INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Solutions containing CO₂/HCO, which act as the pH buffering system, have recently been shown to modulate whole cell calcium as well as sodium currents (Bruehl et al. 2000; Gu et al. 2000). Furthermore, the excitability of neurons in the slice preparation is different when the tissue is superfused with CO₂/HCO-buffered saline instead of a solution buffered with the artificial pH buffer HEPES (Church 1992, 1999; Church and McLennan 1989; Cowan and Martin 1995, 1996; Gu et al. 2000). Moreover, we have demonstrated that the voltage-dependent properties of calcium currents and their maximal conductance are concentration dependent modulated when the amount of bicarbonate ions was raised from 0 up to 37 mM (Bruehl et al. 2000). Gu et al. (2000) showed a strong negative shift in the voltage dependence of voltage-gated sodium currents after the exchange of a HEPES-buffered saline for a solution buffered with 26 mM HCO and 5% CO₂. Moreover, they demonstrated a largely reduced excitability of the CA1 neurons during the CO₂/HCO situation when cells were held in current-clamp mode. Switching from a nominally CO₂/HCO-free HEPES to a CO₂/HCO-containing medium cannot distinguish between bicarbonate or CO₂ as the modulating factor to evoke these effects. Furthermore, it is not clear whether these alterations are an all-or-nothing phenomenon or whether they depend on the concentration/gas-pressure of the modulator (i.e., HCO or CO₂).

Because CO₂/HCO is the mayor pH buffer system in the CNS and alterations in the concentration of both CO₂ and HCO can occur during normal, as well as patho-physiological conditions, it is important to investigate whether and to what extent CO₂/HCO-buffered saline can modulate ion conductances like the whole cell sodium current. Therefore we undertook the present study to unravel which of both components induces the modulation of the sodium current properties and whether there is a concentration dependence as can be seen on whole cell calcium currents.

For this purpose, we used the conventional whole cell voltage-clamp technique on freshly isolated hippocampal CA1 neurons of young Wistar rats.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell preparation

CA1 pyramidal neurons were enzymatically isolated from male Wistar rats (50-75 g) as described in detail previously (Vreugdenhil and Wadman 1992). Slices (500 µm) were cut from both hippocampi, and the CA1 area was dissected. These tissue pieces were incubated for 38 min at 32°C in oxygen-saturated dissociation solution (in mM/l: 120 NaCl, 5 KCl, 1 CaCl₂, 1 MgCl₂, 20 PIPES, 25 D-glucose; pH: 7.0; osmolarity: 295 mosmol) containing 1 mg/ml trypsin (Bovine Type XI). Following enzymatic treatment, tissue was washed twice and kept in the dissociation solution without trypsin at 19°C. Directly before measurements tissue pieces were dispersed in HEPES-buffered bath solution by trituration through Pasteur pipettes with decreasing tip diameter, and cells were allowed to settle in the perfusion chamber.

To assure total solution exchange, we used a bath chamber with a volume of ~120 µl, which was perfused with a constant flow rate of 1 ml/min. Bath solutions contained (in mM/l): 37 NaCl, 5 KCl, 2.5 CaCl₂, 1 MgCl₂, 5 4-aminopyridine (4-AP), 30 TEA-Cl, 72 choline-Cl, and 25 D-glucose and 100 µM CdCl₂; pH was set at 7.3 (unless otherwise stated), with an osmolarity of 318 mosmol/l. For seal formation, the majority of cells were patched in the preceding mentioned solution, plus 10 mM HEPES as the pH-buffer system. Bath solutions containing CO₂/HCO instead of HEPES were thoroughly gassed with different amounts of CO₂ (2.5; 5.0;10%) before the corresponding amount of HCO was added. NaCl was replaced equimolarily by NaHCO, and osmolarity was adjusted with glucose to 318 mosmol/l when necessary. Special care was taken to assure that the solutions were always equilibrated with CO₂ throughout the course of the experiment because otherwise CaCO₃ and CdCO₃ would have precipitated. Liquid-junction potentials may occur at the tip of the patch electrodes due to the different ion compositions of the solutions and may in some cases mislead the measurements. Therefore we measured the electrode potentials in HEPES- and bicarbonate-buffered solutions. The observed junction potentials never exceeded ±0.5 mV and were therefore regarded as negligible.

All chemicals were obtained from Sigma (Deisenhofen, Germany) or Merck (Darmstadt, Germany).

Current recording

Currents were measured under whole cell voltage-clamp conditions at room temperature using patch pipettes of 2-4 M resistance. Electrode solution contained (in mM/l): 5 NaCl, 115 CsF, 2 MgCl₂, 0.5 CaCl₂, 115 TEA-Cl, 10 EGTA, 5 phosphocreatine, 2 MgATP, 0.1 NaGTP, and 0.1 leupeptin (pH set at 7.3) and 50 units/ml phosphocreatine kinase. Osmolarity was adjusted to 300 mosmol/l, when necessary, by adding glucose. The solution was heavily buffered by 50 mM HEPES to minimize intracellular pH changes following the introduction of CO₂/HCO-buffered solution. Currents were measured with an Axopatch 200B amplifier (Axon Instruments) and stored on an Atari ST computer (10-kHz sample frequency). Capacitive transients were corrected and series resistances (<10 M) were compensated for >90% on-line. Data were evaluated off-line using a custom-made computer program. All current traces were corrected for aspecific linear leak (reversal potential: 0 mV) determined at holding potential. Rundown phenomena were never observed during the recording period and any neuron that escape from voltage clamp was rejected from the analysis. Such an escape was characterized by a slow activation of the currents, a delayed achievement of the peak amplitude and/or by the occurrence of more than one current peaks during the voltage step.

Experimental protocols

Sodium currents were activated, following a prepotential of 130 mV (500 ms), using 10-ms voltage steps to voltage levels between 70 and +15 mV (increment: 5 mV). Holding potential was kept at 80 mV. The steady-state inactivation of the sodium current was determined using a standard depolarization to 20 mV after the cell was polarized for 500 ms to various levels between 110 and 45 mV (increment: 5 mV).

First, neurons were normally bathed in HEPES-buffered saline, and the voltage protocols that determine activation and inactivation properties were performed. Second, the HEPES-buffered saline was replaced by CO₂/HCO-buffered saline. During solution changes a test pulse stepping from 130 mV (500 ms) to 20 mV was applied (10 ms), to monitor current amplitude changes and to ensure the stability of the final condition. A 3-min period was allowed to guarantee total solution exchange before the same protocols were applied as with HEPES-containing solution. During this period, the current amplitude increased within seconds after solution exchange and reached its final value within the following minute.

Current analysis

Peak amplitudes of the currents (I) evoked with the activation protocol were plotted as a function of membrane potential (V). The resulting I-V relations were fitted with a combination of a third-order Boltzmann activation function and the Goldman-Hodgkin-Katz (GHK) current-voltage relation (Hille 1992; Kortekaas and Wadman 1997)

(1)

with  = F/RT and g_max =  FP₀ [Na⁺]_out, where g_max is the maximal membrane conductance (which is proportional to the maximal permeability and the extracellular sodium concentration), V_h is the potential of half-maximal activation, and V_c is proportional to the slope of the curve at V_h. F represents the Faraday constant, R the gas constant, P₀ is the maximal permeability, and T the absolute temperature.

The voltage dependence of steady-state inactivation of the sodium current was estimated from the relation of peak current amplitude versus the prepotential. This relation was well described by a Boltzmann function, which also normalized the current

(2)

where N(V) is the level of steady-state inactivation determined from the current amplitude I(V) normalized to I_max, V is the prepulse potential, V_h is the potential of half-maximal inactivation, and V_c is a factor proportional to the slope of the curve at V_h.

Kinetics of the whole cell sodium currents were determined using a fit procedure which implies a third-order exponential term for activation and two exponentials describing the inactivation kinetics. The following function was applied

(3)

where I1 and I2 are the amplitudes of the two current components and _a; _i,1 and _i,2 represent the time constants for activation and inactivation after the start of the voltage step at time t0.

Statistics

Values are presented as the mean ± SE. Statistical comparisons were made with Student's t-test, if not stated otherwise. P < 0.05 was used to indicate significant differences.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Concentration dependent effect of CO₂/HCO₃-buffered solution on whole cell sodium currents

ACTIVATION IN HEPES-BUFFERED SALINE. In a HEPES-buffered saline, whole cell sodium currents could be evoked in all neurons (n = 16) tested (Fig. 1). They showed a fast activation and an almost complete inactivation when elicited from a potential of 130 mV (Fig. 2A, left). A voltage protocol that steps to different voltage levels (between 70 and +15 mV) revealed the typical current voltage relation (Fig. 2B, bottom) with an mean peak amplitude of 5.4 ± 0.5 nA at around 15 mV. When those I-V curves were fitted with Eq. 1, they delivered three variables, i.e., the maximal sodium conductance (g_max), the potential of half-maximal activation (V_h,a), and the slope factor at the point of V_h,a (V_c). For the cells tested in this way, g_max was evaluated to be 203.1 ± 21.0 nS (Fig. 3, left) and 50% of the channels were activated at a voltage of V_h,a:-27.4 ± 1.4 mV with a slope of V_c: 4.3 ± 0.3 mV (Fig. 2B, top). With regards to the different ion compositions and evaluation methods (third- vs. first-order Boltzmann activation functions), these values resembled data previously been published from freshly isolated CA1 neurons (Gu et al. 2000; Ketelaars et al. 2001).

View larger version (28K): [in this window] [in a new window]   Fig. 1. Sodium currents evoked from a CA1 neuron. Activation protocol was preceded by a period of 500 ms at a voltage of -130 mV to remove inactivation totally. Currents were elicited at voltages ranging from -70 to 15 mV (see inset). Within the test period of 10 ms, sodium currents did not fully inactivate.

View larger version (21K): [in this window] [in a new window]   Fig. 2. A: whole cell sodium current evoked by a voltage step from 130 to 20 mV (see inset) of a CA1 neuron. Cell was subsequently bathed in 4 solutions containing either HEPES as the pH buffer, or CO2/HCO with different concentrations. The amplitude of this current increased in parallel with increasing amounts of bicarbonate. B: I-V relationship (bottom) and activation properties (top) of the voltage-gated sodium current in relation to the concentration of bicarbonate and CO2. Currents are activated from a prepotential of 130 mV. The peak amplitude increased and shifted to more negative potentials, when the amount of CO2/HCO was elevated. This phenomenon underlies a negative shift of the Boltzmann curves of activation, which results in an increased driving force for the sodium ions.

View larger version (49K): [in this window] [in a new window]   Fig. 3. Bar-graph showing the maximal conductances (gmax), as were evaluated with Eq. 1 for 3 different bicarbonate and CO2 combinations (see subtitles) and the HEPES-buffered solution. The observed increase in sodium current amplitude was not accompanied by an increase in maximal conductance when neurons (n = 16) were successively bathed in solutions with increasing amounts of CO2/HCO.

INACTIVATION IN HEPES-BUFFERED SALINE. To investigate the steady-state inactivation properties, we evoked currents with a voltage step to 20 mV (10 ms) following a 500-ms period at different prepotentials (110 to 45 mV; increment: 5 mV). When the measured peak amplitudes were plotted against the prepotential voltage, it gives a Boltzmann-like inactivation curve that was best described by Eq. 2. The values received by a fit with this equation; i.e., the half-maximal potential of steady-state inactivation (V_h,i) and the slope of the Boltzmann curve at this potential (V_c) amounted to 53.6 ± 1.8 and 6.1 ± 0.2 mV, respectively.

KINETICS IN HEPES-BUFFERED SALINE. When evoked by potential changes from 130 mV to more positive values, currents activated rapidly and inactivated almost completely within the time window of 20 ms. The kinetic of activation and inactivation could be best described using a combination of a third-order exponential term for the activation of the current and two exponential terms for the inactivating part (Eq. 3). Application of this algorithm usually leads to fit results within the noise level (Fig. 4, top). The evaluation of the time constants was restricted to test voltages of 35 up to 5 mV and delivered time constants for activation, which decreased with more positive voltages, starting with 0.39 ± 0.11 ms at 35 mV and ending with 0.08 ± 0.01 ms (n = 16) at 5 mV. The time constants for inactivation were also voltage dependent and were estimated to be in the range of 7.98 ± 1.46 ms (at 35 mV) and 3.11 ± 0.24 ms (at 5mV) for the slow component and 1.33 ± 0.43 and 0.51 ± 0.03 ms (n = 16), for the fast inactivating component (Fig. 4, ).

View larger version (24K): [in this window] [in a new window]   Fig. 4. Activation and inactivation kinetics of the whole cell sodium current. The original current (; top) was best fitted (scattered line) with a 3rd-order exponential for activation (time constant a) and a double exponential for inactivation (time constants i,1 and i,2). Increasing the amounts of CO2 and HCO led to a decline of the time constant of activation in the voltage range of 35 to 20 mV (left, bottom), while the time constants of inactivation decreased over the entire voltage range with increasing CO2/HCO (middle and right).

ACTIVATION IN CO₂/HCO-BUFFERED SALINE. After performing the described protocols the bath perfusion was switched from HEPES buffer to solutions containing different amounts of CO₂ and HCO. Cells were subjected to solutions containing 2.5% CO₂/5.6 mM HCO, 5% CO₂/18 mM HCO, and 10% CO₂/37 mM HCO, with the same pH of 7.3 for all solutions. Half the cells were tested with solutions in ascending order of the amounts of CO₂ and HCO, while the other half were tested with solutions in descending order to avoid any hysteresis effect induced by the order of solution changes. Because the effects were fully reversible, no difference was found between the two sequences of solution exchanges. Data were pooled and are presented as mean values. No sign of a solution dependent alteration in g_max was observed (Fig. 3) during these experiment. None of the g_max values obtained in the three CO₂/HCO-buffered solutions was significantly different from the value seen in HEPES-buffered solution, which could be shown by one-way ANOVA analysis (Table 1). Changing to a solution with low CO₂/HCO buffering (i.e.: 2.5% CO₂ and 5.6 mM HCO) following the HEPES condition induced only small alterations of the currents (Fig. 2A, top). The mean peak amplitude (at 15 mV) was merely unaffected (Fig. 2B, ) with a value of 5.4 ± 0.6 nA, while the potential of half-maximal activation V_h,a was negatively shifted by 5.8 ± 2.1 mV (Fig. 5B1) to a voltage of 33.2 ± 2.7, and the slope (V_c) was almost unchanged 3.8 ± 0.4 mV. Increasing the amounts of CO₂ and HCO in the bath solution led to a more pronounced negative shift of V_h,a and to a significant increase of the current amplitude. The addition of the 5% CO₂/18 mM HCO-buffered saline resulted in a shift of V_h,a by 13.1 ± 1.8 mV to 40.5 ± 2.2 mV, which was significantly different from the V_h,a measured in HEPES solution (one-way ANOVA: P < 0.05, Fig. 5B1), and in a raise of the peak amplitude to 6.7 ± 0.8 nA (at 25 mV). Again the slope V_c was only slightly affected (3.4 ± 0.4 mV). The shift of voltage dependence and increase in amplitude was most pronounced in the 10% CO₂/37 mM HCO-containing solution, with V_h,a: 19.4 ± 1.8 mV (V_h,a: 46.9 ± 1.8 mV; P < 0.05) and a peak amplitude value of 7.9 ± 1.0 nA (at 35 mV; P < 0.05). No alteration of the slope V_c (3.5 ± 0.6 mV) was observed (Figs. 2B and 5B1). The shift of V_h,a, induced by increasing amounts of CO₂ and HCO, was almost linear within the concentration-range measured with a slope-rate of 0.7 mV/mM HCO. Nevertheless it cannot be excluded that with even higher amounts of CO₂ and HCO, this function will break away from linearity. A subtraction of the I-V curves obtained in HEPES-buffered solution from those in CO₂/HCO-containing solutions revealed an increase in current amplitude that was most prominent in the voltage range of 50 to 30 mV (Fig. 5A1).

View this table: [in this window] [in a new window]   Table 1. Whole cell calcium current properties in HEPES- and CO2/HCO-buffered solutions

View larger version (46K): [in this window] [in a new window]   Fig. 5. A: mean differential current constructed by subtracting I-V curves obtained in HEPES-buffered solution from I-V curves derived with CO2/HCO in the bathing solution. A prominent increase in current amplitude and a shift to more negative potentials was seen when both CO2 and HCO were raised (A1). The huge amplitude in differential current reflects also the negative shift of the potential of half-maximal activation during these experiments. The same alterations take place when only the concentration of bicarbonate was increased and CO2 was held constant at 5% (A2). When CO2 was increased but not HCO (18 mM), no differences between the 3 CO2/HCO-buffered solution could be observed (A3). The maximal increase in current amplitude was observed in the voltage range of -50 to -30 mV in all experiments. B, 1-3: bar-charts showing shifts of the voltage dependence of activation () and inactivation () following superfusion with solutions containing different amounts of CO2/HCO. Both, the potential of half-maximal activation (Vh,a) and inactivation (Vh,i) were negatively shifted with respect to the amount of CO2/HCO (B1). Qualitatively the same shifts of Vh,i and Vh,a were observed when only the concentration of bicarbonate was increased while CO2 was kept stable (B2). There was no correlation in the shifts of the Vh,i and Vh,a when the concentration of CO2 (2.5, 5.0; 10%) was increased and the concentration of bicarbonate was held constant at 18 mM (B3). Insets: numbers of cells measured for the corresponding pairs of data in A and B.

INACTIVATION IN CO₂/HCO-BUFFERED SALINE. Concomitantly with the shift of voltage dependence of activation, also the voltage parameters of the steady-state inactivation were negatively shifted when the concentrations of CO₂/HCO were elevated (Fig. 5B1). The potential of half-maximal inactivation was shifted almost linearly (within the concentration-range measured; slope-rate: 0.6 mV/mM HCO) by 2.8 ± 1.2, 8.8 ± 1.1, and 16.2 ± 1.0 mV when CO₂/HCO was added in ascending order (absolute values: 56.4 ± 1.9 mV, 62.4 ± 2.0 mV, P < 0.05; 69.8 ± 2.1 mV; P < 0.05). As with activation, no alteration of the slope (V_c) at the point V_h,i was observed (Table 1). The calculation of the pairwise difference between the potentials of half-maximal activation and inactivation showed no significant difference among the solutions (-25.4 ± 2.2, 25.0 ± 2.8, 24.0 ± 2.1, and 24.0 ± 2.1 mV; n = 14). This demonstrated that the effect took place to the same degree on the activation as well as on the inactivation properties of the currents.

KINETICS IN CO₂/HCO-BUFFERED SALINE. The prominent shift in voltage dependence of the current characteristics with increasing amounts of CO₂ and HCO were accompanied by alterations of the inactivation and activation kinetics (Fig. 4, bottom). Both time constants for inactivation decreased when CO₂ and HCO were increased, yielding values, in the 10% CO₂/37 mM HCO-containing solution, of 4.13 ± 0.36 and 1.64 ± 0.25 ms (35 and 5 mV; n = 17) for the slow component and 0.64 ± 0.06 and 0.31 ± 0.03 ms for the fast current component. The time constant for activation showed qualitatively the same reduction when cells were exposed to increasing amounts of bicarbonate and CO₂ with minimal values of 0.13 ± 0.01 ms (35 mV) and 0.12 ± 0.01 ms (5 mV) in the solution containing 10% CO₂/37 mM HCO.

Bicarbonate but not CO₂ shifts the voltage dependence of sodium currents

In the first set of experiments, both compounds of the CO₂/HCO buffer were altered to keep the extracellular pH constant. It was therefore not possible to state whether CO₂ or bicarbonate modulates the whole cell sodium current. In a second series of experiments, the concentration of CO₂ was held constant at 5%, and the concentration of bicarbonate was altered (5.6; 18; 26 mM). Under these circumstances, extracellular pH varies considerably between the CO₂/HCO-buffered solutions (pH: 6.96; 7.30; 7.44; respectively). Therefore experiments were carried out in which only one switch, from HEPES-buffered saline to the concerned CO₂/HCO-buffered saline was performed. The pH value of the HEPES- buffered saline was adjusted to match the CO₂/HCO-buffered saline to avoid any effect by a difference in pH_o. The concentration of the high bicarbonate containing solution (i.e., 37 mM HCO) was reduced to 26 mM HCO to prevent precipitation of CaCO₃ and CdCO₃. In total, 58 neurons were investigated.

When I-V curves of the whole cell sodium current were constructed from the data obtained in HEPES-buffered solutions, a depression of the current amplitude over the entire voltage range was observed with more acidic pH_o values (Fig. 6B). This reduction was mainly due to a slight, but not significant, decrease of g_max, concomitantly with the decreased pH_o. As previously observed (Tombaugh and Somjen 1996), only small and not significant shifts in the potentials of half-maximal activation and inactivation could be elicited. The following switch to the corresponding CO₂/HCO-buffered solutions led to a clear evidence for a bicarbonate-concentration-dependent effect on the voltage properties of activation and inactivation (Fig. 5B2). The potential of half-maximal activation and inactivation was negatively shifted, in a significant way, with increasing bicarbonate amounts, when compared with values achieved in the corresponding HEPES-buffered saline (Table 1). Within the concentration range measured the half-maximal potentials shifted by 0.5 mV/mM dissolved HCO. Again, it could be shown by calculating the pairwise difference between V_h,a and V_h,i that both the activation and inactivation properties were effected in the same manner. When the differences of currents obtained in HEPES- and CO₂/HCO-buffered solutions were plotted, it became obvious, that the increase in current amplitude was most prominent in the voltage range between -50 to -30 mV (Fig. 5A2). This effect was similar to the experimental series described before. The slope parameters (V_c) for inactivation and activation remained unchanged and showed no concentration-dependent effect. The same applied to the maximal sodium conductance, which did not significantly increase during these experiments.

View larger version (18K): [in this window] [in a new window]   Fig. 6. The role of intra- and extracellular pH on the modulation of sodium currents. A: the current-voltage relationship of the current was slightly shifted to more negative potentials and amplitude was moderately depressed when neurons were exposed to 23 mM acetate, which induces an intracellular acidification. The differential current showed a different course than was seen after a change to CO2/HCO-buffered solutions because it was negative in the range 50 to 20 mV and became than positive with more positive voltages. B: extracellular pH differences lead to an increase in amplitude with more alkaline pH values, but not to a shift of the I-V curve.

Lack of modulation by CO₂

As a complementary series of experiments, also CO₂/HCO-buffered solutions were tested, which contained equal quantities of bicarbonate (18 mM) but different amounts of CO₂ (2.5, 5.0, and 10%). This was done to justify whether bicarbonate alone can modulate the whole cell sodium currents or whether CO₂ has an additional effect. Again the pH values of the HEPES-buffered solutions were adjusted to fit with the values of the corresponding CO₂/HCO-buffered saline (pH: 7.5, 7.30, 7.0). Briefly, the use of solutions with different CO₂ concentrations, but constant amounts of bicarbonate, did not change any of the current parameters in a concentration- and CO₂-dependent manner. The potential of half-maximal activation V_h,a was negatively shifted by about the same amount in all three solutions (Fig. 5B3) with 11.4 ± 2.5 mV (6), 12.8 ± 1.6 mV (18), and 8.5 ± 1.3 mV (n = 6). A similar result was obtained for the half-maximal potential of inactivation (V_h,i: 9.9 ± 1.3, 9.0 ± 1.1, and 8.2 ± 0.4 mV). Also the slope factors (V_c) of both activation and inactivation and, furthermore, the maximal sodium conductance did not differ between the solutions tested. Finally, the difference (Fig. 5A3) between I-V curves obtained in HEPES- and CO₂/HCO-buffered solutions was similar for all three CO₂ amounts, indicating that variations of CO₂ do not modulate the voltage dependence of the whole cell sodium currents.

Effects of internal pH on whole cell sodium currents

The role of pH in the modulation of sodium currents has previously been demonstrated (Tombaugh and Somjen 1996). To estimate the fraction of effect, which could be attributed to an alteration of intracellular pH, we conducted a small experimental series in which neurons were exposed to bath solutions containing 23 mM Na-acetate. Because the acetic acid crosses the membrane like the carbonic acid, this solution mimics the acid load of the neurons, which can be assumed during exposure to CO₂/HCO-containing solutions. The intracellular pH during these experiments was controlled by only 10 mM HEPES, instead of 50 mM, to show the maximum effect of the intracellular pH changes.

With this weak intracellular pH buffering, several changes of the sodium current properties were observed during the measurements. The maximal conductance was decreased by 23 ± 12% (112.4 ± 6.5 vs. 87.2 ± 14.9 nS; P = 0.11; n = 5), which is opposite to the effects seen with CO₂/HCO-buffered solution, where a small, but not significant, increase was found. Negative shifts of the potentials of half-maximal activation (9.6 ± 1.6 mV; n = 5; P < 0.05) and inactivation (4.3 ± 1.0 mV; P < 0.05) were observed when acetate was introduced. Nevertheless this shift was grossly only half the size of the maximal shift observed with CO₂/HCO-containing solutions. Because the shift in the voltage sensitivity and the decrease of g_max have opposite effects on the peak size of the I-V curve, only a small, but not significant, decrease of the peak amplitude was observed (Fig. 6A). Furthermore the difference current from the I-V curves derived under the acetate and nonacetate condition showed a clearly different course than was seen with CO₂/HCO-buffered solution. Because in the voltage range 40 to 30 mV, the current was negative, like in the other experiments, while in the range more positive than 25 mV, it gained positive values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Bicarbonate modulates sodium current activation and inactivation

The present study indicates the potential role of bicarbonate ions as a modulator of voltage dependent properties of whole cell sodium currents in CA1 neurons of the rat. Bicarbonate, in a concentration-dependent manner, shifts the potential of half-maximal activation and inactivation to more negative voltages, leaving the slopes of the activation and inactivation curves unaffected. Variation of the CO₂ contents of the solutions (from 2.5 to 10%) did not have any additional effect, showing that bicarbonate alone modulates the voltage dependence of the sodium currents. The negative, bicarbonate-dependent shift of V_h,i made the mean peak current amplitude highly sensitive to small voltage changes close to the membrane potential, which is different to the situation in HEPES-buffered saline. Bicarbonate did not change the maximal conductance of the investigated sodium currents. The latter effect is in contrast to the result found on whole cell calcium currents from the same type of neurons. In this case, a prominent reduction of the maximal conductance induced by bicarbonate ions was observed (Bruehl et al. 2000). This difference is responsible for the fact that under certain circumstances, the amplitude of the sodium currents increases with the increase of bicarbonate concentration rather than decreases as demonstrated for calcium currents. The increase of the sodium current amplitude can further be explained by the negative shift of the activation threshold (resp. the V_h,a), which results in an increase of the number of sodium channels being opened at more negative membrane voltages at which the sodium ions are subject to a stronger driving force.

The lack of effect on the sodium conductance and the prominent reduction of the calcium conductance points toward the possibility that bicarbonate ions act in a different way on both channel types.

Sodium current amplitude and amount of bicarbonate

Recently, Gu et al. (2000) demonstrated that CO₂/HCO-buffered saline can shift both the potential of half-maximal activation and inactivation to more negative potentials in mice neurons. This is in line with the present study. In addition, they stated that this saline did not change the total amplitude of the whole cell sodium current. In contrast, the present study demonstrates a strong voltage sensitivity of the sodium current amplitude, depending on the amount of dissolved bicarbonate. Increasing concentrations of bicarbonate shifted the Boltzmann curve of inactivation in negative direction (up to 70 mV) to voltages close to resting membrane potential (i.e., approximately 65 mV). With the steep slope of the curve close to resting potential, it is expected that even small voltage changes strongly alter the number of inactivated sodium channels. Therefore the sodium current amplitude and the resulting action potential will be larger with potentials more negative than about -70 mV and much smaller with potentials positive to this value. This feature of the sodium current in CO₂/HCO-buffered saline is different from the situation in solutions with low concentrations of bicarbonate or even in HEPES-buffered saline. Under these circumstances, the potential of half-maximal inactivation is far away (approximately 54 mV) from the resting potential value, which leaves the sodium current amplitude almost independent from voltage changes around resting potential.

Previous studies have shown that the excitability of the neuronal network is reduced when brain slices or cell cultures are superfused with CO₂/HCO-buffered saline instead of a HEPES-buffered solution (Cowan and Martin 1995, 1996; Gu et al. 2000). This observation fits with the presented datain the case that bicarbonate is low enoughand the previous finding that whole cell calcium currents are reduced (Bruehl et al. 2000), when CO₂/HCO is used as the pH buffer system. The present study also points out that the excitability might be increased under certain conditions when the bicarbonate concentration is sufficiently high (Church 1992, 1999; Church and McLennan 1989) and neurons are temporarily hyperpolarized beyond resting potential. Under these conditions, the shift in activation increases the sodium current amplitude, while the steady-state inactivation has little effect. Regarding the negative shift of the potential of half-maximal activation (V_h,a), one can predict that the action potential threshold also should shift to more negative potentials with increasing concentrations of dissolved bicarbonate. Indeed, such a shift of activation threshold was found by Church and McLennan (1989) on intracellular recorded CA1 neurons in slice preparations. They found the activation threshold lowered 9 mV, when the bath solution was switched from 26 to 72 mM bicarbonate, and a positive shift (16 mV) following a switch to bicarbonate-free (HEPES-buffered) media (Church 1992). Furthermore, both studies demonstrated that neurons that were quiet in low- or bicarbonate-free media became spontaneously firing when higher bicarbonate concentrations were used. This again can be explained by a threshold for action potential firing that is closer to the resting potential.

The present study also demonstrates that the inactivation kinetics of the sodium current is strongly affected by bicarbonate. The time constants of both components decreased over the entire voltage range measured, while the time constant of activation decreased only in the range between 35 up to 25 mV, but remained unchanged at more positive values. Consequently, the inactivation becomes more efficient in terminating the current at potentials more positive to -25 mV, for example during the generation of action potentials. The lack of a so-called overshoot of the action potential, which occasionally occurs in vivo (Witte et al. 1996) but also in vitro (Gu et al. 2000); see there Fig. 2A), may be a consequence of this faster inactivation in CO₂/bicarbonate-buffered solution because the current shuts down before the zero voltage level has been reached.

The concentration values at which bicarbonate acts on the whole cell sodium current are clearly in the physiological range observed in brain in vivo, which has been shown to be 24-26 mM during normal activity (Betz et al. 1989). During pathophysiological processes, like metabolic acidosis, respiratoric alkalosis or hypercapnia this normal level can be changed by 10-15 mM in both directions. Under these circumstances, it is most likely that the excitability of the network changes, not only by pH alterations (Church 1999; Tombaugh and Sapolsky 1990, 1996, 1997) but also by differences in bicarbonate concentration in the brain tissue.

Changes of the intracellular pH

When bicarbonate/CO₂-free bath solutions are exchanged by solutions containing these compounds, the trans-membraneous passage of CO₂ into cells leads, at least transiently, to a decline of the intracellular pH, due to liberation of protons in accordance to the Henderson-Hasselbalch equation. It is conceivable that such changes contributed to the alterations of sodium currents in our experiments. Nevertheless, our data obtained by superfusing neurons with bath solutions containing 23 mM Na-acetate, a solution that mimics the acid introducing properties of HCO/CO₂-buffered media, showed only subtle alterations of the currents. The intracellular pH was only weakly buffered by 10 mM HEPES during these experiments. First, the maximal conductance decreased and the mean peak amplitude was merely unchanged. Second, the potentials of half-maximal activation or inactivation showed smaller potential shifts in negative direction as was seen in HCO/CO₂-buffered solution. Furthermore, Tombaugh and Somjen (1997) have shown that an increase of the intracellular buffering power by enhancing the HEPES concentration blunted pH-related effects on whole cell calcium currents by 50%. These observations indicate that the alteration of sodium currents by bicarbonate-buffered solutions are mainly related to the modulating action of the bicarbonate ions, and are only minimally contaminated by pH-related effects. Further evidence for this conclusion arises from the experiment in which bicarbonate was kept constant (at 18 mM) and CO₂ was altered from 2.5 to 10%. This should alter intracellular pH values and cause cumulative changes of the sodium current properties. However, such changes of the sodium currents were not observed.

As for the modulating action of bicarbonate ions on whole cell calcium currents, we still do not know how these ions can interact with the sodium channel pores. The mechanisms underlying the modulation of sodium currents by protons were discussed earlier in detail (Hille 1992) and might help to understand the way of modulation by bicarbonate ions. First of all, modulation of the channels needs charged ions, which can interfere with the charged domains of the channel proteins. This theoretically rules out any action of the uncharged molecule CO₂. In practice, our data strongly support this hypothesis. Three theories have been proposed of how protons may alter the characteristics of the sodium channels. Two titration theories that assume that protons may titrate negative surface charges or negative acid groups within the channel have been favored. The covering of the sodium ion attracting sites should end up in a decrease of the single channel conductance, which explains the reduced sodium permeability at low extracellular pH. A reduced sodium conductance has not been observed in the present study, when bicarbonate ions act as the modulator. This finding supports the third explanation, the gating theory, because an influence of the modulator on the gating properties of the channels does not result in an alteration of the conductance, but in a shift of voltage dependence of the activation and inactivation. In fact alterations of the gating properties of sodium channels by CO₂/HCO solutions have been demonstrated by Gu et al. 2000. Taken together, the lack of effect on the conductance and the shift in activation and inactivation, shown in the present study, plus the findings of Gu makes the gating theory, the most conceivable mechanism of how bicarbonate ions act on sodium current channels. To substantiate this assumption, further studies on the single channel level and/or binding studies are necessary.