Abstract
5-Hydroxytryptamine (5-HT; serotonin) administration enhances GABAergic synaptic activity recorded in pyramidal neurons of the CA1 region of hippocampus. Previous studies have attributed this effect to the activation of HT-53 receptors located on GABAergic interneurons. During unrelated experiments, we noticed that under our recording conditions, 5-HT can still increase GABAergic synaptic activity after the complete blockade of 5-HT3 receptors. This indicated the involvement of an additional 5-HT receptor subtype. Therefore, we reinvestigated the effects of 5-HT on GABAergic synaptic activity recorded in pyramidal cells of the CA1 region. The ability of 5-HT to increase GABAergic synaptic activity in the presence of 5-HT3 receptor blockade was mimicked by the selective 5-HT2 agonist (±)-1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane and blocked by the selective 5-HT2 antagonist ketanserin. This indicated that the additional 5-HT receptor belongs to 5-HT2 receptor family. 5-HT2 receptor activation resulted in an increase in the frequency of spontaneous inhibitory postsynaptic currents as well as a shift in their amplitude distribution toward larger sizes. These effects were absent in the presence of tetrodotoxin. We interpret these results to indicate that 5-HT2 receptors activate GABAergic interneurons in the slice, leading to an increase in GABAergic synaptic activity onto pyramidal cells of the CA1 region.
The mammalian hippocampus receives a dense serotonergic innervation originating from the midbrain raphe nuclei (Moore and Halaris, 1975;Azmitia and Segal, 1978). In the CA fields, incoming serotonergic fibers innervate both pyramidal cells and inhibitory GABAergic interneurons (Freund et al., 1990). Previous electrophysiological studies have identified a variety of 5-HT-mediated responses in pyramidal neurons (Segal, 1980; Beck and Goldfarb, 1985;Andrade and Nicoll, 1987; Ropert, 1988; Andrade and Chaput, 1991). Less is known however about the effects of 5-HT on GABAergic interneurons.
In the CA1 region, the administration of 5-HT results in a large increase in spontaneous GABAergic synaptic potentials, an effect mediated, at least in part, by the activation of 5-HT receptors of the 5-HT3 subtype (Ropert and Guy, 1991). A quantal analysis of the increased synaptic activity indicates that 5-HT3 receptors elicit this effect by activating GABAergic interneurons (Ropert and Guy, 1991). Direct evidence for the depolarization and activation of GABAergic interneurons by 5-HT3 receptors has been obtained recently in the dentate gyrus (Kawa, 1994). Anatomic studies, however, have reported the expression of additional 5-HT receptor subtypes on hippocampal interneurons (Pompeiano et al., 1994; Wright et al., 1995). Thus, it is likely that multiple 5-HT receptors subtypes are involved in the regulation of these cells. We now present evidence that in addition to 5-HT3 receptors, 5-HT can act on receptors of the 5-HT2 subtype family to depolarize and excite GABAergic interneurons of the CA1 region.
Materials and Methods
Whole-cell recordings were obtained in vitro from pyramidal neurons of the CA1 region in rat hippocampal slices as previously described (Haj-Dahmane and Andrade, 1996; Torres et al., 1996). Briefly, male Sprague-Dawley rats (175–250 g) were killed under halothane anesthesia. The brain was removed and cooled in ice-cold Ringer’s (composition in mM: NaCl 119, KCl 2.5, MgSO4 1.3, CaCl2 2.5, NaH2PO4 1, NaHCO3 26.2, glucose 11), and the right and left hippocampi were dissected. The hippocampi were affixed to a metal block using cyanoacrylate glue and supported with an agar block. Hippocampal slices (400 μm nominal thickness) were cut using a vibratome. The slices were incubated for ≥1 hr in an interface chamber for recovery and storage before recording. One slice was transferred at a time to a recording chamber (Nicoll and Alger, 1981). In the recording chamber, the slice was held submerged between two nylon nets and superfused with Ringer’s bubbled to equilibrium with 95% O2/5% CO2. All experiments were conducted at 30 ± 1°C.
Whole-cell recordings were obtained from pyramidal neurons of the CA1 region using the blind patch technique (Blanton et al., 1989). Patch electrodes were pulled from 1.2 mm o.d. borosylicate glass (Glass Company of America, Bargaintown, NJ) using a Flaming-Brown type horizontal puller (model P-97; Sutter Instruments, Novato, CA) and filled with an intracellular solution containing high chloride (composition in mM: KCl 110, NaCl 5, CaCl2 1, MgCl2 2, EGTA 10, HEPES 10, ATP 2, GTP 0.5, pH 7.35). ECl under these recording conditions is near 0 mV, and cells were held at hyperpolarized potentials such that GABAergic synaptic potentials appear as inward currents. Electrodes filled with this intracellular solution exhibit resistances in the range of 4 to 7 MΩ. Access resistances ranged from 10 to 40 MΩ. In most experiments, the intracellular solution also contained 2 mM QX 314 to block sodium channels. QX 314 also blocks G protein-activated potassium channels and hence GABAB-mediated synaptic potentials (Nathan et al., 1990).
Electrical signals were measured with an Axoclamp 2A (Axon Instruments, Foster City, CA) amplifier operating either in current clamp or continuous voltage clamp mode. In all the voltage clamp experiments, series resistance was compensated by ∼70% using the built-in circuit of the amplifier. Signals were recorded continuously online using a paper chart recorder (model 3200, Gould, Valleyview, OH) and digitized using an Intel 80486-based computer equipped with a 12-bit A/D converter under the control of pClamp 5.5 (Axon Instruments). To conduct a quantitative analysis of the 5-HT-induced synaptic activity, current was filtered at 3 kHz and digitized on line (3.3 kHz sampling rate), and 10 consecutive episodes, each 666 msec long, were repeatedly acquired and saved on disk. The synaptic analysis was based on five consecutive acquisitions (33.3 sec) taken over a period of 3 to 5 min before and after each experimental manipulation. For the data analysis, we used MINI (kindly provided by Dr. J. H. Steinbach), a computer program that detects and measures spontaneous synaptic events based on amplitude, duration and rate of rise criteria. The selection criteria was optimized for each cell, and flagged events were accepted or rejected after visual inspection of the records. This procedure minimized the possibility of interpreting closely spaced IPSCs as a single large amplitude event. Quantitative comparisons of the extracted amplitude and inter-sIPSC distributions were conducted with the K-S test using the statistical software package Statistica (Tulsa, OK). Figures were prepared using the scientific graphic program Origin (Microcal Software, Northampton, MA) and the drawing program CorelDraw (Corel, Ottawa, Canada).
Pyramidal neurons of the CA1 region express 5-HT1A and 5-HT4 receptors (Andrade and Nicoll, 1987; Andrade and Chaput, 1991). Because the postsynaptic effects mediated by these receptors could have complicated the analysis of GABAergic synaptic potentials, all of the experiments described here were conducted in the presence of BMY 7378 (8-[2-[4(2-methoxyphenyl)-1-piperazinyl]ethyl]-8-azaspirol[4,5]decane-7,9-dione dihydrochloride; 10 μM) and GR 113808 ([1-[2-(methylsulfonylamino)ethyl]-4-piperidinyl]methyl-1-methyl-1H-indol-3-carboxylate; 1 μM) to block these receptors (Chaput et al., 1990;Torres et al., 1994). Most experiments were also conducted in the presence of 1 to 2 μM tropisetron [(3-α-tropanyl)-H-indol-3-carboxylic acid ester; ICS 205–930] to block 5-HT3 receptors (Richardsonet al., 1985). For all of the experiments presented in this study, drugs were administered in the bath at known concentrations. In the case of the bicuculline [6-(5,6,7,8-tetrahydro-6-methyl-1,3-dioxolo[4,5-g]isoquinolin-5-yl)furo[3,4-e]-1,3-benzodioxol-8(6H)-one] experiments, a knife cut was made between the CA1 and CA3 regions to prevent epileptiform discharges. This manipulation did not affect the ability of 5-HT to enhance synaptic activity in the CA1 region. Most drugs and reagents used in this study were obtained from Sigma (St. Louis, MO). Tropisetron was a kind gift from Sandoz (Basel, Switzerland). BMY 7378 was from Bristol-Myers-Squibb (Wallingford, CT). GR 113808 was from Glaxo (Greenford, UK). Ketanserin was from Janssen (Beerse, Belgium). TTX was obtained from Calbiochem (La Jolla, CA). APV, CNQX, 2-Me-5-HT, DOI and MDL 72222 (3-tropanyl-3,5-dichlorobenzoate) were from Research Biochemicals (Natick, MA).
Results
When pyramidal neurons of the CA1 region are recorded using high intracellular chloride, one of the most striking observations is the presence of a steady background of spontaneous synaptic events. When these events are recorded in voltage clamp mode, they appear as spontaneous inward currents (fig. 1). The vast majority of these synaptic events are blocked by bicuculline administration (15 μM, fig. 1B, n = 5 cells) or by reducing intracellular chloride and holding the cell near ECl, indicating that these synaptic events represent predominantly GABAA receptor-mediated synaptic potentials (n = 5 cells). We refer to these synaptic currents as sIPSCs.
As previously reported (Ropert and Guy, 1991), the administration of 5-HT (10–30 μM) greatly increases this spontaneous synaptic activity (fig. 1A, n = 6 cells). This increase reflects an enhancement of GABAergic synaptic transmission because the effect of 5-HT is completely blocked by administration of bicuculline (15 μM,n = 3 cells) but persisted in the presence of CNQX and APV (10 and 50 μM, n = 8 cells, not shown).
Previous studies have shown that this 5-HT enhancement of GABAergic synaptic activity is mediated by 5-HT receptors of the 5-HT3 subtype (Ropert and Guy, 1991). Thus, we were surprised to see that even in the presence of the 5-HT3 receptor blocker tropisetron (1–2 μM), 5-HT could still elicit an increase in GABAergic sIPSCs under our recording conditions (n = 16 cells). This effect does not seem to be as robust as that observed in the absence of the 5-HT3 receptor antagonist but is clearly evident in a majority of cells recorded (16 of 21 cells, or 76% of the cells tested). The ability of 5-HT to increase GABAergic synaptic activity exhibited an onset comparable to that of 5-HT1Aor 5-HT4 receptor-mediated responses recorded in this preparation (generally on the order of 1–2 min) (Andrade and Nicoll, 1987; Andrade and Chaput, 1991; Andrade, 1993), but recovery is very slow, usually taking several minutes. This effect of 5-HT was not affected by the administration of CNQX and APV (10 and 50 μM,n = 5 cells).
One possible explanation for the failure of tropisetron to block the effect of 5-HT could be incomplete blockade of the 5-HT3 receptors. This possibility seems unlikely because these experiments used tropisetron at concentrations close to 103 times theKd value for 5-HT3 receptors (Hoyer et al., 1994). Nevertheless, to address this possibility, we first compared the effects of 5-HT and the selective 5-HT3 agonist 2-Me-5-HT. As illustrated in figure 2, in the presence of tropisetron (1 μM), 2-Me-5-HT (35 μM) is without effect on synaptic activity (n = 7 cells), whereas under the same condition, 5-HT is still active. We then examined the effect of coadministration of tropisetron (1 μM) and MDL 72222 (1 μM), a second 5-HT3 receptor antagonist (Hoyer et al., 1994). Under these conditions, 5-HT is still capable of enhancing spontaneous synaptic activity (n = 2 cells, not shown). Thus, 5-HT can act on non-5-HT3receptors to enhance GABAergic synaptic activity in the CA1 region.
To characterize this response, we analyzed quantitatively the effect of 5-HT on sIPSCs. All of these experiments were conducted in the presence of 1 to 2 μM tropisetron to isolate the non-5-HT3 receptor-mediated enhancement in GABAergic synaptic activity. Figure 3 illustrates the effect of 5-HT on sIPSC frequency and amplitude. 5-HT produces a large increase in sIPSC frequency that results in a significant shift in the cumulative distribution curve of inter-sIPSC intervals toward shorter durations (fig. 3, n = 16 cells, P < .05 for each cell, K-S test). This frequency increase is seen in all sIPSC size classes. In addition, there is a broadening of the amplitude distribution to include larger events not present in the sampling obtained under control conditions. As a result, 5-HT also produces a significant shift in the cumulative distribution of sIPSC amplitudes toward larger sIPSC sizes (fig. 3, n = 16 cells, P < .05 for each cell, K-S test).
In situ hybridization studies have shown that a subset of GABAergic interneurons in the CA1 region of hippocampus express mRNA for the 5-HT2A and 5-HT2Creceptor subtypes (Pompeiano et al., 1994; Wright et al., 1995). This raised the possibility that one or both of these receptor subtypes could mediate the ability of 5-HT to enhance GABAergic synaptic activity. To test this possibility, we first examined the effect of the selective 5-HT2agonist DOI. As illustrated in figure 4, bath administration of DOI (30 μM) mimics the effects of 5-HT by eliciting an increase in the frequency of sIPSCs (n = 6 of 9 cells tested, P < .05 for each cell, K-S test) and a shift in the cumulative amplitude distribution of the synaptic currents toward larger sizes (n = 6 of 9 cells tested, P < .05 for each cell, K-S test). This effect of DOI recovers slowly over a period of 15 to 20 min after its removal from the bath.
We next tested the effect of ketanserin, a selective 5-HT2 antagonist. We first administered 5-HT (30 μM) to determine its effect on sIPSC amplitude and frequency; we then applied ketanserin (1 μM), and after 3 to 5 min, we retested 5-HT. As illustrated in figure 5, the effect of 5-HT is blocked by ketanserin (n = 6 cells). Although this experiment suggests that ketanserin antagonizes the effect of 5-HT, the interpretation of the result is complicated by the potential desensitization of 5-HT receptors (Araneda and Andrade, 1991), which could produce a false impression of antagonism. Therefore, to control for this possibility, we also examine whether a second application of 5-HT could still elicit an enhancement in sIPSC activity. In 2 cells tested, no desensitization to 5-HT effect is observed (not shown). The administration of ketanserin also reversed the effect of DOI (n = 3 cells tested).
The observation that 5-HT shifts the sIPSC amplitude distribution toward larger sizes could be explained if 5-HT is increasing mean quantal content (the number of vesicles released to produce each sIPSC). Alternatively, 5-HT could have shifted the distribution by increasing mean quantal size (the mean postsynaptic response produced by the release of the content in a single GABA vesicle). To distinguish between these possibilities, we examine the effects of 5-HT in the presence of TTX (1 μM), which blocks action potential-dependent release of GABA. Under this condition, release events should be composed exclusively of single vesicle release events (miniature IPSCs, mIPSCs) and thus allow us to directly measure quantal size. As illustrated in figure 6 in the presence of TTX, the administration of DOI has no effect on the amplitude distribution of mIPSCs (n = 7 cells). Similarly, in most cells tested (6 of 7), DOI also fails to increase the frequency of mIPSCs observed in the presence of TTX. In the remaining cell, DOI produces a small increase in the frequency of mIPSCs.
Discussion
The administration of 5-HT results in an enhancement of GABAergic synaptic activity recorded in CA1 pyramidal neurons of the rat hippocampus. This response has been attributed previously to the activation of 5-HT receptors of the 5-HT3 subtype located on inhibitory GABAergic interneurons (Ropert and Guy, 1991). In the present study, we observed that 5-HT increases sIPSC activity even after administration of the 5-HT3 receptor blockers tropisetron and MDL 72222. The failure of tropisetron to completely antagonize the effect of 5-HT does not reflect incomplete 5-HT3 receptor blockade because under the same condition, the ability of the selective 5-HT3receptor agonist 2-Me-5-HT to enhance sIPSCs is completely abolished. Thus, one or more additional 5-HT receptor subtypes must contribute to the ability of 5-HT to enhance GABAergic synaptic transmission in the CA1 region. Because 5-HT is still capable of enhancing sIPSCs in the presence of CNQX and APV, which blocks glutamate-mediated synaptic potentials, these receptors must be located on GABAergic interneurons.
Anatomic studies indicate that a subset of hippocampal interneurons in the CA1 region express mRNA for the 5-HT2A and/or 5-HT2C receptor subtypes (Pompeiano et al., 1994; Wright et al., 1995). These observations suggest that the non-5-HT3 receptor capable of enhancing sIPSCs in this area might belong to the 5-HT2 subtype. Consistent with this possibility, the ability of 5-HT to elicit an increase in GABAergic synaptic transmission is mimicked by DOI. DOI is a selective agonist for receptors of the 5-HT2 family and displays low affinity for receptors in the 5-HT1, 5-HT4, 5-HT5, 5-HT6 and 5-HT7 families (Hoyer et al., 1994; Schoeffter and Waeber, 1994). As such, its ability to mimic the effect of 5-HT on sIPSCs argues for the involvement of receptors belonging to the 5-HT2family.
The 5-HT-induced enhancement of GABAergic synaptic activity is also blocked by the potent 5-HT2 antagonist ketanserin. At the concentration used in the present study, this blocker is expected to have little or no effect on a serotonergic response mediated by most non-5-HT2 receptors. Aside from receptors in the 5-HT2 family, only 5-HT7 receptors exhibit submicromolar affinity for ketanserin (pKd = 6.69; Hoyeret al., 1994). Under our experimental conditions (1 μM ketanserin), we would expect only slight antagonism of 5-HT7 receptors (<20% of a maximal response) but close to complete antagonism of a response mediated by 5-HT2 receptors. We observed complete blockade of the response to 5-HT. Thus, from the agonist and antagonist results, we conclude that 5-HT enhances GABAergic synaptic transmission in the CA1 region most likely by activating 5-HT receptors of the 5-HT2 family.
5-HT2 receptors could enhance GABAergic synaptic transmission through a presynaptic or a postsynaptic mechanism. In this study, 5-HT2 receptor activation is found to enhance the frequency of sIPSCs and shift their amplitude distribution toward larger sizes. In contrast, activation of these receptors has little if any effect in the presence of TTX. These results indicate that 5-HT acts presynaptically to increase GABA release.
Several mechanisms could account for the increase in GABA release. The simplest mechanism is that 5-HT acts on 5-HT2receptors to depolarize and excite GABAergic interneurons. The increase in interneuron firing would explain the increase in the frequency of sIPSCs recorded on pyramidal neurons. It could also explain the observed shift in the sIPSC amplitude distribution. The recorded population of IPSCs represent a mixture of action potential-dependent IPSCs and mIPSCs. 5-HT, by increasing firing of the interneurons, could increase the contribution of action potential-dependent (multiquantal) sIPSCs to the overall population of recorded synaptic events. This would result in a shift the sIPSC amplitude distribution (Mintz and Korn, 1991). Alternatively, it is possible that 5-HT2 receptors could shift in the sIPSC amplitude distribution and elicit the appearance of large-size sIPSC by also enhancing the probability of GABA release from axon terminals (Mintz and Korn, 1991). However, because 5-HT is essentially without effect on mIPSCs frequency, this possibility would require that 5-HT act on axonal terminals to preferentially enhance action potential-dependent release of GABA. Of course, 5-HT2 receptors could act cooperatively at the somatodendritic level and at axonal terminals to enhance GABAergic synaptic transmission in the CA1 region. Finally, it is possible that the shift in the amplitude distribution might represent recruitment of GABAergic neurons with a large quantal size. Further studies will be required to distinguish among these possibilities.
GABAergic interneurons express 5-HT2A and/or 5-HT2C receptors (Pompeiano et al., 1994; Wright et al., 1995). In the present study, we did not try to distinguish between these receptor subtypes. It is difficult to establish reliable concentration-response relationships for sIPSCs in this slice preparation presumably because the 5-HT-induced increase in sIPSPs is contingent on a limited numbers of interneurons reaching firing threshold. This, coupled to the limited selectivity of the currently available antagonists, makes it impractical to try to distinguish between the involvement of 5-HT2A and 5-HT2C receptors. Studies involving direct recording from inhibitory interneurons will be required to determine the roles of these different receptors in mediating the effects of 5-HT on GABAergic interneurons in the CA1 region.
Several distinct classes of GABAergic interneurons can be distinguished in the hippocampus based on morphological, biochemical and functional criteria (Buhl et al., 1994). Previous studies have shown a strong serotonergic innervation of at least one subpopulation of GABAergic interneurons in this region (Freund et al., 1990). Consistent with this innervation, 5-HT2A and 5-HT2C receptor mRNAs are expressed in neurons, presumably interneurons, in the stratum oriens of the CA1–CA3 field. Similarly, a population of presumed GABAergic interneurons in the stratum lacunosum moleculare expresses 5-HT2Areceptor mRNA (Pompeiano et al., 1994). It is tempting to speculate that the expression of 5-HT2 receptors might define functional class or classes of interneurons. Interestingly, regulation of GABAergic inhibitory interneurons by receptors of the 5-HT2 family is not restricted to hippocampus. In pyriform cortex (Sheldon and Aghajanian, 1990;Sheldon and Aghajanian, 1991) and dentate gyrus (Piguet and Galvan, 1994), 5-HT has also been reported to excite GABAergic interneurons by acting on receptors of the 5-HT2 family. Equally interesting, it appears that 5-HT activates hippocampal GABAergic interneurons by acting on both ionotropic (5-HT3) and metabotropic (5-HT2) receptors. This contrasts with the effects of 5-HT on pyramidal neurons, which appear to be mediated solely by metabotropic 5-HT receptors (Andrade and Nicoll, 1987; Ropert, 1988; Andrade and Chaput, 1991; Beck et al., 1992). It will be interesting to find out whether 5-HT2 and 5-HT3 receptors are coexpressed in a common subpopulation of interneurons or in segregated populations.
In conclusion, the administration of 5-HT enhances GABAergic synaptic activity recorded on pyramidal neurons of the CA1 region. This effect is mediated by receptors of the 5-HT2 family, in addition to 5-HT3 receptors. 5-HT may activate these receptors in the somatodendritic region to depolarize and excite GABAergic interneurons; it might also act on interneuron terminals to increase the probability of GABA release on the arrival of action potentials.
Acknowledgments
We thank Dr. S. Haj-Dahmane for experimental advice and Dr. C. L. Arfken for reading the manuscript.
Footnotes
-
Send reprint requests to: Dr. Rodrigo Andrade, Department of Psychiatry and Behavioral Neurosciences, Wayne State University School of Medicine, 2309 Scott Hall, 540 E. Canfield, Detroit, MI 48201. E-mail randrade{at}med.wayne.edu
-
↵1 This work was supported by NIH grants MH43985 (R.A) and AA09829 (R.S.).
- Abbreviations:
- GABA
- γ-aminobutyric acid
- 5-HT
- 5-hydroxytryptamine
- mIPSC
- miniature inhibitory postsynaptic current
- K-S
- Kolmogorov-Smirnov
- sIPSC
- spontaneous inhibitory postsynaptic current
- TTX
- tetrodotoxin
- APV
- DL-2-aminophosphonovaleric acid
- CNQX
- 6-cyano-7-nitro-quinoxaline-2,3-dione
- 2-Me-5-HT
- 2-methyl-5-hydroxytryptamine maleate
- DOI
- (±)-1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane
- Received March 31, 1997.
- Accepted January 15, 1998.
- The American Society for Pharmacology and Experimental Therapeutics