Action of serotonin on the hyperpolarization-activated cation current (I h ) in rat CA1 hippocampal neurons

Transcription

1 European Journal of Neuroscience, Vol. 11, pp. 3093±3100, 1999 ã European Neuroscience Association Action of serotonin on the hyperpolarization-activated cation current (I h ) in rat CA1 hippocampal neurons Sonia Gasparini* and Dario DiFrancesco UniversitaÁ degli Studi di Milano, Dipartimento di Fisiologia e Biochimica Generali, Elettro siologia, Via Celoria 26, Milano, Italy Keywords: 5-HT, camp, hippocampus, hyperpolarization-activated cation current (I h ) Abstract We studied the effects of serotonin (5-HT) on hippocampal CA1 pyramidal neurons. In current-clamp mode, 5-HT induced a hyperpolarization and reduction of excitability due to the opening of inward recti er K + channels, followed by a late depolarization and partial restoration of excitability. These two components could be dissociated, as in the presence of BaCl 2 to block K + channels, 5-HT induced a depolarization accompanied by a reduction of membrane resistance, whereas in the presence of ZD 7288 [4-(N-ethyl-Nphenylamino)-1,2-dimethyl-6-(methylamino) pyrimidinium chloride], a selective blocker of the hyperpolarization-activated cation current (I h ), 5-HT only hyperpolarized neurons. We then studied the action of 5-HT on I h in voltage-clamp conditions. 5-HT increased I h at ±90 mv by % and decreased the time constant of activation by % (n = 16), suggesting a shift in the voltage dependence of the current towards more positive potentials; however, the fully activated current measured at ±140 mv also increased (by %, n = 14); this increase was blocked by ZD 7288, implying an effect of 5-HT on the maximal conductance of I h. Both the shift of activation curve and the increase in maximal conductance were con rmed by data obtained with ramp protocols. Perfusion with the membrane-permeable analogue of camp, 8-bromoadenosine 3 5 -cyclic monophosphate (8-Br-cAMP), increased I h both at ±90 and ±140 mv, although the changes induced were smaller than those due to 5-HT. Our data indicate that 5-HT modulates I h by shifting its activation curve to more positive voltages and by increasing its maximal conductance, and that this action is likely to contribute to the 5-HT modulation of excitability of CA1 cells. Introduction Correspondence: Dr S. Gasparini, International School for Advanced Studies (SISSA), Biophysics Sector, Via Beirut 2-4, Trieste, Italy. gaspari@sissa.it *Present address: International School for Advanced Studies (SISSA), Biophysics Sector, Via Beirut 2-4, Trieste, Italy. Received 24 November 1998, revised 23 March 1999, accepted 21 April 1999 The modulation of ionic currents by neurotransmitters plays an important role in the control of activity of cells and neurons. In the hippocampus, for example, several neurotransmitters mediate the ow of information through the hippocampal circuitry by modulating conductances involved in the generation of the action potential and its repolarization, and in the control of the resting membrane potential or accommodation of spike frequency during prolonged stimulations (Brown et al., 1990), thus contributing to different states of excitability of the pre- or postsynaptic neurons at different times. Three major currents have been reported to control the resting membrane potential of CA1 pyramidal neurons. Two of these are K + currents: a voltage-insensitive leakage K + current modulated in opposite ways by somatostatin and muscarinic agonists (Madison et al., 1987; Schweitzer et al., 1998) which sets the resting potential close to E K (Storm, 1990), and a K + current [muscarinic K + current, I K(M) ] inhibited by acetylcholine, which is activated from a threshold of ±60 mv and thus can dampen depolarizing inputs without contributing to the resting potential (Halliwell & Adams, 1982; Madison et al., 1987). A third component is a mixed hyperpolarization-activated Na + /K + current (I h ), active at potentials more negative than ±40 mv; this inward current contributes a moderate depolarization (» 5 mv) at rest (Maccaferri et al., 1993; Gasparini & DiFrancesco, 1997). A hyperpolarization-activated cation current (I h ) with similar characteristics is expressed in a variety of cells (for review see DiFrancesco, 1993; Pape, 1996), and plays an important role in the generation and control of rhythmic oscillatory activity in the pacemaker region of the heart (where it has been rst reported and named I f, Brown et al., 1979), in the thalamus (Pape & McCormick, 1989; McCormick & Bal, 1997), in the hypothalamic suprachiasmatic nucleus (Akasu et al., 1993) and in inferior olivary neurons (Bal & McCormick, 1997). The interest in this conductance and its functional role has been renewed by the recent cloning of the genes encoding pacemaker channels (HCN family, Clapham, 1998). I h has been reported to be modulated by acetylcholine in hippocampal CA1 pyramidal neurons (Colino & Halliwell, 1993), by noradrenaline in pyramidal neurons and inhibitory interneurons (Pedarzani & Storm, 1995; Bergles et al., 1996; Maccaferri & McBain, 1996), and by opioids in interneurons (Svoboda & Lupica, 1998). The hippocampus is richly supplied with serotonin (5-HT)- containing afferents, and 5-HT itself has been shown to affect the activity of the hippocampus through different receptors (Segal, 1989). 5-HT has been shown to modulate I h in different areas of the central nervous system (Bobker & Williams, 1989; Pape & McCormick, 1989; Larkman & Kelly, 1992), but a modulation of I h by 5-HT in the hippocampus has not yet been reported. We thus focused our attention on a possible action of 5-HT on I h in hippocampal CA1 pyramidal neurons, to determine whether I h might mediate some of the actions of the transmitter on hippocampal excitability.

2 3094 S. Gasparini and D. DiFrancesco We report here that, following the early hyperpolarization due to the opening of inward recti er K + channels (Andrade & Nicoll, 1987; Colino & Halliwell, 1987), 5-HT induces a depolarization of resting membrane potential of hippocampal CA1 neurons. We investigated this second depolarizing phase and found that it is attributable to the activation of the I h current. Materials and methods Slice preparation and solutions Transverse hippocampal slices were obtained from young male Wistar rats (17±25 days, Charles River, Calco, Italy) as described elsewhere (Maccaferri et al., 1993). Brie y, rats were anaesthetized with ether, killed by decapitation, and the brain was rapidly dissected out in an ice-cold solution containing (in mm): NaCl, 120; KCl, 3.1; MgCl 2, 4; CaCl 2,1;KH 2 PO 4, 1.25; NaHCO 3, 26; glucose, 10. Slices (300 mm thick) were cut by a vibratome (Series 1000, Technical Products International, St. Louis, MO, USA) and stored at room temperature in a holding bath containing the same saline solution as above. After a recovery period of at least 1 h, an individual slice was transferred to the recording chamber where it was held by a mesh formed by nylon threads glued to a U-shaped platinum bar, fully submerged and continuously superfused with an arti cial cerebrospinal uid (ACSF) composed of (in mm): NaCl, 120; KCl, 3.1; MgCl 2, 1; CaCl 2,2;KH 2 PO 4, 1.25; NaHCO 3, 26; glucose, 10. To block all conductances except I h and to improve clamp conditions, in voltageclamp experiments, ACSF was replaced by a control solution composed of (in mm): NaCl, 110; NaHCO 3, 24, KCl, 5; MgCl 2,1, BaCl 2, 2, CdCl 2, 0.1±0.3, tetraethylammonium chloride (TEA-Cl), 10; 4-aminopyridine (4-AP), 2; tetrodotoxin (TTX), 0.001; glucose, 10. In this modi ed solution, KH 2 PO 4 was not included (and therefore KCl raised) to avoid precipitation. All solutions were FIG. 1. Effect of 5-HT on the membrane potential and conductance of a CA1 neuron in control conditions. Top: biphasic response of the resting membrane potential, consisting of an early hyperpolarization followed by a late depolarization, induced by 5-HT (50 mm) in current-clamp conditions. Bottom: time course of the membrane conductance change, monitored by injecting a constant hyperpolarizing current (±75 pa for 300 ms, every 4 s). equilibrated with 95% O 2 ±5% CO 2 to a ph of 7.4. Stock solutions of 5-HT (10 mm) were prepared weekly and stored frozen in 200-mL vials to avoid oxidization, and were added to solutions as required just before the experiment. Test solutions contained 30±50 mm 5-HT, concentrations known to elicit reversible responses in CA1 neurons (Andrade & Nicoll, 1987). All drugs were obtained from Sigma- Aldrich (Milano, Italy), except for TTX which was supplied by Calbiochem-Inalco (Milano, Italy). ZD 7288 [4-(N-ethyl-N-phenylamino)-1,2-dimethyl-6-(methylamino) pyrimidinium chloride] was a generous gift from Zeneca Pharmaceuticals, Maccles eld, UK (Dr I. Briggs). Electrophysiological recordings Whole-cell recordings from single CA1 pyramidal neurons were performed at room temperature using the patch-clamp technique, by using an Axopatch 200 ampli er (Axon Instruments, Foster City, CA, USA). Seals were obtained using the blind technique (Blanton et al., 1989), with the recording electrode positioned by visual localization in the stratum pyramidale. After impalement, cells were routinely identi ed as pyramidal neurons by observation of their ring pattern in response to a depolarizing pulse (100 pa for 400 ms). The pipette was lled with a solution containing (in mm): K- gluconate, 100; MgCl 2, 5; adenosine trisphosphate (ATP), 2; guanosine trisphosphate (GTP), 0.5; N-2-hydroxyethylpiperazine- N -2-ethanesulphonic acid (HEPES), 0.5; ethylenebis (oxonitrilo) tetra-acetic acid (EGTA), 40; ph 7.2 (resistance 3±8 MW). A superfusing pipette was positioned close to the recording neuron immediately after the impalement to allow a fast exchange of solutions. Leakage and capacitative components were corrected for either by subtraction of traces in the presence of CsCl (5 mm) or ZD 7288 (100 mm), or of current traces recorded during 10-mV steps (from ±35 to ±45 mv) after averaging (n = 10) and scaling up. Cell capacitance was obtained by integrating the capacitative transient during 10 mv hyperpolarizations and scaling to the voltage step. Data presented here represent mean 6 SEM, and were obtained from 49 slices. Results Effects of 5-HT on the membrane potential of CA1 neurons We performed a rst set of experiments investigating the effect of 5- HT in current-clamp conditions. When recording membrane potentials from CA1 pyramidal cells, superfusion with 5-HT (50 mm) typically yielded a biphasic response (Fig. 1, top), consisting of an early hyperpolarization followed by a slow recovery phase toward the resting level, observed while still in the presence of 5-HT. Both these phases were accompanied by an increase in membrane conductance, measured as the voltage displacement due to injection of a constant hyperpolarizing current pulse (Fig. 1, bottom). Membrane potential and conductance recovered slowly back to control values upon washout (Fig. 1). The early hyperpolarization may result from the direct activation of an inward recti er K + channel by a G-protein, following activation of 5-HT 1A receptors as previously reported (Andrade & Nicoll, 1987; Colino & Halliwell, 1987). This component is abolished by perfusion of BaCl 2 (2 mm), through channel block, or perfusion of a 5-HT 1A receptor antagonist (Andrade & Nicoll, 1987). In agreement with this interpretation, block of the early hyperpolarization was obtained in all cells in the presence of BaCl 2 (n = 5). We then routinely used BaCl 2 (2 mm) to block the G-protein-activated K + component, and performed another set of experiments to analyse any residual action

3 Modulation of I h by 5-HT in hippocampus 3095 of 5-HT. Perfusion with Ba 2+ caused a depolarization of the resting membrane potential, as expected from the block of K + currents active at rest [I K(M) and the leak K + current (Storm, 1990)]. We compensated for this depolarization by injecting a tonic hyperpolarizing current, until the resting membrane potential (» ±60 mv) was restored. Under these conditions, only a small depolarization was recorded during perfusion with 5-HT (from ±60.1 mv in control conditions to ±57.8 mv in the presence of 5-HT, for the cell in Fig. 2), accompanied by an increase of membrane conductance (from 4.20 to 5.05 ns, %, Fig. 2, bottom). In n = 3 experiments, we found a depolarization of mv with an increase in membrane conductance of %. Partial recovery from the early hyperpolarization seen in control solution was thus attributable to a slow depolarizing phase which could be observed separately from the early hyperpolarization. Both the depolarization and the increase in membrane conductance observed in the presence of Ba 2+ could result from the activation of previously inactive channels or the increase of an inward current active at rest. Because one of the actions of 5-HT in CA1 pyramidal cells is to activate adenylate cyclase (Torres et al., 1995), we hypothesized that the component involved in this response could be the hyperpolarization-activated, camp-gated I h current, which activates at voltages more negative than ±40 mv in CA1 pyramidal cells (Maccaferri et al., 1993; Gasparini & DiFrancesco, 1997). We therefore repeated the protocol of Fig. 1 above in the presence of ZD 7288, a selective blocker of I h (BoSmith et al., 1993; Harris & Constanti, 1995; Gasparini & DiFrancesco, 1997). The perfusion of ZD 7288 (100 mm) caused the membrane potential of CA1 neurons to hyperpolarize and the membrane conductance to increase as expected from the block of an inward current active at rest, as previously reported (Gasparini & DiFrancesco, 1997); a tonic depolarizing current (25 pa) was thus injected to set the membrane potential at ±60 mv. Under these conditions, 5-HT (50 mm) produced a fast hyperpolarization of the membrane potential associated with an increase in membrane conductance, but no depolarizing phase during 5-HT perfusion. The changes induced by 5-HT were both reversible upon washout (Fig. 3). Similar effects of 5-HT were observed in a total of six cells in the presence of ZD These results indicate that the late depolarizing phase observed in physiological conditions during 5-HT perfusion can be attributed to the modulation of I h. Though small, the late depolarization was superimposed on the early hyperpolarization generated by the opening of the inward recti er K + channels and was thus able to partially compensate for it. We studied the effects of this late depolarization on the excitability of CA1 neurons by measuring the ring generated by injection of depolarizing current steps of different intensity (20, 60, 100 pa, Fig. 4). In the cell in Fig. 4, the resting potential was ±57 mv and a 60-pA step was required to elicit repetitive ring in control conditions (Fig. 4a). At» 15 s, 5-HT hyperpolarized the cell to ±61 mv and decreased its excitability such that even 100-pA steps were unable to induce a regenerative response (Fig. 4b). After 1 min of 5-HT perfusion, the resting membrane potential partially recovered to a more depolarized value, and 100-pA steps were able to bring the neuron to the ring threshold (Fig. 4c). These observations suggest that the late depolarization due to 5-HT induced a partial recovery of neuronal excitability. A complete recovery of neuronal activity was then obtained upon washout (Fig. 4d). Qualitatively similar results were found in four more neurons. Effects of 5-HT on the hyperpolarization-activated (I h ) current We then studied the effects of 5-HT on I h in voltage-clamp conditions by activating the current with hyperpolarizing voltage steps applied every 6 s (Fig. 5a). In this cell, 5-HT was applied after a period of stable recording in control conditions, during which the current was activated by a hyperpolarizing step from ±35 mv to ±90 mv. 5-HT increased the size and accelerated activation of I h at ±90 mv. Each trace in Fig. 5a represents the average of three original recordings, and time constants were calculated by single-exponential ttings of FIG HT-induced change of the membrane potential of a CA1 neuron in the presence of Ba 2+ (2 mm). Top: in the presence of Ba 2+ (2 mm), 5-HT (50 mm) induced a depolarization of the membrane potential, accompanied by an increase of the membrane conductance (bottom) measured as in Fig. 1. FIG. 3. Effect of 5-HT on resting membrane potential and conductance of CA1 neurons in the presence of the selective blocker of I h, ZD 7288 (100 mm). Top: the ZD 7288-induced hyperpolarization was compensated by the injection of tonic depolarizing current (25 pa); note that 5-HT (50 mm, bar) induced a fast hyperpolarization of membrane potential associated with an increase in membrane conductance (bottom), but not a late depolarization as in Fig. 1.

4 3096 S. Gasparini and D. DiFrancesco FIG. 4. Action of 5-HT (50 mm) on the ring of a CA1 neuron in response to 450-ms current steps of 20, 60 and 100 pa amplitude. (a) In control conditions, 60 pa were required to evoke ring. (b) 5-HT-induced early hyperpolarization depressed neuronal excitability such that 100-pA steps were no more able to elicit action potentials. (c) After 1-min perfusion with 5-HT, following late depolarization, excitability was partially recovered and 100-pA steps were again suf cient to induce ring. Excitability was fully restored upon washout (d). Action potentials truncated at 0 mv. the averaged traces. I h increased from ±224.6 pa in control conditions to ±283.0 pa in the presence of 5-HT (50 mm, +26.0%), while the time constant of activation decreased from to ms (±25.4%). The average I h amplitude at ±90 mv increased by % and the time constant decreased by % (n = 16). 5-HT-induced changes in the amplitude and the time constant of activation of I h had similar time courses and were completely reversible upon washout (Fig. 5b and c). Shifts of the voltage dependence of gating are known to occur as a result of neurotransmitter-induced modulation of hyperpolarizationactivated currents in cardiac (DiFrancesco et al., 1986; DiFrancesco & Tromba, 1988) and neuronal cells (McCormick & Pape, 1990; Pedarzani & Storm, 1995). As both the amplitude and time constant of activation of I h are voltage dependent, this behaviour could be attributed to a shift of the activation curve of I h toward more positive potentials. We tested this hypothesis by using a two-pulse protocol such as the one in Fig. 6. This consists of a 1-s step to ±90 mv followed by a 450-ms step to ±140 mv applied rst to half-activate the current and then to fully activate it. 5-HT increased both the component recorded at ±90 mv and that at ±140 mv (Fig. 6a). I h recorded at ±90 mv was ±263.6 and ±353.2 pa in control conditions and in the presence of 5-HT (30 mm), respectively (+34.0%), while the current at ±140 mv increased from ±618.7 to ±717.5 pa (+16.0%). In n = 14 experiments, the steady-state current recorded at ±140 mv increased by % in the presence of 5-HT. The current increase at ±140 mv could be due to an increased I h conductance or the activation of a different 5-HT-dependent current at hyperpolarized potentials. To verify whether 5-HT activated a component other than I h, we repeated the protocol in Fig. 6 in the presence of ZD HT was rst perfused in control conditions, to verify the presence of the typical response at both potentials (Fig. 7a). I h was subsequently blocked by perfusing ZD 7288 (100 mm) and 5-HT reapplied; under these conditions, no variations were observed at either voltage (Fig. 7b). A similar lack of effect of 5- HT was observed after ZD 7288 in a total of n = 4 cells, and after I h block by Cs + (2 mm) inn = 4 cells (not shown). This nding con rms that the effects of 5-HT at ±140 mv are due to an increase of I h, possibly involving, as well as a shift of the activation curve of the FIG. 5. Effect of 5-HT (50 mm) on I h in voltage-clamp conditions. (a) I h was activated by 1-s steps to ±90 mv from a holding potential of ±35 mv applied at the frequency of 1/6 Hz. 5-HT increased I h amplitude and accelerated its activation. Single exponential tting of current activation at ±90 mv yielded a time constant of 214, 160 and 212 ms in control conditions, during and after 5- HT perfusion, respectively. The time course of I h amplitude change upon perfusion with 5-HT (b) was similar to that of the time constant of activation of the current (c). current toward more positive potentials, an increase in I h conductance. To better characterize the effect of 5-HT on the voltage dependence and conductance of I h, we used 1-min duration ramp protocols from ±40 to ±130 mv (Fig. 8) to measure the steady-state conductance of I h. The conductance curves were tted by using the Boltzmann equation (see Fig. 8 legend). In the example of Fig. 8, the half-activation voltages (V 1/2 ) were ±89.34 and ±86.38 mv (with a shift of 2.96 mv towards positive potentials), and the inverse slope factors (s) were and mv for the control and 5-HT curves, respectively; g h,max was found to be 5.04 ns in control conditions and 5.57 ns during the perfusion of 5-HT (+10.5%). In a total of n =4 neurons, 5-HT shifted V 1/2 by mv and increased the fully activated conductance by %. The inverse slope factor was not altered by the perfusion of 5-HT ( and mv in control conditions and in the presence of 5-HT, respectively). Second messengers involved in the response of I h to 5-HT We further investigated the response of I h to 5-HT with the aim of identifying the second messengers involved. In particular, as it is known that the position of the activation curve of the hyperpolarization-activated current on the voltage axis depends upon the intracellular concentration of camp in the heart (DiFrancesco & Tortora, 1991) and in neurons (Wang & Aghajanian, 1987; Bobker & Williams, 1989; McCormick & Pape, 1990; Pedarzani & Storm, 1995), and that 5-HT itself is able to increase intracellular camp levels in the hippocampus (Bockaert et al., 1990), we set out to establish if the modulation of I h by 5-HT is due to an increase in camp. After a period of stable recording in control conditions, a membrane-permeable analogue of camp, 8-bromoadenosine 3 5 cyclic monophosphate (8-Br-cAMP, 250 mm) was perfused, immediately followed by 5-HT. In these experiments, 8-Br-cAMP enhanced the current at ±90 mv (Fig. 9). Perfusion with 5-HT (30 mm) immediately after 8-Br-cAMP was able to further increase the current. At ±90 mv, 8-Br-cAMP increased I h by %, while

5 Modulation of I h by 5-HT in hippocampus 3097 FIG. 6. Effect of 5-HT on the voltage dependency and conductance of I h. Eight hundred-millisecond steps to ±90 mv (to half-activate I h ), followed by 450-ms steps to ±140 mv (to fully activate I h ) were applied every 10 s from a holding potential of ±35 mv in one cell. 5-HT (30 mm) increased the components recorded both at ±90 (by 34.0%) and ±140 mv (by 16.0%). 5-HT induced an increase of % (n = 5). From the leakagesubtracted values of I h measured at ±140 mv, we obtained a normalized slope conductance of ps/pf for control conditions, ps/pf during perfusion of 8-Br-cAMP and ps/pf in the presence of 5-HT [calculated assuming a reversal potential of ±17 mv (Maccaferri et al., 1993)]. The effects of both drugs were reversible upon washout. We then investigated if the membrane-permeable analogue of camp also affected the maximal conductance of I h, by using a twostep protocol (Fig. 10). 8-Br-cAMP (250 mm) increased I h recorded at ±90 mv, but also the fully activated current at ±140 mv. In the example of Fig. 10, 8-Br-cAMP increased I h by 12.3% at ±90 mv and by 6.0% at ±140 mv. In a total of n = 7 cells, the increase due to the perfusion of 8-Br-cAMP (250 mm) was % at ±90 mv and % at ±140 mv. Discussion In this work, we have shown that 5-HT, in addition to the inhibitory action previously described (Andrade & Nicoll, 1987), exerts a separate late effect on neuronal resting potential and excitability of hippocampal CA1 neurons, attributable to the modulation of the hyperpolarization-activated I h current. The best-described action of 5- HT in CA1 cells is the activation of an inward recti er K + channel through a 5-HT 1A receptor (Andrade & Nicoll, 1987; Colino & Halliwell, 1987) and the depression of neuronal excitability by the consequent hyperpolarization. We have shown that the early hyperpolarization is balanced by a secondary late depolarization, which appears some 10±15 s from the beginning of the perfusion and is characterized by a decreased input resistance relative to control conditions. This secondary effect of 5-HT is unlikely to be due to activation of ionotropic 5-HT 3 receptors, as in the hippocampus these receptors are expressed prevalently in inhibitory interneurons and not in pyramidal cells (Tecott et al., 1993; Morales et al., 1996), and because the activation of ionotropic receptors should be faster than activation of 5-HT 1A metabotropic receptors. Although a depolarization has been previously reported in the presence of 5-HT, it has been attributed to the block of K + channels, as it was associated with a decrease of membrane conductance (Andrade & Nicoll, 1987). In addition to the inhibitory effect mediated by 5-HT 1A receptors, 5-HT is known to have other actions on different K + currents in the hippocampus, e.g. the Ca 2+ -activated K + current which mediates after-hyperpolarization (I AHP ) (Colino & FIG HT action on the conductance of I h. Following the typical response to 5-HT (30 mm) during a two-pulse protocol (upper, see Fig. 6), ZD 7288 (100 mm) was perfused to block I h. No current changes were observed under these conditions upon perfusion of 5-HT (lower). Halliwell, 1987), the K + current inhibited by muscarinic receptors (I K(M) ) (Colino & Halliwell, 1987; but see Andrade & Nicoll, 1987); the inhibition of some of these conductances could determine a depolarization of the membrane potential. In order to dissect the late from the initial action of 5-HT, we studied the effect of 5-HT after block of the early hyperpolarization with Ba 2+ (2 mm). In the presence of Ba 2+, a non-speci c blocker of K + currents, a slow depolarization associated with an increased conductance was still observed upon perfusion of 5-HT, suggesting an increase of an inward component, e.g. I h. The depolarizing action of 5-HT was abolished by ZD 7288, a selective blocker of I h, suggesting that the 5-HT-dependent current underlying the secondary depolarization could be I h. This depolarization is unlikely to occur because of activation of I h at more hyperpolarized potential, as the time course of activation of the current at the voltages involved is too fast to be compatible with it, and has therefore to be attributed to an increase of I h by 5-HT. The late depolarization was quantitatively small ( mv, n = 3), but functionally important in recovering part of the hyperpolarization from the original resting potential and thus the excitability of CA1 neurons, as shown by the effect on neuronal ring (Fig. 4). We therefore studied the effect of 5-HT on the kinetic properties of I h in voltage-clamp conditions. Using a typical voltage protocol to activate I h, we found that 5-HT increased I h by acting on both the current±voltage dependence and maximal conductance. At ±90 mv, 5-HT increased current amplitude and accelerated its activation, suggesting a depolarizing shift of the voltage dependence of gating (Pape, 1996). The shift measured by voltage ramp protocols was mv, in agreement with the moderate depolarizing shifts observed under conditions of stimulation of adenylate cyclase in other neuronal types, e.g. thalamic neurons (McCormick & Pape, 1990) and facial motoneurons (Larkman & Kelly, 1992). Hyperpolarization-activated cation currents are known to be modulated by neurotransmitters through the modi cation of the activity of

6 3098 S. Gasparini and D. DiFrancesco FIG. 8. Effect of 5-HT on I h conductance/voltage [g h (V)] relationship obtained by a ramp protocol. (a and b) Voltage ramp protocol consisted of a 1-min duration hyperpolarizing ramp from ±40 to ±130 mv applied from a holding potential of ±40 mv. This yielded the current traces shown in (b), in control conditions and during perfusion of 5-HT (30 mm). Linear leakage was digitally subtracted. In both panels, time runs backwards to allow for a direct comparison with the curves in (c). (c) Conductance curves calculated from the relation I h = g h,max *(V ± V f )*y`, where y` is the steady-state activation parameter, V f the reversal potential and g h,max the fully activated conductance; the steady-state conductance±voltage relation, g h (V)=g h,max *y` was calculated as the ratio between steady-state current in (b) and V ± V f. V f was set to ±17 mv (after Maccaferri et al., 1993). Curves were tted by the Boltzmann equation, g h (V)=g h,max /{1 + exp[(v ± V 1/2 )/s]}, which yielded values of 5.04 and 5.75 ns for g h,max, ±89.34 and ±86.38 mv for the mid-activation voltage (V 1/2 ), and and mv for the inverse slope factor (s), in control conditions and in the presence of 5-HT, respectively. adenylate cyclase in the heart (DiFrancesco, 1993; 1995) and in neurons (Pape, 1996), by an action involving a shift of the voltage dependence of activation without modi cation of maximal conductance, which is due to the direct binding of camp to a cytoplasmic site of the channel (DiFrancesco & Tortora, 1991; Pedarzani & Storm, 1995; Ingram & Williams, 1996; Larkman & Kelly, 1997). A more surprising and unusual nding was that, when I h was maximally activated by using a two-pulse protocol, 5-HT also increased the current recorded at ±140 mv. This effect was abolished by the perfusion of the selective I h blocker ZD 7288 (BoSmith et al., 1993; Harris & Constanti, 1995; Gasparini & DiFrancesco, 1997), indicating an effect of 5-HT on the maximal conductance of the current. The increase of I h maximal conductance due to 5-HT was % (n = 14) as obtained from two-pulse protocols, and % (n = 4) as obtained from ramp protocols. Neurotransmitter-induced modulation of I h by an increase of its conductance is, to our knowledge, a novel nding. In ventral tegmental neurons, dopamine D 2 and GABA B (g-aminobutyric acid) receptors have been shown to induce a reduction in I h conductance without any change in the voltage dependence and with no apparent FIG. 9. Comparison of the effects of 8-Br-cAMP (250 mm) and 5-HT (30 mm) on I h. I h was activated by steps to ±90 mv applied from ±35 mv every 10 s (a). 8-Br-cAMP was superfused for 90 s and increased I h moderately. In n = 5 cells, the increase at ±90 mv was %, P < 0.05, histogram in (b). 5-HT added after 8-Br-cAMP perfusion further increased I h ( % of control current after 90 s perfusion with 5-HT, n =5, P < 0.05). involvement of camp (Jiang et al., 1993), although it has been proposed that the action of baclofen on I h in substantia nigra neurons is secondary to activation of a K + current (Watts et al., 1996); also, a decrease of I h conductance following activation of protein kinase C (PKC) by neurotensin has been recently reported in neurons of substantia nigra pars compacta (Cathala & Paupardin-Tritsch, 1997). Changes in conductance have been reported for the hyperpolarization-activated (I f ) current in cardiac preparations in the presence of inhibitors of protein phosphatase, e.g. calyculin A (Accili et al., 1997), or of tyrosine kinase (Wu & Cohen, 1997). Interestingly, 5-HT has been reported to induce a partial reduction of phosphatase activity in colliculi neurons through the camp pathway (Ansanay et al., 1995). The current increase we observed, however, cannot be explained only in terms of an increased conductance, because as well as a faster activation of the current at ±90 mv during 5-HT perfusion, we always observed an increase in the ratio of I h activated at ±90 and ±140 mv ( in control versus in the presence of 5-HT, 30 mm, n = 13, P < 0.001). As discussed above, the use of ramp protocols further con rmed the presence of a positive shift of the current activation curve. Investigation of the ability of intracellular camp to mediate the action of 5-HT on I h indicated that 8-Br-cAMP does induce an increase of I h at both ±90 and ±140 mv, but does not completely occlude the action of 5-HT, which is always able to further increase I h when perfused after 8-Br-cAMP. Because the effects of these two drugs are qualitatively similar but quantitatively different, it is possible that the intracellular concentration achieved by extracellular perfusion of 8-Br-cAMP was smaller or differently localized compared with that achieved by stimulation of 5-HT receptors. The increase of the maximal conductance in the presence of the membrane-permeable analogue of camp suggests that this action

7 Modulation of I h by 5-HT in hippocampus 3099 FIG. 10. Action of 8-Br-cAMP (250 mm) on the maximal conductance of I h.a two-pulse protocol was applied to ±90/±140 mv in order to activate I h by different degrees. Note that 8-Br-cAMP augmented not only the current recorded at ±90 mv (by 12.3%), but also at ±140 mv (by 6.0%), indicating an increase in the fully activated conductance. of 5-HT is also camp dependent. Because changes in the conductance of the f/h-channel are found in the presence of inhibitors of kinases and phosphatases (Accili et al., 1997; Wu & Cohen, 1997), or following activation of protein kinase C by neurotensin (Cathala & Paupardin-Tritsch, 1997), our results raise the possibility that I h conductance is modi ed by 5-HT through a camp-dependent alteration of phosphorylation balance in the cell, due to activation of protein kinase A or reduction of phosphatase activity in a way similar to that found in colliculi neurons (Ansanay et al., 1995). We did not investigate which receptor is involved in the response of I h to 5-HT; however, it is known that at least two types of 5-HT receptors activating adenylate cyclase are expressed in the CA1 region of the hippocampus, e.g. 5-HT 4 (Torres et al., 1995) and 5-HT 6 (Gerard et al., 1997). In summary, our data are consistent with the view that 5-HT modulates I h by activation of adenylate cyclase as well as by a different mechanism, possibly involving inhibition of phosphatase activity. Acknowledgements We wish to thank E. Cherubini for carefully reading the manuscript and A. Ferroni for discussion. This work was supported by MURST-COFIN (no ) and Telethon (no. 971) grants to D.D. Abbreviations 5-HT, serotonin; 8-Br-cAMP, 8-bromoadenosine 3'5'-cyclic monophosphate; ACSF, arti cial cerebrospinal uid; EGTA, ethylenebis (oxonitrilo) tetraacetic acid; HEPES, N-2-hydroxyethylpiperazine-N -2-ethanesulphonic acid; I h, hyperpolarization-activated cation current; I K(M), muscarinic K + current; PKC, protein kinase C; TTX, tetrodotoxin; ZD 7288, 4-(N-ethyl-N-phenylamino)-1,2-dimethyl-6-(methylamino) pyrimidinium chloride. References Accili, E.A., Redaelli, G. & DiFrancesco, D. (1997) Differential control of the hyperpolarization-activated current (I f ) by camp gating and phosphatase inhibition in rabbit sino-atrial node myocytes. J. Physiol. (Lond.), 500, 643± 651. Akasu, T., Shoji, S. & Hasuo, H. (1993) Inward recti er and low-threshold calcium currents contribute to the spontaneous ring mechanism in neurons of the rat suprachiasmatic nucleus. P uègers Arch., 425, 109±116. Andrade, R. & Nicoll, R.A. (1987) Pharmacologically distinct actions of serotonin on single pyramidal neurones of the rat hippocampus recorded in vitro. J. Physiol. (Lond.), 394, 99±124. Ansanay, H., Dumuis, A., Sebben, M., Bockaert, J. & Fagni, L. (1995) campdependent, long lasting inhibition of a K + current in mammalian neurons. Proc. Natl Acad. Sci. USA, 92, 6635±6639. Bal, T. & McCormick, D.A. (1997) Synchronized oscillations in the inferior olive are controlled by the hyperpolarization-activated cation current I h. J. Neurophysiol., 77, 3145±3156. Bergles, D.E., Doze, V.A., Madison, D.V. & Smith, J.S. (1996) Excitatory actions of norepinephrine on multiple classes of hippocampal CA1 interneurons. J. Neurosci., 16, 572±585. Blanton, M.G., Loturco, J.J. & Kriegstein, A.R. (1989) Whole-cell recording from neurons of reptilian and mammalian cerebral cortex. J. Neurosci. Meth., 30, 203±210. Bobker, D.H. & Williams, J.T. (1989) Serotonin augments the cationic current I h in central neurons. Neuron, 2, 1533±1540. Bockaert, J., Sebben, M. & Dumuis, A. (1990) Pharmacological characterization of 5-hydroxytryptamine4 (5-HT 4 ) receptors positively coupled to adenylate cyclase in adult guinea pig hippocampal membranes: effect of substituted benzamide derivatives. Mol. Pharmacol., 37, 408±411. BoSmith, R.E., Briggs, I. & Sturgess, N.C. (1993) Inhibitory actions of Zeneca ZD 7288 on whole-cell hyperpolarization activated inward current (I f )in guinea-pig dissociated sino-atrial node cells. Br. J. Pharmacol., 110, 343± 349. Brown, H.F., DiFrancesco, D. & Noble, S.J. (1979) How does adrenaline accelerate the heart? Nature, 280, 235±236. Brown, D.A., GaÈhwiler, B.H., Grif th, W.H. & Halliwell, J.V. (1990) Membrane currents in hippocampal neurons. Progr. Brain Res., 83, 141± 160. Cathala, L. & Paupardin-Tritsch, D. (1997) Neurotensin inhibition of the hyperpolarization-activated cation current (I h ) in the rat substantia nigra pars compacta implicates the protein kinase C pathway. J. Physiol. (Lond.), 503, 87±97. Clapham, D.E. (1998) Not so funny anymore: pacing channels are cloned. Neuron, 21, 5±7. Colino, A. & Halliwell, J.V. (1987) Differential modulation of three separate K-conductances in hippocampal CA1 neurons by serotonin. Nature, 327, 73±77. Colino, A. & Halliwell, J.V. (1993) Carbachol potentiates Q current and activates a calcium-dependent non speci c conductance in rat hippocampus in vitro. Eur. J. Neurosci., 5, 1198±1209. DiFrancesco, D. (1993) Pacemaker mechanisms in cardiac tissue. Annu. Rev. Physiol., 55, 451±467. DiFrancesco, D. (1995) The onset and autonomic regulation of cardiac pacemaker activity: relevance of the f-current. Cardiovascular Res., 29, 449±456. DiFrancesco, D., Ferroni, A., Mazzanti, M. & Tromba, C. (1986) Properties of the hyperpolarizing-activated current (I f ) in cells isolated from the rabbit sino-atrial node. J. Physiol. (Lond.), 377, 61±88. DiFrancesco, D. & Tortora, P. (1991) Direct activation of cardiac pacemaker channels by intracellular cyclic AMP. Nature, 351, 145±114. DiFrancesco, D. & Tromba, C. (1988) Inhibition of the hyperpolarizingactivated current, i f, induced by acetylcholine in rabbit sino-atrial node myocytes. J. Physiol. (Lond.), 405, 477±491. Gasparini, S. & DiFrancesco, D. (1997) Action of the hyperpolarizationactivated current (I h ) blocker ZD 7288 in hippocampal CA1 neurons. P uègers Arch., 435, 99±106. Gerard, C., Martres, M.P., Lefevre, K., Miquel, M.C., Verge, D., Lanfumey, L., Doucet, E., Hamon, M. & el Mestikawy. S. (1997) Immunolocalization of serotonin 5-HT 6 receptor-like material in the rat central nervous system. Brain Res., 746, 207±219. Halliwell, J.V. & Adams, P.R. (1982) Voltage-clamp analysis of muscarinic excitation in hippocampal neurons. Brain Res., 250, 71±92. Harris, N.C. & Constanti, A. (1995) Mechanism of block by ZD 7288 of the hyperpolarization-activated inward rectifying current in guinea pig substantia nigra neurons in vitro. J. Neurophysiol., 74, 2366±2378. Ingram, S.L. & Williams, J.T. (1996) Modulation of the hyperpolarizationactivated current (I h ) by cyclic nucleotides in guinea-pig primary afferent neurons. J. Physiol. (Lond.), 492, 97±106. Jiang, Z.G., Pessia, M. & North, R.A. (1993) Dopamine and baclofen inhibit the hyperpolarization-activated cation current in rat ventral tegmental neurones. J. Physiol. (Lond.), 462, 753±764. Larkman, P.M. & Kelly, J.S. (1992) Ionic mechanisms mediating 5-

8 3100 S. Gasparini and D. DiFrancesco hydroxytriptamine- and noradrenaline-evoked depolarization of adult rat facial motoneurones. J. Physiol. (Lond.), 456, 473±490. Larkman, P.M. & Kelly, J.S. (1997) Modulation of I h by 5-HT in neonatal rat motoneurones in vitro: mediation through a phosphorylation independent action of camp. Neuropharmacology, 36, 721±733. Maccaferri, G., Mangoni, M., Lazzari, A. & DiFrancesco, D. (1993) Properties of the hyperpolarization-activated current in rat hippocampal CA1 pyramidal cells. J. Neurophysiol., 69, 2129±2136. Maccaferri, G. & McBain, C.J. (1996) The hyperpolarization-activated current (I h ) and its contribution to pacemaker activity in rat CA1 hippocampal stratum oriens-alveus interneurones. J. Physiol. (Lond.), 497, 119±130. Madison, D.V., Lancaster, B. & Nicoll, R.A. (1987) Voltage-clamp analysis of cholinergic action in the hippocampus. J. Neurosci., 7, 733±741. McCormick, D.A. & Bal, T. (1997) Sleep and arousal: thalamocortical mechanisms. Annu. Rev. Neurosci., 20, 185±215. McCormick, D.A. & Pape, H.-C. (1990) Noradrenergic and serotonergic modulation of a hyperpolarization-activated cation current in thalamic relay neurones. J. Physiol. (Lond.), 431, 319±342. Morales, M., Battenberg, E., Delecea, L. & Bloom, F.E. (1996) The type 3 serotonin receptor is expressed in a subpopulation of GABAergic neurons in the rat neocortex and hippocampus. Brain Res., 731, 199±202. Pape, H.-C. (1996) Queer current and pacemaker: the hyperpolarizationactivated cation current in neurons. Annu. Rev. Physiol., 58, 299±327. Pape, H.-C. & McCormick, D.A. (1989) Noradrenaline and serotonin selectively modulate thalamic burst ring by enhancing a hyperpolarization-activated current. Nature, 340, 715±718. Pedarzani, P. & Storm, J.F. (1995) Protein kinase A-independent modulation of ion channels in the brain by cyclic AMP. Proc. Natl Acad. Sci. USA, 92, ± Schweitzer, P., Madamba, S.G. & Siggins, G.R. (1998) Somatostatin increases a voltage-insensitive K + conductance in rat CA1 hippocampal neurons. J. Neurophysiol., 79, 1230±1238. Segal, M. (1989) Serotonin modulation of hippocampal activity. In Chan- Palay, V. & KoÈhler, C. (eds), The Hippocampus. New Vistas. Alan R. Liss, New York, pp. 307±315. Storm, J.F. (1990) Potassium currents in hippocampal pyramidal cells. Progr. Brain Res., 83, 161±187. Svoboda, K.R. & Lupica, C.R. (1998) Opioid inhibition of hippocampal interneurons via modulation of potassium and hyperpolarization-activated cation (I h ) currents. J. Neurosci., 15, 7084±7098. Tecott, L.H., Maricq, A.V. & Julius, D. (1993) Nervous system distribution of the serotonin 5-HT 3 receptor mrna. Proc. Natl Acad. Sci. USA, 90, 1430± Torres, G.E., Chaput, Y. & Andrade, R. (1995) Cyclic AMP and protein kinase A mediate 5-hydroxytryptamine type 4 receptor regulation of calciumactivated potassium current in adult hippocampal neurons. Mol. Pharmacol., 47, 191±197. Wang, Y.-Y. & Aghajanian, G.K. (1987) Excitation of locus coeruleus neurons by an adenosine 3,5 -cyclic monophosphate-activated inward current: extracellular and intracellular studies in rat brain slices. Synapse, 1, 481±487. Watts, A.E., Williams, J.T. & Henderson, G. (1996) Baclofen inhibition of the hyperpolarization-activated cation current I h in rat substantia nigra compacta neurones may be secondary to potassium current activation. J. Neurophysiol., 76, 2262±2270. Wu, J.Y. & Cohen, I.S. (1997) Tyrosine kinase inhibition reduces I f in rabbit sino-atrial node myocytes. P uègers Arch., 434, 509±514.