INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The initial studies of the relation between the intrinsic properties of motoneurons and their input-output behavior focused on the contributions of their passive membrane properties and the conductances underlying the postspike afterhyperpolarization (cf. Binder et al. 1996; Burke 1981; Henneman and Mendell 1981; Kernell 1983; Powers and Binder 2001). Subsequent work revealed that motoneuron behavior is also strongly influenced by voltage-dependent persistent (i.e., non- or slowly inactivating) inward currents first activated in the voltage range below the threshold for spike initiation (e.g., Hounsgaard et al. 1984; Hounsgaard and Kiehn 1989; Lee and Heckman 1998a,b; Schwindt and Crill 1982). Under voltage-clamp conditions, these persistent inward currents can lead to a region of negative slope conductance in the steady-state current-voltage (I-V) relation (e.g., Hsiao et al. 1998;Lee and Heckman 1998a; Nishimura et al. 1989; Schwindt and Crill 1982). Under current clamp conditions, these currents are thought to contribute to a number of behaviors, including an increase in the slope of the frequency-current relation (e.g., Bennett et al. 1998; Lee and Heckman 1998a; Schwindt and Crill 1982), firing rate acceleration (Hounsgaard et al. 1988; Lee and Heckman 1998a), bistable discharge behavior (Hounsgaard et al. 1984, 1988; Lee and Heckman 1998a), and amplification of synaptic current (Lee and Heckman 2000; Powers and Binder 2000; Prather et al. 2001).

Persistent inward currents in spinal motoneurons are thought to be mediated by membrane channels located primarily on the dendrites (Carlin et al. 2000b; Hounsgaard and Kiehn 1993). One indication of the dendritic location of these channels is a clockwise hysteresis in the whole-cell inward currents recorded in response to ascending and then descending voltage ramp commands, i.e., the deactivation of the inward current on the descending ramp occurs at a lower somatic voltage than activation on the ascending ramp (Carlin et al. 2000b; Lee and Heckman 1998b; Svirskis and Hounsgaard 1997). This phenomenon is thought to arise from the fact that the voltage-dependent channels carrying the inward current are electrically distant from the soma and not under voltage-clamp control. As a result, the distal channels are activated at relatively high somatic depolarizations during the ascending voltage ramp command but then continue to supply current to the soma during the descending ramp command (cf. Booth et al. 1997; Carlin et al. 2000b; Lee and Heckman 1998b). Other signs that distal channels may make a significant contribution to inward voltage-clamp currents are a delayed current onset in response to a depolarizing voltage-clamp step (Booth et al. 1997; Carlin et al. 2000b; Muller and Lux 1993) and a prolonged decay following the step offset (Carlin et al. 2000b).

The sources of persistent inward currents may vary among motoneurons innervating different muscles, motoneurons in different species, and motoneurons at different developmental stages. In both turtle and mouse spinal motoneurons, persistent inward currents are mediated primarily by L-type calcium channels (Carlin et al. 2000a,b; Hounsgaard and Mintz 1988). In guinea pig brain stem motoneurons, sodium (Nishimura et al. 1989) or a mixture of sodium and L-type calcium channels (Hsiao et al. 1998) mediates the persistent inward current. In neonatal rat facial and hypoglossal motoneurons, these currents are mediated predominantly by calcium channels, but L-type channels carry only about 5-7% of the total calcium current measured at the soma (Plant et al. 1998; Umemiya and Berger 1994). However, the small proportion of L-type channels in neonatal motoneurons may not reflect the adult state, since the density of L-type channels increases during development (Jiang et al. 1999; Perrier et al. 2000).

The purpose of the present study was to characterize persistent inward currents in juvenile (10- to 24-day-old) rat hypoglossal motoneurons. We chose to study these particular motoneurons for several reasons. First, many of their basic electrophysiological properties, as well as their modulation by different neurotransmitters, have been well characterized (cf. Bayliss et al. 1997). Second, unlike spinal motoneurons (see above), there is no evidence that hypoglossal motoneurons exhibit bistable discharge behavior (Bayliss et al. 1997). Finally, the dendritic trees of hypoglossal motoneurons are oriented primarily in the transverse plane, so that most of the dendritic trees are preserved in transverse brain stem slices (Nunez-Abades et al. 1994), maximizing the likelihood that dendritic currents can be measured from whole-cell recordings made at the soma.

We found that outward currents dominated the whole-cell voltage-clamp records obtained from the somata of hypoglossal motoneurons. However, there was often a region of negative slope on the rising phase of the response to a slow, triangular voltage-clamp command, indicating the presence of persistent inward currents. We characterized these inward currents following block of potassium currents with internal cesium and external 4-aminopyridine (4-AP) and tetraethylammonium (TEA). TTX-sensitive sodium channels mediated a small portion of the persistent inward current, whereas calcium channels mediated the majority of the current. We found roughly the same relative contributions of P-, N-, and L-type channels to the calcium currents recorded in the somata of these more mature hypoglossal motoneurons as previously found in neonatal cells (Umemiya and Berger 1994). A substantial clockwise hysteresis in the I-V relation and long tail currents following voltage-clamp steps were observed in about one-third of the cells, consistent with a prominent contribution from dendritic channels. However, in contrast to previous reports on turtle (Hounsgaard and Mintz 1988) and mouse motoneurons (Carlin et al. 2000a,b), blocking L-type calcium channels did not eliminate these presumed dendritic currents.

A preliminary account of some of these data has been presented (Powers et al. 2001).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Slice preparation

Rat hypoglossal motoneurons were studied in 300-µm-thick brain stem slices obtained from 10- to 24-day-old Sprague-Dawley rats. Following induction of anesthesia with an intramuscular injection of a mixture of ketamine (68 mg kg¹) and xylazine (4 mg kg¹), the animals were decapitated and a section of brain stem was removed and glued to a Plexiglas tray filled with cooled, modified, artificial cerebrospinal fluid (ACSF). A DSK microslicer was used to cut a series of transverse slices throughout the length of the hypoglossal nucleus. The slices were transferred to a holding chamber and incubated at room temperature (19-21°C) for 30 min in the modified ACSF for 30 min, followed by 30 min incubation in standard ACSF.

Solutions and chemicals

To minimize neural activity during the initial preparation of the slices, two different modified ACSF solutions were used. For slices obtained from younger animals (10-16 days), we used a low Ca²⁺, high Mg²⁺ solution [Low Ca-ASCF (in mM): 132 NaCl, 3 KCl, 1.25 NaH₂PO₄, 26 NaHCO₃, 5 MgCl₂, 1 CaCl₂, and 10 D-glucose], whereas in older animals we used a sucrose-based solution (S-ACSF: same as Low-Ca ACSF, except 220 mM sucrose was substituted for NaCl and concentrations of MgCl₂ and CaCl₂ were both 2 mM). Kynurenic acid (1 mM) and lactate sodium (4 mM) were also added to the initial incubation medium to improve cell viability. The standard ASCF was identical to that of the S-ACSF except that 132 mM NaCl was substituted for sucrose. The pH of the ASCF solutions ranged from 7.3 to 7.4 and their measured osmolalities ranged from 310 to 320 mOsm.

In some experiments MnCl₂ was substituted for CaCl₂. Phosphate was omitted from ACSF solutions containing Mn²⁺. Variations in NaCl content were made to effect similar osmolalities when used with 1) added K⁺ channel blockers or 2) K⁺ channel blockers added in conjunction with neurotransmitter antagonists to reduce synaptic noise during recording. This was unnecessary when using the sodium channel blocker TTX. The NaCl concentration was 120 mM when combined with 4 mM 4-AP and 10 mM tetraethylammonium chloride (TEACl) to block potassium channels. The NaCl concentration was 115.5 mM when K⁺ channel blockers were supplemented with the following neurotransmitter blockers: 25 µM D()AP-5, 10 µM bicuculline (methochloride or methobromide salt), 10 µM DNQX, and 10 µM strychnine HCl. Synaptic blockers were prepared as stock solution aliquots (in Pi-free recording ACSF, except DMSO for DNQX) stored at 20°C. Care was taken to protect preparations containing bicuculline and strychnine from light.

A number of different channel blockers or agonists were added to the standard ASCF. FPL 64176 (10 µM prepared in DMSO; Sigma/RBI), nifedipine (10-20 µM prepared in absolute ethanol; Sigma), and TTX (1 µM prepared in Pi-free recording ACSF; Molecular Probes) were applied as stock solution aliquots stored at 20°C. Again, care was taken to protect preparations containing FPL 64176 and nifedipine from light. Agatoxin IVA (Alomone Labs) and -conotoxin GVIA (Bachem), with 7- to 10-fold molar excess cytochrome c to reduce peptide loss due to adherence to container walls, were prepared in dry form by vacuum centrifugation of aqueous solution aliquots and stored at 20°C. -Agatoxin IVA was applied at 480 nM to the ACSF recording solution in the slice recording chamber. Perfusion flow was interrupted for 10 min to allow peptide binding to cells before resumption of flow through the chamber. -Conotoxin GIVA (4 µM) was applied similarly. The toxin concentrations have been found to be saturating in this preparation (Umemiya and Berger 1994).

To study the entire ensemble of whole-cell currents, we used a patch-recording solution containing the following (in mM): 146 KCH₃SO_4, 5 KCl, 2 MgCl_2, 2 EGTA, 10 MOPS, 2 Na₂ATP, and 0.2 Na₃GTP. KOH/HCl were added to bring the pH to 7.2. The osmolality of this solution was 310 mOsm. To study inward currents in isolation, the pipette solution was composed of the following (in mM): 100 CsCl, 20 TEACl, 5 MgCl_2, 2 BAPTA, 10 HEPES, 5 Na₂ATP, 0.5 Na₃GTP, and CsOH/HCl for pH 7.3. Sucrose was added to bring the osmolality to 305 mOsm. Patch solution aliquots were stored at 20°C until time of use. Preparations containing BAPTA were protected from light.

Whole-cell recordings

Whole-cell recordings were obtained from the somata of rat hypoglossal motoneurons under visual control using a Zeiss Axioskop equipped with Nomarski optics for differential interference contrast (DIC) and infrared video recording. The patch electrodes were glass pipettes with tip diameters of 1-2 µm and resistances of 2-4 M when filled with the pipette solution. Electrical recordings were made initially with an Axopatch 200B amplifier and digitized at 10 kHz using an Instrutech A/D board connected to a Macintosh PowerPC. Series resistance was typically 5-15 M, following 40-60% compensation by the Axopatch circuitry. In some of the experiments, we used an Axon Instruments Multiclamp 700A amplifier. Data acquisition and voltage-clamp commands were controlled by custom software routines running in Igor (WaveMetrics).

Following the establishment of whole-cell recording, the membrane potential was clamped at 70 mV. Whole-cell currents were measured in response to slow (typically 14 mV/s), triangular voltage-clamp commands from 70 to 0 mV and back and to a family of 1-s voltage-clamp steps of different amplitudes, also starting at 70 mV. In most of the illustrated current records, the linear leak component has been subtracted either on-line by the amplifier's circuitry or off-line based on scaling the responses to voltage changes within 10 mV of the holding potential. Following a change in the perfusion solution, 10 to 15 minutes were allowed to elapse before obtaining additional recordings.

The effects of channel agonists and antagonists on peak currents are expressed as a percentage of the control currents. Whenever possible the effects of reversible agents are compared with the average of the control recordings obtained before and after their application. The data are presented as means ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results presented below were obtained from 62 hypoglossal motoneurons recorded in brain stem slices taken from rats aged 10 to 24 days (15 ± 4; mean ± SD). In 11 cells, no potassium channel blockers were present in either the internal or external solution. In the remaining 51 cells potassium channel blockers were present in the internal solution and in most cases (40/51) in the external solution as well.

Control I-V relations

In the absence of potassium channel blockers, large net outward membrane currents were recorded. The I-V relations were characterized by a relatively small, linear increase in outward current as the membrane voltage was increased from 70 to about 30 mV, followed by a sharp increase in outward current at higher levels of depolarization. However, most of the cells (8/11) showed a region of decreased slope conductance between about 50 and 40 mV, indicative of the activation of persistent inward currents. Figure 1 illustrates an example of this phenomenon. Figure 1A shows the triangular voltage-clamp command (bottom trace) along with the recorded current (top trace). Figure 1B displays the leak-subtracted current as a function of membrane voltage and illustrates the region of decreased slope conductance (vertical arrows) between 50 and 40 mV. This region was only present on the response to the rising phase of the voltage-clamp command, suggesting slow inactivation of the responsible channels.

View larger version (13K): [in this window] [in a new window]   Fig. 1. Voltage-clamp current recorded in a hypoglossal motoneuron in response to a triangular voltage-clamp command (70 to 0 mV). A: total membrane current (top trace) recorded in response to the voltage command shown in the bottom trace. B: relation between membrane current and voltage command. The linear leak component of the response has been subtracted. Large diagonal arrows indicate the direction of voltage change, and the small vertical arrows indicate the onset (left) and peak (right) of a negative-slope region in the leak-subtracted current-voltage (I-V) relation.

Persistent sodium currents

Net inward currents that typically activated around 40 to 30 mV characterized the I-V relations of hypoglossal motoneurons obtained in the presence of internal and external potassium channel blockers. In most cells, there was also a low-threshold, TTX-sensitive component that was activated around 60 to 50 mV on the ascending phase of the triangular voltage-clamp command but was minimal or absent on the descending phase. Figure 2, A and B, illustrates I-V relations in a cell with a particularly prominent TTX-sensitive component obtained before (thick lines) and after TTX application (thin lines). The records are averages of four responses to the ascending phase of triangular voltage-clamp commands at two different ramp rates: our standard rate of 14 mV/s (A) and a faster rate of 70 mV/s (B). The peak inward current was similar at the two ramp speeds, but the TTX-sensitive component was more prominent at the faster ramp rate (peak TTX-sensitive current = 304 pA at 70 mV/s vs. 159 pA at 14 mV/s). This low-threshold, TTX-sensitive current was present in five of six cells tested with our standard ramp rate and ranged in amplitude from 53 to 159 pA (85 ± 43 pA). In six additional cells, both potassium and calcium channels were blocked (by replacing calcium with manganese). Under these conditions, TTX-sensitive currents were detected in four of six cells (range, 22-83 pA, mean ± SD = 61 ± 29 pA; no significant difference from cells studied without potassium and calcium channel blockade).

View larger version (15K): [in this window] [in a new window]   Fig. 2. Persistent sodium currents in hypoglossal motoneurons in the presence of internal and external potassium channel blockers. Average, leak-subtracted currents during the ascending phase of a triangular voltage command plotted as a function of voltage for voltage ramp rates of 14 mV/s (A) and 70 mV/s (B). Responses obtained before TTX application are shown with thick lines; those after TTX application with thin lines. Arrows indicate the peak magnitude of the TTX-sensitive current, which is larger at the faster ramp rate (304 pA at 70 mV/s vs. 159 pA at 14 mV/s).

Characteristics of persistent calcium currents

The calcium currents recorded following application of both potassium channel blockers and TTX exhibited two qualitatively different patterns of activation in different cells. In most of the hypoglossal motoneurons (19/29) the peak inward current recorded during the ascending phase of the triangular voltage-clamp command was clearly larger than that recorded during the descending phase, and the voltages at which the current turned on and off were similar. Figure 3A illustrates the total current response in one of these cells and Fig. 3B shows the leak-subtracted current (digitally low-pass- filtered at 100 Hz) in the same cell as a function of voltage. The somatic voltage threshold for the inward current as well as the voltage at which the peak inward current occurs were similar for the ascending and descending phases of the response, but the peak current was slightly larger on the ascending phase. For the group of cells showing this type of response, the peak ascending current was on average about twice as large as the peak descending current (422.4 ± 352.6 pA, range: 114.4 to 1721.7 vs. 218.0 ± 117.1 pA, range: 59.7 to 433.0), a difference that was statistically significant (paired t-test = 2.543, P < 0.05).

View larger version (23K): [in this window] [in a new window]   Fig. 3. Different patterns of calcium current activation in hypoglossal motoneurons. Calcium currents recorded in two different motoneurons, one with (D-F) and one without (A-C) evidence of a prominent contribution from dendritic channels. A and D: total currents (top traces) recorded during triangular voltage-clamp commands (bottom traces). Each trace is the average of 4 responses. B and E: leak-subtracted, low-pass-filtered (100 Hz) currents plotted as a function of the command voltage. Arrows indicate direction of voltage change. C and F: leak-subtracted, low-pass-filtered (100 Hz) currents (top traces) recorded in response to a 1-s voltage-clamp step from 70 to 10 mV. Capacitative current transients have been removed for clarity.

Figure 3C shows the filtered, leak-subtracted current recorded in the same motoneuron in response to a 1-s somatic voltage-clamp step from 70 to 10 mV. A large, partially inactivating inward current is evoked during the step, but the current quickly decays to baseline following the offset of the voltage step. A series of voltage clamp steps were applied in 16 of 19 cells exhibiting the type of response to the triangular voltage command shown in Fig. 3, A and B. In 8 of these cells, the membrane current exhibited a rapid return to baseline following the offset of depolarizing voltage steps. In the other 8 cells, small, but persistent (duration > 0.5 s) current tails were observed following the offset of voltage steps that activated inward currents.

About one-third (10/29) of the motoneurons we studied in the presence of potassium channel blockers and TTX exhibited a "clockwise" hysteresis (cf. Carlin et al. 2000b) in the current response to the triangular voltage-clamp command. As shown in Fig. 3, D and E, this type of response is characterized by the activation of inward current at lower levels of depolarization on the descending phase of the response and a peak inward current on the descending phase that is similar in amplitude or sometimes larger than that recorded on the ascending phase. In cells showing this clockwise hysteresis, the average peak current amplitudes were similar on the ascending and descending phases of the response (ascending: 517.7 ± 263.3, range 159.9 to 1067.9; descending: 453.6 ± 209.6, range 187.9 to 777.4). Long tails of inward current following depolarizing voltage steps were present in all of the cells exhibiting clockwise hysteresis, and in many cases these tail currents were quite large (e.g., Fig. 3F).

The pattern of responses illustrated in Fig. 3, D-F are consistent with the idea that dendritic channels that are not under adequate voltage-clamp control make a prominent contribution to the persistent inward current (see INTRODUCTION). Depolarization of the soma membrane should be associated with a delayed and decreased depolarization of the dendrites that leads to delayed activation of dendritic channels. When the soma is repolarized, dendritic channels continue to supply inward current to the soma, resulting in a clockwise hysteresis in the I-V relation and prolonged inward tail currents following the offset of a depolarizing voltage step. The presence of this pattern in some hypoglossal motoneurons but not others may reflect differences in the portion of the dendritic tree remaining intact in the slice, since the holding current, leak conductance, and age of the animals were not significantly different in cells with and without hysteresis.

The amplitude and duration of inward tail currents depended on the amplitude and duration of somatic depolarization. Figure 4A shows tail currents following 1-s steps from 70 mV to four different voltages. A step to 40 mV elicited a net outward current and a small tail current that quickly decayed to baseline. In contrast, a step to 30 mV elicited a large net inward current, followed by a larger tail current that increased over 200 ms following the step offset and reached a steady level that lasted for well over a second. Larger depolarizing voltage steps were followed by slowly decaying tail currents, with a rate of decay that increased with increasing depolarization. Figure 4B illustrates tail currents recorded in another motoneuron in response to voltage steps of the same magnitude (70 to 30 mV) but different duration (0.05 to 1 s). As the duration of somatic depolarization was increased, the amplitude of the tail current increased and its rate of decay decreased. These data are consistent with the delayed development of calcium-mediated persistent depolarizations in the dendrites that act as sources of inward current following the offset of somatic depolarization.

View larger version (15K): [in this window] [in a new window]   Fig. 4. Time and voltage dependence of inward tail currents. A: tail currents recorded following 1-s voltage-clamp steps of different amplitudes. (Only the last 500 ms of the voltage steps are shown.) B: tail currents following 40-mV voltage-clamp steps of different duration.

We often observed other signs consistent with a remote site for the generation of the observed inward currents. For example, in some cells we noted that voltage-steps that were just threshold for producing a net inward current could be associated with a delayed current onset or an inflection point in the rising phase of the inward current, as reported by Carlin et al. (1997b). In other cases, the inward currents could appear in an almost "all-or-none" fashion, i.e., in response to small increments in the amplitude of a voltage-clamp step, the inward current could increase from near zero to near its maximum amplitude. However, we found that these other indications of a distal site for current generation were not nearly as consistent as the appearance of tail currents and I-V hysteresis, so we did not study them systematically.

Relative contribution of different types of channels to persistent calcium currents recorded at the soma

Studies in both turtle and mouse spinal motoneurons have indicated that L-type calcium channels are primarily responsible for sustained inward currents of dendritic origin (Carlin et al. 2000b; Hounsgaard and Kiehn 1993; Hounsgaard and Mintz 1988; Perrier and Hounsgaard 2000). This hypothesis is based on the finding that these currents are abolished following application of the L-type channel antagonists nifedipine and nimodipine. In contrast, we have found that L-type calcium channels make a relatively small contribution to the calcium current recorded at the soma in rat hypoglossal motoneurons and do not appear to be necessary for the calcium-mediated persistent depolarizations in the dendrites of these cells. In 12 of the motoneurons we studied, the calcium currents recorded before, during, and after the addition of 10-20 µM nifedipine to the extracellular solution were compared. Whenever possible, we compared the currents measured in the presence of nifedipine to the average of responses taken before nifedipine application and after washout, to control for the effects of slow, time-dependent changes in the amplitude of inward currents.

Figure 5A illustrates I-V relations obtained in one motoneuron in the presence and absence of 20 µM nifedipine. Each trace is the leak-subtracted average of four responses to the ascending portion of a triangular voltage-clamp command (70 to 0 mV). The peak current recorded during the first application of nifedipine (green trace, 2) is nearly identical to that recorded prior to nifedipine application (black trace, 1). The peak current increases after washing out nifedipine (blue trace, 3) and decreases again during the second nifedipine application (red trace, 4). Similarly, in the other cells tested, nifedipine typically produced either no effect or only a relatively small decrease in the peak inward current recorded at the soma. Nifedipine produced an average decrease in peak current of about 7% (6.7 ± 14.7, range 26.6 to 40, n = 24) compared with bracketing responses without nifedipine.

View larger version (19K): [in this window] [in a new window]   Fig. 5. Effects of L-type calcium channel antagonist (A) and agonist (B) on calcium currents. A: leak-subtracted currents recorded on the ascending phase of a triangular voltage-clamp command in the presence (2: green and 4: red) and absence (1: black and 3: blue) of 20 µM nifedipine. Nifedipine produced either no change in the amplitude of the peak inward current (traces 2 vs. 1) or a slight decrease (traces 4 vs. 3). B: leak-subtracted current-voltage relations obtained before (black) and after (red) application of 4 µM FPL, an L-type channel agonist. FPL had little effect on the peak inward current amplitude but shifted the current activation and deactivation thresholds to more hyperpolarized voltages.

Although these results are consistent with a relatively small contribution from L-type channels to the total calcium current measured at the soma, they do not rule out the presence of L-type calcium channels on the dendrites. To enhance the contribution of L-type channels, we applied the L-type channel activator FPL-64176 (4 µM) in two cells. FPL did not produce a large change in the peak current, but in one of the two cells it lead to a lower activation threshold and a more prominent hysteresis in the I-V relation as shown in Fig. 5B (black trace, control; red trace, FPL). This supports some role for L-type channels in the generation of calcium currents in the dendrites of hypoglossal motoneurons. However, blockade of L-channels does not necessarily abolish hysteresis in the I-V relations or the long tail currents following voltage-clamp steps. Four of 12 cells tested with nifedipine exhibited both long tail currents and hysteresis in their responses to triangular voltage-clamp commands. In two of these cells, nifedipine eliminated hysteresis but not the long tail currents, whereas, in the other two, neither phenomenon was abolished by nifedipine. Figure 6 illustrates the results obtained in the motoneuron with the most prominent I-V hysteresis and tail currents of those tested with nifedipine. Figure 6A shows averaged, leak-subtracted, current responses to the triangular voltage command (lower trace), taken before (black trace, 1) and during nifedipine application (blue trace, 2). Both responses exhibit larger peak inward currents on the descending limb of the voltage ramp command. Figure 6B illustrates the corresponding I-V relations, both of which exhibit clockwise hysteresis. Figure 6C shows tail currents following 1-s commands from 70 to 10 mV (top traces), 0 mV (middle traces), and +10 mV (bottom traces). Long tail currents following the voltage pulses were not abolished by nifedipine, although their peak amplitudes could be either larger or smaller than the corresponding responses before nifedipine, depending on the amplitude of the voltage pulse.

View larger version (27K): [in this window] [in a new window]   Fig. 6. Contrasting effects of nifedipine and -conotoxin on hysteresis and tail currents. A: leak-subtracted currents recorded in the absence of calcium channel blockers (1: black) and in the presence of nifedipine (2: blue) and nifedipine and -conotoxin (3: red). B: same currents plotted as a function of command voltage. C: tail currents recorded following 1-s voltage-clamp steps from 70 mV to the voltage indicated, before nifedipine application (1: black), during nifedipine (2: blue), and during nifedipine and -conotoxin (3: red).

These results suggest that calcium channels other than L-type are likely to make a more prominent contribution to persistent calcium currents recorded at the soma of rat hypoglossal motoneurons. Both the N-type channel blocker, -conotoxin GVIA (Reynolds et al. 1986), and the P-type blocker, -agatoxin IVA (Mintz et al. 1992), produced more prominent reductions in calcium currents measured at the soma than did nifedipine. Conotoxin produced an average decrease in peak current of 41.7 ± 16.5% (range, 16.9 to 60.4, n = 5), whereas agatoxin produced an average decrease of 36.6 ± 22.9% (range, 14.7 to 68.9, n = 4). Both conotoxin and agatoxin were tested in two cells exhibiting I-V hysteresis. Conotoxin was applied after nifedipine in the cell shown in Fig. 6 (red traces, 3). Conotoxin produced a large decrease in calcium current and eliminated both I-V hysteresis and long current tails. The further addition of agatoxin produced an additional decrease in the peak inward current of about 50 pA (data not shown). Figure 7 shows the effects of conotoxin and agatoxin on the other motoneuron exhibiting I-V hysteresis. Prior to application of the toxins (black traces, 1), the peak current was largest on the ascending phase of the voltage-clamp command, but the I-V relation still exhibited clockwise hysteresis, since the current on the descending phase deactivated at a more hyperpolarized voltage than that required for activation (Fig. 7B). Although the inward currents were reduced by both agatoxin (blue traces, 2) and conotoxin (red traces, 3), the I-V relations still exhibited a clockwise hysteresis. Substitution of manganese for calcium in the bathing solution eliminated all of the residual inward currents in almost every case (data not shown).

View larger version (20K): [in this window] [in a new window]   Fig. 7. Effects of -agatoxin and -conotoxin on calcium currents. A: averaged, leak-subtracted current responses (top traces) to triangular voltage-clamp command (bottom trace), in the absence of calcium channel blockers (1: black), in the presence of agatoxin (2: blue), and in the presence of -agatoxin and -conotoxin (3: red). B: same responses plotted as a function of command voltage. In this example, the control I-V relation exhibits a small degree of hysteresis that is not abolished by -agatoxin or -conotoxin.

These results support the contribution of several types of calcium channels to the persistent inward currents presumed to originate from the dendrites. The contribution of N- and P-type channels to the somatically recorded calcium currents appears to be more prominent in hypoglossal motoneurons than that of L-type channels, suggesting either a higher density and/or a more proximal location. The potential functional consequences of the presence of multiple types of dendritic calcium channels will be considered in the DISCUSSION.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The input-output behavior of motoneurons is strongly influenced by the presence of voltage-activated persistent inward currents (see INTRODUCTION) (Binder et al. 1996; Perrier and Hounsgaard 2000; Powers and Binder 2001). The present results show that both sodium and calcium channels generate persistent inward currents in juvenile rat hypoglossal motoneurons. For the inward currents recorded during slow (14 mV/ms) voltage-command ramps, the sodium (TTX-sensitive) component was about 20% of the amplitude of the calcium component but was activated at more negative command voltages (60 to 50 mV). The sodium current was generally only apparent on the ascending phase of a triangular voltage clamp command, which is consistent with its slow inactivation at more depolarized voltages. Nonetheless, since the persistent sodium current is activated in the voltage range between resting potential and spike threshold and may be only partially inactivated during the brief depolarizations occurring during action potentials, it may make an important contribution to the control of repetitive discharge (Lee and Heckman 2001). Due to the slow inactivation of the sodium current, it is difficult to determine whether the responsible channels are all on or close to the soma or located more distally as well.

The calcium currents recorded during triangular voltage-clamp commands typically exhibited larger peak currents during the ascending phase of the command, suggesting partial inactivation. However, in about one-third of the cells, the calcium current deactivated at lower command voltages on the descending phase than its activation threshold on the ascending phase of the command, and in many of these cases the peak inward currents were larger on the descending phase. This clockwise hysteresis in the I-V relation was typically associated with long inward tail currents following depolarizing voltage steps. Both of these phenomena could arise from sustained depolarizations in regions of the dendrites beyond voltage-clamp control (cf. Booth et al. 1997; Carlin et al. 2000b; Lee and Heckman 1998b).

The I-V hysteresis (>30 mV) we observed in the hypoglossal motoneurons was substantially greater than that seen in cat (Lee and Heckman 1998b) or mouse (Carlin et al. 2000b) spinal motoneurons, as were the relative magnitudes of the sustained tail currents (>1 nA). We now have a number of preliminary observations that support the hypothesis that substantial portions of these currents are generated by current flowing through voltage-sensitive calcium conductances in the dendrites. We have observed large increases in intracellular calcium at both proximal and distal dendritic locations during somatic voltage-clamp ramps and steps using multiphoton fluorescence microscopy (D. Margolis, C. Davenport, R. Powers, M. Binder, and P. Detwiler, unpublished data). We do not think that the tail currents are produced by a Ca²⁺-activated nonselective cation current because they were insensitive to reducing the sodium concentration in the bathing solution (e.g., Perrier and Hounsgaard 1999). Moreover, we do not think that the tail currents are mediated by a Ca²⁺-activated chloride channel (e.g., Ward and Kenyon 2000) because they persisted when 2 mM Ba²⁺ was substituted for Ca²⁺ in the bathing solution and when the Cl equilibrium potential was matched to the resting potential (70 mV) by reducing the Cl in the patch solution to 9 mM.

An alternative explanation for many of the features of calcium currents that we observed is depolarization-induced facilitation of calcium channels (Svirskis and Hounsgaard 1997). For example, depolarization during the rising phase of a voltage-clamp ramp could act to convert L-type channels to a "willing" state (Bean 1989; Delgado-Lezama and Hounsgaard 1999) in which they can open at lower levels of depolarization on the descending phase of the ramp. Although depolarization-induced facilitation of L-type channels can be produced by relatively modest levels of depolarization (Svirskis and Hounsgaard 1997), it typically requires larger depolarizations than those used in the present study (Kammermeir and Jones 1998; Song and Surmeir 1996). Moreover, we found that the L-channel blocker nifedipine did not block the long tail currents following depolarizing current steps nor did it always eliminate hysteresis in the I-V relation (Fig. 6). Depolarization-induced faciliation of N- and P-type channels has also been reported, but this too requires high levels of depolarization (e.g., Herlitze et al. 2001).

The relative contributions of different channel types to the calcium current recorded in the somata of juvenile rat hypoglossal motoneurons is similar to that reported for neonatal rat hypoglossal motoneurons (Umemiya and Berger 1994), i.e., the majority of the persistent current measured at the soma is carried by N- and P- type channels, with only a small contribution from L-type channels. Since the current recorded during somatic voltage clamp is likely to be derived from both somatic and dendritic channels, this suggests that N- and P-type channels may also be the primary carriers of dendritic calcium currents or that they are located more proximally than L-type channels. The block of I-V hysteresis and tail currents by conotoxin (Fig. 6) provides direct evidence for the contribution of N-type channels to dendritic calcium currents.

At present, one can only speculate as to the functional roles that the different types of calcium channels on the dendrites of motoneurons might play. Although it has been previously demonstrated that the dendrites of turtle motoneurons are endowed with at least two types of calcium channels (Hounsgaard and Kiehn 1993), most experimental and theoretical studies have concentrated on the role of L-type calcium channels (e.g., Booth et al. 1997, Carlin et al. 2000b). In particular, the dendritic L-channels are thought to underlie bistable discharge behavior and plateau potentials (for review see Perrier and Hounsgaard 2000). In the case of rat hypoglossal motoneurons, there is no evidence for bistable discharge behavior, although their properties are affected by a number of neuromodulators (Bayliss et al. 1997). However, in the presence of potassium channel blockade, we found persistent inward currents that could exceed 1 nA, an input sufficient to generate changes in repetitive firing >20 impulses/s (cf. Sawczuk et al. 1995). Thus it is conceivable that the persistent inward currents on the dendrites of motoneurons might act primarily to augment synaptic currents en route to the soma (Binder 2002; Lee and Heckman 2000; Powers and Binder 2000, 2001; Prather et al. 2001). In support of this hypothesis, we have recently found that synaptic currents evoked by iontophoretic application of glutamate to the dendrites of hypoglossal motoneurons are enhanced when the cell is depolarized to activate dendritic calcium channels (Binder et al. 2002). Having different types of calcium channels on the dendrites of motoneurons, with a variety of activation thresholds, may extend the range of membrane potentials over which persistent inward currents can augment synaptic inputs (cf. Bernander et al. 1994). The next decade should witness the full elaboration of the role that persistent inward currents play in synaptic integration and shaping the input-output functions of motoneurons.