0306-4522/92 $5.00 + 0.00 Pergamon Press plc ii-8 1991 IBRO
NeuroscienceVol. 46, No. 3, pp. 51 I-518, 1992 Printed in Great Britain
SEROTONIN BLOCKS THE LONG-TERM POTENTIATION INDUCED BY PRIMED BURST STIMULATION IN THE CA1 REGION OF RAT HIPPOCAMPAL SLICES R. Dipartimento
CORRADETTI,*
di Farmacologia
L. BALLERINI, A. M. ~UCLIESE and G. PEPEU
F’reclinica e Clinica, Universiti 50134 Firenze,
di Firenze,
Viale G.B.
Morgagni
65,
Italy
Abstract-The effect of 5-hydroxytryptamine on the induction of long-term potentiation by a train of high frequency pulses (100 Hz; 1 s) or by a stimulation consisting of one burst of five pulses at 100 Hz delivered 170 ms after a single pulse (primed burst) was investigated in the CA1 region of the rat hippocampal slice in vitro with extracellular recordings. Superfusion with Shydroxytryptamine (3-30pM) produced a concentration-dependent decrease in amplitude of the population spikes evoked by test stimuli. The presence of 5-hydroxytryptamine (30pM) did not affect the magnitude of long-term potentiation produced by the high-frequency stimulation but it prevented the long-term potentiation induced by a primed burst. The action of 5-hydroxytryptamine was mimicked by the Shydroxytryptamine,, agonist 5-carboxamidotryptamine (0.3 PM) and blocked by the 5-hydroxytryptamineJ5-hydroxytryptamine,. antagonist spiperone (3 y M) or by the 5-hydroxytryptamine,/5-hydroxytryptamine* antagonist methiothepin (l-10 PM). The selective 5_hydroxytryptamine, antagonist ritanserin (1 PM) did not antagonize the
block of long-term potentiation produced by 5-hydrixytryptamine. antagonists (3-tropanyl)-IH-indole-3-carboxylic acid ester (ES
The selective 5-hydroxytryptamine, 205-930; 1nM) and ondansetron
(GRI38032; 30nG) did. not affect the reduction in the population spike produced by application of 5-hydroxytryptamine. In contrast, a primed burst delivered at the fifth minute of 5-hydroxytryptamine application in the presence of a 5_hydroxytryptamine, antagonist induced a long-term potentiation. It is concluded that activation of 5-hydroxytryptamine,. and 5_hydroxytryptamine, receptors blocked the induction of long-term potentiation induced by primed burst stimulation but not that induced by a 1s high-frequency train. Given the heterogeneous localization of these receptor subtypes, it is suggested that the overall action of 5-hydroxytryptamine was exerted by hyperpolarizing pyramidal cells via 5-hydroxytryptamine,. receptors and by increasing spontaneous discharges of GABAergic interneurons via stimulation of 5_hydroxytryptamine, receptors.
The hippocampal CA1 region is enriched with 5hydroxytryptamine (5-HT; serotonin) receptors” and SHT-containing terminals” which originate from cells located in the raphe nuclei.30 These observations, together with biochemical and electrophysiological data,2.2’.2b.29 support the hypothesis that 5-HT is a neurotransmitter released by the raphe-hippocampal pathway. The effects of S-HT on pyramidal neurons of the CA1 region have been investigated in vivo25,26and in vitro using the hippocampal slice preparation. ~~~~~~~~~~~~~~~~~~~ At least two distinct receptor subtypes appear to be involved in the postsynaptic action of 5-HT on CA1 pyramidal cells. One of these receptor classes appears to involve 5-HT,, receptors, which mediate cell hyperpolarization.‘,’ A second receptor class responsible for the block of calcium-mediated afterhyperpolarization and slow depolarization’,9 has not been pharmacologically characterized.
Little attention has, however, been directed to a possible modulatory role of 5-HT on the induction of the sustained increase in hippocampal synaptic efficacy, known as long-term potentiation (LTP), produced by short trains of high-frequency stimulation. This phenomenon, first described in the rabbit perforant pathway in viva,’ has also been observed in isolated hippocampal slices.6 A form of LTP can be produced either by two bursts of a few pulses at 100 Hz with an interbust interval of 200 ms”.” or by a single burst of five pulses delivered 170 ms after a “priming” single stimulus (primed burst, PB).“.‘4 In both cases the pattern of stimulation intends to replicate the sequence of complex spike bursting and theta rhythm characteristic of the hippocampus. The aim of the present study was to investigate any modulatory action of 5-HT on the LTP induced by high-frequency stimulation for 1 s or by a PB. A preliminary account of part of the present data has been communicated.”
*To whom correspondence should bc addressed. Ahhre~~iutions: aCSF. artificial cerebrospinal fluid; 5-CT, 5-carboxamidotryptamine; CR-38032, 1,2,3,9-tetrahydro-9-methyl-3-[2-methyl-IH-imidazol-l-yl)-methyl]4H-carbazol-4-one (ondansetron); 5-HT, 5-hydroxytryptamine (serotonin); ICS 205-930, (3-tropanyl)-lHindole-3.carboxylic acid ester; LTP. long-term -potentiation; PB, primed burst.
EXPERIMENTAL
PROCEDCRES
Experiments were conducted using the in oitro hippocampal slice preparation obtained as described previously.” Charles River male Wistar rats of 150-200 g body weight were anaesthetized with ether and decapitated. Then hippocampi were rapidly dissected in 0 ‘C oxygenated (95% 511
0,/5% CO,) artificial cerebrospinal fluid (aCSF) of the following composition (mM): NaCl 124, KC1 3.33, KHIPO, 1.25, MgSO, 2, CaCI, 2, NaHCO, 25, D-ghCOSe IO(pH 7.3). Slices (400pm thick) were cut using a McIIwain tissue chopper and kept in oxygenated aCSF for at least I h at room temperature. Slices were obtained from the dorsal hippocampus where S-HT has been described to exert a predominantly hyperpolarizing effect on CA 1 pyramidal neurons.” A single slice was then placed on a nylon mesh, completely submerged in a small chamber and superfused with oxygenated aCSF (3435°C) at a constant flow rate of 2-3 ml/min. Drugs were applied via a three-way tap with complete exchange of the chamber volume in less than I min. Test pulses (80-I 10~s duration, 0.017 Hz) were delivered through bipolar nichrome electrodes positioned in the stratum radiatum. Evoked potentials were extracellularly recorded with 3 M NaCl-filled electrodes (2.-10 M a) placed in the CAI region of the stratum pyramidale, which contains primarily the cell bodies of the pyramidal neurons. Responses were amplified (Neurolog NL 104, Digitimer Ltd.), displayed on the storage oscilloscope, digitized (sample rate 30 kHz), and stored on floppy disk for later analysis (DATA 6000, Analogic; pClamp, Axon). Input-output curves were constructed at the beginning of each experiment by gradual increases in stimulus strength. The stimulus strength of the test pulses was adjusted to produce a population spike whose amplitude was 3040% (1.5-2 mV) of the maximum and, unless otherwise stated, was kept constant throughout the experiment. After 30 min equilibration a 25-min control period was implemented. Whenever these responses did not change during this period, the last IO min was then used to generate baseline values before experimental tests. The stratum radiatum was stimulated with one of the following patterns: (i) a high-frequency train (100 Hz, 1 s) at control stimulus strength; (ii) a PB consisting of a pulse followed (after an interval of 170 ms) by a burst of five pulses at 100 Hz. The stimulus strength during the PB stimulation was either kept at the same intensity as the one used for the test responses (weak PB) or raised to a near maximal strength (strong PB); (iii) five pulses at 100 Hz (unprimed burst). When appropriate, agonists were applied for 5 min and the stimulation pattern aimed to induce LTP was delivered during the fifth minute of application of the drug, which was washed immediately after the train. Antagonists were allowed to equilibrate for at least 15 min before adding 5-HT and were left for at least I5 min after the high-frequency stimulation and washout of the agonist. The amplitude of the population spike of the test responses (measured from peak to peak of the first negative phase of the population spike) was monitored for at least 1 h after the highfrequency stimulation. Unless otherwise stated, values given in the text are means + S.E.M. of measurements from different experiments. For statistical analysis the population strike amplitude was measured at fixed times according to the various experimental protocols used: (i) at the fifth minute after the start of bath application of an agonist; (ii) at the 120th minute after a train of high-frequency stimulation; (iii) at the 60th minute after PB or unprimed burst stimulation. Paired or unpaired Student’s r-test, as appropriate, was employed, and a value of P < 0.05was considered as indicative of a statistically significant difference.
Drugs used were: 5-hydroxytryptamine hydrochloride or creatinin sulphate (RBI), 5-carboxamidotryptamine maleate (RBI), ritanserin (RBI), spiperone hydrochloride (RBI); (3-tropanyl)-IH-indole-3-carboxylic acid ester (ICS 205-930, gift of Dr P. Herrling, Sandoz); ondansetron (GR-38032; 1,2,3,9-tetrahydro-9-methyl-3-[(2-methyl-lH-
imidazol-l-yl)-methyl]-4Ncarbazol-Cone; giti 01 Dr P Blandina); methiothepin maleate (gift of Dr M. DaPrada. Hoffmann-LaRoche). RESULTS
The present study is based on extracellular recordings from the pyramidal cell layer of the CA1 region of I85 slices taken from the dorsal hippocampus of 110 fats. Effects of serotonin on the long-term potentiation produced by a high-frequency train or by a primed burst stimulation
In a first set of experiments we investigated the effects of 5-HT on the induction of LTP evoked by a train (1 s) of stimuli at 100 Hz delivered to the stratum radiatum. Electrophysiological responses were monitored every 60 s throughout the experimental period. In control conditions the train of highfrequency stimulation produced a sustained (> 2 h) increase in the amplitude of the population spike evoked by the test stimulus (127 k 22%; n = 7). When 5-HT (30 /*M) was bath-applied for 5 min prior to such a high-frequency stimulation, the population spike of the test response was significantly (P < 0.01) decreased in amplitude versus the non-
A
-i+
1
LTP
5-HT
control
-f--
0
.
0
-.
‘1
10
20
30
Time
40
(min)
50
60
70
Fig. I. Serotonin (5-HT) does not modify the magnitude of LTP induced by a train of high-frequency stimulation. (A) In control conditions (left trace) stimulation with a test pulse (12 V; 100 ps duration; arrowheads) to the stratum radiaturn evokes in the CA1 pyramidal cell layer an excitatory postsynaptic potential and a population spike (sharp negative deflection); superfusion of 5-HT (30 PM) for 5 min reduces the amplitude of population spike and excitatory postsynaptic potential (middle trace); 60 min after a train of high-frequency stimulation (100 Hz; 1s) delivered at the fifth minute of 5-HT application (washed immediately thereafter), the test pulse evoked a potentiated response (right trace). (B) Mean (f S.E.M.) of amplitudes of test population spikes after high-frequency stimulation (arrowhead) delivered either in control conditions (n = 7) or at the fifth minute of 5-HT application (30pM; solid bar; n = 7). In this and in following figures, unless otherwise stated, the amplitude of the test responses is expressed as a pemntage of the average of the population spike amplitudes obtained during a period of 10min before starting the experiment (baseline).
5-HT blocks primed burst-induced
tetanized controls (-41 rl: 7%). Nonetheless, in the presence of 5-HT the stimulation train was still able to produce a long-lasting (>2 h) increase in the population spike amplitude (113 + 54%; n = 7; Fig. 1). These results indicated that the LTP produced by this type of high-frequency stimulation was not significantly affected by 5-HT. Since stimulation of the afferent pathway for 1 s at 100 Hz is unlikely to occur under physiological conditions,’ we consider the possibility that such a high degree of pyramidal cell excitation had surmounted any modulatory action of 5-HT. In order to induce LTP with a pattern of CA1 pyramidal cell activation closer to that observed in vim, we used a shorter high-frequency train, such as the PB stimulation described by Rose and Dunwiddie.z4 Furthermore, since the effect of 5-HT on the population spikes might be influenced by the intensity of afferent activation, we investigated the effects of 5-HT (in various concentrations) on input-output curves, as shown in Fig. 2. The reduction in population spike amplitude produced by S-HT (3-30pM) was evident at strengths which evoked small amplitude population spikes. The effect of 5-HT decreased at higher strengths of stimulation until it disappeared at strengths producing near maximal amplitude (2 80%) population spikes (Fig. 2). These results led us to use two stimulation protocols: in the first one the stimulus was set at a strength which elicited 30.-40% of the maximal response and this strength of stimulation was kept constant both
7
F
41
Control
S-HT
Stimulus
strength
(V)
Fig. 2. The effect of serotonin (S-I-IT) on the evoked population spike depends on stimulus strength. Input--output curves were constructed in control solution f O-0) by increasing the strength of the pulse delivered to the stratum radiatum (100 ps duration; 0.05 Hz). They were repeated starting from the fifth minute of application of 5-HT in increasing concentrations: 3 PM (A-A); 5 PM (0-U); 10 PM (II-m) or 30 PM (A-A). The drug was washed for 15 min between applications. Inset: responses obtained from a different slice by stratum radiatum stimulation a& 22 V or 8 V (arrow) which evoked the maximal or the 40% maximal response, respectively (left superimposed traces). In the presence of 5-HT (30pM; 5 min; right superimposed traces) the population spike evoked by an 8 V (arrow) stimulation is reduced, whereas that produced by a 22V stimulus is not. Arrowheads indicate time of stimulation. Data are representative of six experiments which gave similar results.
LTP
primed
OkrYz7 Time
40
(min)
Fig. 3. Stimulation of the stratum radiatum with a PB induces an LTP of the evoked population spike. (A) Mean (+S.E.M.) amplitude of population spikes evoked by test pulses after stimulation (arrowhead) with a weak PB (for de~nition of PB stimulus parameters see methods; n = I I), with a strong PB (n = IO). or with a weak unprimed burst (n = 8). (B) Left traces: control responses following stimulation of the stratum radiatum (arrowheads) in two preparations. Right traces: corresponding responses recorded 60 min after stimulation with a weak PB (upper) or a strong unprimed burst (lower); the drawings in the middle schematize the two patterns of stimulation with an arbitrary time base. Arrowheads indicate time of stimulation. filibrdtion bars apply to all recordings.
during test responses and during the PB (“weak PB”). In the second protocol the stimulus strength during the PB was increased to that producing 90% of maximal response (“strong PB”). Stimulation of the stratum radiatum with a priming pulse followed at a 170-ms interval by a burst of five lOO-Hz pulses invariably induced an enduring enhancement of the test response for longer than 1 h. Figure 3 illustrates the time course of the changes in amplitude of test responses in the first 30 min after a PB has been delivered. The increase in response produced by weak PBS is compared to that induced by strong PBS. In both cases a significant increase in the amplitude of the test response was obtained. The transient effect of five pulses at 100 Hz not preceded by the priming pulse (unprimed burst) is also shown in Fig. 3. Superfusion of 5-HT (30 /tM) for 5 min significantly (P < 0.01) reduced the amplitude of the test population spike (-51 k 6%; n = 47). A weak PB, delivered at the fifth minute of 5-HT application, did not produce any long-lasting enhancement of test responses in about 80% of the preparations (38 out of 47). Figure 4 shows the time course of the responses before and after PB stimulation either in the control solution or in the presence of 5-HT. It is clear that application of 5-HT fully prevented the rise in amplitude of the population spike. Stimulation with a second weak PB 40 min after wash of 5-HT elicited an LTP similar to that obtained under control conditions (50 & 13; P < 0.01;Fig. 4B). This indicated that the slices were indeed able to develop LTP and that the block of LTP produced by 5-HT was reversible. When 5-HT was bath-applied to reach the
514
CORRADETTI
rt uf.
--
-. 10
0
_ 20
30
Time
0
10 20 30 40 50 60 70 10 90 Time
(min)
Fig. 4. Serotonin (S-HT) btocks the induction of LTP by weak PB stimulation. (A) Mean (*S.E.M.) amplitude popuiation spikes evoked by test pulses after stimulation (arrowhead) with a weak PB delivered in control aCSF (n = lo), or at the fifth minute of bath application of 5-HT (30 PM; solid bar; n = 11). (B) A weak PB delivered at the fifth minute of bath application (solid bar) of 30 p M 5-HT does not potentiate the test responses (mean & S.E.M.; n = S), whereas a second weak PB delivered 4Omin after washout of S-HI evoked an LTP of the test response. Arrowheads indicate time of PB stimulation.
preparation within 10s after the weak PB had been delivered, it still reduced the population spike amplitude (- 69 f 6%) at the fifth minute of application in comparison with the response 10 s after PB stimulation (P < 0.01; n = 6). Once 5-HT was washed out, within 10 min the amplitude of the test response returned to a potentiated value (63 f 15; 60 min after the PB) similar to that obtained with a weak PB in control slices (60 & 7; n = 10). This indicated that the action(s) of S-HT to suppress LTP had to be exerted during the PB stimulation. In a second group of experiments the stimulus strength during the PB was set at a value which evoked 90% of the maximal response in the control input-output curves. Under these conditions a strong PB enhanced test responses which, after an initial decay, reached a stable degree of potentiation (67 Ltr: 9%; 60 min after the PB; n = 10). When delivered in the presence of 30 PM 5-HT (n = 9), the strong PB enhanced test responses for about 30 min although to a lesser extent than in control conditions. Then the amplitude of responses gradually declined (Fig. 5) and, 60 min after PB stimulation, test responses returned to values (16 _I 6%) not significantly different from baseline. E@ct of the Shydroxytryptumine,, 5-carboxamidoiryp~am~~e
receptor agonist
Among the various actions exerted by S-HT on CA1 pyramidal cells, neuronal hyperpolarization is known to be mediated by the 5-HT,,, subtype of S-HT receptors.‘,3 For this receptor subtype 5-carboxamidotryptamine (S-CT) appears to be the most selective and rapidly reversible full agonist.3*4
_. 40
_. 50
60
70
(min)
Fig. 5. Strong PB stimulation in the presence of serotonin (.S-HT) produces a short-lasting enhancement of evoked responses. Mean (& S.E.M.) amplitude of population spikes evoked by test pulses after stimulation (arrowhead) with a strong PB delivered either in control aCSF (n = 10) or at the fifth minute of bath appli~tion of S-HT (30 FM; solid bar; n =9).
5-CT (0.3 PM; superfused for 5 min) significantly (P < 0.01) decreased the evoked test response ( - 87 f 3%; n = 8) and blocked the induction of LTP by a weak PB delivered at the fifth minute of drug application (6 + 6%; Fig. 6A). However, a second weak PB delivered 1 h after wash of .5-CT elicited an LTP similar to that seen under control conditions (60 + 19%; n = 4; data not shown). Effeccts of antagon~?s 5-hydroxy~ryptamine*
of 5-hyd~oxy~rypt~~e~ receptors
and
To investigate further the involvement of 5-HT,, receptors in the observed block of LTP, we tested the actions of spiperone, a 5-HT,,/5-HT2 receptor antagonist, on the effect of 5-HT. Superfusion with 5-HT
!i 0
10
20
30
Time
40
50
60
70
(min)
Fig. 6. The effect of serotonin (5-HT) on the induction of LTP by a weak PB is mimicked by S-CT, and is antagonized by spiperone. (A) Mean (&S.E.M.; n = 8) amplitude of ~pulation spikes evoked by test pulses after stimulation with a weak PB (arrowhead) delivered at the fifth minute of bath application of 5-CT (0.3 PM; open bar). (B) When .5-HT (30 PM; solid bar) is bath-applied in the presence of spiperone (3 pM, allowed to equilibrate for 1h), the evoked test response increased and a weak PB (arrowhead) induced an LTP of the test response. Values are mean + S.E.M. from four experiments.
5-HT blocks primed burst-induced
(30 PM for 5 min) in the presence of spiperone (3 PM, allowed to equilibrate for more than 1 h) significantly (P < 0.01) increased the amplitude of untetanized test responses (161 + 24%; n = 4). Subsequent stimulation with a weak PB produced a further, long-lasting enhancement of these responses (Fig. 6B). This finding was compatible with the involvement of 5-HT,, receptors on the block of PB-induced LTP. However, since spiperone also blocks dopamine receptors’* and produces an excitatory action not yet characterized in terms of receptor pharmacology.‘,4,9 we tried a different 5-HT,/5-HT, receptor antagonist, namely methiothepin. In the presence of methiothepin (10 PM), which per se had no detectable effects on the evoked test response, 5-HT did not depress either the test population spike or the long-lasting enhancement of the response following a weak PB (Fig. 7; n = 3). Similar results were obtained with 1 PM methiothepin, which, however, did not completely block the reduction in the population spike induced by 5-HT (- 23 f 9%; P < 0.05; n = 7; see Fig. 7B). During the first few minutes of wash from methiothepin a transient decrease in amplitude of the test response was observed. Such a phenomenon was not apparent after washout of other antagonists (e.g. Fig. 8). Since both spiperone and methiothepin are mixed S-HT,/5-HT, antagonists, it seems possible that their action was through blockade of 5-HT, receptors. We therefore investigated the action of the selective 5-HT2 receptor antagonist ritanserin. In six control slices 5-HT (30 PM) was applied for 5 min and a weak PB was delivered. 5-HT was then washed out and the test responses showed that the PB stimulation failed to produce a persistent potentiation (17 + 10%; 60 min after the PB). Ritanserin (1 PM) was then allowed to equilibrate for 15 min before repeating the sequence of 5-HT application and weak PB stimulation. In the presence of ritanserin 5-HT did not decrease the amplitude of the responses at the fifth minute of application ( - 11 + 10% versus - 40 + 9% in the absence of ritanserin; P < 0.05). In spite of the reduced effects of 5-HT on test responses, a weak PB delivered in the presence of 5-HT and ritanserin did not produce a sustained LTP (14 + 9% of baseline in ritanserin; 60 min after the PB) when compared with that obtained under control conditions or in the presence of 5-HT and methiothepin (53 + 11%; 60 min after the PB). This suggested that 5-HT2 receptors were not responsible for the block of LTP exerted by 5-HT. E@cts
of 5-hydroxytryptamine, receptor antagonists
To extend the characterization of the receptor(s) involved in the effects of 5-HT on PB-induced LTP, we tested the action of the selective 5-HT, receptor antagonist ICS 205-930 (3-tropanyl)- 1H-indole-3carboxylic acid ester). ICS 205-930 (1 nM; superfused for
15 min)
did not
modify
the amplitude
of evoked
515
LTP
Fig. 7. Methiothepin antagonized the block by serotonin (5-HT) of the LTP induced by a weak PB. (A) Time course of the amplitude (in mV) of the evoked somatic response in a representative experiment. Bath application of 5-HT (30 PM, solid bar) reversibly reduced the amplitude of the population spike, an effect blocked in the presence of methiothepin (10 PM, open bar). A weak PB (arrowhead) delivered at the fifth minute of 5-HT application in the presence of methiothepin induced an LTP of the evoked test response. (B) Mean (& S.E.M.; n = 8) amplitudes of population spikes evoked by test pulses after stimulation (arrowhead) with a weak PB delivered at the fifth minute of 5-HT application (30pM; solid bar) in the presence of methiothepin (1 p M; open bar). Note that the effect of 5-HT on the amplitude of test responses is not completely antagonized by this concentration of methiothepin. Note also the transient decrease in response amplitude immediately after removal of methiothepin from aCSF.
population spikes (107 k 3% of control). Addition of 5-HT (30 PM) plus the antagonist produced a decrease in amplitude of the non-tetanized response (-42 + 12%; P < 0.05; n = 7; Fig. 8A) similar to that obtained with 5-HT alone. On the other hand, when a weak PB was delivered to the stratum radiaturn in the presence of both ICS 205-930 and 5-HT, a long-lasting increase in response amplitude was recorded (+ 52 + 13%; Fig. 8A, B). The action of ondansetron (GR-38032), another selective blocker of the 5-HT, receptor, was also tested. The presence of GR-38032 (30 nM; n = 3) which per se had no effect on the test response, did not antagonize the reduction in amplitude of population spike brought about by addition of 5-HT (Fig. 8A). However, as in the case of ICS 205-930, a weak PB delivered in the presence of both GR-38032 and 5-HT induced a long-lasting increase in the amplitude of the responses (Fig. 8A). These findings indicated that 5-HT, receptor activation was also a necessary step for suppressing LTP. DISCUSSION
The major finding of our experiments is that, in rat hippocampus, 5-HT acting via 5-HT,, and 5-HT, receptors blocked the induction of LTP produced by PB patterns, but not that induced by longer highfrequency trains. The PB stimulation consisting of a
516
K. COKRAIIETTI rl nl.
A
I”
KS
205.-930
Control
--s--3-in
W-38032
5-m
wsok
PB
“r---
A
_12m”
on its maintenance. Evidence has been provrded that the induction of LTP requires events occurrmg in the postsynaptic neurons, whereas presynaptic mechanisms appear to be involved mostly in the maintenance of LTP (for a review see Ref. 6). We will therefore focus our discussion on the effects of 5-HT which, converging onto pyramidal neurons, might be responsible for the failure of LTP initiation. Evidence for the involvement qJ’5-hydroxytryptumine,,, receptors in the action of serotonin
At the concentration (30pM) used in this study, S-HT produces a marked hyperpolarization of pyramidal cells, mediated by 5-HT,A receptors’.3.‘0 and through this mechanism it might reduce neuronal response to weak stimuli. It has been demonstrated Fig. 8. KS 205-930 or ondansetron (GR-38032) prevent the that hyperpolarization of a pyramidal cell during S-HT-induced block of LTP induced by a weak PB. (A) Two high-frequency stimulation blocks the development representative experiments conducted in the presence of ICS 205-930 (1 nM; upper traces) or ondansetron (30 nM; n = 3; of the LTP in the recorded cell.” Thus, hyperpolarizlower traces). Stimulation (arrowheads) of the stratum ation of pyramidal cells via activation of S-HT,* radiatum with test pulses in the presence of either antagonist receptors could account for the failure of PB stimu(left) evoked an excitatory postsynaptic potential and a lation to induce LTP. population spike which were reduced after 5 min application of S-HT (middle). Stimulation with a weak PB at the fifth However, this explanation appears fully tenable minute of 5-HT application induced a potentiation of the only for the action of the 5-HT,, agonist 5-CT, since test response which lasted longer than 6Omin (right). (B) 5-HT also exerts other actions, notably the block of Mean (+S.E.M.; n = 7) amplitude population spikes calcium-mediated afterhyperpolarization, and acevoked by test pulses after stimulation (arrowhead) with a commodation which may counteract the effects of weak PB delivered at the fifth minute of S-HT application (30 FM; solid bar) in the presence of ICS 205-930 (1 nM; neuron hyperpolarization during the high-frequency open bar). afferent stimulation.’ Nonetheless, the fact that 5-CT mimicked the effect of 5-HT suggested an involvement of 5-HT,, receptors in the block of LTP by PB stimulation. The antagonism of 5-HT action by conditioning pulse followed by a burst of five stimuli at 100 Hz, was aimed at inducing the discharge of spiperone (5-HT, > 5-HT,, antagonist) and methiothepin (5-HT,, > 5-HT, antagonist) but not by CA1 pyramidal cells with temporal constraints which ritanserin (selective 5-HT, antagonist) validates the stimulated the theta rhythm and complex spike activity peculiar to the hippocampus. Previous work’2*24 notion that activation of 5-HT,, receptors WdS necessary for the action of 5-HT on PB-induced LTP. demonstrated that the persistent enhancement of the synaptic activity produced by PB stimulation shares Evidence for the involvement of’ 5_hydroxytryptamine, some properties (e.g. duration, sensitivity to Nreceptors in the action of serotonin methyl-D-aspartate receptor antagonists) with those An additional mechanism, mediated by 5-HT, reof LTP induced by longer trains at 100 Hz in the same ceptors, appeared to be involved in the action of hippocampal region. However, it appears from our 5-HT. In fact, the selective block of 5-HT, receptors results that the enhancement induced by PB stimuby ICS 205-930 or ondansetron (GR-38032) also lation was sensitive to S-HT modulatory action, antagonized the effect of 5-HT on the induction of whereas the LTP produced by a longer highLTP by a PB. frequency train was not. It is possible that the The finding was unexpected since none of the high-frequency train represents a type of afferent stimulation so powerful that it can overcome the effects of 5-HT so far described in CA1 pyramidal cells is selectively affected by 5-HT, antagonists,‘~9~‘0 inhibitory actions of 5-HT on LTP induction. In our experiments S-HT reduced the evoked popuand in particular no reduction of the 5-HT-induced hyperpolarization has been observed with concenlation spike in extracellular recordings from the pyratrations of ICS 205-930 higher than those used in the midal cell layer. Previous studies have shown that the present study. The presence of 5-HT, receptors medieffect of S-HT is restricted to this cell layer since no ating fast cationic inward currents in hippocampal change in the dendritic field response or presynaptic (possibly non-pyramidal) neurons has been demonfibre activity could be detected with very large (mM) concentrations of S-HT. I4 strated.” It has recently been consistently shown that Our data demonstrated that S-HT, applied im- 5-HT can enhance the spontaneous unitary GABAmediated inhibitory postsynaptic potentials recorded mediately after the PB stimulation, did not suppress from CA1 pyramidal cell~.‘~This effect is mediated by the LTP once produced. This indicates that the action 5-HT, receptors probably located on GABAergic of 5-HT was exerted on the induction of LTP and not
5-HT blocks primed burst-induced
interneurons. Anatomical data demonstrate the existence of 5-HT-containing terminals on GABAergic interneurons in the hippocampus.“. The above mentioned data showing an increase in the activity of GABAergic interneurons produced by 5-HT13 suggest that such a phenomenon may reinforce the SHTelicited inhibition of pyramidal cell activity to block the PB-induced LTP. Physiological
implications
The hippocampus displays distinct neurophysiological patterns of activity: the complex spike and the theta rhythm. These patterns have been correlated to “intelligent” behaviour.33 to some forms of learning, and to LTP.’ Although the relationship among these correlates of theta rhythm remains to be elucidated, it is remarkable that LTP induction is facilitated during the positive phase of theta rhythm2’ and that patterns of stimulation which reproduce complex spike activity and theta rhythm induce stable potentiation of synaptic transmission both in vitro’2.‘7.24 and in aivo.‘2~2” It deserves mention that in vivo the expression of LTP in the hippocampus is dependent on the integrity of its subcortical afferents, including those originating from raphe nucleC8 Anatomical data suggest that different populations of serotoninergic cells project from the rdphe nuclei to different cellular groups in the cerebral cortex and in the hippocampus. ‘b.3’ In particular, one subset of termi-
517
LTP
nals appears to innervate non-pyramidal, predominantly GABAergic cells.” Thus the activity of raphe nuclei may influence cell discharge of both pyramidal and non-pyramidal cells, possibly independently.
CONCLUSION
We propose that 5-HT, through membrane hyperpolarization, produces a reduction in the response of pyramidal neurons to weak and scattered inputs. The consequence would be to affect greatly the probability to “prime” the response of pyramidal cells to a single burst of high-frequency action potentials. leaving, however, the “priming” still achievable by stronger, or repetitive inputs. The action of 5-HT, released by possibly independent raphe projections onto non-pyramidal neurons, will further reduce the responsiveness of pyramidal cells to inputs offered with theta patterns. This could be obtained by increasing the discharge of GABAergic interneurons via 5-HT, receptors. The overall action of 5-HT would therefore counteract the expression of LTP following behaviourally “non-meaningful” inputs to avoid saturation of LTP mechanisms. Acknowiedgements~We thank Dr T. Gessi and Dr A. Nistri for their helpful suggestions. This work was supported by grantsfromM.U.R.S.T.(60%)andC.N.R.(90.03188.CT04 and 91.00595.CT04) to R.C.
REFERENCES
1. Andrade R. and Nicoll R. A. (1987) Pharmacologically distinct actions of serotonin on single pyramidal neurones of the rat hippocampus recorded in vitro. J. Physiol. 394, 9--l 24. 2. Assaf S. Y. and Miller J. J. (1978) The role of a raphe serotonin system in the control of septal unit activity and hippocampal desynchronization. Neuroscience 3, 539-550. 3. Beck S. G. (1989) 5-Carboxyamidotryptamine mimics only the 5-HT elicited hyperpolarization of hippocampal pyramidal cells via 5-HT,, receptor. Neurosci. Left. 99, 101-106. 4. Beck S. G.. Clarke W. P. and Goldfarb J. (1985) Spiperone differentiates multiple 5-hydroxytryptamine responses in rat hippocampal slices in vitro. Eur. J. Pharmcrc. 116, 195-197. 5. Bliss T. V. P. and Lsmo T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anesthetized rabbit following stimulation of the perforant path. J. Phvsiol. 232, 331-336. 6. Bliss T. V. P. and Lynch M. A. (1988) Long-term potentiation of synapiic transmission in the hippocampus: properties and mechanism. In Long-term Potentiation: From Biophysics IO Behavior (eds Landfield P. W. and Deadwyler S. A.), pp. 3-72. Alan Liss, New York. 7. Buzsriki G. (1989) Two-stage model of memory trace formation: a role for “noisy” brain states. Neuroscience 31, 55 l-570. 8. Buz&ki G. and Gage F. H. (1989) Absence of long-term potentiation in the subcortically deafferented dentate gyrus. Bruin Res. 484, 94-101. 9. Chaput Y., Araneda R. C. and Andrade R. (1990) Pharmacological and functional analysis of a novel serotonin receptor in the rat hippocampus. Eur. J. Pharmac. 182, 441456. 10. Colino A. and Halliwell J. V. (1987) Differential modulation of three separate K-conductances in hippocampal CA1 neurons by serotonin. Nature 328, 73-77. 11. Corradetti R., Moneti G., Moroni F., Pepeu G. and Wieraszko A. (1983) Electrical stimulation of the stratum radiatum increases the release and neosynthesis of aspartate. glutamate and y-aminobutyric acid in rat hippocampal slices. J. Neurochem. 41, 1518-1525. 12. Diamond D. M., Dunwiddie T. V. and Rose G. M. (1988) Characteristics of hippocampal primed burst potentiation in d-o and in the awake rat. J. Neurosci. 8, 40794088. 13. Guy N. and Ropert N. (1990) Serotonin facilitates GABAergic inhibition on hippocampal CA1 neurones in vitro. J. Physiol. 423, 94P. 14. Hiner B. C., Mauk M. D., Peroutka S. J. and Kocsis J. D. (1988) Buspirone, 8-OH-DPAT and ipsapirone: effects on hippocampal cerebellar and sciatic fiber excitability. Brain Res. 461, 1 9. 15. Kiihler C. (1984) The distribution of serotonin binding sites in the hippocampdl region of the rat brain. An autoradiographic study. Neuroscience 13, 667480. 16. Kosofsky B. E. and Molliver M. E. (1987) The serotoninercic innervation of cerebral cortex: different classes of axon terminals arise from dorsal and median raphe nuclei. Svnapse 1, 153-168.
518
R. COKKADETTI et ol.
17. Larson J. and Lynch G. (1986) Induction of synaptic potentiation in hippocampus by patterned stimulation involve< two events. Science 232, 9855988. 18. Leysen J. E., Niemegeers C. J. E., Tollenaere J. P. and Laduron P. M. (1978) Serotonergic component OI neuroleptic receptors. Nature 272, 1688171. 19. Malinow R. and Miller J. P. (1986) Postsynaptic hyperpolarization during conditioning reversibly blocks induction o! long-term potentiation. Nature 320, 529-530. 20. Oleskevich S. and Descarries L. (1990) Quantified distribution of the serotonin innervation in adult rat hippocampus. Neuroscience 34, 19-33. 21. Pavlides C., Greenstein Y. J., Grudman M. and Winson J. (1988) Long-term potentiation in the dentate gyrus is induced preferentially on the positive phase of Q-rhythm. Brain Res. 439, 383-387. R. and Pepeu G. (1990) 5Hydroxytryptamine blocks the long-term potentiation 22. Pugliese A. M., Ballerini L., Corradetti induced by primed bursts in the CA1 region of rat hippocampal slices. Pharmac. Res. 22 (Suppl. 2). 416. 23. Ropert N. (1988) Inhibitory action of serotonin in CA1 hippocampal neurons in vitro. Neuroscience 26, 69-81. 24. Rose G. M. and Dunwiddie T. V. (1986) Induction of hippocampal long-term potentiation using physiologically patterned stimulation. Neurosci. Left. 69, 244-248. 25. Segal M. (1975) Physiological and pharmacological evidence for a serotonergic projection to the hippocampus. Brain Res. 94, 115-131. 26. Segal M. (1976) 5-HT antagonists in rat hippocampus. Brain Res. 103, 161-166. 27. Segal M. (1980) The action of serotonin in the rat hippocampal slice preparation. J. Physiol. 303, 423439. 28. Staubli U. and Lynch G. (1987) Stable hippocampal long-term potentiation elicited by “theta” pattern stimulation. Brain Res. 435, 227-234. 29. Storm-Mathisen J. (1977) Localization of transmitter candidates in the brain: the hippocampal formation as a model. Prog. Neurobiol. 8, 119-181. 30. Swanson L. W., Kijhler C. and Bjorklund A. (1987) The limbic region. I: The septohippocampal system. In Handbook of ChemicalNeuroanatomy (eds Bjiirklund A., Hiikfelt T. and Swanson L. W.), Vol. 5, Part I, pp. 125-277. Elsevier, Amsterdam. 31. Tork I. and Homung J.-P. (1988) Interaction of serotonin and GABA in cerebra1 cortex and hippocampus. Sot. Neurosci. Abstr. 14, 274.6. 32. Tiirk I., Hornung J.-P. and Somogyi P. (1988) Serotoninergic innervation in the cerebra1 cortex. Neurosci. Left. Suppl. 30, Sl31. 33. Vandervolf C. H. and Baker G. B. (1986) Evidence that serotonin mediates non-cholinergic neocortical low voltage fast activity, non-cholinergic hippocampal rhythmical slow activity and contributes to intelligent behavior. Bruin Res. 374, 342-356. 34. Yakel J. L. and Jackson M. B. (1988) 5-HT3 receptors mediate rapid responses in cultured hippocampus and a clonal cell line. Neuron 1, 615621. (Accepted 17 July 1991)
NeuroscienceVol. 46, No. 3, pp. 51 I-518, 1992 Printed in Great Britain
SEROTONIN BLOCKS THE LONG-TERM POTENTIATION INDUCED BY PRIMED BURST STIMULATION IN THE CA1 REGION OF RAT HIPPOCAMPAL SLICES R. Dipartimento
CORRADETTI,*
di Farmacologia
L. BALLERINI, A. M. ~UCLIESE and G. PEPEU
F’reclinica e Clinica, Universiti 50134 Firenze,
di Firenze,
Viale G.B.
Morgagni
65,
Italy
Abstract-The effect of 5-hydroxytryptamine on the induction of long-term potentiation by a train of high frequency pulses (100 Hz; 1 s) or by a stimulation consisting of one burst of five pulses at 100 Hz delivered 170 ms after a single pulse (primed burst) was investigated in the CA1 region of the rat hippocampal slice in vitro with extracellular recordings. Superfusion with Shydroxytryptamine (3-30pM) produced a concentration-dependent decrease in amplitude of the population spikes evoked by test stimuli. The presence of 5-hydroxytryptamine (30pM) did not affect the magnitude of long-term potentiation produced by the high-frequency stimulation but it prevented the long-term potentiation induced by a primed burst. The action of 5-hydroxytryptamine was mimicked by the Shydroxytryptamine,, agonist 5-carboxamidotryptamine (0.3 PM) and blocked by the 5-hydroxytryptamineJ5-hydroxytryptamine,. antagonist spiperone (3 y M) or by the 5-hydroxytryptamine,/5-hydroxytryptamine* antagonist methiothepin (l-10 PM). The selective 5_hydroxytryptamine, antagonist ritanserin (1 PM) did not antagonize the
block of long-term potentiation produced by 5-hydrixytryptamine. antagonists (3-tropanyl)-IH-indole-3-carboxylic acid ester (ES
The selective 5-hydroxytryptamine, 205-930; 1nM) and ondansetron
(GRI38032; 30nG) did. not affect the reduction in the population spike produced by application of 5-hydroxytryptamine. In contrast, a primed burst delivered at the fifth minute of 5-hydroxytryptamine application in the presence of a 5_hydroxytryptamine, antagonist induced a long-term potentiation. It is concluded that activation of 5-hydroxytryptamine,. and 5_hydroxytryptamine, receptors blocked the induction of long-term potentiation induced by primed burst stimulation but not that induced by a 1s high-frequency train. Given the heterogeneous localization of these receptor subtypes, it is suggested that the overall action of 5-hydroxytryptamine was exerted by hyperpolarizing pyramidal cells via 5-hydroxytryptamine,. receptors and by increasing spontaneous discharges of GABAergic interneurons via stimulation of 5_hydroxytryptamine, receptors.
The hippocampal CA1 region is enriched with 5hydroxytryptamine (5-HT; serotonin) receptors” and SHT-containing terminals” which originate from cells located in the raphe nuclei.30 These observations, together with biochemical and electrophysiological data,2.2’.2b.29 support the hypothesis that 5-HT is a neurotransmitter released by the raphe-hippocampal pathway. The effects of S-HT on pyramidal neurons of the CA1 region have been investigated in vivo25,26and in vitro using the hippocampal slice preparation. ~~~~~~~~~~~~~~~~~~~ At least two distinct receptor subtypes appear to be involved in the postsynaptic action of 5-HT on CA1 pyramidal cells. One of these receptor classes appears to involve 5-HT,, receptors, which mediate cell hyperpolarization.‘,’ A second receptor class responsible for the block of calcium-mediated afterhyperpolarization and slow depolarization’,9 has not been pharmacologically characterized.
Little attention has, however, been directed to a possible modulatory role of 5-HT on the induction of the sustained increase in hippocampal synaptic efficacy, known as long-term potentiation (LTP), produced by short trains of high-frequency stimulation. This phenomenon, first described in the rabbit perforant pathway in viva,’ has also been observed in isolated hippocampal slices.6 A form of LTP can be produced either by two bursts of a few pulses at 100 Hz with an interbust interval of 200 ms”.” or by a single burst of five pulses delivered 170 ms after a “priming” single stimulus (primed burst, PB).“.‘4 In both cases the pattern of stimulation intends to replicate the sequence of complex spike bursting and theta rhythm characteristic of the hippocampus. The aim of the present study was to investigate any modulatory action of 5-HT on the LTP induced by high-frequency stimulation for 1 s or by a PB. A preliminary account of part of the present data has been communicated.”
*To whom correspondence should bc addressed. Ahhre~~iutions: aCSF. artificial cerebrospinal fluid; 5-CT, 5-carboxamidotryptamine; CR-38032, 1,2,3,9-tetrahydro-9-methyl-3-[2-methyl-IH-imidazol-l-yl)-methyl]4H-carbazol-4-one (ondansetron); 5-HT, 5-hydroxytryptamine (serotonin); ICS 205-930, (3-tropanyl)-lHindole-3.carboxylic acid ester; LTP. long-term -potentiation; PB, primed burst.
EXPERIMENTAL
PROCEDCRES
Experiments were conducted using the in oitro hippocampal slice preparation obtained as described previously.” Charles River male Wistar rats of 150-200 g body weight were anaesthetized with ether and decapitated. Then hippocampi were rapidly dissected in 0 ‘C oxygenated (95% 511
0,/5% CO,) artificial cerebrospinal fluid (aCSF) of the following composition (mM): NaCl 124, KC1 3.33, KHIPO, 1.25, MgSO, 2, CaCI, 2, NaHCO, 25, D-ghCOSe IO(pH 7.3). Slices (400pm thick) were cut using a McIIwain tissue chopper and kept in oxygenated aCSF for at least I h at room temperature. Slices were obtained from the dorsal hippocampus where S-HT has been described to exert a predominantly hyperpolarizing effect on CA 1 pyramidal neurons.” A single slice was then placed on a nylon mesh, completely submerged in a small chamber and superfused with oxygenated aCSF (3435°C) at a constant flow rate of 2-3 ml/min. Drugs were applied via a three-way tap with complete exchange of the chamber volume in less than I min. Test pulses (80-I 10~s duration, 0.017 Hz) were delivered through bipolar nichrome electrodes positioned in the stratum radiatum. Evoked potentials were extracellularly recorded with 3 M NaCl-filled electrodes (2.-10 M a) placed in the CAI region of the stratum pyramidale, which contains primarily the cell bodies of the pyramidal neurons. Responses were amplified (Neurolog NL 104, Digitimer Ltd.), displayed on the storage oscilloscope, digitized (sample rate 30 kHz), and stored on floppy disk for later analysis (DATA 6000, Analogic; pClamp, Axon). Input-output curves were constructed at the beginning of each experiment by gradual increases in stimulus strength. The stimulus strength of the test pulses was adjusted to produce a population spike whose amplitude was 3040% (1.5-2 mV) of the maximum and, unless otherwise stated, was kept constant throughout the experiment. After 30 min equilibration a 25-min control period was implemented. Whenever these responses did not change during this period, the last IO min was then used to generate baseline values before experimental tests. The stratum radiatum was stimulated with one of the following patterns: (i) a high-frequency train (100 Hz, 1 s) at control stimulus strength; (ii) a PB consisting of a pulse followed (after an interval of 170 ms) by a burst of five pulses at 100 Hz. The stimulus strength during the PB stimulation was either kept at the same intensity as the one used for the test responses (weak PB) or raised to a near maximal strength (strong PB); (iii) five pulses at 100 Hz (unprimed burst). When appropriate, agonists were applied for 5 min and the stimulation pattern aimed to induce LTP was delivered during the fifth minute of application of the drug, which was washed immediately after the train. Antagonists were allowed to equilibrate for at least 15 min before adding 5-HT and were left for at least I5 min after the high-frequency stimulation and washout of the agonist. The amplitude of the population spike of the test responses (measured from peak to peak of the first negative phase of the population spike) was monitored for at least 1 h after the highfrequency stimulation. Unless otherwise stated, values given in the text are means + S.E.M. of measurements from different experiments. For statistical analysis the population strike amplitude was measured at fixed times according to the various experimental protocols used: (i) at the fifth minute after the start of bath application of an agonist; (ii) at the 120th minute after a train of high-frequency stimulation; (iii) at the 60th minute after PB or unprimed burst stimulation. Paired or unpaired Student’s r-test, as appropriate, was employed, and a value of P < 0.05was considered as indicative of a statistically significant difference.
Drugs used were: 5-hydroxytryptamine hydrochloride or creatinin sulphate (RBI), 5-carboxamidotryptamine maleate (RBI), ritanserin (RBI), spiperone hydrochloride (RBI); (3-tropanyl)-IH-indole-3-carboxylic acid ester (ICS 205-930, gift of Dr P. Herrling, Sandoz); ondansetron (GR-38032; 1,2,3,9-tetrahydro-9-methyl-3-[(2-methyl-lH-
imidazol-l-yl)-methyl]-4Ncarbazol-Cone; giti 01 Dr P Blandina); methiothepin maleate (gift of Dr M. DaPrada. Hoffmann-LaRoche). RESULTS
The present study is based on extracellular recordings from the pyramidal cell layer of the CA1 region of I85 slices taken from the dorsal hippocampus of 110 fats. Effects of serotonin on the long-term potentiation produced by a high-frequency train or by a primed burst stimulation
In a first set of experiments we investigated the effects of 5-HT on the induction of LTP evoked by a train (1 s) of stimuli at 100 Hz delivered to the stratum radiatum. Electrophysiological responses were monitored every 60 s throughout the experimental period. In control conditions the train of highfrequency stimulation produced a sustained (> 2 h) increase in the amplitude of the population spike evoked by the test stimulus (127 k 22%; n = 7). When 5-HT (30 /*M) was bath-applied for 5 min prior to such a high-frequency stimulation, the population spike of the test response was significantly (P < 0.01) decreased in amplitude versus the non-
A
-i+
1
LTP
5-HT
control
-f--
0
.
0
-.
‘1
10
20
30
Time
40
(min)
50
60
70
Fig. I. Serotonin (5-HT) does not modify the magnitude of LTP induced by a train of high-frequency stimulation. (A) In control conditions (left trace) stimulation with a test pulse (12 V; 100 ps duration; arrowheads) to the stratum radiaturn evokes in the CA1 pyramidal cell layer an excitatory postsynaptic potential and a population spike (sharp negative deflection); superfusion of 5-HT (30 PM) for 5 min reduces the amplitude of population spike and excitatory postsynaptic potential (middle trace); 60 min after a train of high-frequency stimulation (100 Hz; 1s) delivered at the fifth minute of 5-HT application (washed immediately thereafter), the test pulse evoked a potentiated response (right trace). (B) Mean (f S.E.M.) of amplitudes of test population spikes after high-frequency stimulation (arrowhead) delivered either in control conditions (n = 7) or at the fifth minute of 5-HT application (30pM; solid bar; n = 7). In this and in following figures, unless otherwise stated, the amplitude of the test responses is expressed as a pemntage of the average of the population spike amplitudes obtained during a period of 10min before starting the experiment (baseline).
5-HT blocks primed burst-induced
tetanized controls (-41 rl: 7%). Nonetheless, in the presence of 5-HT the stimulation train was still able to produce a long-lasting (>2 h) increase in the population spike amplitude (113 + 54%; n = 7; Fig. 1). These results indicated that the LTP produced by this type of high-frequency stimulation was not significantly affected by 5-HT. Since stimulation of the afferent pathway for 1 s at 100 Hz is unlikely to occur under physiological conditions,’ we consider the possibility that such a high degree of pyramidal cell excitation had surmounted any modulatory action of 5-HT. In order to induce LTP with a pattern of CA1 pyramidal cell activation closer to that observed in vim, we used a shorter high-frequency train, such as the PB stimulation described by Rose and Dunwiddie.z4 Furthermore, since the effect of 5-HT on the population spikes might be influenced by the intensity of afferent activation, we investigated the effects of 5-HT (in various concentrations) on input-output curves, as shown in Fig. 2. The reduction in population spike amplitude produced by S-HT (3-30pM) was evident at strengths which evoked small amplitude population spikes. The effect of 5-HT decreased at higher strengths of stimulation until it disappeared at strengths producing near maximal amplitude (2 80%) population spikes (Fig. 2). These results led us to use two stimulation protocols: in the first one the stimulus was set at a strength which elicited 30.-40% of the maximal response and this strength of stimulation was kept constant both
7
F
41
Control
S-HT
Stimulus
strength
(V)
Fig. 2. The effect of serotonin (S-I-IT) on the evoked population spike depends on stimulus strength. Input--output curves were constructed in control solution f O-0) by increasing the strength of the pulse delivered to the stratum radiatum (100 ps duration; 0.05 Hz). They were repeated starting from the fifth minute of application of 5-HT in increasing concentrations: 3 PM (A-A); 5 PM (0-U); 10 PM (II-m) or 30 PM (A-A). The drug was washed for 15 min between applications. Inset: responses obtained from a different slice by stratum radiatum stimulation a& 22 V or 8 V (arrow) which evoked the maximal or the 40% maximal response, respectively (left superimposed traces). In the presence of 5-HT (30pM; 5 min; right superimposed traces) the population spike evoked by an 8 V (arrow) stimulation is reduced, whereas that produced by a 22V stimulus is not. Arrowheads indicate time of stimulation. Data are representative of six experiments which gave similar results.
LTP
primed
OkrYz7 Time
40
(min)
Fig. 3. Stimulation of the stratum radiatum with a PB induces an LTP of the evoked population spike. (A) Mean (+S.E.M.) amplitude of population spikes evoked by test pulses after stimulation (arrowhead) with a weak PB (for de~nition of PB stimulus parameters see methods; n = I I), with a strong PB (n = IO). or with a weak unprimed burst (n = 8). (B) Left traces: control responses following stimulation of the stratum radiatum (arrowheads) in two preparations. Right traces: corresponding responses recorded 60 min after stimulation with a weak PB (upper) or a strong unprimed burst (lower); the drawings in the middle schematize the two patterns of stimulation with an arbitrary time base. Arrowheads indicate time of stimulation. filibrdtion bars apply to all recordings.
during test responses and during the PB (“weak PB”). In the second protocol the stimulus strength during the PB was increased to that producing 90% of maximal response (“strong PB”). Stimulation of the stratum radiatum with a priming pulse followed at a 170-ms interval by a burst of five lOO-Hz pulses invariably induced an enduring enhancement of the test response for longer than 1 h. Figure 3 illustrates the time course of the changes in amplitude of test responses in the first 30 min after a PB has been delivered. The increase in response produced by weak PBS is compared to that induced by strong PBS. In both cases a significant increase in the amplitude of the test response was obtained. The transient effect of five pulses at 100 Hz not preceded by the priming pulse (unprimed burst) is also shown in Fig. 3. Superfusion of 5-HT (30 /tM) for 5 min significantly (P < 0.01) reduced the amplitude of the test population spike (-51 k 6%; n = 47). A weak PB, delivered at the fifth minute of 5-HT application, did not produce any long-lasting enhancement of test responses in about 80% of the preparations (38 out of 47). Figure 4 shows the time course of the responses before and after PB stimulation either in the control solution or in the presence of 5-HT. It is clear that application of 5-HT fully prevented the rise in amplitude of the population spike. Stimulation with a second weak PB 40 min after wash of 5-HT elicited an LTP similar to that obtained under control conditions (50 & 13; P < 0.01;Fig. 4B). This indicated that the slices were indeed able to develop LTP and that the block of LTP produced by 5-HT was reversible. When 5-HT was bath-applied to reach the
514
CORRADETTI
rt uf.
--
-. 10
0
_ 20
30
Time
0
10 20 30 40 50 60 70 10 90 Time
(min)
Fig. 4. Serotonin (S-HT) btocks the induction of LTP by weak PB stimulation. (A) Mean (*S.E.M.) amplitude popuiation spikes evoked by test pulses after stimulation (arrowhead) with a weak PB delivered in control aCSF (n = lo), or at the fifth minute of bath application of 5-HT (30 PM; solid bar; n = 11). (B) A weak PB delivered at the fifth minute of bath application (solid bar) of 30 p M 5-HT does not potentiate the test responses (mean & S.E.M.; n = S), whereas a second weak PB delivered 4Omin after washout of S-HI evoked an LTP of the test response. Arrowheads indicate time of PB stimulation.
preparation within 10s after the weak PB had been delivered, it still reduced the population spike amplitude (- 69 f 6%) at the fifth minute of application in comparison with the response 10 s after PB stimulation (P < 0.01; n = 6). Once 5-HT was washed out, within 10 min the amplitude of the test response returned to a potentiated value (63 f 15; 60 min after the PB) similar to that obtained with a weak PB in control slices (60 & 7; n = 10). This indicated that the action(s) of S-HT to suppress LTP had to be exerted during the PB stimulation. In a second group of experiments the stimulus strength during the PB was set at a value which evoked 90% of the maximal response in the control input-output curves. Under these conditions a strong PB enhanced test responses which, after an initial decay, reached a stable degree of potentiation (67 Ltr: 9%; 60 min after the PB; n = 10). When delivered in the presence of 30 PM 5-HT (n = 9), the strong PB enhanced test responses for about 30 min although to a lesser extent than in control conditions. Then the amplitude of responses gradually declined (Fig. 5) and, 60 min after PB stimulation, test responses returned to values (16 _I 6%) not significantly different from baseline. E@ct of the Shydroxytryptumine,, 5-carboxamidoiryp~am~~e
receptor agonist
Among the various actions exerted by S-HT on CA1 pyramidal cells, neuronal hyperpolarization is known to be mediated by the 5-HT,,, subtype of S-HT receptors.‘,3 For this receptor subtype 5-carboxamidotryptamine (S-CT) appears to be the most selective and rapidly reversible full agonist.3*4
_. 40
_. 50
60
70
(min)
Fig. 5. Strong PB stimulation in the presence of serotonin (.S-HT) produces a short-lasting enhancement of evoked responses. Mean (& S.E.M.) amplitude of population spikes evoked by test pulses after stimulation (arrowhead) with a strong PB delivered either in control aCSF (n = 10) or at the fifth minute of bath appli~tion of S-HT (30 FM; solid bar; n =9).
5-CT (0.3 PM; superfused for 5 min) significantly (P < 0.01) decreased the evoked test response ( - 87 f 3%; n = 8) and blocked the induction of LTP by a weak PB delivered at the fifth minute of drug application (6 + 6%; Fig. 6A). However, a second weak PB delivered 1 h after wash of .5-CT elicited an LTP similar to that seen under control conditions (60 + 19%; n = 4; data not shown). Effeccts of antagon~?s 5-hydroxy~ryptamine*
of 5-hyd~oxy~rypt~~e~ receptors
and
To investigate further the involvement of 5-HT,, receptors in the observed block of LTP, we tested the actions of spiperone, a 5-HT,,/5-HT2 receptor antagonist, on the effect of 5-HT. Superfusion with 5-HT
!i 0
10
20
30
Time
40
50
60
70
(min)
Fig. 6. The effect of serotonin (5-HT) on the induction of LTP by a weak PB is mimicked by S-CT, and is antagonized by spiperone. (A) Mean (&S.E.M.; n = 8) amplitude of ~pulation spikes evoked by test pulses after stimulation with a weak PB (arrowhead) delivered at the fifth minute of bath application of 5-CT (0.3 PM; open bar). (B) When .5-HT (30 PM; solid bar) is bath-applied in the presence of spiperone (3 pM, allowed to equilibrate for 1h), the evoked test response increased and a weak PB (arrowhead) induced an LTP of the test response. Values are mean + S.E.M. from four experiments.
5-HT blocks primed burst-induced
(30 PM for 5 min) in the presence of spiperone (3 PM, allowed to equilibrate for more than 1 h) significantly (P < 0.01) increased the amplitude of untetanized test responses (161 + 24%; n = 4). Subsequent stimulation with a weak PB produced a further, long-lasting enhancement of these responses (Fig. 6B). This finding was compatible with the involvement of 5-HT,, receptors on the block of PB-induced LTP. However, since spiperone also blocks dopamine receptors’* and produces an excitatory action not yet characterized in terms of receptor pharmacology.‘,4,9 we tried a different 5-HT,/5-HT, receptor antagonist, namely methiothepin. In the presence of methiothepin (10 PM), which per se had no detectable effects on the evoked test response, 5-HT did not depress either the test population spike or the long-lasting enhancement of the response following a weak PB (Fig. 7; n = 3). Similar results were obtained with 1 PM methiothepin, which, however, did not completely block the reduction in the population spike induced by 5-HT (- 23 f 9%; P < 0.05; n = 7; see Fig. 7B). During the first few minutes of wash from methiothepin a transient decrease in amplitude of the test response was observed. Such a phenomenon was not apparent after washout of other antagonists (e.g. Fig. 8). Since both spiperone and methiothepin are mixed S-HT,/5-HT, antagonists, it seems possible that their action was through blockade of 5-HT, receptors. We therefore investigated the action of the selective 5-HT2 receptor antagonist ritanserin. In six control slices 5-HT (30 PM) was applied for 5 min and a weak PB was delivered. 5-HT was then washed out and the test responses showed that the PB stimulation failed to produce a persistent potentiation (17 + 10%; 60 min after the PB). Ritanserin (1 PM) was then allowed to equilibrate for 15 min before repeating the sequence of 5-HT application and weak PB stimulation. In the presence of ritanserin 5-HT did not decrease the amplitude of the responses at the fifth minute of application ( - 11 + 10% versus - 40 + 9% in the absence of ritanserin; P < 0.05). In spite of the reduced effects of 5-HT on test responses, a weak PB delivered in the presence of 5-HT and ritanserin did not produce a sustained LTP (14 + 9% of baseline in ritanserin; 60 min after the PB) when compared with that obtained under control conditions or in the presence of 5-HT and methiothepin (53 + 11%; 60 min after the PB). This suggested that 5-HT2 receptors were not responsible for the block of LTP exerted by 5-HT. E@cts
of 5-hydroxytryptamine, receptor antagonists
To extend the characterization of the receptor(s) involved in the effects of 5-HT on PB-induced LTP, we tested the action of the selective 5-HT, receptor antagonist ICS 205-930 (3-tropanyl)- 1H-indole-3carboxylic acid ester). ICS 205-930 (1 nM; superfused for
15 min)
did not
modify
the amplitude
of evoked
515
LTP
Fig. 7. Methiothepin antagonized the block by serotonin (5-HT) of the LTP induced by a weak PB. (A) Time course of the amplitude (in mV) of the evoked somatic response in a representative experiment. Bath application of 5-HT (30 PM, solid bar) reversibly reduced the amplitude of the population spike, an effect blocked in the presence of methiothepin (10 PM, open bar). A weak PB (arrowhead) delivered at the fifth minute of 5-HT application in the presence of methiothepin induced an LTP of the evoked test response. (B) Mean (& S.E.M.; n = 8) amplitudes of population spikes evoked by test pulses after stimulation (arrowhead) with a weak PB delivered at the fifth minute of 5-HT application (30pM; solid bar) in the presence of methiothepin (1 p M; open bar). Note that the effect of 5-HT on the amplitude of test responses is not completely antagonized by this concentration of methiothepin. Note also the transient decrease in response amplitude immediately after removal of methiothepin from aCSF.
population spikes (107 k 3% of control). Addition of 5-HT (30 PM) plus the antagonist produced a decrease in amplitude of the non-tetanized response (-42 + 12%; P < 0.05; n = 7; Fig. 8A) similar to that obtained with 5-HT alone. On the other hand, when a weak PB was delivered to the stratum radiaturn in the presence of both ICS 205-930 and 5-HT, a long-lasting increase in response amplitude was recorded (+ 52 + 13%; Fig. 8A, B). The action of ondansetron (GR-38032), another selective blocker of the 5-HT, receptor, was also tested. The presence of GR-38032 (30 nM; n = 3) which per se had no effect on the test response, did not antagonize the reduction in amplitude of population spike brought about by addition of 5-HT (Fig. 8A). However, as in the case of ICS 205-930, a weak PB delivered in the presence of both GR-38032 and 5-HT induced a long-lasting increase in the amplitude of the responses (Fig. 8A). These findings indicated that 5-HT, receptor activation was also a necessary step for suppressing LTP. DISCUSSION
The major finding of our experiments is that, in rat hippocampus, 5-HT acting via 5-HT,, and 5-HT, receptors blocked the induction of LTP produced by PB patterns, but not that induced by longer highfrequency trains. The PB stimulation consisting of a
516
K. COKRAIIETTI rl nl.
A
I”
KS
205.-930
Control
--s--3-in
W-38032
5-m
wsok
PB
“r---
A
_12m”
on its maintenance. Evidence has been provrded that the induction of LTP requires events occurrmg in the postsynaptic neurons, whereas presynaptic mechanisms appear to be involved mostly in the maintenance of LTP (for a review see Ref. 6). We will therefore focus our discussion on the effects of 5-HT which, converging onto pyramidal neurons, might be responsible for the failure of LTP initiation. Evidence for the involvement qJ’5-hydroxytryptumine,,, receptors in the action of serotonin
At the concentration (30pM) used in this study, S-HT produces a marked hyperpolarization of pyramidal cells, mediated by 5-HT,A receptors’.3.‘0 and through this mechanism it might reduce neuronal response to weak stimuli. It has been demonstrated Fig. 8. KS 205-930 or ondansetron (GR-38032) prevent the that hyperpolarization of a pyramidal cell during S-HT-induced block of LTP induced by a weak PB. (A) Two high-frequency stimulation blocks the development representative experiments conducted in the presence of ICS 205-930 (1 nM; upper traces) or ondansetron (30 nM; n = 3; of the LTP in the recorded cell.” Thus, hyperpolarizlower traces). Stimulation (arrowheads) of the stratum ation of pyramidal cells via activation of S-HT,* radiatum with test pulses in the presence of either antagonist receptors could account for the failure of PB stimu(left) evoked an excitatory postsynaptic potential and a lation to induce LTP. population spike which were reduced after 5 min application of S-HT (middle). Stimulation with a weak PB at the fifth However, this explanation appears fully tenable minute of 5-HT application induced a potentiation of the only for the action of the 5-HT,, agonist 5-CT, since test response which lasted longer than 6Omin (right). (B) 5-HT also exerts other actions, notably the block of Mean (+S.E.M.; n = 7) amplitude population spikes calcium-mediated afterhyperpolarization, and acevoked by test pulses after stimulation (arrowhead) with a commodation which may counteract the effects of weak PB delivered at the fifth minute of S-HT application (30 FM; solid bar) in the presence of ICS 205-930 (1 nM; neuron hyperpolarization during the high-frequency open bar). afferent stimulation.’ Nonetheless, the fact that 5-CT mimicked the effect of 5-HT suggested an involvement of 5-HT,, receptors in the block of LTP by PB stimulation. The antagonism of 5-HT action by conditioning pulse followed by a burst of five stimuli at 100 Hz, was aimed at inducing the discharge of spiperone (5-HT, > 5-HT,, antagonist) and methiothepin (5-HT,, > 5-HT, antagonist) but not by CA1 pyramidal cells with temporal constraints which ritanserin (selective 5-HT, antagonist) validates the stimulated the theta rhythm and complex spike activity peculiar to the hippocampus. Previous work’2*24 notion that activation of 5-HT,, receptors WdS necessary for the action of 5-HT on PB-induced LTP. demonstrated that the persistent enhancement of the synaptic activity produced by PB stimulation shares Evidence for the involvement of’ 5_hydroxytryptamine, some properties (e.g. duration, sensitivity to Nreceptors in the action of serotonin methyl-D-aspartate receptor antagonists) with those An additional mechanism, mediated by 5-HT, reof LTP induced by longer trains at 100 Hz in the same ceptors, appeared to be involved in the action of hippocampal region. However, it appears from our 5-HT. In fact, the selective block of 5-HT, receptors results that the enhancement induced by PB stimuby ICS 205-930 or ondansetron (GR-38032) also lation was sensitive to S-HT modulatory action, antagonized the effect of 5-HT on the induction of whereas the LTP produced by a longer highLTP by a PB. frequency train was not. It is possible that the The finding was unexpected since none of the high-frequency train represents a type of afferent stimulation so powerful that it can overcome the effects of 5-HT so far described in CA1 pyramidal cells is selectively affected by 5-HT, antagonists,‘~9~‘0 inhibitory actions of 5-HT on LTP induction. In our experiments S-HT reduced the evoked popuand in particular no reduction of the 5-HT-induced hyperpolarization has been observed with concenlation spike in extracellular recordings from the pyratrations of ICS 205-930 higher than those used in the midal cell layer. Previous studies have shown that the present study. The presence of 5-HT, receptors medieffect of S-HT is restricted to this cell layer since no ating fast cationic inward currents in hippocampal change in the dendritic field response or presynaptic (possibly non-pyramidal) neurons has been demonfibre activity could be detected with very large (mM) concentrations of S-HT. I4 strated.” It has recently been consistently shown that Our data demonstrated that S-HT, applied im- 5-HT can enhance the spontaneous unitary GABAmediated inhibitory postsynaptic potentials recorded mediately after the PB stimulation, did not suppress from CA1 pyramidal cell~.‘~This effect is mediated by the LTP once produced. This indicates that the action 5-HT, receptors probably located on GABAergic of 5-HT was exerted on the induction of LTP and not
5-HT blocks primed burst-induced
interneurons. Anatomical data demonstrate the existence of 5-HT-containing terminals on GABAergic interneurons in the hippocampus.“. The above mentioned data showing an increase in the activity of GABAergic interneurons produced by 5-HT13 suggest that such a phenomenon may reinforce the SHTelicited inhibition of pyramidal cell activity to block the PB-induced LTP. Physiological
implications
The hippocampus displays distinct neurophysiological patterns of activity: the complex spike and the theta rhythm. These patterns have been correlated to “intelligent” behaviour.33 to some forms of learning, and to LTP.’ Although the relationship among these correlates of theta rhythm remains to be elucidated, it is remarkable that LTP induction is facilitated during the positive phase of theta rhythm2’ and that patterns of stimulation which reproduce complex spike activity and theta rhythm induce stable potentiation of synaptic transmission both in vitro’2.‘7.24 and in aivo.‘2~2” It deserves mention that in vivo the expression of LTP in the hippocampus is dependent on the integrity of its subcortical afferents, including those originating from raphe nucleC8 Anatomical data suggest that different populations of serotoninergic cells project from the rdphe nuclei to different cellular groups in the cerebral cortex and in the hippocampus. ‘b.3’ In particular, one subset of termi-
517
LTP
nals appears to innervate non-pyramidal, predominantly GABAergic cells.” Thus the activity of raphe nuclei may influence cell discharge of both pyramidal and non-pyramidal cells, possibly independently.
CONCLUSION
We propose that 5-HT, through membrane hyperpolarization, produces a reduction in the response of pyramidal neurons to weak and scattered inputs. The consequence would be to affect greatly the probability to “prime” the response of pyramidal cells to a single burst of high-frequency action potentials. leaving, however, the “priming” still achievable by stronger, or repetitive inputs. The action of 5-HT, released by possibly independent raphe projections onto non-pyramidal neurons, will further reduce the responsiveness of pyramidal cells to inputs offered with theta patterns. This could be obtained by increasing the discharge of GABAergic interneurons via 5-HT, receptors. The overall action of 5-HT would therefore counteract the expression of LTP following behaviourally “non-meaningful” inputs to avoid saturation of LTP mechanisms. Acknowiedgements~We thank Dr T. Gessi and Dr A. Nistri for their helpful suggestions. This work was supported by grantsfromM.U.R.S.T.(60%)andC.N.R.(90.03188.CT04 and 91.00595.CT04) to R.C.
REFERENCES
1. Andrade R. and Nicoll R. A. (1987) Pharmacologically distinct actions of serotonin on single pyramidal neurones of the rat hippocampus recorded in vitro. J. Physiol. 394, 9--l 24. 2. Assaf S. Y. and Miller J. J. (1978) The role of a raphe serotonin system in the control of septal unit activity and hippocampal desynchronization. Neuroscience 3, 539-550. 3. Beck S. G. (1989) 5-Carboxyamidotryptamine mimics only the 5-HT elicited hyperpolarization of hippocampal pyramidal cells via 5-HT,, receptor. Neurosci. Left. 99, 101-106. 4. Beck S. G.. Clarke W. P. and Goldfarb J. (1985) Spiperone differentiates multiple 5-hydroxytryptamine responses in rat hippocampal slices in vitro. Eur. J. Pharmcrc. 116, 195-197. 5. Bliss T. V. P. and Lsmo T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anesthetized rabbit following stimulation of the perforant path. J. Phvsiol. 232, 331-336. 6. Bliss T. V. P. and Lynch M. A. (1988) Long-term potentiation of synapiic transmission in the hippocampus: properties and mechanism. In Long-term Potentiation: From Biophysics IO Behavior (eds Landfield P. W. and Deadwyler S. A.), pp. 3-72. Alan Liss, New York. 7. Buzsriki G. (1989) Two-stage model of memory trace formation: a role for “noisy” brain states. Neuroscience 31, 55 l-570. 8. Buz&ki G. and Gage F. H. (1989) Absence of long-term potentiation in the subcortically deafferented dentate gyrus. Bruin Res. 484, 94-101. 9. Chaput Y., Araneda R. C. and Andrade R. (1990) Pharmacological and functional analysis of a novel serotonin receptor in the rat hippocampus. Eur. J. Pharmac. 182, 441456. 10. Colino A. and Halliwell J. V. (1987) Differential modulation of three separate K-conductances in hippocampal CA1 neurons by serotonin. Nature 328, 73-77. 11. Corradetti R., Moneti G., Moroni F., Pepeu G. and Wieraszko A. (1983) Electrical stimulation of the stratum radiatum increases the release and neosynthesis of aspartate. glutamate and y-aminobutyric acid in rat hippocampal slices. J. Neurochem. 41, 1518-1525. 12. Diamond D. M., Dunwiddie T. V. and Rose G. M. (1988) Characteristics of hippocampal primed burst potentiation in d-o and in the awake rat. J. Neurosci. 8, 40794088. 13. Guy N. and Ropert N. (1990) Serotonin facilitates GABAergic inhibition on hippocampal CA1 neurones in vitro. J. Physiol. 423, 94P. 14. Hiner B. C., Mauk M. D., Peroutka S. J. and Kocsis J. D. (1988) Buspirone, 8-OH-DPAT and ipsapirone: effects on hippocampal cerebellar and sciatic fiber excitability. Brain Res. 461, 1 9. 15. Kiihler C. (1984) The distribution of serotonin binding sites in the hippocampdl region of the rat brain. An autoradiographic study. Neuroscience 13, 667480. 16. Kosofsky B. E. and Molliver M. E. (1987) The serotoninercic innervation of cerebral cortex: different classes of axon terminals arise from dorsal and median raphe nuclei. Svnapse 1, 153-168.
518
R. COKKADETTI et ol.
17. Larson J. and Lynch G. (1986) Induction of synaptic potentiation in hippocampus by patterned stimulation involve< two events. Science 232, 9855988. 18. Leysen J. E., Niemegeers C. J. E., Tollenaere J. P. and Laduron P. M. (1978) Serotonergic component OI neuroleptic receptors. Nature 272, 1688171. 19. Malinow R. and Miller J. P. (1986) Postsynaptic hyperpolarization during conditioning reversibly blocks induction o! long-term potentiation. Nature 320, 529-530. 20. Oleskevich S. and Descarries L. (1990) Quantified distribution of the serotonin innervation in adult rat hippocampus. Neuroscience 34, 19-33. 21. Pavlides C., Greenstein Y. J., Grudman M. and Winson J. (1988) Long-term potentiation in the dentate gyrus is induced preferentially on the positive phase of Q-rhythm. Brain Res. 439, 383-387. R. and Pepeu G. (1990) 5Hydroxytryptamine blocks the long-term potentiation 22. Pugliese A. M., Ballerini L., Corradetti induced by primed bursts in the CA1 region of rat hippocampal slices. Pharmac. Res. 22 (Suppl. 2). 416. 23. Ropert N. (1988) Inhibitory action of serotonin in CA1 hippocampal neurons in vitro. Neuroscience 26, 69-81. 24. Rose G. M. and Dunwiddie T. V. (1986) Induction of hippocampal long-term potentiation using physiologically patterned stimulation. Neurosci. Left. 69, 244-248. 25. Segal M. (1975) Physiological and pharmacological evidence for a serotonergic projection to the hippocampus. Brain Res. 94, 115-131. 26. Segal M. (1976) 5-HT antagonists in rat hippocampus. Brain Res. 103, 161-166. 27. Segal M. (1980) The action of serotonin in the rat hippocampal slice preparation. J. Physiol. 303, 423439. 28. Staubli U. and Lynch G. (1987) Stable hippocampal long-term potentiation elicited by “theta” pattern stimulation. Brain Res. 435, 227-234. 29. Storm-Mathisen J. (1977) Localization of transmitter candidates in the brain: the hippocampal formation as a model. Prog. Neurobiol. 8, 119-181. 30. Swanson L. W., Kijhler C. and Bjorklund A. (1987) The limbic region. I: The septohippocampal system. In Handbook of ChemicalNeuroanatomy (eds Bjiirklund A., Hiikfelt T. and Swanson L. W.), Vol. 5, Part I, pp. 125-277. Elsevier, Amsterdam. 31. Tork I. and Homung J.-P. (1988) Interaction of serotonin and GABA in cerebra1 cortex and hippocampus. Sot. Neurosci. Abstr. 14, 274.6. 32. Tiirk I., Hornung J.-P. and Somogyi P. (1988) Serotoninergic innervation in the cerebra1 cortex. Neurosci. Left. Suppl. 30, Sl31. 33. Vandervolf C. H. and Baker G. B. (1986) Evidence that serotonin mediates non-cholinergic neocortical low voltage fast activity, non-cholinergic hippocampal rhythmical slow activity and contributes to intelligent behavior. Bruin Res. 374, 342-356. 34. Yakel J. L. and Jackson M. B. (1988) 5-HT3 receptors mediate rapid responses in cultured hippocampus and a clonal cell line. Neuron 1, 615621. (Accepted 17 July 1991)
