the stimulation intensity of the hippocampus was gradually enhanced and the amplitude of the EPSP increased while its duration decreased (Fig. 3A,B). The EPSP evoked spikes in only 19% of the cells. When cells were depolarized by the application of a continuous intracellular current of weak intensity, the early EPSP evoked by hippocampal stimulation induced a discharge composed of one or two spikes in all responding cells (Fig. 2A3, B3, C3). The amplitude of the EPSP increased when the cell was hyperpolarized and decreased with depolarization (Fig. 4). The early EPSP presented similar characteristics in the three electrophysiological classes of pyramidal neurons.
In addition to the early EPSP, a late EPSP with an apparent mean latency of 37.3 ± 9.3 ms was observed in 55% of responding pyramidal cells (Fig. 2C). When the hippocampal stimulation intensity was increased, the latency of the late EPSP slightly varied and its amplitude decreased (Fig. 3A,B). In contrast to the early EPSP, the late EPSP did not trigger spikes even during application of a prolonged intracellular depolarizing current pulse or during periods of spontaneous depolarization (Figs 2C and 3D). The late EPSP amplitude decreased when the cell was depolarized and increased when the cell was hyperpolarized (Figs 3D and 4C). Finally, the late EPSP was observed more frequently in RS cells (53%) than in NIB (34%) or IB (37%) cells.
The hyperpolarizing phase that followed the EPSP(s) was of long duration (mean duration of 300 ± 106 ms) and small amplitude at resting membrane potential (Fig. 2). The hyperpolarization phase became obvious when the neuron was depolarized by application of an intracellular current and appeared to comprise an early and a late component (Fig. 2A3, B3, C3). Increasing the intensity of hippocampal stimulation resulted in a larger amplitude and longer duration of these inhibitory components (Fig. 3C). The increased amplitude of the early IPSP was concomitant with a decrease in the late EPSP amplitude as shown in Figure 3A,B.
Hyperpolarizing and depolarizing intracellular current pulses of different intensities were applied in order to measure the reversal potential of the inhibitory components. The mean reversal potential of the early IPSP was −68.9 ± 4.2 mV (n = 19, Fig. 4A). For the cells presenting an early and a late EPSP, the reversal potential of the early IPSP, measured between the early and late EPSP, had a similar value (Fig. 4C). When the recording microelectrode was filled with potassium chloride instead of potassium acetate, the early IPSP was depolarizing and its reversal potential was shifted to a less negative value (Fig. 4B). Even though the late hyperpolarizing component hardly reversed, its amplitude reached zero at −77 ± 5.5 mV and this value was not modified by the replacement of potassium acetate by potassium chloride in the microelectrode (Fig. 4B). These data are consistent with previous studies performed in the neocortical slices indicating that evoked inhibitory responses were composed of an early IPSP largely resulting from a Cl− current through GABAA receptors and of a late IPSP due to K+ currents gated by GABAB receptors (Avoli, 1986; Connors et al., 1988; Hablitz and Sutor, 1990; Higashi et al., 1991; Cox et al., 1992).
Finally, the inhibitory phase of the response presented similar characteristics in the different classes of pyramidal cells and was followed by a depolarizing rebound in 73% of the cells. This rebound triggered spikes in 70% of IB and NIB cells (Fig. 2B,C) and only in 28% of RS cells.
Synaptic Responses Evoked by Paired-pulse Stimulation of the Hippocampus
In most tested cells (n = 67/71, 94%), for increasing inter-stimulus intervals (ISI) from 40 to 800 ms, paired-pulse stimulation (PP) induced a depression and/or a facilitation of the early EPSP. Based on the variation in the amplitude of the early EPSP, three types of cells were distinguished: the PP-D cells (n = 37, 52%) presented a depression (Fig. 5), the PP-DF cells (n = 20, 28%) displayed a depression at short ISIs and a facilitation at longer ISIs (Fig. 6) and finally the PP-F cells (n = 10, 14%) showed a facilitation (Fig. 7). There was no apparent correlation between the type of cell (PP-D, PP-DF and PP-F cells) and the main electrophysiological classes of pyramidal cells (RS, IB, NIB; Table 1). The other components of the synaptic response were modified in some cases, as described below.
In the 37 PP-D cells, the early EPSP showed a maximal depression at ISIs of 40–80 ms and recovered at ISIs ranging from 100 to 400 ms. A complete suppression of the early EPSP was observed in 16 cells, and in the other 21 cells the maximal decrease in the amplitude of the early EPSP was of 68 ± 30% (range, 18–149%, P < 0.001). In the example of the PP-D cell shown in Figure 5A, the maximal decrease of the EPSP amplitude occurred at an ISI of 40 ms, and a progressive recovery was observed at an ISI of up to 240 ms. The depression of the early EPSP was sufficient to block a supra-liminar response at short ISIs (40–80 ms, Fig. 5D).
The late EPSP, that followed the early EPSP in 17 cells, was completely suppressed at short ISIs ranging from 40 to 90–160 ms, and progressively recovered the control amplitude at longer ISIs (180–580 ms, Fig. 5B).
The amplitude and duration of the hyperpolarization phase was decreased in 26 cells and not modified in the 11 remaining cells. The depression of the IPSP occurred at ISIs ranging from 40 ms to 160–400 ms and was maximal (61 ± 17%, range 28–86%, P < 0.0001) at 120–160 ms ISIs (Fig. 5A,C). In some cells, the depression of the IPSP was concomitant with a broadening of the early EPSP. This likely resulted from the reduction of the electric shunt induced by the inhibitory conductance.
In the 20 PP-DF cells, the early EPSP displayed depression at short ISIs and facilitation at longer ISIs. For 40–80 ms ISIs, the early EPSP was suppressed in nine cells and a maximal decrease of 58 ± 38% in amplitude (range 16–284%, P < 0.005) was observed in the 11 other cells. Following recovery of the depression (100–160 ms ISIs), the amplitude of the early EPSP increased by 40 ± 33% (range 14–102%, P < 0.0001) above the control value (160–300 ms ISIs) and then progressively recovered at 300–800 ms ISIs. An example of such a PP-DF cell is illustrated in Figure 6A.
The late EPSP that followed the early EPSP in seven cells was suppressed at short ISIs (40 to 80–160 ms) and recovered at ISIs ranging from 160 to 800 ms. In most cases, the late EPSP recovered at ISIs at which the early EPSP was already potentiated (Fig. 6B). In three cells, a subsequent increase in the amplitude of the late EPSP occurred at ISIs ranging from 180 to 450 ms.
In nine cells, the inhibitory component of the response was depressed at ISIs of 40 ms to 200–350 ms, with a maximal decrease of the IPSP amplitude of 70 ± 23% (range 50–100%, P < 0.0001) at 60–160 ms ISIs (Fig. 6C). The depression of the IPSP occurred both during the periods of depression and facilitation of the early EPSP. In the 11 remaining cells, no obvious modification of the inhibitory component was observed.
In the 10 PP-F cells, a facilitation of the early EPSP was observed at ISIs of 40 ms to 180–400 ms. The maximal increase in the EPSP amplitude (40–180 ms ISIs) was of 54 ± 25% (range 18–87%, P < 0.0005). In the example illustrated in Figure 7, the early EPSP presented a maximal increase in amplitude at an ISI of 120 ms and recovered at an ISI of 400 ms. The late EPSP, that followed the early EPSP in four cells, was suppressed at short ISIs (40 to 80–120 ms) as shown in Figure 7. For longer ISIs (120–160 ms), the late EPSP recovered and could show facilitation at ISIs of up to 300–400 ms. No obvious modification in the inhibitory component was observed except for two cells that presented a small increase in the amplitude of the IPSP for ISIs ranging from 40 ms to 160–300 ms.
Synaptic Responses Evoked after Tetanic Stimulation of the Hippocampus
Twelve neurons were held long enough to test the effect of tetanic stimulation of the hippocampus on the synaptic responses evoked by single pulse stimulation. A long-term potentiation (LTP) of the early EPSP occurred in eight cells (66%) and no obvious modification of the synaptic responses was observed in the four other cells. Among the eight cells exhibiting LTP, seven were RS cells and one was an IB cell. Among the four cells that did not exhibit LTP, three were RS cells and one was a NIB cell. The amplitude of the early EPSP evoked by single pulse stimulation typically presented a sustained increase from 15 min up to at least 60 min after tetanic stimulation (Fig. 8), without modification of the resting membrane potential. The mean increase of the EPSP amplitude was of 37 ± 5% (n = 8) at 60 min.
The IPSP that followed the EPSP showed a rapid and sustained increase in amplitude and duration in four of these cells and was not obviously modified in the four other cells. A marked enhancement of the IPSP amplitude (149% ± 22%, n = 4) was observed from 5 min up to at least 60 min after the tetanic stimulation (Fig. 8). The initial increase in the IPSP was concomitant with a decrease in the EPSP duration suggesting that the IPSP partly shunted the EPSP.
The present in vivo intracellular study shows that, in the PL/MO areas of the rat PFC, the vast majority of pyramidal cells (91%) located in layers II to VI receive a complex synaptic influence from the hippocampus. Synaptic responses evoked by single pulse stimulation of the ventral CA1/subiculum region consisted of an early EPSP and in some cases of a late EPSP that were followed in all cells by a prolonged IPSP. These synaptic responses were observed in the three main electrophysiological classes of pyramidal cells (RS, IB, NIB cells) recently distinguished in the rat PFC (Dégenètais et al., 2002). In addition, paired-pulse and tetanic stimulation of the hippocampus can induce short-term and long-term modifications of the synaptic responses, respectively.
Synaptic Responses Evoked by Single Pulse Stimulation of Hippocampo-PFC Pathway in Pyramidal Cells
The hippocampo-PFC pathway, which originates from the CA1/subiculum region of the hippocampus, is glutamatergic and innervates the different layers of PL/MO areas of the PFC (Férino et al., 1987; Jay et al., 1989, 1992; Jay and Witter, 1991). Hippocampal nerve terminals form primarily asymmetric synapses on dendritic spines and shafts of presumed PFC pyramidal cells (Carr and Sesack, 1996).
The present study further indicates that the three main classes of pyramidal cells of the PL/MO areas receive a monosynaptic excitatory input from the hippocampus. Indeed, the latency of the early EPSP evoked by single pulse stimulation of the CA1/subiculum region in RS, IB and NIB cells is compatible with the conduction time of the hippocampo-PFC pathway (Férino et al., 1987) and does not vary with the strength of the hippocampal stimulation. Excitatory responses evoked in the PFC by low frequency stimulation of the hippocampus are primarily blocked by the iontophoretic application of CNQX, an antagonist of AMPA receptors (Jay et al., 1992), suggesting that the early EPSP is mediated by the activation of AMPA receptors. In addition, the early EPSP decreased in amplitude with depolarization and increased in amplitude with hyperpolarization of the membrane potential, a property that has been described for AMPA mediated early EPSPs in cortical slices (Jones and Baughman, 1988; Hablitz and Sutor, 1990; Higashi et al., 1991; Cox et al., 1992; Hwa and Avoli, 1992; Metherate and Ashe, 1994).
The late EPSP that occurred in a subset of pyramidal cells (55%) is likely distinct from the early EPSP. Indeed, these two EPSPs were separated by an early IPSP, which reversal potential was similar to that measured in cells that did not present a late EPSP. Since cortical pyramidal neurons have an extensive local collateral network, the late EPSP could result from the activation of recurrent collaterals of pyramidal cells that receive a direct excitatory input from the hippocampus. However, it cannot be excluded that the late EPSP also result from the activation of a reverberating circuit within the hippocampus or from the activation of an indirect pathway. The amplitude of the late EPSP greatly decreased when the cell was depolarized and increased when the cell was hyperpolarized. In cortical slices a late EPSP with similar characteristics has been shown to likely result from the activation of NMDA receptors (Sutor and Hablitz, 1989b; Hwa and Avoli, 1992).
EPSPs evoked in cortical pyramidal cells by either stimulation of afferent cortical pathways in vivo (Szente et al., 1988; Agmon and Connors, 1992; Baranyi et al., 1993; Nuñez et al., 1993), stimulation of the white matter or stimulation of the cerebral cortex in vitro (Connors et al., 1982; Howe et al., 1987; Sutor and Hablitz, 1989a; Chagnac-Amitai et al., 1990; de la Peña and Geijo-Barrientos, 1996) are followed by a long-lasting inhibition consisting of a fast and a slow inhibitory component that have been attributed to the activation of GABAergic interneurons. It has been shown that the fast IPSP results from an increase of Cl− conductance through activation of GABAA receptors and the slow IPSP is generated by an increase of a K+ conductance through the activation of GABAB receptors (Kelly and Krnjevic, 1969; Avoli, 1986; Connors et al., 1988; Karlsson et al., 1988; Deisz and Prince, 1989). Similarly, in the present study, excitatory responses were followed by an early and a late component probably corresponding to a fast and a slow IPSP, respectively. Reversal potentials of the early and late components suggest that they result from the activation of GABAA and GABAB receptors, respectively. However, it cannot be excluded that the late inhibitory component was partly due to a disfacilitation process since the early inhibition can induce a decrease of excitatory inputs in pyramidal cells. Such a process could explain why applying hyperpolarizing currents did not reverse the late inhibitory phase.
With increasing intensities of the hippocampal stimulation, an enhancement of the early IPSP and a concomitant decrease of the late EPSP were observed, suggesting that the late EPSP can be partially shunted by the chloride conductance associated with the early IPSP (Metherate and Ashe, 1994). This process could also, in part, explain the absence of a late EPSP in some pyramidal cells.
A recent immunocytochemical and ultrastructural study indicates that axon terminals originating from CA1 establish asymmetrical synaptic junctions with parvalbumine immunoreactive GABA-containing interneurons in the prelimbic area of the PFC (Gabbott et al., 2002). Since these interneurons innervate the soma and proximal dendrites of pyramidal cells (Gabbott et al., 2002), IPSPs evoked in pyramidal cells by hippocampal stimulation likely result from a direct activation of a subpopulation of GABA interneurons. However, it cannot be excluded that cortical interneurons have also been indirectly driven through the activation of the recurrent collateral network of pyramidal cells that receive a direct hippocampal input.
Short-term Modifications of Synaptic Responses Induced by Paired-pulse Stimulation
PP-F of the hippocampo-PFC pathway has been previously described using single unit or field potential recordings (Laroche et al., 1990; Mulder et al., 1997). Confirming these observations, the present study shows that paired-pulse stimulation of the hippocampus can produce short-term facilitation of the monosynaptic EPSP in PFC pyramidal cells. In addition, our results revealed that paired-pulse stimulation can also produce short-term depression of the EPSP as well as modifications of the polysynaptic IPSP.
Synaptic responses evoked in cortical and hippocampal slices presented PP-F of the early EPSP that was concomitant with a depression of the fast IPSP, suggesting that the EPSP facilitation can result from the IPSP depression (Nathan and Lambert, 1991; Metherate and Ashe, 1994; Buonomano and Merzenich, 1998). In contrast, in our study the IPSP was either not modified or varied in the same direction as the EPSP suggesting that the depression or facilitation of the EPSP did not result from a modification of the hyperpolarizing phase.
Facilitation or depression of synaptic responses are thought to be due to calcium-dependent changes in the probability of transmitter release (Katz and Miledi, 1969; Fisher et al., 1997; Zucker, 1999) and can be observed both at excitatory and inhibitory synapses. In the neocortex, dual intracellular recordings have shown that fast EPSPs elicited by a pyramidal neuron in an interneuron exhibited preferentially PP-F (Thomson et al., 1993; Thomson and Deuchars, 1997) and that fast IPSP induced by an interneuron in a pyramidal cell typically displayed PP-D (Thomson et al., 1996; Thomson and Destexhe, 1999). In both cases, the release probability was considered as a major factor that determines the direction of short-term plasticity (facilitation or depression), low probability synapses displaying predominantly facilitation and high probability synapses exhibiting predominantly depression (Thomson et al., 1995; Thomson and Destexhe, 1999; Thomson, 2000). Short-term modification of synaptic transmission is also ‘target-cell-specific’ (Katz et al., 1993; Markram et al., 1998; Reyes et al., 1998; Scanziani et al., 1998). For example, repetitive action potentials of a given pyramidal cell can induce facilitation or depression of the EPSP in two different target cells: an interneuron and a pyramidal cell (Markram et al., 1998) or in two distinct classes of interneurons (Reyes et al., 1998).
In the present study, following paired-pulse stimulation of the hippocampo-prefrontal pathway, depression or facilitation of the monosynaptic EPSP was observed in PFC pyramidal cells. However, no apparent correlation between the direction of the short-term synaptic plasticity and the electrophysiological classes of pyramidal cells could be established. Thus, different factors such as the characteristics of the presynaptic axon terminals, the localization of the synapses on the dendrites, or their density can determine the direction of short-term plasticity of the hippocampo-PFC transmission in a given pyramidal cell.
Paired-pulse stimulation of the hippocampus also produced short-term facilitation or depression of the polysynaptic components (late EPSP and IPSP) of the responses in some pyramidal cells. Since the late EPSP is likely due to activation of the recurrent collateral network of pyramidal cells, facilitation or depression of this component may result from modifications at the level of the direct hippocampo-prefrontal synapses and/or at the level of synapses established by recurrent collaterals of the pyramidal neurons. Similarly, the IPSP results from the activation of cortical interneurons driven either directly by the hippocampo-prefrontal pathway and/or indirectly through recurrent axon collaterals of the pyramidal cells. Thus, paired-pulse modifications of the IPSP can result from facilitation or depression of the synaptic transmission at different levels of this circuit. Finally, it cannot be excluded that decrease or increase in the early EPSP can produce changes in the recruitment of pyramidal cells and/or interneurons involved in the local network, resulting in modifications of the late EPSP and of the inhibitory phase of the hippocampal responses.
Long-term Modifications of Synaptic Responses Following Tetanic Stimulation
Since the first report of Bliss and Lomo (Bliss and Lomo, 1973) showing that brief high-frequency stimulation produces long-term potentiation of synaptic transmission, this experimental protocol has been largely used to study the processes leading to long-term potentiation. At glutamatergic synapses, high-frequency stimulation of presynaptic fibers can produce, through the activation of NMDA receptors, a postsynaptic Ca2+ influx that initiates a series of biochemical reactions leading to long-term modification of synaptic strength [for a review, see Malenka and Nicoll (Malenka and Nicoll, 1999)]. Evidence for long-term potentiation of hippocampo-PFC transmission was first provided by studies showing that tetanic stimulation of CA1/subiculum induces a sustained increase in the field potential evoked in the PFC by single pulse stimulation (Laroche et al., 1990). As expected, the induction of LTP in the glutamatergic hippocampal-PFC pathway was shown to be an NMDA receptor-dependent process (Jay et al., 1995). Further confirming that the excitatory hippocampo-PFC pathway can support long-term potentiation, our results show at a cellular level that hippocampal tetanic stimulation induces a sustained enhancement of the monosynaptic EPSP evoked in PFC pyramidal cells. The enhancement of the EPSP likely does not result from a concomitant depression of the IPSP since no change or a long-lasting increase in the amplitude of the IPSP was observed during the maintenance phase. However, it cannot be excluded that a transient depression of the inhibitory component contributes to the LTP induction (Davies et al., 1991).
The processes leading to the sustained and large enhancement in the IPSP observed in four pyramidal cells could result from an increased activation of GABAergic interneurons due to a long-term potentiation of excitatory synapses on these interneurons. Alternatively, since long-term plasticity of synaptic transmission in inhibitory connections has also been described (Buzsáki and Eidelberg, 1982; Kairiss et al., 1987; Taube and Schwartzkroin, 1987; Komatsu, 1994; Shew et al., 2000), increased activation of GABAergic interneurons could also lead to long-term potentiation of the inhibitory connections between interneurons and pyramidal cells. Finally, since interneurons can be driven by the recurrent collaterals of the pyramidal cells, the long-lasting enhancement of the IPSP can also result from an increased recruitment of pyramidal cells due to the potentiation of the early EPSP induced in pyramidal cells following tetanic stimulation of the hippocampus.
The present study shows that most PFC pyramidal cells receive a direct excitatory synaptic influence from the CA1/subiculum region. Also suggesting that hippocampus exerts a major excitatory influence on PFC pyramidal cells, recent data indicate that membrane potential of PFC pyramidal cells was affected following lesion of the ventral hippocampus. Indeed, plateau depolarization of membrane potential (‘up state’) that occur in chloral hydrate anesthetized rats were no longer observed (O’Donnell et al., 2002). On the other hand, our study shows that the excitatory influence of the hippocampo-PFC pathway on pyramidal cells is limited by a simultaneous activation of the GABA interneurons, a property that could lead to a more transient response of pyramidal cells to excitatory inputs. This would allow for spatial focalization of hippocampal excitatory signals to a limited group of PFC cells by suppressing the propagation of intracortical excitations through the recurrent collateral network of pyramidal cells.
Synapses from the hippocampo-PFC pathway displayed alterations in strength during paired-pulse stimulation. Indeed, the monosynaptic EPSP evoked in pyramidal cells by hippocampal stimulation exhibited depression, facilitation or both depression and facilitation. As recently proposed, such short-term plasticity may contribute to the temporal filtering of information (Fortune and Rose, 2000, 2001). By attenuating high frequency influx but not isolated or low frequency impulses, short-term depression of EPSP could play the role of a low-pass filter. In contrast, short-term facilitation would favor high frequency influx.
Short-term modifications of the inhibitory phase were also observed in some pyramidal cells. Amplitude changes of the EPSP and IPSP generally occurred in the same direction, suggesting that the balance between excitation and inhibition is preserved. Thus, the processing of hippocampal information by the PFC is complex and depends on both the pattern of discharge of efferent hippocampal cells (single spike to burst firing) and the interplay between facilitation and depression of synaptic strength within the cortical network.
Finally, high frequency stimulation of the hippocampus can elicit a long-term potentiation of the direct synaptic excitatory transmission and in some cases of the indirect inhibitory transmission via local interneurons. Synchronized high frequency discharge of the output neurons of CA1/subiculum occurs during consummatory behavior and slow-wave sleep (Buzsáki and Eidelberg, 1983; Wilson and McNaughton, 1994; Chrobak and Buzsáki, 1996). As recently described, a subset of neurons in CA1 coactivated during a given behavior were selectively reactivated during subsequent periods of slow-wave sleep and in particular during sharp waves (Pavlides and Winson, 1989; Wilson and McNaughton, 1994; Qin et al., 1997; Kudrimoti et al., 1999; Nádasdy et al., 1999; Lee and Wilson, 2002). Furthermore, the existence of temporal correlations between hippocampal ripples and cortical spindles has been described during slow-wave sleep (Siapas and Wilson, 1998). Thus, discharge of CA1/subiculum neurons during these high frequency oscillations may provide, through the hippocampo-prefrontal pathway, the depolarizing force needed to produce long-term synaptic alterations in the PFC network. Through this process, the hippocampo-PFC network can participate in the formation and consolidation of memories. The hippocampus plays a crucial role in the formation and encoding of memories, its involvement is only temporary as memories are gradually translated to long-term neocortical sites (Squire, 1992). Therefore, in this context, the hippocampo-PFC pathway occupies a key position in learning and memory.
In conclusion, the characteristics of the synaptic influence of the hippocampo-PFC pathway on pyramidal cells further support the existence of a cooperative relationship between two structures known to be implicated in higher cognitive functions and particularly in memory processes.
The authors would like to thank A.M. Godeheu and A. Menetrey for histological assistance and P. Tierney for critical reading of the manuscript. E. Dégenètais is recipient of a fellowship from the Ministère de l’Enseignement supérieur et de la Recherche. This work was supported by the Institut National de la Santé et de la Recherche Médicale (INSERM).
Address correspondence to Dr Y. Gioanni, Chaire de Neuropharmacologie, INSERM U114, Collège de France, 11 place Marcelin Berthelot, 75231 Paris Cedex 05, France. Email: email@example.com.
|Electrophysiological classes||n||PP-D||PP-DF||PP-F||No change|
|Electrophysiological classes||n||PP-D||PP-DF||PP-F||No change|