The Journal of Neurophysiology Vol. 78 No. 1 July 1997, pp. 177-186
Copyright ©1997 by the American Physiological Society

Hyperpolarization-Activated Currents in the Growth Cone and Soma of Neonatal Rat Dorsal Root Ganglion Neurons in Culture

Z. Wang, R. J. Van Den Berg, and D. L. Ypey

Department of Physiology, Leiden University, 2300 RC Leiden, The Netherlands

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Wang, Z., R. J. Van Den Berg, and D. L. Ypey. Hyperpolarization-activated currents in the growth cone and soma of neonatal rat dorsal root ganglion neurons in culture. J. Neurophysiol. 78: 177-186, 1997. Dissociated dorsal root ganglion neuron growth cones and somata from neonatal rats were voltage and current clamped with the use of the perforated-patch whole cell configuration to study the occurrence and properties of slow hyperpolarization-activated currents (I_h) at both regions. Under voltage-clamp conditions I_h, blockable by 2 mM extracellular CsCl, was present in 33% of the growth cones tested. Its steady-state activation as a function of voltage could be fitted with a single Boltzmann function with a midpoint potential of 97 mV. The time course of current activation could be best described by a double-exponential function. The magnitude of the fully activated conductance was 3.5 nS and the reversal potential amounted to 29 mV. At the soma, I_h was found in 80% of the somata tested, which is much higher than occurrence at the growth cone. The steady-state activation curve of I_h at the soma, fitted with a single Boltzmann function, had a midpoint potential of 92 mV, which was more positive than that in the growth cone. The double-exponential activation of the current was faster than in the growth cone. The fully activated conductance of 5.1 nS and the reversal potential of 27 mV were not significantly different from the values obtained at the growth cone. Membrane hyperpolarization by current-clamp pulses elicited depolarizing sags in 30% and 78% of the tested growth cones and somata, respectively, which is in agreement with our voltage-clamp findings. Termination of the hyperpolarizing current pulse evoked a transient membrane depolarization or an action potential at both sites. Application of 2 mM extracellular CsCl hyperpolarized the membrane potential reversibly by ~5 mV and blocked the depolarizing sags and action potentials following the current injections at these regions. Thus I_h contributes to the resting membrane potential and modulates the excitability of both the growth cone and the soma. Intracellular perfusion with the second messenger adenosine 3,5-cyclic monophosphate (cAMP) was only possible at the soma by the use of the conventional whole cell configuration. Addition of 100 µM cAMP to the pipette solution shifted the midpoint potential of the I_h activation curve from 108 to 78 mV. The current activation time course was also accelerated. The reversal potential and the fully activated conductance underlying I_h were not changed by cAMP. These results imply that cAMP primarily affects the gating kinetics of I_h. Our results show for the first time quantitative differences in I_h properties and occurrence at the growth cone and soma membrane. These differences may reflect differences in intracellular cAMP concentration and in the expression of I_h.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Neuron somata are able to grow new neurites in culture after mechanical dissociation from the intact tissue. The mechanism of nerve regeneration and neurite outgrowth is poorly understood. Neurite formation and outgrowth can be manipulated by both blockers and openers of voltage-dependent Na⁺, K⁺, and Ca²⁺ channels (Berdan and Easaw 1992). In neuroblastoma cells it has been found that the inwardly rectifying K⁺ current I_Ki is essential for neuron growth, because blockage of this current completely and reversibly suppressed neurite outgrowth (Arcangeli et al. 1994). Thus evidence is growing that ion channels are functionally involved in the process of neurite outgrowth. In addition, second messengers, like adenosine 3,5-cyclic monophosphate (cAMP), influence neuron regeneration probably through actions on ion channels (Lankford and Letourneau 1991). It is, therefore, of interest to study the role of cAMP-dependent ionic conductances in neuron regeneration. For example, the possible involvement of the cAMP-modulated and hyperpolarization-activated current I_h in neurite outgrowth has not yet been studied.

The presence of I_h has been described in different neuronal somata (Destexhe et al. 1993; Halliwell and Adams 1982; Maccaferri et al. 1993; McCormick and Pape 1990; Solomon and Nerbonne 1993; Tokimasa and Akasu 1990), including dorsal root ganglion (DRG) neurons of the mouse (Mayer and Westbrook 1983). From these studies it appeared that both the resting membrane potential and excitability are influenced by I_h. Because we have shown that at different regions of regenerating neurons, the membrane properties may vary considerably (Wang et al. 1994a), we wondered whether I_h might be differentially expressed in the neuron soma and growth cone. To gain insight into this, we have investigated the presence of I_h in isolated growth cones and in neuron somata without processes. Up to now, the occurrence of I_h and its properties in growth cones have not been studied yet. Because the modulation of I_h by cAMP in rat DRG neurons has not been examined either, we also investigated the effect of cAMP on I_h in the neuron soma.

Our results show that I_h was less frequently found in the growth cone than in the soma. On the other hand, an inwardly rectifying current, similar to I_Ki (Arcangeli et al. 1994), occurred more frequently in the growth cone than in the soma. The time course of activation of I_h in the growth cone was slower than that in the soma, probably related to modulating actions by intracellular cAMP. Some of these results have been published in abstract form (Wang et al. 1994b).

	METHODS

Abstract Introduction Methods Results Discussion References

DRG neurons were mechanically dissociated from 1-day-old neonatal Wistar rats and cultured on poly-D-lysine-coated (Sigma, St. Louis, MO) coverslips in Dulbecco's modified Eagle's medium enriched with 10% fetal calf serum (Wang et al. 1994a) or F14 medium (Imperial Lab) with 10% horse serum (Gibco) (Wood et al. 1988).

Both perforated-patch (Horn and Marty 1988) and conventional whole cell patch-clamp configurations (Hamill et al. 1981) were used. The first technique allowed us to study I_h at undisturbed intracellular conditions, whereas the second technique permitted us to vary intracellular concentrations of, e.g., cAMP. Growth cones with neurite stumps 100 µm were separated from 1-day (24-36 h)-old cells. This was achieved by placing a patch pipette on the neurite at a site 50-100 µm from the growth cone and aspirating the neurite (cf. Mayers 1993). Another technique was to destroy the soma in the whole cell mode by applying high pressure through the pipette. Somata studied were either isolated freshly (4-5 h) or after 1 day in culture, when they had no processes. Recordings were accepted only from growth cones and somata that had a steady resting membrane potential for >20 min and could fire action potentials on stimulation. A series of depolarizing current pulses, consisting of 15 pulses of 200 ms in steps of 20 pA, was applied to determine the firing threshold. Hyperpolarizing current injections, consisting of 15 current pulses of 200 ms to 1 s in steps of 20 or 50 pA, were used to measure membrane resistance and to record membrane potential responses. Under voltage-clamp conditions the holding potential was set at 70 or 80 mV. Hyperpolarizing pulses from 65 to 145 mV and lasting 1-3 s were applied to activate I_h. Small patch pipettes with a resistance of ~10-20 M were used to patch the growth cone and large patch pipettes with a resistance of ~2 M were used to patch the soma. The series resistances were compensated for 50-80%. Because usually the magnitude of I_h was <1 nA, the residual series resistance could give rise to small voltage errors only (<5 mV). A reasonable control of voltage of the growth cone extension can be expected, because the space constant of our neurites is estimated to be ~1 mm (Wang et al. 1994a).

Patch pipettes were filled with an intracellular-like solution (ICS) composed of (in mM) 140 KCl, 10 NaCl, 1.0 CaCl₂, 2.0 MgCl₂, 10 ethylene glycol-bis(-aminoethyl ether)-N,N,N,N-tetraacetic acid, and 10 N-2-hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES)/KOH, pH 7.2. To achieve the perforated-patch configuration, pipettes were filled with ICS with addition of 250 µg/ml of the antibiotic nystatin (Sigma). The effect of cAMP was studied by adding 100 µM cAMP and 5 mM ATP (Sigma) to the ICS, or the membrane-permeable dibutyryl-cAMP (1 mM) was used in an extracellular-like solution (ECS) containing (in mM) 140 NaCl, 5.0 KCl, 1.0 CaCl₂, 1.0 MgCl₂, 6.0 glucose, and 10 HEPES/NaOH, pH 7.2. To minimize the breakdown of dibutyryl-cAMP, the solution containing the drug was freshly made and light exposure was minimized by covering the tube with aluminium foil. Neurons were incubated with this drug for 10-20 min. Glass coverslips with attached DRG neurons were mounted in an open-bottom Teflon culture dish, which was placed on the stage of an inverted microscope (Zeiss). To block I_h and to measure the leakage current, 2 mM CsCl was added in the ECS (Cs-ECS). The inward rectifier I_Ki was blocked by 2 mM BaCl₂ (Rudy 1988).

For stimulus protocols and data acquisition, the software package pClamp (version 5.6 and 6.01, Axon Instruments, Burlingame, CA) was used. Recordings were made with the use of an EPC-7 amplifier (List Electronic, Darmstadt, Germany) and stored after A/D conversion (TL-100). The data were analyzed with the software packages pClamp and FigP (version 6.0, Biosoft, Cambridge, UK) and the results are presented as means ± SE n, with n being the number of cells investigated. The Student's t-test (Asystant, version 1.10) and ² test on 2 × 2 table, with Yates correction (SPSS/PC+, version 5.0), were used to compare values and percentages, respectively, obtained under different conditions. The F test for the residual sum of squares (RSS) described by Ljung (1987) was used to compare the different mathematical models, fitted to our experimental data. The F ratio, given by [(RSS_j  RSS_k)·df_k]/[RSS_k·(df_j  df_k)], where df = N  p, with N being the number of data points and p the number of free parameters, and the subscripts j and k representing the simple and more complex models, respectively, was compared with a table of critical values for the F distribution with (df_j  df_k), df_k degrees of freedom. The asymptotic information criterion for the RSS was also used for model discrimination. The significance level was chosen as P = 0.05.

	RESULTS

Abstract Introduction Methods Results Discussion References

Perforated-patch current-clamp measurements on the isolated growth cone

Patching growth cones was difficult because their lamellipodia were flat (<2 µm) and small (<20 µm), resulting in a success rate of <10%. Accepted growth cones exhibited resting membrane potentials of 57 ± 1 mV and membrane resistances of 921 ± 80 M (n = 16) as measured close (10 mV) to the resting potential. Injection of depolarizing current pulses evoked action potentials in all growth cones tested (n = 16) and 63% of them fired repetitively. The firing threshold was equal to 20 ± 2 mV, the action potential duration at half-height (half-width) was 10 ± 1 ms, and the overshoot was 33 ± 3 mV (n = 16). On small (20-pA) hyperpolarizing current pulses, the voltage transients [V(t)] were fitted with a two-exponential function (Wang et al. 1994a; cf. Rall 1969)

(1)

where V(t) is the membrane potential transient (10 mV) due to the current step, C₀ and C₁ are constants, _m is the passive membrane time constant of the growth cone, and ₁ is the first equalizing time constant due to the charging of the axonal compartment. However, curve fitting showed that our growth cone responses were best described by a single-exponential function, indicating a negligible contribution of the neurite stump. The value of _m was found to be 32 ± 3 ms (n = 16). The membrane capacitance, calculated from _m/R_m, where R_m is the steady-state membrane resistance, amounted to 35 ± 4 pF (n = 16).

On larger (>100-pA) hyperpolarizing current injections, the membrane potential exponentially relaxed toward a new steady-state level (not shown) in 11 of 16 growth cones tested. However, in five growth cones (31%) the initial membrane hyperpolarization was followed by a slow depolarizing sag of the membrane potential to a less hyperpolarized level (Fig. 1A). Interestingly, just after the applied current pulses, in particular in the smaller cones, the membrane potential first decayed toward a depolarized subthreshold level and then back to the resting membrane potential (Fig. 1A). In response to larger hyperpolarizing current injections, the amplitude of the membrane afterdepolarizations increased from subthreshold to suprathreshold levels, resulting in action potentials (Fig. 1A). The appearance of the depolarizing sag suggested the presence of a slow hyperpolarization-activated conductance. Indeed, extracellular application of 2 mM CsCl reversibly blocked the depolarizing sags and also the action potentials following the hyperpolarizing current injections (Fig. 1, B and C). In the presence of Cs⁺ the membrane resistance was significantly increased from 754 ± 67 M to 1,034 ± 89 M (P = 0.02, n = 5) in a reversible manner and the resting membrane potential was reversibly hyperpolarized by ~7 ± 2 mV (P = 0.01). This indicates that Cs⁺ blocked a conductance, associated with a reversal potential that is positive with respect to the resting potential and not an inward rectifier type of K⁺ conductance. Apparantly, the blocked conductance underlies the hyperpolarization-activated current I_h.

View larger version (9K): [in this window] [in a new window]   FIG. 1. Perforated-patch current-clamp recordings from an isolated growth cone of a rat dorsal root ganglion (DRG) neuron with a resting membrane potential of 58 mV. Inset: applied current protocol. Increment in current was 50 pA (B) or 150 pA (A and C). A: under control conditions, membrane hyperpolarization evoked membrane depolarizing sags in 5 of 16 growth cones tested, indicating activation of hyperpolarization-activated current Ih. Following current injections, membrane depolarizations and action potentials were observed (). B: extracellular application of 2 mM CsCl blocked depolarizing sags and action potentials following current injections. Because CsCl hyperpolarized the membrane potential by 7 mV, a depolarizing holding current was injected to bring membrane potential close to original resting membrane potential. C: washout recordings, obtained 5 min after replacement of extracellular-like solution containing 2 mM CsCl (Cs-ECS) with control extracellular-like solution (ECS).

In conclusion, two types of membrane responses were found in the growth cone on hyperpolarizing current injections: exponential relaxations (69%) and hyperpolarizations followed by depolarizing sags (31%). Thus about one-third of the investigated growth cones possessed a hyperpolarization-activated conductance, contributing to the resting membrane potential. Blockage of action potentials by Cs⁺ strongly suggested that the hyperpolarization-activated conductance is able to modulate the excitability of these growth cones.

Perforated-patch voltage-clamp measurements on the isolated growth cone

In response to a series of hyperpolarizing voltage pulses in the range from 80 to 140 mV, three distinct types of current responses could be observed in individual growth cones (Fig. 2). First was a slowly activating current, which did not show inactivation (Fig. 2A), probably identical to the hyperpolarization-activated current I_h (Maccaferri et al. 1993; Mayer and Westbrook 1983; Solomon and Nerbonne 1993). This current was found in 33% of the growth cones tested (n = 18). Second, a fast-activating and subsequently partly inactivating current (Fig. 2B), similar to the inward rectifier K⁺ current I_Ki (cf. Arcangeli et al. 1994), was present in 39% of the growth cones. Finally, the remaining 28% of the growth cones exhibited a leakage or small background current (Fig. 2C). The properties of the hyperpolarization-activated currents will be further examined in this paper.

View larger version (24K): [in this window] [in a new window]   FIG. 2. Voltage-clamp recordings of hyperpolarization-activated currents from isolated growth cones in perforated-patch configuration. A-C were obtained from different growth cones; applied voltage protocol is shown in C. Three types of currents were observed in growth cones on hyperpolarizing voltage steps. A: slowly activating Ih was observed in 33% of growth cones tested. B: fast-activating current similar to the inwardly rectifying K+ current IKi was observed in 39% of growth cones tested. C: remaining 28% of growth cones exhibited leakage or small background currents. At the most negative steps there were probably dielectric breakdowns of membrane. D: plot of mean amplitude of Ih as a function of time since start of perforated patch.

First, the stability of I_h during 20 min was tested by applying a hyperpolarizing pulse from 70 to 120 mV every 2 min. The amplitude of I_h is plotted as a function of time in Fig. 2D. Linear regression of pooled amplitudes on time yielded regression coefficients of 0.6 ± 0.7 pA/min(n = 5), which is not significantly different from zero. Thus under our experimental conditions no significant current rundown occurred during the observation period. The leak currents (I_L) were measured in the presence of Cs-ECS as a function of voltage V. The I_L-V relation could be fitted by a straight line according to I_L = g_L(V  V_L), where g_L is the leakage conductance and V_L is the reversal potential of the leakage current with g_L = 5 ± 2 nS and V_L = 41 ± 4 mV (n = 6), respectively. Steady-state activation of the hyperpolarization-activated current as a function of membrane voltage was determined from the peak amplitude of the inward tail current at 70 mV following a series of hyperpolarizing prepulses (V_pp) ranging from 70 to 140 mV (Fig. 3A). To remove capacitive and leakage currents from the instantaneous component of the tail current, we either subtracted the currents in Cs-ECS from the currents in ECS or we corrected for these currents digitally, yielding the Cs⁺-sensitive tail current I_t. I_t was normalized with respect to the maximal tail current I_t,max, obtained after the most negative prepulses (130 to 145 mV). The relationship between the normalized tail currents (I_t/I_t,max) and V_pp could be described by a Boltzmann equation (Fig. 3B)

(2)

where V_0.5 is the midpoint potential at which the conductance (g_h) underlying the hyperpolarization-activated current I_h is half-activated and k is the slope factor. Curve fitting according to Eq. 2 yielded the parameters V_0.5 and k, which amounted to 97 ± 3 mV and 12 ± 3 mV (n = 6), respectively.

View larger version (18K): [in this window] [in a new window]   FIG. 3. Properties of Ih at isolated growth cone and soma under conditions of perforated-patch configuration. A: growth cone tail currents at 70 mV on a series of hyperpolarizing prepulses from 70 to 140 mV. Instantaneous tail current It reflects degree of activation of Ih at the various potentials. Normalized tail currents It/It,max were used to obtain activation curves. Inset: applied voltage protocol. B: activation curves of Ih measured at growth cone () and soma (). Curves: fitted Boltzmann distributions (Eq. 2), yielding V0.5 = 97 mV at growth cone and V0.5 = 88 mV at soma. A clear difference between activation curves existed at these 2 sites. C: activation of Ih at soma on membrane hyperpolarization. This type of current was found in 81% of somata tested. Inset: applied voltage protocol. Note time scale difference between C and Fig. 2A. D: growth cone current responses on voltage steps from 120 to 70 mV after prehyperpolarization to 140 mV, recorded for measurement of It. Leakage current was digitally subtracted. Inset: voltage protocol. E: determination of fully activated It-V relation and Vrev of Ih at growth cone and soma. : It at growth cone. : It at soma. : fits according to Eq. 3, yielding gh,max = 3.5 nS and Vrev = 29 mV at growth cone, and gh,max = 5.1 nS and Vrev = 27 mV at soma. Values of gh,max and Vrev at these 2 sites were not significantly different from each other. Conductance densities, obtained by dividing gh,max by membrane capacitance, were 0.13 and 0.15 nS/pF, respectively, at growth cone and soma, values that were not significantly different.

To measure the fully activated amplitude and the reversal potential (V_rev) of I_h, a voltage protocol, shown in Fig. 3D, was applied after I_h was activated at 140 mV for 1 s. The tail currents were recorded at the moment when the membrane potential was stepped to test voltages (V) ranging from 120 to 70 mV. This voltage range was chosen to avoid the activation of depolarization-activated conductances. The value of the instantaneous component of the tail current I_t at a test potential is the fully activated I_h at that potential. The I_t-V relation could be described by a straight line according to

(3)

where g_h,max is the maximal conductance underlying I_h, attained at fully activated current levels (Fig. 3E). Fitting Eq. 3 to the data points yielded g_h,max = 3.5 ± 0.7 nS andV_rev = 29 ± 4 mV (n = 5).

The activation time course of I_h was clearly voltage dependent. Either one- or multiexponential functions were fitted to the measured currents to decide which function yielded the best fit. A two-exponential function fitted better than a single-exponential one in every case (F test for RSS: P  0.02, cf. Fig. 4, A and B). A three-exponential function resulted in poor fits (F test for RSS: P > 0.06). Thus the current traces were best fitted with a double-exponential function, given by

(4)

where I₀ is the instantaneous value of I_h, and I_h,f and I_h,s are the steady-state levels of the fast and slow components, respectively, with _f and _s as activation time constants. In Fig. 4C it is shown that both _f and _s were dependent on membrane voltage V, with _f = 221 ± 43 ms and _s =952 ± 98 ms at 100 mV, and _f = 88 ± 15 ms and _s = 428 ± 74 ms at 130 mV (n = 6). The magnitudes of _f and _s were significantly decreased (P  0.03) at larger hyperpolarizing voltages. The amplitude of the fast component I_h,f was, as expected, voltage dependent, with a value of 249 ± 63 pA at 100 mV and with an increased value of 466 ± 106 pA at 130 mV (P = 0.01, n = 6). Contrary to our expectation, the amplitude of the slow component I_h,s was not voltage dependent, with a value of about 173 ± 40 pA (n = 6) in the voltage range of 130 to 100 mV (Fig. 4D). This unusual finding has also been reported for rat auditory brain stem cells with the use of the same kinetic model (Banks et al. 1993).

View larger version (22K): [in this window] [in a new window]   FIG. 4. Kinetics of Ih in growth cone and soma in perforated-patch configuration. A: examples of growth cone Ih responses to 4 potential steps from a holding potential of 70 mV (10-mV increments). : 1/16 of sample points. A single-exponential function was fitted to the data points and is represented by the solid line. Asymptotic information criterion (AIC) for residual sum of squares (RSS), given by Nln RSS + 2p, with N being number of data points and p being number of free parameters (Ljung 1987), amounted to 220 (3 free parameters). B: currents from A were fitted with a 2-exponential function according to Eq. 4. The 2-exponential function fit is significantly (F test, P  0.02) better than the 1-exponential fit. AIC amounted to 710 (5 free parameters). Thus the best fit, corresponding to lowest value of AIC, was obtained by Eq. 4, in agreement with the F test. C: activation time constants f (, ) and s (, ) are voltage dependent at the growth cone (, ) and soma (, ). On membrane hyperpolarization both f and s decreased at these 2 sites. At growth cone f and s were significantly (P = 0.02) larger than those at soma. D: amplitudes of 2 components, Ih,f (, ) and Ih,s (, ) at growth cone (, ) and soma (, ) as a function of membrane potential. Magnitude of Ih,f is voltage dependent, whereas that of Ih,s is not voltage dependent in voltage range tested.

In summary, we demonstrated for the first time that I_h was present in 33% of the investigated growth cones. This percentage is close to that found under current clamp (31%), for neurons exhibiting depolarizing sags in response to hyperpolarizing currents. The properties of I_h are comparable with those in many other neuronal somata (Banks et al. 1993; Maccaferri et al. 1993; Mayer and Westbrook 1983; McCormick and Pape 1990; Solomon and Nerbonne 1993).

Perforated-patch current-clamp measurements at the soma

Under control conditions, the resting membrane potential and membrane resistance of somata were found to be60 ± 5 mV and 645 ± 56 M (n = 14), respectively. Action potentials could be evoked in all the somata investigated (n = 14) on suprathreshold current injections, and 21% (n = 3) of the somata fired repetitively. The firing threshold was 28 ± 5 mV, half-width was 4 ± 1 ms, and overshoot was 40 ± 8 mV (n = 14). On small (20-pA) hyperpolarizing current pulses, the voltage transients (10 mV) were fitted with Eq. 1. Here too, curve fitting showed that a single-exponential function fitted the traces best (cf. Wang et al. 1994a). The values of _m and membrane capacitance (_m/R_m) were found to be 26 ± 3 ms and 40 ± 8 pF (n = 14), respectively.

On larger (>100-pA) hyperpolarizing current injections, the membrane potential exponentially relaxed toward a new steady-state level in 3 of 14 somata tested (not shown). However, in the remaining 11 somata (79%) slow depolarizing sags of membrane potential were found to be present. On termination of the first relatively small current pulses, the membrane potential first increased toward a depolarized level and then back to the resting membrane potential. On larger hyperpolarizing current injections, the amplitude of these membrane depolarizations was increased from subthreshold to suprathreshold levels, resulting in action potentials similar to those in the growth cones, as shown in Fig. 1A. The appearance of the depolarizing sag was consistent with the presence of a slow hyperpolarization-activated conductance. Application of extracellular 2 mM CsCl reversibly blocked the depolarizing sags and action potentials following the current injections (cf. Fig. 1). Furthermore, the membrane potential was reversibly hyperpolarized by ~5 ± 3 mV (P = 0.02) and the membrane resistance was increased to 861 ± 89 M (P = 0.03). These findings indicated that Cs⁺ blocked a conductance in series with a reversal potential that is positive with respect to the resting potential. Thus the blocked conductance is not of the inward rectifier type.

In conclusion, responses on hyperpolarizing currents were either exponential relaxations (21%) or hyperpolarizations followed by depolarizing sags (79%). Thus a much larger fraction of the somata exhibited a hyperpolarization-activated conductance as compared with that of the growth cones (31%). This conductance clearly influences the soma resting membrane potential and excitability.

I_h properties at the soma in the perforated-patch configuration

In voltage-clamp mode, I_h was present in the voltage range from 65 to 145 mV (Fig. 3C) in 81% of the somata investigated (n = 26). Extracellular application of 2 mM CsCl blocked I_h in a reversible manner. Another hyperpolarization-activated current, similar to I_Ki (cf. Arcangeli et al. 1994), was found in 4% of the somata (1 of 26), whereas the remaining 15% of the somata exhibited a leakage current only (cf. Fig. 2, A-C).

The steady-state activation curve of I_h was obtained in a similar way to that of the growth cone. The relationship between the normalized instantaneous component of the tail current (I_t/I_t,max) and V_pp was fitted by Eq. 2 (Fig. 3B), yielding V_0.5 = 88 ± 2 mV and k = 11 ± 1 mV (n = 11). The relationship between the fully activated current I_t and V could be fitted according to Eq. 3, yielding g_h,max =5.1 ± 1 nS and V_rev = 27 ± 3 mV (n = 7, Fig. 3E).

The time courses of activation of I_h at the soma could be best fitted by a double-exponential function (Eq. 4), yielding _f and _s as a function of V (Fig. 4C). At a membrane voltage of 100 mV, _f and _s amounted to 158 ± 27 ms and 660 ± 60 ms, respectively, whereas at 130 mV, the respective time constants were 62 ± 16 ms and 223 ± 54 ms (n = 7). Both _f and _s were decreased (P  0.01) on the larger membrane hyperpolarization. In Fig. 4D the relationship between I_h,f, I_h,s and V is depicted. The amplitude of I_h,f increased from 453 ± 52 pA at 100 mV to614 ± 47 pA at 130 mV (P = 0.01, n = 7). On the contrary, the magnitude of I_h,s remained at ~221 ± 43 pA (n = 7) through the voltage range tested (Fig. 4C).

In summary, as compared with the growth cone, the incidence of I_h at the soma (80%) was higher (P = 0.001) and the midpoint potential was more positive (P = 0.02). The activation time constants (_f and _s) obtained at the soma were faster (P = 0.02) than those at the growth cone. However, the reversal potential and the fully activated current at both regions were not significantly different. We conclude that at growing neurite tips the incidence and gating of I_h were different from those measured at neuron somata.

Modulation of I_h by cAMP at the soma in the conventional whole cell configuration

To study the effect of intracellular cAMP on the properties of I_h, we applied the conventional whole cell voltage clamp to the soma perfused with high (100 µM) and low (0 µM) cAMP concentrations, a measurement mode impossible to realize for the growth cone because of the very small patch pipette tips. We compared the activation curves of I_h, the fully activated current amplitudes, reversal potentials, and kinetics at both concentrations.

Activation curves of I_h in the absence and the presence of 100 µM cAMP are shown in Fig. 5A. When Eq. 2 was fitted to the normalized tail currents (I_t/I_t,max) the midpoint potential was found to be 108 ± 7 mV (n = 12) and78 ± 4 mV (n = 8), respectively, with 0 and 100 µM cAMP in the pipette. These values were statistically different (P = 0.001). Thus 100 µM cAMP shifted the I_h activation curve 30 mV toward more depolarized voltages as compared with the absence of cAMP. The slope factors at 100 and 0 µM cAMP were 11 ± 2 mV (n = 8) and 11 ± 1 mV (n = 12), respectively, values that were not significantly different.

View larger version (23K): [in this window] [in a new window]   FIG. 5. Effects of adenosine 3,5-cyclic monophosphate (cAMP) on soma Ih in conventional whole cell configuration. A: activation curves of Ih with 0 () and 100 µM () cAMP in the pipette, respectively. Curves: fitted Boltzmann distributions (Eq. 2), yielding V0.5 = 108 mV at 0 µM cAMP and V0.5 = 78 mV at 100 µM cAMP. B: fully activated It-V relation in absence () and presence () of 100 µM cAMP in the pipette. : fits according to Eq. 3, yielding gh,max = 3.6 nS and Vrev = 27 mV at 0 µM cAMP and gh,max = 4.3 nS and Vrev = 26 mV at 100 µM cAMP. Results imply that fully activated conductance underlying Ih and value of Vrev are not affected by intracellular cAMP. C: time constants f (, ) and s (, ) with 0 (n = 7, , ) and 100 µM (n = 6, , ) cAMP in pipette. Both f and s are voltage dependent. At 0 µM cAMP, magnitudes of f and s are significantly (P  0.01) larger than those at 100 µM cAMP. D: amplitudes of Ih,f (, ) and Ih,s (, ) in absence (, ) and presence (, ) of 100 µM cAMP in pipette. Magnitude of Ih,f is voltage dependent, contrary to that of Ih,s, in voltage range tested.

The relationship between the fully activated current amplitude I_t and V was fitted by Eq. 3, yielding g_h,max = 3.6 ± 0.6 nS and V_rev = 27 ± 2 mV (n = 8) in the absence and g_h,max = 4.3 ± 1 nS and V_rev = 26 ± 2 mV (n = 6) in the presence of 100 µM cAMP. These values were not significantly different (Fig. 5B). Thus intracellular cAMP did not change the fully activated conductance and the selectivity of the ion channels, underlying I_h.

At both cAMP concentrations, the current recordings in response to hyperpolarizing voltage steps could be best fitted with the use of a double-exponential function according to Eq. 4. In Fig. 5C it is shown that at both cAMP concentrations the activation time constants _f and _s were dependent on membrane voltage V. The magnitudes of _f and _s at 100 µM cAMP were significantly smaller (P  0.01) than those at 0 µM cAMP at all voltages tested. At both cAMP concentrations, the magnitudes of I_h,f increased with voltage in the range from 100 to 130 mV, whereas the magnitudes of I_h,s remained approximally constant (Fig. 5D).

Because the steady-state amplitude of I_h varied considerably from cell to cell, the membrane-permeable dibutyryl-cAMP (1 mM) was applied extracellularly to investigate the effect of cAMP on the amplitude of I_h from the same cell in conventional whole cell configuration with 0 µM cAMP in the pipette. Figure 6A shows that at a voltage close to V_0.5 of 110 mV, the amplitude of I_h was significantly increased (P = 0.005) from 176 ± 12 pA to 236 ± 33 pA (n = 4). However, at the voltage of 130 mV, the current amplitudes before and after addition of dibutyryl-cAMP were 294 ± 56 and 283 ± 77 pA (n = 4), respectively, not significantly different. This suggested a shift of the activation curve. Indeed, the midpoint potential of I_h activation was significantly (P = 0.005) shifted from 106 ± 1 mV to 97 ± 1 mV (n = 4, Fig. 6B). Thus 1 mM dibutyryl-cAMP shifted the activation curve ~10 mV toward depolarizing voltages. Furthermore, the time courses of the current were accelerated by dibutyryl-cAMP (Fig. 6A, P = 0.03), in agreement with our preceding conclusions.

View larger version (11K): [in this window] [in a new window]   FIG. 6. Effect of dibutyryl-cAMP on soma Ih in conventional whole cell configuration with 0 µM cAMP in the pipette. We added 2 mM BaCl2 extracellularly to block IKi. A: current traces, before and after (*) application of 1 mM dibutyryl-cAMP, at 2 voltages (as indicated). Dibutyryl-cAMP increased currents at a voltage of 110 mV, but current amplitude at 130 mV was not significantly changed. Note that activation time courses were accelerated in the presence of dibutyryl-cAMP. B: activation curve measured in absence and presence of dibutyryl-cAMP. Solid lines: fits according to Eq. 2, yielding V0.5 = 106 mV and 97 mV (n = 4), before and after addition of the drug, respectively. Slope factors were not changed.

From all these results we suggest that cAMP primarily affects the gating kinetics of the channels underlying I_h.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

In the present work we show that under perforated-patch conditions, hyperpolarization-activated currents and depolarizing membrane potential sags were found in ~30% of the isolated growth cones, whereas at the somata the incidences were 80%. On the contrary, 39% of the growth cones possessed an I_Ki-like current, whereas 4% of the somata exhibited this type of current. Assuming no difference in selection of the cells for growth cones or somata measurements, these findings imply that the density of ion channels carrying I_h and I_Ki varied at the two regions of the neuron. A similar difference in I_h and I_Ki expression at the soma was also found in rat dorsal motor nucleus of the vagus nerve neurons, where 64% of the cells exhibited I_h and 28% showed I_Ki (Travagli and Gillis 1994). The regional differences in channel occurrence in our preparation suggest differential functions of I_h and I_Ki-like currents at the soma and the growth cone. It has been proposed that I_Ki is essential for neurite outgrowth in neuroblastoma cells (Arcangeli et al. 1994). We do not yet know, however, whether I_Ki is important for the growth of rat DRG neurons as well. At present the functions of I_h are not well understood. Extracellular application of Cs⁺ hyperpolarizes the membrane potentials at the growth cone and soma, which demonstrates that I_h contributes to the setting of resting membrane potential at the growing neurite tip and at the cell body of the rat DRG neuron. Furthermore, we showed that I_h is able to influence the excitability of the growth cone and soma whenever these two sites have been hyperpolarized. Our previous work has demonstrated that an increase of intracellular inositol trisphosphate concentration hyperpolarizes the soma membrane potential up to 30 mV from the resting potential (Wang et al. 1992). Under these conditions and dependent on the concentration of intracellular cAMP (see below), I_h channels could be activated and as a consequence influence cell functions. Because extracellular application of 2 mM CsCl inhibits neurite sprouting (Wang et al. 1995), I_h may be functionally involved in neurite outgrowth as well. Another possibility is that I_h is not fully expressed in the growth cone membrane until the growth cone contacts its target cells from maturation to a sensory nerve terminal requiring I_h for repetitive firing.

At both growth cone and soma, the time course of I_h was voltage dependent and could be best described by a double-exponential function, consistent with findings in rat auditory brain stem (Banks et al. 1993), mouse DRG neuron somata (Mayer and Westbrook 1983), and rat superior colliculus-projecting neurons (Solomon and Nerbonne 1993). It has been suggested that the multiexponential kinetics may originate from channels with multiple open or closed states or from distinct populations of fast and slow channels (Banks et al. 1993; Solomon and Nerbonne 1993). In agreement with results for rat auditory brain stem neurons (Banks et al. 1993), in our preparation the amplitude of the fast component was voltage dependent, whereas the amplitude of the slow component was not. The voltage independence of the slow component amplitude suggests that the conductance underlying the slow component decreases or the current saturates with membrane hyperpolarization (Banks et al. 1993). Furthermore, it has been shown that in some neurons there is a delay following the decay of the capacitive current transient before the time-dependent increase in current (Hille 1992; Solomon and Nerbonne 1993). Various models have been used to account for the delay (Solomon and Nerbonne 1993; and see references therein). We characterized the delay by fitting a power function according to Hodgkin and Huxley (1952) to the measured activation of the current. The value of the exponent of this function varied from 0.98 to 1.4, with a mean value of 1.1 ± 0.1 (n = 16). Thus, although a delay in the onset of I_h is present in some cells, conventional Hodgkin-Huxley-type kinetics cannot be used to describe the delay. Further kinetic analyses may help to understand the origin of the delay.

The effect of the second messenger cAMP could be well studied at the soma and it affected several properties of I_h. First, the I_h activation curve was shifted by intracellular cAMP to more depolarized voltages, in agreement with findings in other neurons and in heart cells (Banks et al. 1993; DiFrancesco and Tortora 1991; Tokimasa and Akasu 1990; Tsien 1974). The 30-mV shift in response to cAMP change from 0 to 100 µM in our preparation is close to the 27-mV shift in Purkinje fibers (Tsien 1974), but larger than the14-mV shift found in the auditory brain stem (Banks et al. 1993). However, it was nearly 3 times larger than the shift of 11 mV in sinoatrial node myocytes (DiFrancesco and Tortora 1991). The activation curve obtained during the perforated-patch mode was halfway between the curves obtained during conventional whole cell mode with 0 and 100 µM cAMP in the pipette. Thus alterations in the intracellular cAMP concentration may cause considerable changes in the activation of I_h. Second, both _f and _s were decreased by cAMP. The shift of the activation curve and the acceleration of the activation time course are consistent with an action of cAMP on channel kinetics. Third, the fully activated amplitude of I_h was independent of the intracellular cAMP level. These results are consistent with the findings in heart cells (DiFrancesco and Tortora 1991), but in contrast to those obtained in sympathetic and auditory brain stem neurons (Banks et al. 1993; Tokimasa and Akasu 1990), where extracellular application of 8-bromoadenosine 3,5-cyclic monophosphate increased the fully activated amplitude of I_h. Finally, the reversal potential of I_h was not altered, implying that the ionic selectivity is independent of cAMP. Thus, in the somata of rat DRG neurons, intracellular cAMP most likely affects solely the gating of channels underlying the hyperpolarization-activated I_h.

At the isolated growth cone the resting and firing properties were similar to our previous results from growth cones of intact neurons (Wang et al. 1994a). This suggested that the small isolated growth cones were healthy during our measurements (20 min). Longer observation times probably can only be realized with large-diameter (80-100 µm) growth cones from other species, which proved to be viable for several hours after isolation (Haydon et al. 1987; Shaw and Bray 1977; Wessels et al. 1978). Under our experimental conditions, we found a low incidence of I_h at the growth cone as compared with the soma. One possible reason may be the loss of intracellular factors or inflow of Ca²⁺ during the growth cone separation procedure. However, this alternative seemed unlikely, because we do not expect fast diffusion of factors or of Ca²⁺ through a narrow (1-µm) and long (100-µm) neurite. On the other hand, we cannot exclude a Ca²⁺ increase shortly after the growth cone separation due to stretch of the membrane (Sigurdson et al. 1993), which may cause growth cone degeneration and influence ion channel expression. Furthermore, we used the perforated-patch configuration to avoid the loss of growth cone intracellular substances. Thus the properties of I_h obtained within our measuring period probably represent those of undisturbed growth cones. Under these conditions, we found for the growth cone I_h that the midpoint potential was more negative and that the activation time constants were larger than those measured in the soma. Because we showed that lowering the intracellular cAMP in the soma shifted V_0.5 to hyperpolarizing values and increased _f and _s, our results suggest that the cAMP concentration in the growth cone is lower than that in the soma. We may speculate that a lower intracellular cAMP concentration at the growth cone may be important for its behaviour, because increased intracellular levels at this site inhibited neurite outgrowth (Lankford and Letourneau 1991).

In conclusion, our present work demonstrates for the first time differences in I_h properties and distribution at growing rat DRG neuron growth cones and somata. These differences either reflect developmental stages during neurite outgrowth or are causally involved in neurite sprouting, elongation, or directional growth.

	ACKNOWLEDGEMENTS

  The authors thank M. Deenen for excellent technical assistance and E. Olofsen for help with mathematical model discrimination. We also thank Dr. D. Büsselberg for advice on cell culture procedures and Dr. A. F. Weidema for support with the perforated-patch experiments.

	FOOTNOTES

   Present address of Z. Wang: Lausanne University, Dept. of Physiology, 7 Rue du Bugnon, CH-1005, Lausanne, Switzerland.

  Address for reprint requests: R. J. Van den Berg, Dept. of Physiology, Leiden University, PO Box 9604, 2300 RC Leiden, The Netherlands.

  Received 15 October 1996; accepted in final form 13 March 1997.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• ARCANGELI, A., BECCHETTI, A., MANNINI, A., MUGNAI, G., DE FILIPPI, P., TARONE, G., DEL BENE, M. R., BARLETTA, E., WANKE, E., OLIVOTTO, M. Integrin-mediated neurite outgrowth in neuroblastoma cells depends on the activation of potassium channels. J. Cell Biol. 122: 1131-1143, 1994.[Abstract/Free Full Text]
• BANKS, M. I., PEARCE, R. A., SMITH, P. H. Hyperpolarization-activated cation current (I_h) in neurons of the medial nucleus of the trapezoid body: voltage-clamp analysis and enhancement by norepinephrine and cAMP suggest a modulatory mechanism in the auditory brain stem. J. Neurophysiol. 70: 1420-1432, 1993.[Abstract/Free Full Text]
• BERDAN, R. C., EASAW, J. C. Modulation of sprouting in organ culture after axotomy of an identified molluscan neuron. J. Neurobiol. 23: 433-450, 1992.[Medline]
• DESTEXHE, A., BABLOGANTZ, A., SEJNOWSKI, J. Ionic mechanism for intrinsic oscillations in thalamic relay neurons. Biophys. J. 65: 1538-1552, 1993.[Medline]
• DIFRANCESCO, D., TORTORA, P. Direct activation of cardiac pacemaker channels by intracellular cyclic AMP. Nature Lond. 351: 145-147, 1991.[Medline]
• HALLIWELL, J. V., ADAMS, P. R. Voltage-clamp analysis of muscarinic excitation in hippocampal neurons. Brain Res. 250: 71-92, 1982.[Medline]
• HAMILL, O. P., MARTY, E., NEHER, E., SAKMANN, B., SIGWORTH, F. J. Improved patch-clamp techniques for high resolution current recording from cells and cell free membrane patches. Pfluegers Arch. Eur. J. Physiol. 391: 85-100, 1981.[Medline]
• HAYDON, P. G., MCCOBB, D. P., KATER, S. B. The regulation of neurite outgrowth, growth cone motility, and electrical synaptogenesis by serotonin. J. Neurobiol. 18: 197-215, 1987.[Medline]
• HILLE, B. In: Ionic Channels of Excitable Membrane., . Sunderland, MA: Sinauer, 1992
• HODGKIN, A. L., HUXLEY, A. F. A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol. Lond. 117: 500-544, 1952.
• HORN, R., MARTY, A. Muscarinic activation of ionic currents measured by a new whole-cell recording method. J. Gen. Physiol. 92: 145-159, 1988.[Abstract/Free Full Text]
• LANKFORD, K. L., LETOURNEAU, P. C. Roles of actin filaments and three second-messenger systems in short-term regulation of chick dorsal root ganglion neurite outgrowth. Cell Motil. Cytoskeleton 20: 7-29, 1991.[Medline]
• LJUNG, L. In: System Identification: Theory for the User., . Englewood Cliffs, NJ: Prentice-Hall, 1987
• MACCAFERRI, G., MANGONI, M., LAZZARI, A., DIFRANCESCO, D. Properties of the hyperpolarization-activated current in rat hippocampal CA1 pyramidal cells. J. Neurophysiol. 69: 2129-2136, 1993.[Abstract/Free Full Text]
• MAYER, M. L., WESTBROOK, G. L. A voltage-clamp analysis of inward (anomalous) rectification in mouse spinal sensory ganglion neurons. J. Physiol. Lond. 340: 19-45, 1983.[Abstract/Free Full Text]
• MAYERS, D.E.R. Distribution of ionic currents in the soma and growing region of an identified peptidergic neuron in a defined culture. J. Neurophysiol. 69: 406-415, 1993.[Abstract/Free Full Text]
• MCCORMICK, D. A., PAPE, H. C. Properties of a hyperpolarization-activated cation current and its role in rhythmic oscillation in thalamic relay neurons. J. Physiol. Lond. 431: 291-318, 1990.[Abstract/Free Full Text]
• RALL, W. Time constants and electrotonic length of membrane cylinders and neurons. Biophys. J. 9: 1483-1509, 1969.
• RUDY, B. Diversity and ubiquity of K channels. Neurosci. 25: 729-749, 1988.[Medline]
• SHAW, G., BRAY, D. Movement and extension of isolated growth cones. Exp. Cell Res. 104: 55-62, 1977.[Medline]
• SIGURDSON, W. J., SACHS, F., DIAMOND, S. L. Mechanical perturbation of cultured human endothelial cells causes rapid increase of intracellular calcium. Am. J. Physiol. 262: H1745-H1752, 1993.
• SOLOMON, J. S., NERBONNE, J. M. Two kinetically distinct components of hyperpolarization-activated current in rat superior colliculus-projecting neurons. J. Physiol. Lond. 469: 291-313, 1993.[Abstract/Free Full Text]
• TOKIMASA, T., AKASU, T. Cyclic AMP regulates an inward rectifying sodium-potassium current in dissociated bull-frog sympathetic neurons. J. Physiol. Lond. 420: 409-429, 1990.[Abstract/Free Full Text]
• TRAVAGLI, R. A., GILLIS, R. A. Hyperpolarization-activated currents, IH and IKIR, in rat dorsal motor nucleus of the vagus neurons in vitro. J. Neurophysiol. 71: 1308-1317, 1994.[Abstract/Free Full Text]
• TSIEN, R. W. Effects of epinephrine on the pacemaker potassium current of cardiac Purkinje fibres. J. Gen. Physiol. 64: 293-319, 1974.[Abstract/Free Full Text]
• WANG, Z., VAN DEN BERG, R. J., YPEY, D. L. Resting membrane potentials and excitability at different regions of rat dorsal root ganglion neurons in culture. Neuroscience 60: 245-254, 1994a.[Medline]
• WANG, Z., VAN DEN BERG, R. J., YPEY, D. L. Hyperpolarization activated current and cAMP in rat neonatal dorsal root ganglion neurons.J. Physiol. Lond. 479, Suppl.: 113, 1994b.
• WANG, Z., VAN DEN BERG, R. J., YPEY, D. L. Do hyperpolarization-activated currents I_h and I_Ki play a role in neurite sprouting in sensory neurons in culture? Neurosci. Res. Commun. 17: 53, 1995.
• WANG, Z., YPEY, D. L., VAN DUIJN, B. Inositol trisphosphate-induced hyperpolarization in rat dorsal root ganglion neurons. FEBS. Lett. 304: 124-128, 1992.[Medline]
• WESSELS, N. K., JOHNSON, S. R., NUTTALL, R. P. Axon initiation and growth cone regeneration in cultured motor neurons. Exp. Cell Res. 117: 335-345, 1978.[Medline]
• WOOD, J. N., WINTER, J., JAMES, I. F., RONG, H. P., TEATS, J., BEVAN, S. Capsaicin-induced ion fluxes in dorsal root ganglion cells in culture. J. Neurosci. 8: 3208-3220, 1988.[Abstract]

0022-3077/97 $5.00 Copyright ©1997 The American Physiological Society