5-HT Modulation of Multiple Inward Rectifiers in Motoneurons in Intact Preparations of the Neonatal Rat Spinal Cord OLE KJAERULFF AND OLE KIEHN Division of Neurophysiology, Department of Medical Physiology, The Panum Institute, University of Copenhagen, DK-2200 Copenhagen, Denmark Received 19 June 2000; accepted in final form 3 October 2000

Kjaerulff, Ole and Ole Kiehn. 5-HT modulation of multiple inward rectifiers in motoneurons in intact preparations of the neonatal rat spinal cord. J Neurophysiol 85: 580 –593, 2001. This study introduces novel aspects of inward rectification in neonatal rat spinal motoneurons (MNs) and its modulation by serotonin (5-HT). Whole cell tight-seal recordings were made from MNs in an isolated lumbar spinal cord preparation from rats 1–2 days of age. In voltage clamp, hyperpolarizing step commands were generated from holding potentials of ⫺50 to ⫺40 mV. Discordant with previous reports involving slice preparations, fast inward rectification was commonly expressed and in 44% of the MNs co-existed with a slow inward rectification related to activation of Ih. The fast inward rectification is likely caused by an IKir. Thus it appeared around EK and was sensitive to low concentrations (100 –300 ␮M) of Ba2⫹ but not to ZD 7288, which blocked Ih. Both IKir and Ih were inhibited by Cs2⫹ (0.3–1.5 mM). Extracellular addition of 5-HT (10 ␮M) reduced the instantaneous conductance, most strongly at membrane potentials above EK. Low [Ba2⫹] prevented the 5-HT–induced instantaneous conductance reduction below, but not that above, EK. This suggests that 5-HT inhibits IKir, but also other instantaneous conductances. The biophysical parameters of Ih were evaluated before and under 5-HT. The maximal Ih conductance, Gmax, was 12 nS, much higher than observed in slice preparations. Gmax was unaffected by 5-HT. In contrast, 5-HT caused a 7-mV depolarizing shift in the activation curve of Ih. Doubleexponential fits were generally needed to describe Ih activation. The fast and slow time constants obtained by these fits differed by an order of magnitude. Both time constants were accelerated by 5-HT, the slow time constant to the largest extent. We conclude that spinal neonatal MNs possess multiple forms of inward rectification. Ih may be carried by two spatially segregated channel populations, which differ in kinetics and sensitivity to 5-HT. 5-HT increases MN excitability in several ways, including inhibition of a barium-insensitive leak conductance, inhibition of IKir, and enhancement of Ih. The quantitative characterization of these effects should be useful for further studies seeking to understand how neuromodulation prepares vertebrate MNs for concerted behaviors such as locomotor activity.

The discharge pattern of motoneurons (MNs) in the spinal cord depends on the characteristic of each of a number of ionic currents selectively channeled through membrane proteins (Kiehn 1991b; Rekling et al. 2000; Schwindt and Crill 1984). A subset of currents preferably flow at hyperpolarized membrane potentials and are therefore known as inward rectifiers

(Hille 1992). Dependent on the speed of current activation, inward rectification (IR) is seen as a downward bend in either the instantaneous or the steady-state current-voltage (I-V) relationship, or both. The presence of one particular type of inward rectifier causes a slow sag toward the original membrane potential when the cell is hyperpolarized from the resting membrane potential. In spinal cord MNs, Ito and Oshima (1965) originally described this slowly developing inward rectification in the cat. Later studies in a slice preparation of the neonatal rat spinal cord suggested that the underlying current is a slow, noninactivating inward cation current with a reversal potential more positive than rest (Takahashi 1990a). This current corresponds to the hyperpolarization-activated inward mixed cation current, or Ih, found in many vertebrate and invertebrate neurons (Pape 1996). The slow activation of Ih explains the time dependency of the sag. Moreover, since Ih also deactivates slowly when the hyperpolarization is released, this current contributes to the formation of a long-lasting rebound excitation. Another form of inward rectification is found in various types of excitable cells including neurons. This “instantaneous” current, IKir, is carried by potassium ions (Constanti and Galvan 1983; Katz 1949; Nichols and Lopatin 1997; Standen and Stanfield 1979; Travagli and Gillis 1994; Williams et al. 1997; Yamoah et al. 1998). IKir, as well as Ih, typically contribute to the resting membrane potential. This is one important reason why inward rectifiers play essential roles for the excitability and response properties of nerve cells. Moreover, the contribution of both IKir and Ih to the resting membrane potential makes these currents prime targets for neuromodulators. One such neuromodulator is serotonin (5HT), which is often used to induce locomotion in the isolated spinal cord of the neonatal rat (Cazalets et al. 1992; Kiehn and Kjaerulff 1996; Kjaerulff and Kiehn 1996, 1997). 5-HT strongly excites spinal MNs in the rat (Berger and Takahashi 1990; Kiehn 1991a; Takahashi and Berger 1990; Wang and Dun 1990). A part of this excitation has been ascribed to an enhancement of Ih (Takahashi and Berger 1990). However, the exact mechanism behind the enhancement has not been determined. We have started to investigate the contribution of inward rectification to the discharge pattern in neonatal rat lumbar MNs during transmitter-induced rhythmic locomotor activity

Address for reprint requests: O. Kiehn, Section of Neurophysiology, Dept. of Medical Physiology, The Panum Institute, Blegdamsvej 3, DK-2200 Copenhagen, Denmark (E-mail: [email protected]).

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

INTRODUCTION

580

0022-3077/01 $5.00 Copyright © 2001 The American Physiological Society

www.jn.physiology.org

5-HT EFFECTS ON INWARD RECTIFICATION

(Kiehn et al. 2000). A complete understanding of how inward rectification contributes to the discharge patterns in neonatal rat lumbar MNs requires knowledge of the biophysical parameters of the underlying currents. Furthermore, earlier studies have generally been performed on MNs located superficially in thin slice preparations (e.g., Takahashi 1990a). Despite the importance of the information obtained under such circumstances, it has hardly been possible to avoid elimination of part of the dendritic arbor. This might lead to important qualitative and quantitative deviations from the membrane current characteristic in the more intact MNs located in whole spinal cords used for locomotor studies. Our study shows that, in addition to the slowly developing inward rectification caused by activation of Ih, most neonatal rat MNs possess an instantaneous rectifier with kinetics and pharmacology similar to IKir . We show that 5-HT enhances Ih by causing a depolarizing shift in the activation curve. In contrast, the fast inward rectifier IKir is inhibited by 5-HT. In addition, 5-HT appears to inhibit other instantaneous currents, possibly “leak” currents, as well. These effects of 5-HT enhance resting MN excitability and cause a theoretical enhancement of rebound firing and phase-transition during rhythmic motor discharges. Preliminary results from these experiments have been presented in abstract form (Kjaerulff and Kiehn 1998). METHODS

Dissection The procedure for isolating the spinal cord has been described in detail earlier (Kjaerulff and Kiehn 1996, 1997) and is only partially summarized here. Neonatal rat pups (1–2 days old) were anesthetized with ether and quickly decapitated. The spinal cord was isolated and a piece of cord comprising approximately the Th12-S1 segments was placed in a recording chamber perfused with standard Krebs solution containing (in mM) 128 NaCl, 4.69 KCl, 25 NaHCO3, 1.18 KH2PO4, 1.25 MgSO4, 2.52 CaCl2, and 22 glucose, aerated with 5% CO2 in O2. The composition of this extracellular solution is identical to that used in our previous locomotor studies (e.g., Kiehn and Kjaerulff 1996; Kjaerulff and Kiehn 1997) (see INTRODUCTION) and was selected also for the present study to be able to directly relate the findings on the inward rectifier currents to the results of the locomotor experiments.

Recordings Recordings from MNs were made with patch electrodes pulled from thick-walled borosilicate glass (OD: 1.5 mm; ID: 1.0 mm, Clark Electromedical Instruments, Pangbourne, England) to a final resistance of 4 –7 M⍀. Recordings were made in the whole cell configuration (Hamill et al. 1981) using an Axoclamp 2B amplifier (Axon Instruments, Foster City, CA) in the continuous single-electrode voltage-clamp (cSEVC) mode. With maximal phase-advance, it was usually possible to increase the clamp gain to 100 nA/mV (maximal gain). However, the gain was generally kept slightly lower (⬃90 nA/mV) to prevent fatal oscillations of the clamp circuit induced by changes in the access resistance. pCLAMP (Axon Instruments) was used for data acquisition and off-line current measurements. Current signals were low-pass filtered at 1–2 kHz and digitized at 2.667 kHz. In most experiments the spinal cord was pinned down with the ventral side up and the electrode lowered blindly into the ventrolateral region of the L4 (occasionally L5) segment through a slit cut in the white matter. In a few cases the ventral part of the cord was isolated (Kjaerulff and Kiehn 1996), pinned down dorsal side up, and the electrode lowered into the motor column via the dorsal cut surface.

581

The pipette electrode solution contained (in mM) 128 potassium gluconate, 4 NaCl, 1 glucose, 10 HEPES, 0.0001 CaCl2, 4 ATP-Mg, and 0.3 GTP-Li. The pH was adjusted to 7.3 with KOH. Recordings were made at room temperature.

Correction of the voltage In general, the access resistance, Ra, was within 10 –30 M⍀. Since the current could reach amplitudes of 1 nA or more, the predicted voltage drop over Ra was not negligible. To compensate for this error, we monitored Ra pertaining to individual voltage steps and corrected the membrane potential off-line using the relation V ⫽ V cmd ⫺ R a I

(1)

where V is the corrected membrane potential, Vcmd is the command potential, and I is the current (Finkel and Redman 1984). Equation 1 implies that the changes in voltage-sensitive currents evoked by the step command will be mirrored by changes in V. This fluctuation in the effective command was generally below 10 mV, and hence it was ignored in the data analysis. In the relevant figures, however, rather than the ideal constant step command, we display the trajectories of the membrane potential calculated using Eq. 1, since this is expected to more accurately represent the effective command. JUNCTION POTENTIAL. Voltages were not corrected for the liquid junction potential (Neher 1992), which was within the range 6 – 8 mV (pipette interior negative) with the intracellular and various extracellular solutions used in the present experiments. Correcting potentials would lead to more hyperpolarized values than those reported in the text.

Database Recordings from 138 MNs were included in this study. Cells were identified as MNs when ventral root stimulation through a suction electrode evoked an antidromic action potential. For a MN to be included in the database, the holding current corresponding to a holding potential of ⫺50 mV had to be more positive than 0 pA.

Pharmacological compounds Serotonin creatine sulfate (5-HT, 10 ␮M), tetrodotoxin (TTX, 0.2– 0.3 ␮M), tetraethylamminium-chloride (TEA 20 –30 mM), and 4-aminopyridine (4-AP, 2– 4 mM) were from Sigma (St. Louis, MO). The 5-HT concentration chosen is at the plateau of the dose-response curve for activation of the previously described 5-HT–induced inward current, I5-HT, in neonatal rat MNs (Takahashi and Berger 1990). D,L-2-amino-5-phosphonovaleric acid (APV, 20 ␮M) and 6-cyano-7nitroquinoxaline-2,3-dione (CNQX, 10 ␮M) were from RBI (Natick, MA) and 4-(N-ethyl-N-phenylamino)-1,2-dimethyl-6-(methylamino) pyrdinium chloride (ZD 7288, 100 ␮M) from Tocris Cookson (Bristol, UK). In experiments involving the extracellular use of BaCl2 (100 –300 ␮M), KH2PO4 was omitted and MgSO4 replaced with an equimolar concentration of MgCl2.

Statistical analysis Commercially available software (Statistica; StatSoft, Tulsa, OK) was used to test for differences between the amplitude of the difference current, Iss ⫺ Iin, before, during, and after application of 5-HT in the same MNs. Data were transformed by taking the square root of the absolute current amplitudes prior to significance testing (Altman 1991) to conform with the assumptions underlying the statistical model used, i.e., repeated measures ANOVA. This model was also used to evaluate changes in holding current induced by 5-HT in normal medium. Duncans multiple range test was employed for post hoc comparisons.

582

O. KJAERULFF AND O. KIEHN

INSTANTANEOUS I-V RELATIONSHIP. We used a statistical criterion to help decide whether rectification was present in the instantaneous I-V relationship. If the I-V relationship was judged by eye to be linear, this was accepted without further analysis. However, if the I-V curve appeared to bend around a given voltage, Vbend, linear regression was used to determine Gdep and Ghyp, i.e., the average slope conductance in the voltage range more depolarized and hyperpolarized, respectively, than Vbend (see Figs. 3A and 5, A and B, for indications of Vbend, Gdep, and Ghyp). The 97.5% confidence interval for Gdep was then compared with the 97.5% confidence interval for Ghyp. Fast rectification was set to be present if the two confidence intervals did not overlap. The bend in the instantaneous I-V curve was generally sharp enough to permit the identification of Vbend with an uncertainty below ⬃10 mV. Pharmacologically induced changes in Gdep, Ghyp, and in fast rectification were analyzed using paired t-tests. Since the distributions of the instantaneous conductance data were positively skewed, data were transformed by taking the logarithm before testing (Altman 1991). Reported summary statistics (mean ⫾ SE) were, however, calculated from the raw data. The level of significance was set to 5%. RESULTS

Generally, experiments were done in the presence of TTX to block the fast Na⫹ current, INa (Takahashi 1990b). TTX also blocks synaptically mediated indirect effects of 5-HT on MNs caused by spike activation of presynaptic interneurons (Wang and Dun 1990). In most experiments, the glutamate receptor antagonists APV and CNQX were also added to reduce spontaneous synaptic noise. Current responses in normal medium LINEAR. In some MNs, when hyperpolarizing voltage steps from a holding potential of ⫺50 mV were generated, the size of the instantaneous current response, Iinst, was proportional to the size of the voltage step (Fig. 1, A and B). In normal medium (with or without TTX/APV/CNQX), such a linear instantaneous I-V relationship was observed in 30 of 72 MNs (41.7%). FAST INWARD RECTIFICATION. In a larger fraction of the MNs, however, the Iinst elicited by large hyperpolarizing voltage steps was disproportionally larger than Iinst evoked by smaller hyperpolarizing steps. Thus these MNs showed fast IR (Fig. 1, C and D). Typically, the membrane potential at which the increase in conductance occurred was rather sharply defined, producing an inflection point on the instantaneous I-V relationship at 79.8 ⫾ 1.1 mV (mean ⫾ SE; Vbend, see METHODS). A comparison of the 97.5% confidence interval of the average instantaneous slope conductance at voltages above Vbend, termed Gdep, with that of the conductance, Ghyp, below Vbend confirmed the presence of fast IR in a high number of MNs (35 of 72 MNs, 48.6%). OUTWARD RECTIFICATION. In relatively few MNs, the Iinst elicited by large hyperpolarizing voltage steps was disproportionally smaller than the Iinst evoked by smaller hyperpolarizing steps (7 of 72 MNs, 9.7%; not shown). Thus fast outward rectification (OR) was also observed, but only in a minority of the MNs. The average Vbend for fast OR was ⫺83.3 ⫾ 2.8 mV. SLOWLY DEVELOPING INWARD RECTIFICATION, Ih. An inwardly increasing current usually followed the instantaneous current jump. The rate of inward increase was fast initially but leveled off with time (inward relaxation). The steady-state current, Iss, was reached only after several seconds (e.g., Fig. 1, A and C).

FIG. 1. Variability in fast and slowly developing inward rectification (IR) among spinal neonatal rat motoneurons (MNs). A: current responses (top) evoked by a family of hyperpolarizing voltage steps (bottom) from a holding potential of ⫺50 mV. Capacitive transients related to changes in the voltage command have been blanked. Note instantaneous current (●) proportional to the voltage step amplitude and the pronounced slowly developing inward current relaxation caused by Ih. Ih approached steady state only after several seconds (E). B: current-voltage (I-V) plots derived from A showing the absence of fast IR (linear, ●) and the strong steady-state IR caused by Ih (E). Measurements were taken from A at the time points indicated by the symbols identical to those used for plotting. C and D: different MN showing fast IR (●) and relatively weak Ih-dependent steady-state IR (E). Holding current, 61 pA in A and B; 71 pA in C and D.

The activation rate and the amplitude of this slowly developing inward relaxation increased with increasing hyperpolarization. The amplitude increase lead to inward rectification in the steady-state I-V relationship (steady-state IR; Fig. 1, B and D). The slowly developing inward rectification shares voltage and time dependency with the hyperpolarization-activated cation current, Ih, previously described by Takahashi (1990a), who named it IIR. Here we will use the term Ih. In normal medium, an Ih was observed unambiguously in 65 of the 72 MNs mentioned above (90%). However, the distinctness of Ih and consequently of the steady-state IR could vary considerably (compare in Fig. 1, A with C, and B with D). In current clamp, Ih was seen as a slowly developing, depolarizing sag in the membrane potential in the face of a constant hyperpolarizing current injection (data not shown). Multiple forms of inward rectification caused by distinct conductances co-exist in MNs In roughly half of the MNs, Ih appeared to be the only form of inward rectifying current present, since these cells showed no fast IR (33 of 72 MNs, 45.8%; Fig. 1B). However, fast IR co-existed with the steady-state IR indicative of Ih in a substantial number of MNs (32 of 72 MNs, 44.4%; Fig. 1, C and D). In contrast to Ih, fast IR has not previously been reported to be present in spinal rat MNs. It therefore became important to PERSISTENCE OF FAST INWARD RECTIFICATION IN ZD 7288.

5-HT EFFECTS ON INWARD RECTIFICATION

583

determine whether the two forms of inward rectification are mediated by different conductances, or whether they are different expressions of the same conductance, e.g., Ih. For this purpose, we employed ZD 7288, an established blocker of Ih (BoSmith et al. 1993; Harris and Constanti 1995; Hughes et al. 1998; Khakh and Henderson 1998; Williams et al. 1997). Figure 2A shows a MN, in which both fast IR and the slowly developing IR indicative of Ih was present in control conditions. ZD 7288 reduced both the instantaneous and the steadystate conductance (Fig. 2B). Moreover, ZD 7288 abolished the slow inward relaxation. In contrast, fast IR persisted (Fig. 2B; Fig. 2C, compare E and F). In current clamp (not shown; see Kiehn et al. 2000) ZD 7288 removed the sag response, again indicating an efficient blockade of Ih. As in voltage clamp, however, fast IR persisted. Similar results were obtained in a total of five MNs (2 in voltage clamp, 3 in current clamp). The fact that ZD 7288 blocked Ih but not fast IR suggests that the fast IR is caused by a different inward rectifier than Ih. In accordance with the known sensitivity of Ih to Cs2⫹ (Pape 1996; Takahashi 1990a), the slow inward relaxation under voltage clamp, or the sag under current clamp, were blocked in a concentration-dependent manner by adding CsCl2 to the superfusate (0.3–1.5 mM, n ⫽ 7 MNs; data not shown). The fast IR was also strongly affected by Cs2⫹ (see the following section). Identity of the fast inward rectifying current Several of our observations suggest that at least part of the fast IR in spinal neonatal rat MNs is caused by an IKir, i.e., an inward (“anomalous”) rectifying current carried by potassium ions (Constanti and Galvan 1983; Hagiwara et al. 1976; Katz 1949; Nichols and Lopatin 1997; Standen and Stanfield 1979; Travagli and Gillis 1994). First, IKir is generally associated with fast or “instantaneous” IR. Second, fast IR in the rat MNs occurred around a Vbend of ⫺80 mV on average; this voltage is close to the K⫹ equilibrium potential (EK, predicted to be ⫺79 mV with the intra- and extracellular solutions used). This similarity suggests that fast IR, like IKir, is largely K⫹ dependent. Third, we sometimes observed an apparent inactivation during strong hyperpolarizing steps (Fig. 4). This behavior of IKir has been reported previously in a number of cell types (Biermans et al. 1987; Constanti and Galvan 1983; Hagiwara et al. 1976; Silver and DeCoursey 1990; Yamoah et al. 1998) and has generally been attributed to an external voltage- and timedependent block by Na⫹ ions rather than true inactivation of the IKir channels themselves. EFFECTS OF DIVALENT CATIONS ON FAST IR. If the fast IR in spinal neonatal rat MNs is caused by IKir, it should be sensitive to Cs2⫹ concentrations in the millimolar range (Benson and Levitan 1983; Constanti and Galvan 1983; Hagiwara et al. 1976; Standen and Stanfield 1979; Travagli and Gillis 1994). Indeed, we found that fast IR was reduced or blocked by CsCl2 (0.3–1.5 mM, n ⫽ 6 MNs; not shown). IKir is known to be reduced by low concentrations (few hundred micromolar) of Ba2⫹, in contrast to Ih (e.g., Hagiwara et al. 1978). We therefore investigated the effect of Ba2⫹ on the instantaneous I-V relationship in those MNs that showed fast IR. The difference ⌬Ginst ⬅ Ghyp ⫺ Gdep was determined in control and after addition of Ba2⫹. From its definition, ⌬Ginst is positive when fast IR is present, and negative in the face of

FIG. 2. The Ih antagonist ZD 7288 blocks slowly developing IR but not fast IR. A and B: current responses (top) to hyperpolarizing voltage steps from ⫺50 mV (bottom) before (A) and during (B) application of ZD 7288 (100 ␮M). Note that in ZD 7288 Ih was abolished while the fast IR remained, as in control. Holding current, ⫺120 pA in A; ⫺4 pA in B. C: instantaneous I-V plots based on measurements from A and B at the time points indicated by the symbols also used for plotting. Note persistence of fast IR in ZD 7288. Dashed lines were added for better appreciation of the curve bending related to fast IR.

instantaneous OR. In all MNs, Ba2⫹ reversed ⌬Ginst from positive (fast IR) to negative (fast OR), the difference being highly significant (4.5 ⫾ 1.5 nS, control; ⫺1.5 ⫾ 0.5 nS, Ba2⫹; n ⫽ 5, P ⬍ 0.0001). This conversion of fast IR to fast OR is illustrated for selected MNs in Figs. 3 and 4. ⌬Ginst changed back from negative (fast OR) to positive (fast IR) on washout

584

O. KJAERULFF AND O. KIEHN

voltage range, in which IKir is strong. For example, a more equal reduction in Ghyp and Gdep would have been expected if barium had mainly acted on leak currents. Some effect of Ba2⫹ on leak currents cannot, however, be excluded. An outward current relaxation was sometimes observed in response to large hyperpolarizing voltage steps (i.e., beyond approximately ⫺80 mV; Fig. 4A). This outward relaxation, generally preceded by a brief inward relaxation, was slow, and typically did not reach steady state before the voltage step was terminated, typically after 3 s. Since Ih or its related currents do not inactivate (DiFrancesco 1985; Pape 1996), the outward relaxation in response to extreme hyperpolarizing steps cannot be caused by Ih inactivation. However, it may be caused by the “inactivation” of IKir (Biermans et al. 1987; Yamoah et al. 1998). In accordance with this notion, addition of the IKir antagonist Ba2⫹ (300 ␮M, Fig. 4B) profoundly altered the response by abolishing the outward relaxation. This effect was reversible on washout of Ba2⫹ (Fig. 4C). BLOCKING IKir ENHANCES Ih. The inactivation of IKir can efficiently mask Ih. This important point is illustrated in Fig. 4. Before Ba2⫹, the inward relaxation characteristic of Ih was inconspicuous during voltage steps of small/medium amplitude (Fig. 4A). Furthermore, during stronger hyperpolarizations, it was largely obscured by the persistent outward relaxation that we ascribe to the inactivation of IKir. These observations provide the false impression that Ih was largely absent in this MN. However, this was not the case, since an inward relaxation with a voltage dependency of amplitude and activation rate characteristic of Ih was observed after the Ba2⫹ block of IKir (Fig. 4B). This and similar findings in other MNs points out that to study the behavior of Ih in detail, it is necessary to first eliminate IKir. We conclude that in neonatal rat spinal motor neurons a fast inward rectifier very likely to be a potassium-dependent IKir often co-exists with the slow mixed cation-dependent inward rectifier Ih. OUTWARD CURRENT RELAXATION BLOCKED BY Ba2⫹.

2⫹ FIG. 3. Ba blocks fast IR in a manner compatible with an involvement of IKir. A: instantaneous I-V relationships obtained in normal medium (●), during 2⫹ addition of Ba (300 ␮M, E), and following Ba2⫹ washout (Œ). Note reversible conversion of fast IR to fast outward rectification (OR). The potential at which the I-V curves bend maximally is indicated (Vbend, approximately ⫺80 mV); this potential defines the voltage zones relating to the “hyperpolarized” (Ghyp) and “depolarized” (Gdep) instantaneous conductances (see text). B: the barium-sensitive current (IBa2⫹) calculated by subtracting in A either the control curve or the wash curve from the curve obtained in Ba2⫹ (symbol equations). The approximate reversal potential for IBa2⫹ is indicated [E(IBa2⫹), ⫺82 mV]. Note the similarity of this value with Vbend in A.

of Ba2⫹ (⫺1.4 ⫾ 0.4 nS, Ba2⫹; 9.5 ⫾ 4.9 nS, wash; n ⫽ 3, P ⬍ 0.05; Fig. 3). The reduction in ⌬Ginst caused by Ba2⫹ was due to a dominant reduction of Ghyp (19.6 ⫾ 6.7 nS, control; 11.8 ⫾ 3.6 nS, Ba2⫹, n ⫽ 5, P ⬍ 0.05). Gdep also showed a reduction, which, however, was smaller and not statistically significant (15.1 ⫾ 5.4 nS, control; 13.3 ⫾ 3.9 nS, Ba2⫹, P ⬎ 0.5). REVERSAL POTENTIAL FOR IBa2⫹. The reversal potential for the current blocked by Ba2⫹, IBa2⫹, was determined as the intersection between the instantaneous I-V curves obtained in the same cells in control and during Ba2⫹ application (Fig. 3). The reversal potential for IBa2⫹ was ⫺75.0 ⫾ 3.4 mV (n ⫽ 5); this is close to the potassium equilibrium potential (i.e., ⫺79 mV). IBa2⫹ was generally of small, albeit not negligible, amplitude at membrane potentials depolarized to its reversal potential. These data suggest that Ba2⫹ blocks a fast inward rectifying potassium current (IKir) in neonatal rat spinal MNs. Furthermore, compared with blockade of other potassium currents, the blockade of IKir appears to be the dominant action of Ba2⫹ in the low concentration used here. This notion is based on the observed selective conductance reduction in the hyperpolarized

5-HT influences the instantaneous I-V relation The discovery of IKir in neonatal rat spinal MN prompted an investigation of the possible influence of 5-HT on IKir and other instantaneous currents. By separately analyzing the effect of 5-HT on the “depolarized” and the “hyperpolarized” instantaneous conductances (Gdep and Ghyp, respectively; see METHODS), we found that 5-HT reduces the instantaneous conductance in a voltage-dependent manner. 5-HT reduced Gdep from 21.4 ⫾ 3.0 nS (control) to 15.8 ⫾ 2.1 nS (5-HT). This reduction was significant (P ⬍ 0.00001, n ⫽ 24 MNs, paired t-test; Fig. 5A) and partially reversible, since Gdep rose again on washout of 5-HT (16.7 ⫾ 4.4 nS, 5-HT; 19.8 ⫾ 5.1 nS, wash, n ⫽ 9, P ⬍ 0.05). In the same MNs, 5-HT also significantly reduced Ghyp, from 28.2 ⫾ 4.6 nS in control to 24.4 ⫾ 3.9 nS in 5-HT (P ⬍ 0.005, Fig. 5A). This effect is probably normally reversible, although we were not able to demonstrate this in the present experiments (Ghyp ⫽ 24.8 ⫾ 8.6 nS, 5-HT; 23.5 ⫾ 7.3 nS, P ⬎ 0.5). Thus 5-HT reduced the instantaneous conductance over the entire voltage range tested. Interestingly, this reduction was significantly stronger in the depolarized voltage range than in the hyperpolarized range (reduction in Gdep ⫽

5-HT EFFECTS ON INWARD RECTIFICATION

585

DEDUCED EFFECTS OF 5-HT ON IKir AND OTHER INSTANTANEOUS

Comparing the results of 5-HT application on the instantaneous currents in normal medium and in barium leads to the following conclusions. Since barium, in concentrations favoring the antagonism of IKir, prevented the 5-HT– dependent reduction in Ghyp, we suggest that the Ghyp reduction stems from 5-HT diminishing IKir. In contrast, the 5-HT– dependent reduction of Gdep observed in normal medium is unlikely to strongly involve IKir, since this effect was not prevented by the IKir antagonist barium. Rather, 5-HT appears to reduce other instantaneous conductances in addition to IKir (see DISCUSSION). CURRENTS.

Reversal potential of Ih

2⫹ FIG. 4. Unmasking of Ih by blocking the IKir with Ba . A and B, control; C and D, Ba2⫹; E and F, following Ba2⫹ wash-off. A, C, and E: current responses (top traces) evoked by a family of voltage steps (bottom traces) from a holding potential of ⫺50 mV. In the corresponding I-V relationships (B, D, and F), the instantaneous (●) and steady-state (E) I-V curves are taken at the time points indicated by the same symbols in A, C, and E. In B, D, and F, the difference current (see text; Œ) results from subtracting the instantaneous current (●) from the steady-state current (E). Note that fast IR appears in control and wash but is absent in Ba2⫹ (compare ● in B and F with those in D). On the other hand, Ih seems to be absent in control and wash but is unmasked by Ba2⫹ (compare Œ in B and F with those in D). Holding current (pA), 94 in A; 59 in C; 179 in E. Scale bars in E apply for A, C, and E.

⫺5.7 ⫾ 1.3 nS vs. reduction in Ghyp ⫽ ⫺3.9 ⫾ 2.0 nS; n ⫽ 24; P ⬍ 0.05). Stated equivalently, 5-HT enhanced the inward bend in the instantaneous I-V relationship. These data suggest that the fast IR is under neuromodulatory control in spinal MNs. EFFECTS OF 5-HT ON INSTANTANEOUS CURRENTS IN BARIUM. It has been argued above that low concentrations of barium provide a relatively selective blockade of the putative IKir. In an attempt to extract the effect of 5-HT on IKir from effects on other instantaneous currents, we repeated the experiments described in the previous paragraph, but this time with barium present both before and during the addition of 5-HT, to permanently block IKir. In Ba2⫹ (100 –300 ␮M), 5-HT still reduced Gdep (28.3 ⫾ 0.8 nS, Ba2⫹ alone; 24.3 ⫾ 0.9 nS, Ba2⫹ plus 5-HT, n ⫽ 17, P ⬍ 0.001, Fig. 5B). In contrast, the 5-HT–induced reduction of Ghyp observed in normal medium was not seen in Ba2⫹ (21.9 ⫾ 0.6 nS, Ba2⫹ alone; 23.5 ⫾ 0.7 nS, Ba2⫹ and 5-HT, P ⬎ 0.10, Fig. 5B).

In the following we discuss the effects of 5-HT on Ih. To permit the evaluation of these effects in terms of conductance rather than current alone (see below), we determined Eh, the reversal potential of Ih. We used the method previously employed by others (Bayliss et al. 1994; Mayer and Westbrook 1983; Takahashi 1990a), where the intercept of two instantaneous I-V curves evoked from two different holding potentials is used to estimate Ih (not shown) (see Fig. 6 in Takahashi 1990a for an illustration of a similar protocol). The holding potentials were kept below the threshold for depolarization-induced currents. Eh was determined in 125 experiments (50 MNs) under various pharmacological conditions. Due to the presence of some clear outliers, the median rather than the mean was chosen for calculations (see below) of the Ih conductance. The median Eh was ⫺33 mV, well in agreement with previous measurements in the same MNs (Takahashi 1990a). 5-HT effects on Ih in barium The fact that Ba2⫹ blocks IKir (Figs. 3 and 4) allowed us to isolate the effect of 5-HT on Ih. Ih was elicited by hyperpolarizing voltage steps before and after the addition of 5-HT, in the constant presence of 200 –300 ␮M Ba2⫹ to eliminate IKir. Measurements in 5-HT were taken as soon as possible after control measurements, to minimize the impact of time-dependent rundown (see later). In the example MN in Fig. 6, the access resistance, Ra, associated with the recording was low, so that voltage steps of the same nominal amplitude were similar even after correcting the membrane potential for the voltage drop over Ra (see METHODS). In this MN, therefore, the current responses to voltage steps of the same order, elicited before and during 5-HT, could be compared directly (Fig. 6A). Responses were identical down to ⫺66 mV. Beginning at ⫺72 mV, however, the Ih amplitude was larger in 5-HT than in control. This difference (in steady state) increased with further hyperpolarization, peaked at ⫺90 mV, and then gradually declined again to reach zero at ⫺107 mV. The I-V curve in Fig. 6B summarizes the 5-HT enhancement of the steady-state Ih, calculated as the difference between the steady-state current, Iss, and the instantaneous current, Iin. The average amplitude measured at ⫺80 mV was 144 ⫾ 36 pA in barium alone, and 252 ⫾ 62 pA in barium plus 5-HT (P ⬍ 0.001, n ⫽ 16 MNs, Fig. 6C). Thus the enhancement ENHANCEMENT OF THE Ih AMPLITUDE.

586

O. KJAERULFF AND O. KIEHN

(Larkman and Kelly 1992) but is in contrast to the situation in guinea pig trigeminal MNs where 5-HT increases Gmax (Hsiao et al. 1997). The slope factor was also unaffected by 5-HT. It was 8.5 ⫾ 0.8 mV⫺1 in Ba2⫹ alone, and 7.9 ⫾ 0.4 mV⫺1 in Ba2⫹ plus 5-HT (P ⬎ 0.3, n ⫽ 16 MNs, paired t-test). The pooled value was 8.2 ⫾ 0.5 mV⫺1. The potential at which the Gh conductance is half-maximally activated, V1/2, was depolarized 7 mV by 5-HT (⫺93.0 ⫾ 3.0 mV, Ba2⫹ alone; ⫺85.6 ⫾ 2.5 mV, Ba2⫹ ⫹ 5-HT; P ⬍ 0.005, n ⫽ 16 MNs, paired t-test, Fig. 7C). A similar, albeit less pronounced, depolarizing shift in V1/2 has been observed in adult rat facial MNs (3.4 mV depolarization) (Larkman and Kelly 1992). These data show that 5-HT enhances Ih by selectively imposing a depolarizing shift in the activation curve. 5-HT effects in normal medium

FIG. 5. Effects of serotonin (5-HT) on fast IR. A: instantaneous I-V relationship derived from current responses evoked by a family of voltage steps from ⫺50 mV before (normal medium, ●) and after (E) application of 10 ␮M 5-HT. Note pronounced reduction of the depolarized conductance (Gdep) and weaker reduction in the hyperpolarized conductance (Ghyp). In combination, these conductance changes enhanced bend already present in control. B: same as in A but experiment carried out in the constant presence of 300 ␮M Ba2⫹. Note that 5-HT again reduced Gdep while there was no change in Ghyp. The fast OR induced by Ba2⫹ (see text) was reduced. The approximate Vbend is indicated in A and B, which represent different MNs.

of the Ih amplitude in the intermediate voltage range was a general finding in the barium-containing solution. 5-HT ENHANCEMENT OF V1/2. The estimated Eh permitted us to convert steady-state Ih-V curves (Fig. 6B) to Gh activation curves (Fig. 7A), using the equation G h ⫽ I h Ⲑ共V ⫺ E h 兲

(2)

The data were fitted to a Boltzmann function of the form G ⫽ G max Ⲑ兵1 ⫹ exp关共V ⫺ V1Ⲑ2 兲Ⲑk兴其

(3)

where Gmax is the maximal conductance, k is a slope factor, and V1/2 is the potential at which the conductance is half-maximally activated (Fig. 7A). The maximal conductance in Ba2⫹ before and after adding 5-HT was 11.8 ⫾ 1.5 nS and 12.2 ⫾ 1.7 nS, respectively. These values were not significantly different (P ⬎ 0.5, n ⫽ 16 MNs, paired t-test, Fig. 7B). We therefore pooled the data obtained without 5-HT with those obtained in the presence of the drug, thereby obtaining a Gmax of 12.0 ⫾ 1.5 nS. This Gmax value for Ih is substantially higher than previously reported in neonatal spinal MNs (Takahashi 1990a). The lack of effect of 5-HT on Gmax agrees with the findings in adult rat facial MNs

To test whether the 5-HT enhancement of Ih is a robust phenomenon present under standard conditions and not only under circumstances aimed at pharmacologically isolating Ih, experiments similar to those described in the previous section were performed in standard Krebs solution, i.e., without barium. These data include measurements taken during wash-off of 5-HT. As in barium, 5-HT also enhanced the difference current, Iss – Iin under standard conditions. The effect of 5-HT was transient, such that after prolonged exposure (10 –15 min) to 5-HT the difference current Iss – Iin again began to decline, despite the continuous presence of the drug (not shown). Furthermore, the difference current was typically smaller in wash than in control. We attribute these effects to rundown of Ih (Takahashi and Berger 1990), a current known in these and other cells to be modulated by diffusible second messengers (Hille 1992; Larkman et al. 1995; Pape 1996). Iss – Iin at ⫺80 mV was determined 1) before 5-HT (Fig. 8A, Control), 2) between 5 and 10 min following the application of 5-HT (Early 5-HT), and 3) either more than 15 min after adding 5-HT, or in washout of 5-HT (Late 5-HT/wash). The pooling of data obtained late in 5-HT with those obtained in wash was justified by the observation that there was no substantial differences between measurements in these two conditions, presumably as a result of rundown. The mean of (Iss – Iin) showed a 19.0% average increase in 5-HT compared with control. This is close to significance at the 5% level (P ⫽ 0.07, n ⫽ 14). A 31.4% reduction in Iss – Iin compared with control was seen in the wash/late 5-HT group (P ⬍ 0.001), reflecting the rundown effect. Results obtained in normal medium should be interpreted with caution due to the presence of IKir (Fig. 4). However, even in normal medium Ih often appeared to form the main component of the difference current, Iss – Iin, despite the contribution from IKir. Therefore these results confirm that the 5-HT enhancement of Ih also takes place in standard conditions. Furthermore, in view of the observed rundown, it seems likely that our results may well underestimate the degree of the 5-HT enhancement occurring under physiological circumstances.

5-HT EFFECTS ON INWARD RECTIFICATION

587

2⫹ FIG. 6. 5-HT enhancement of Ih in Ba . A: superimposition of corresponding current traces in response to voltage steps from ⫺40 mV in the constant presence of 300 ␮M Ba2, before (bold traces) and after (light traces) adding 10 ␮M 5-HT. The holding current was 817 and 497 pA before and after adding 5-HT, respectively. Numbers to the right indicate the level of the voltage step in mV. B: plots of Ih against the step potential before (●) and after 5-HT (E). Ih was calculated as the difference between the steady state and the instantaneous currents. Note 5-HT–induced depolarizing shift in current activation. C: box-and-whisker plots summarizing Ih measurements at ⫺80 mV in 16 MNs. Note 5-HT enhancement of Ih. The P value refers to testing the difference in the means of square root transformed data (control vs. 5-HT; see METHODS).

Enhanced rundown of Ih in a poly-channel antagonist solution In some experiments, the standard Krebs solution was replaced with a zero- or low (10%) Ca2⫹ solution with an elevated [Mg2⫹] to which was added TEA (20 mM) and 4-AP (2 mM). Such a “poly-channel antagonist” solution is conventionally used to pharmacologically isolate Ih from voltage-gated Ca2⫹ currents, the delayed rectifier (IK), and the A current (IA) (Takahashi 1990b). Ih was measured at ⫺80 mV in 13 MNs (Fig. 8B), under the same conditions as in the experiments made in standard Krebs solution. Despite the application of 5-HT, a progressive decline of the Ih amplitude was typically observed in the poly-channel antagonist solution. On average, the Ih amplitude in 5-HT was reduced to 56% compared with control (P ⬍ 0.05, Fig. 8B) and after wash only 40% of the control Ih remained (P ⬍ 0.001). This observation is in sharp contrast to the findings in Ba2⫹ (Fig. 6C) and in normal medium (Fig. 8A), where 5-HT enhanced Ih. These data suggest that the poly-channel

antagonist solution enhances Ih rundown. The underlying mechanism is currently unknown. 5-HT accelerates Ih activation The results presented in this section all originate from the experiments done in the Ba2⫹-containing Krebs solution. Current activation in response to a family of hyperpolarizing voltage steps was fitted to a single-exponential function of the form I t ⫽ I ss ⫹ I h exp共⫺tⲐ␶ 兲

(4)

where It is the current at time t, Iss is the steady-state current reached at t ⫽ ⬁, Ih is the difference between Iss and the instantaneous current, and ␶ is a time constant. We also used a double-exponential function of the form I t ⫽ I ss ⫹ I fast exp共⫺tⲐ␶fast兲 ⫹ Islow exp共⫺tⲐ␶slow 兲

(5)

588

O. KJAERULFF AND O. KIEHN

amplitude. However, when the membrane potential was stepped to more hyperpolarized levels than approximately ⫺80 mV, double exponentials clearly fitted the current responses better than single exponentials. The fast and slow time constants, ␶fast and ␶slow, were extracted from the double-exponential fits (Fig. 9, A and B) and plotted against the membrane potential, V (Fig. 9, C and D). V was taken as the average between Vin and Vss (see METHODS). In 6 of 16 MNs (37.5%), there was no obvious relationship between V and ␶fast. This was also the case for smallamplitude steps in the remaining 10 MNs (62.5%, not shown). However, for larger steps, ␶fast systematically decreased with increasing hyperpolarization (Fig. 9C). The relationship between V and ␶fast was well described by the hyperbolic function ␶fast ⫽ 1Ⲑ共aV ⫹ b 兲

(6)

where a and b are constants (Fig. 9, C–F). Analysis of the impact of 5-HT on Ih kinetics relied on reciprocal transforma-

2⫹ FIG. 7. Effects of 5-HT on Gh in Ba . A: Gh activation curves calculated from the steady-state Ih currents in Fig. 6B using the equation Gh ⫽ Ih/(V ⫺ Eh), with Eh ⫽ ⫺33 mV (see text). Boltzmann functions were fitted to the data. Note that compared with control (●) 5-HT (E) did not markedly change the maximal conductance (Gmax) but caused a depolarizing shift in the activation curve. B and C: box-and-whisker plots summarizing the 5-HT effects on Gmax (B) and V1/2 (C) in the constant presence of Ba2⫹. n.s., not significant.

This function defines two amplitude components of Ih, decaying exponentially from the initial constants Ifast and Islow with the time constants ␶fast and ␶slow, respectively. The goodness-of-fit of single- and double-exponential functions did not differ substantially for voltage steps of low

FIG. 8. Effects of 5-HT on the steady-state Ih in normal medium (A) and in a poly-channel antagonist solution aimed at pharmacologically isolating Ih from other currents (B). A: 5-HT enhanced the difference current Iss – Iin in normal medium, in the face of some rundown. Under these conditions Ih generally dominates the difference current, although other currents may also contribute (see text). B: Ih was subject to severe rundown in the poly-channel antagonist solution and showed a monotonous decrease despite the application of 5-HT. See text for results and statistical analysis of the data in A and B.

5-HT EFFECTS ON INWARD RECTIFICATION

589

FIG. 9. 5-HT speeds up the activation of Ih. A and B: current responses (thick black lines) to a family of hyperpolarizing voltage steps from a holding potential of ⫺40 mV, evoked in Ba2⫹ without (A) and with 5-HT (B). Only traces well fitted to double exponential functions (white lines) are shown. The number to the right of each trace represents the level of the corresponding voltage step (in mV). C and D: voltage dependency of the fast and slow time constants, respectively, extracted from the double exponential curve fitting in A and B. The curves were drawn according to the hyperbolic function ␶ ⫽ 1/(aV ⫹ b), where ␶ represents ␶fast (C) or ␶slow (D), V is the step potential and a and b are constants. In D, the ␶slow value in Ba2⫹ at ⫺82 mV was treated as an outlier and omitted from the plot. E and F: summary of similar experiments in 10 MNs. Solid curves were drawn according to the same hyperbolic function as in C and D, but using the means of a and b obtained in individual experiments. The distance to the dashed curves indicates SE.

tion of ␶fast. Thus plots of kfast ⬅ 1/␶fast against V were fitted to the linear function k fast ⫽ aV ⫹ b

(7)

to estimate a and b (not shown). While the addition of 5-HT did not change a significantly (⫺0.18 ⫾ 0.03 s⫺1 mV⫺1, Ba2⫹; ⫺0.20 ⫾ 0.02 s⫺1 mV⫺1, Ba2⫹ and 5-HT; n ⫽ 10, P ⬎ 0.1, paired t-test), b was increased by 5-HT (⫺13.1 ⫾ 1.8 s⫺1, Ba2⫹; ⫺14.4 ⫾ 2.1 s⫺1, Ba2⫹ and 5-HT, n ⫽ 10, P ⬍ 0.01). This increase in b implies (Eq. 7) that 5-HT reduced ␶fast. The dependence of ␶slow on V also largely (10 of 16 MNs) obeyed Eq. 6, although the datapoints were more scattered than those related to ␶fast. Again, 5-HT did not change a (⫺0.022 ⫾ 0.004 s⫺1 mV⫺1, Ba2⫹; ⫺0.026 ⫾ 0.003 s⫺1 mV⫺1, Ba2⫹ and 5-HT; n ⫽ 10, P ⬎ 0.2, paired t-test) but increased b significantly (⫺1.5 ⫾ 0.3 s⫺1, Ba2⫹; ⫺1.7 ⫾ 0.3 s⫺1, Ba2⫹ and 5-HT, P ⬍ 0.00001, n ⫽ 10). Thus like ␶fast, ␶slow was decreased by 5-HT. These data show that activation of Ih is accelerated by 5-HT, which also appears from direct inspection of the raw data (Fig. 9A, see also Fig. 6A). Figure 9, E and F, shows the overall effect of 5-HT on ␶fast and ␶slow, respectively. When comparing the two time constants, 5-HT exerted its strongest effect on ␶slow. This was most obvious in the hyperpolarized region. We were not able to determine with certainty the time constants in the voltage region more depolarized than approx-

imately ⫺70 mV. This was partly due to weak activation of Ih but was also related to the fact that the fitted parameters were contaminated by the decay of the capacitive currents related to generation of the voltage steps. 5-HT effect on the holding current At a holding potential of ⫺50 mV, the holding current in normal medium (including TTX/CNQX/APV) was 80 ⫾ 23 pA (n ⫽ 12 MNs). Five minutes after adding 5-HT, the holding current measured in the same MNs displayed a strong inward shift, reaching ⫺51 ⫾ 27 pA (P ⬍ 0.0005, data not shown). This level did not change significantly during the next 5 min (5 min vs. 10 min in 5-HT, P ⬎ 0.3). We made the qualitative observation, however, that the holding current began to shift back in the outward direction after ⬃15–20 min in 5-HT. The transient effect of 5-HT on the holding current matches the clear but fleeting 5-HT enhancement of the Ih amplitude described in a previous section. An inward shift in holding current was also induced by 5-HT in the experiments done in the constant presence of Ba2⫹ (see above). The holding potential was ⫺40 mV, while the holding current was 440 ⫾ 47 pA in Ba2⫹ and 239 ⫾ 46 pA in Ba2⫹ plus 5-HT (P ⬍ 0.00005, n ⫽ 14 MNs, paired t-test). An inward shift in holding current was observed in a total of 26 MNs. Both enhancing Ih (which has a reversal potential more

590

O. KJAERULFF AND O. KIEHN

positive than the holding potential used in the present experiments) and inhibiting IKir and potassium-dependent leak currents (with a reversal potential more negative than the holding potential) leads to an increase in inward current. Therefore our consistent observation of an inward shift in the holding current is in accordance with our conclusions earlier in the paper that 5-HT enhances Ih and inhibits potassium-dependent currents including IKir. The frequent occurrence of the 5-HT–induced inward current suggests that this feature is important for normal motor function in the neonatal rat. DISCUSSION

Takahashi (1990a) reported the presence of a slowly activating, mixed Na⫹/K⫹-dependent inward current in neonatal rat spinal MNs. While he preferred to denote this current IIR, we have used the term Ih, which is now a more common name for this class of currents in neurons (Pape 1996). The Ih in MNs of the neonatal rat is enhanced by 5-HT (Larkman and Kelly 1997; Takahashi and Berger 1990). Some aspects of this neuromodulatory effect, especially the pharmacology, have been closely studied (Larkman and Kelly 1997; Larkman et al. 1995; Takahashi and Berger 1990). However, a detailed quantitative investigation of how 5-HT affects the biophysical parameters of Ih has been missing in mammalian MNs. Furthermore, previous studies on neonatal rat MNs have often involved recordings from MNs in thin slices. The MNs recorded from were situated superficially in the slices and were mechanically cleared to allow visual inspection of the cells (Larkman and Kelly 1998; Takahashi and Berger 1990). It seems likely that this procedure may have lead to a partial pruning of the dendritic tree, which could alter the quantitative features of whole cell currents. In contrast, here we present the results of whole cell tight-seal voltage-clamp recordings from MNs in the intact spinal cord. Too strong emphasis on the quantitative aspects of our data should be avoided due to the imperfections of the whole cell voltage-clamp technique, including the absence of ideal space clamp and the rundown of the secondmessenger– dependent currents such as Ih. Nevertheless, our results provide a necessary framework for future investigations, e.g., of 5-HT–mediated control of MN behavior during integrated motor tasks in intact preparations, including locomotor activity (Kiehn et al. 2000). INSTANTANEOUS INWARD RECTIFICATION. During experiments aimed at quantifying the effects of 5-HT on the time-dependent inward rectifier, Ih, we discovered that fast inward rectification in neonatal rat spinal MNs is quite common and often coexists with Ih. We treat this finding first, partly because it is novel and of independent interest, and partly because fast inward rectification complicated the analysis of neuromodulation of Ih, as discussed below. We found that half (49%) of the spinal MNs in the neonatal rat display fast IR. Several observations support the idea that this property is due to a potassium-dependent current belonging to the IKir class (Constanti and Galvan 1983; Hille 1992; Katz 1949; Nichols and Lopatin 1997; Standen and Stanfield 1979). First, the general swift activation of IKir fits well with our observation that the inward rectification not related to Ih was “instantaneous,” occurring immediately after settling of the capacitive transient associated with the change in voltage command. Second, since IKir is carried by potassium ions, the

voltage area in which inward rectification appears would be expected to be close to the potassium equilibrium potential, EK. This is in accordance with our finding that the inflection point in the instantaneous I-V relationships occurred at an average of ⫺80 mV; this value is essentially identical to the predicted EK of ⫺79 mV. Third, the fast IR in neonatal rat MNs was abolished by Ba2⫹ in concentrations of a few hundred micromolar, in agreement with the known Ba2⫹ sensitivity of IKir (Constanti and Galvan 1983; Hagiwara et al. 1978; Yamoah et al. 1998). A voltage dependency of the block imposed by Ba2⫹ on IKir has been suggested for supra-spinal MNs, with the potency of the ion block increasing with increasing hyperpolarization (Larkman and Kelly 1998). This conforms to our analysis of the effect of Ba2⫹ on the instantaneous conductance, which showed a significant reduction in Ghyp but no change in Gdep. However, it cannot be excluded that the apparent voltage-dependent effect of Ba2⫹ was simply due to the voltage dependency of IKir itself. Depending on the strength of inward rectification, the outward IKir flowing in the depolarized voltage range may be weak, in which case blocking it might have produced a conductance change too small to be detected. To summarize, we consider the evidence presented here strong enough to conclude that spinal neonatal rat MNs are endowed with an IKir. In view of our finding that half of the spinal neonatal MNs show fast IR, it is surprising that fast IR has not already been demonstrated directly in these cells despite several previous studies of their active membrane properties. In rat facial MNs (Larkman and Kelly 1997, 1998), the presence of an IKir subject to modulation by 5-HT has been inferred mainly from pharmacological evidence (see also below). However, fast IR in the I-V relationships was not reported in these studies. Indeed, these authors found that the K⫹dependent component of the 5-HT–sensitive current displayed a linear, i.e., nonrectifying, I-V relationship. One explanation for the absent demonstration of fast IR in previous studies may be that the extracellular potassium concentration was commonly raised (e.g., to 12 mM) (Larkman et al. 1995; Takahashi 1990a) to depolarize EK and thereby enhance the driving force of Ih. For IKir, the voltage associated with the most pronounced fast IR is identical to EK. This potential may have been depolarized beyond the range covered by the voltage-clamp commands, so that fast IR eluded observation in the previous work. Also, antagonists used to pharmacologically isolate Ih may have obscured the fast IR in previous studies. Specifically, we obtained some evidence suggesting that TEA occludes a Ba2⫹-induced positive shift in the holding current (unpublished observations). This may indicate that the IKir in the neonatal rat, besides being sensitive to barium, is also TEA sensitive, as described for other cells (Bobker and Williams 1995; Hille 1992; Travagli and Gillis 1994). The clear presence of fast IR in neonatal rat MNs in our experiments as opposed to those of others could also be related to differences in the recording circumstances (see above), especially if one assumes a dendritic predominance of IKir, relative to linear leak or outward rectifying currents. Finally, the different experimental approaches in this study compared with that of others might have led to a biased selection of MNs with a pronounced fast IR.

5-HT EFFECTS ON INWARD RECTIFICATION

5-HT modulation of the instantaneous conductance We found that 5-HT reduced the instantaneous conductance of spinal MNs in conjunction with a negative (depolarizing) shift in the holding current at ⫺50 mV. This effect may be decomposed into suppression of at least two currents, which dominate at different voltage ranges. Low concentrations of Ba2⫹, which were sufficient to block IKir, specifically prevented the 5-HT– dependent conductance reduction at hyperpolarized potentials (below approximately ⫺80 mV; Fig. 5); this 5-HT effect may therefore be explained at least partly by blockade of IKir. A previous demonstration of such a mechanism was made in nucelus accumbens neurons, where an instantaneous inward rectifying K⫹ conductance is reduced by 5-HT (North and Uchimura 1989). Also, 5-HT reduces K⫹ currents both in adult (Larkman and Kelly 1992) and neonatal (Larkman and Kelly 1998) rat facial MNs; one of these currents has been suggested to be an IKir mainly on pharmacological grounds. In the depolarized voltage range (above approximately ⫺80 mV), Ba2⫹ did not prevent the 5-HT–induced reduction in the instantaneous conductance in our recordings. Therefore a different current than IKir must be involved in this phenomenon. An interesting possibility is a background K⫹ current present in the neonatal rat spinal MNs (Fisher and Nistri 1993). This current is pharmacologically clearly different from IKir, since it is insensitive to Cs2⫹ (2 mM), and, although reduced in high (1.5 mM) [Ba2⫹], it is not affected by Ba2⫹ in low concentration (200 ␮M). The background current is inhibited by the neuropeptides TRH and substance P. Co-existence of these transmitters with 5-HT in fibers innervating spinal motor nuclei has been described (Arvidsson et al. 1992). Other currents may also be involved, however. In neonatal rat facial MNs, 5-HT reduces a potassium current, which is insensitive to antagonists of IKir, but sensitive to 4-AP (Larkman and Kelly 1998). It appears likely that a similar current exists in the spinal MNs. This current could contribute to the 5-HT effect provided it is not fully inactivated at the resting membrane potential. The reduction of the instantaneous conductance and the outward shift in holding current caused by ZD 7288 suggest that Ih also contributes to the resting conductance at potentials around ⫺50 mV. We found that 5-HT enhances this depolarizing contribution of Ih to the membrane potential (see also Kiehn et al. 2000). Provided that IKir allows current flow at potentials more positive that EK, it will generate an outward current that decreases MN excitability. Inhibition of IKir by 5-HT in the depolarizing voltage range will therefore increase MN excitability. From the present series of experiments, it appears that there is little steady activation of IKir at potentials more positive than EK (see previous paragraph), suggesting that the 5-HT modulation of IKir in the depolarizing voltage range plays a minor role for MN firing. Ih characteristics We have established the biophysical parameters of Ih in MNs in the intact cord in detail, and in this section compare them with the parameters previously determined in acute slices (Takahashi 1990a). Since the co-existence of IKir with

591

Ih in many MNs could obscure Ih (Fig. 4), we blocked the IKir with low concentrations of Ba2⫹. The maximal conductance, Gmax, was determined from Boltzmann fits to Gh/V plots and found to be 12.0 ⫾ 1.5 nS. Takahashi (1990a) reported a half-maximally activated chord conductance of about 0.8 nS. Hence his Gmax value (⬃1.6 nS) would be about an order of magnitude smaller than ours. A likely explanation for this difference is that we have recorded from presumably more intact MNs. This technical aspect would be of particular relevance if the Ih channels in neonatal rat MNs are abundant in the distal membrane, as has been reported for hippocampal CA1 pyramidal neurons, where the density of Ih in the most distal regions is roughly sevenfold higher than in the somatic region (Magee 1998). It is also possible that our experimental approach for an as yet undetermined reason has lead to a selection of large MNs compared with those of previous experimenters. We observed that for hyperpolarizing steps to approximately ⫺80 mV or beyond, the sum of two exponentials was necessary to obtain good fits to Ih activation, with ␶fast being measured in hundreds of milliseconds and ␶slow in seconds. Double-exponential activation of Ih in the neonatal rat MNs may indicate that two kinetically distinct Ih channel populations are present in these neurons (cf., Solomon and Nerbonne 1993). Interestingly, Takahashi (1990a) reported that a single, slow time constant was sufficient to describe the activation of Ih. This apparent discrepancy can be resolved if it is assumed that the two kinetically distinct Ih channel populations are spatially segregated, so that the slow channels are concentrated proximally while the fast channels mainly localize to the distal dendrites that were presumably better conserved in our experiments (see above). 5-HT modulation of Ih It has been reported previously that 5-HT increases a conductance very similar to Gh in spinal MNs (Takahashi and Berger 1990), but the biophysical nature of this effect was not determined. We found that this enhancement consists of a depolarizing shift in the activation curve for Ih, amounting to 7 mV on average. In contrast, 5-HT caused no systematic change in Gmax. Results from other MN types have been mixed. In adult rat facial MNs, 5-HT depolarizes V without changing the maximum tail current amplitude of Ih, i.e., similar to our findings (Larkman and Kelly 1992). On the other hand, 5-HT increases Ih in guinea pig trigeminal MNs without shifting the (normalized) activation curve, suggesting that this enhancement is related to an increase in Gmax (Hsiao et al. 1997). Apparently, the biophysical mechanism underlying the 5-HT enhancement of Ih in mammalian MNs is heterogeneous. In our experiments, 5-HT shortened both the fast and the slow time constant. However, ␶slow was clearly stronger affected that ␶fast (Fig. 9, compare E with F). In combination with the considerations in the previous section, this observation leads us to suggest that neonatal rat spinal rat MNs may be endowed with two kinetically, spatially and pharmacologically distinct Ih channel populations. One of these populations predominates in the soma and proximal dendrites, activates slowly on hyperpolarization, and is strongly enhanced by 5-HT. The second population, although also represented proximally, pre-

592

O. KJAERULFF AND O. KIEHN

dominates in the distal dendrites, activates quickly, and is more moderately enhanced by 5-HT. More experiments are necessary to further substantiate this idea. We thank M. E. Denton and B. Johnson for participating in initial experiments in this study. This work was supported by the NOVO Foundation and the Danish Medical Research Council. Present address of O. Kjaerulff: The Nobel Institute for Neurophysiology, Dept. of Neuroscience, The Karolinska Institute, S-171 77 Stockholm, Sweden (E-mail: [email protected]).

REFERENCES ALTMAN DG. Practical Statistics for Medical Research. London: Chapman and Hall, 1991. ARVIDSSON U, ULFHAKE B, CULLHEIM S, SHUPLIAKOV O, BRODIN E, FRANCK J, ¨ BENNET GW, FONE KC, VISSER TJ, AND HOKFELT T. Thyrotropin-releasing hormone (TRH)–like immunoreactivity in the grey monkey (Macaca fascicularis) spinal cord and medulla oblongata with special emphasis on the bulbospinal tract. J Comp Neurol 322: 293–310, 1992. BAYLISS DA, VIANA F, BELLINGHAM MC, AND BERGER AJ. Characteristics and postnatal development of a hyperpolarization-activated inward current in rat hypoglossal motoneurons in vitro. J Neurophysiol 71: 119 –128, 1994. BENSON JA AND LEVITAN BL. Serotonin increases an anomalously rectifying K⫹ current in the Aplysia neuron R15. Proc Natl Acad Sci USA 80: 3522–3525, 1983. BERGER AJ AND TAKAHASHI T. Serotonin enhances a low-voltage activated calcium current in rat spinal motoneurons. J Neurosci 10: 1922–1928, 1990. BIERMANS G, VEREECKE J, AND CARMELIET E. The mechanism of the inactivation of the inward-rectifying K current during hyperpolarizing steps in guinea-pig ventricular myocytes. Pfl¨ugers Arch 410: 604 – 613, 1987. BOBKER DH AND WILLIAMS JT. The serotonergic inhibitory postsynaptic potential in prepositus hypoglossi is mediated by two potassium currents. J Neurosci 15: 223–229, 1995. BOSMITH RE, BRIGGS I, AND STURGESS NC. Inhibitory actions of ZENECA ZD 7288 on whole-cell hyperpolarization activated inward current (If) in guinea pig dissociated sinoatrial node cells. Br J Pharmacol 110: 343–349, 1993. CAZALETS JR, SQUALLI-HOUSSAINI Y, AND CLARAC F. Activation of the central pattern generators for locomotion by serotonin and excitatory amino-acids in an in vitro new born spinal cord preparation. J Physiol (Lond) 455: 187–204, 1992. CONSTANTI A AND GALVAN M. Fast inward-rectifying current accounts for anomalous rectification in olfactory cortex neurones. J Physiol (Lond) 385: 153–178, 1983. DIFRANCESCO D. The cardiac hyperpolarization-activated current, If. Origins and developments. Prog Biophys Mol Biol 46: 163–183, 1985. FINKEL AS AND REDMAN SJ. Theory and operation of a single microelectrode voltage clamp. J Neurosci Methods 11: 101–107, 1984. FISHER ND AND NISTRI A. A study of the barium-sensitive and -insensitive components of the action of thyrotropin-releasing hormone on lumbar motoneurons of the rat isolated spinal cord. Eur J Neurosci 5: 1360 –1369, 1993. HAGIWARA S, MIYAZAKI S, AND ROSENTHAL NP. Blocking effects of barium and hydrogen ions on the potassium current during anomalous rectification of the egg cell membrane in the starfish egg. J Physiol (Lond) 279: 167–185, 1978. HAGIWARA S, MIYAZAKI S, AND ROSENTHAL P. Potassium current and the effect of cesium on this current during anomalous rectification of the egg cell membrane of a starfish. J Gen Physiol 67: 632– 638, 1976. HAMILL OP, MARTY A, NEHER E, SAKMANN B, AND SIGWORTH FJ. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pfl¨ugers Arch 391: 85–100, 1981. HARRIS NC AND CONSTANTI A. 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, 1995. HILLE B. Ionic Channels of Excitable Membranes. Sunderland, MA: Sinauer, 1992. HSIAO CF, TRUEBLOOD PR, LEVINE MS, AND CHANDLER SH. Multiple effects of serotonin on membrane properties of trigeminal motoneurons in vitro. J Neurophysiol 77: 2910 –2924, 1997.

HUGHES SW, COPE DW, AND CRUNELLI V. Dynamic clamp study of Ih modulation of burst firing and ␦ oscillations in thalamocortical neurons in vitro. Neuroscience 87: 541–550, 1998. ITO M AND OSHIMA T. Electrical behaviour of the motoneurone membrane during intracellularly applied current steps. J Physiol (Lond) 180: 607– 635, 1965. KATZ B. Les constance e´ lectriques de la membrane du muscle. Arch Sci Physiol 51: 196S, 1949. KHAKH BS AND HENDERSON G. Hyperpolarization-activated cationic currents (Ih) in neurones of the trigeminal mesencephalic nucleus of the rat. J Physiol (Lond) 510: 695–704, 1998. KIEHN O. Electrophysiology of 5-HT on vertebrate motoneurones. In: Aspects of Synaptic Transmission: LTP, Galanin, Opiods, Autonomic and 5-HT, edited by Stone T. London: Taylor, 1991a, vol. 1, p. 527–555. KIEHN O. Plateau potentials and active integration in the ‘final common pathway’ for motor behaviour. Trends Neurosci 14: 68 –73, 1991b. KIEHN O AND KJAERULFF O. Spatiotemporal characteristics of 5-HT and dopamine-induced rhythmic hindlimb activity in the in vitro neonatal rat. J Neurophysiol 75: 1472–1482, 1996. KIEHN O, KJAERULFF O, TRESCH MC, AND HARRIS-WARRICK RM. Contributions of intrinsic motor neuron properties to the production of rhythmic motor output in the mammalian spinal cord. Brain Res Bull. In press. KJAERULFF O AND KIEHN O. Distribution of networks generating and coordinating locomotor activity in the neonatal rat spinal cord in vitro: a lesion study. J Neurosci 16: 5777–5794, 1996. KJAERULFF O AND KIEHN O. Crossed rhythmic synaptic input to motoneurons during selective activation of the contralateral spinal locomotor network. J Neurosci 17: 9433–9447, 1997. KJAERULFF O AND KIEHN O. Quantitative effects of 5-HT on inward rectification in neonatal rat spinal motoneurons. Abstr Scand Phys Soc 26: 689, 1998. LARKMAN PM AND KELLY JS. Ionic mechanisms mediating 5-hydroxytryptamine- and noradrenaline-evoked depolarization of adult facial motoneurones. J Physiol (Lond) 456: 473– 490, 1992. LARKMAN PM AND KELLY JS. Modulation of Ih by 5-HT in neonatal rat motoneurones in vitro: mediation through a phoshorylation indepent action of cAMP. Neuropharmacology 36: 721–733, 1997. LARKMAN PM AND KELLY JS. Characterization of 5HT-sensitive potassium conductances in neonatal rat facial motoneurones in vitro. J Physiol (Lond) 508: 67– 81, 1998. LARKMAN PM, KELLY JS, AND TAKAHASHI T. Adenosine 3⬘: 5⬘-cyclic monophosphate mediates a 5-hydroxytryptamine–induced response in neonatal rat motoneurones. Pfl¨ugers Arch 430: 763–769, 1995. MAGEE JC. Dendritic hyperpolarization-activated currents modify the integrative properties of hippocampal CA1 pyramidal neurons. J Neurosci 18: 7613–7624, 1998. MAYER ML AND WESTBROOK GL. A voltage-clamp analysis of inward (anomalous) rectification in mouse spinal sensory ganglion neurons. J Physiol (Lond) 340: 19 – 45, 1983. NEHER E. Correction for liquid junction potentials in patch clamp experiments. Methods Enzymol 207: 123–131, 1992. NICHOLS CG AND LOPATIN AN. Inward rectifier potassium channels. Annu Rev Physiol 57: 171–191, 1997. NORTH RA AND UCHIMURA N. 5-Hydroxytryptamine acts at 5-HT2 receptors to decrease potassium conductance in rat nucleus accumbens neurones. J Physiol (Lond) 417: 1–12, 1989. PAPE H-C. Queer current and pacemaker: the hyperpolarization-activated cation current in neurons. Annu Rev Physiol 58: 299 –327, 1996. REKLING JC, FUNK GD, BAYLISS DA, DONG XW, AND FELDMAN JL. Synaptic control of motoneuronal excitability. Physiol Rev 80: 767– 852, 2000. SCHWINDT PC AND CRILL WE. Membrane properties of cat spinal motoneurones. In: Handbook of the Spinal Cord, edited by Davidoff RA. New York: Dekker, 1984, p. 199 –242. SILVER MR AND DECOURSEY TE. Intrinsic gating of inward rectifier in bovine pulmonary artery endothelial cells in the presence or absence of internal Mg2⫹. J Gen Physiol 96: 109 –133, 1990. SOLOMON JS AND NERBONNE JM. Two kinetically distinct components of hyperpolarization-activated current in rat superior colliculus-projecting neurons. J Physiol (Lond) 469: 291–313, 1993. STANDEN NB AND STANFIELD PR. Potassium depletion and sodium block of potassium currents under hyperpolarization in frog sartorius muscle. J Physiol (Lond) 294: 497–520, 1979.

5-HT EFFECTS ON INWARD RECTIFICATION TAKAHASHI T. Inward rectification in neonatal rat spinal motoneurones. J Physiol (Lond) 423: 47– 62, 1990a. TAKAHASHI T. Membrane currents in visually identified motoneurones of neonatal rat spinal cord. J Physiol (Lond) 423: 27– 46, 1990b. TAKAHASHI T AND BERGER AJ. Direct excitation of rat spinal motoneurones by serotonin. J Physiol (Lond) 423: 63–76, 1990. TRAVAGLI RA AND GILLIS RA. Hyperpolarization-activated currents Ih and IKir in rat dorsal motor nucleus of the vagus neurons in vitro. J Neurophysiol 71: 1308 –1317, 1994.

593

WANG MY AND DUN NJ. 5-hydroxytryptamine responses in neonate rat motoneurones in vitro. J Physiol (Lond) 430: 87–103, 1990. WILLIAMS SR, TURNER JP, HUDGES SW, AND CRUNELLI V. On the nature of anomalous rectification in thalamocortical neurones of the cat ventrobasal thalamus in vitro. J Physiol (Lond) 505: 727–747, 1997. YAMOAH EN, MATZEL L, AND CROW T. Expression of different types of inward rectifier currents confers specificity of light and dark responses in type A and B photoreceptors of Hermissenda. J Neurosci 18: 6501– 6511, 1998.