Aqueous suspensions of the local anesthetic n-butyl-p-aminobenzoate (BAB), epidurally applied in terminal cancer patients, resulted in a sensory blockade, lasting up to several months. To investigate the mechanism of action on the cellular level, the effect of 100 microM BAB on Na+ action potentials and on Na+ currents in dorsal root ganglion neurons from neonatal rats was studied.


Small neurons grown in cell culture were selected for patch-clamp measurements. Both Na+ action potentials, evoked by current pulses of increasing amplitude (current clamp) and Na+ currents, activated at different membrane potentials (voltage clamp), were investigated in the absence and presence of 100 microM BAB. The local anesthetic was applied by external perfusion for 2 or 10 min.


In the presence of 100 microM BAB, either the firing threshold was raised or the action potential was abolished. The maximal peak conductances, underlying the fast sodium current INa,F and the slow sodium current INa,5, were not changed. However, the inactivation of INa,F was increased by BAB. The sigmoid inactivation curve shifted 12 mV toward hyperpolarizing membrane voltages, whereas no changes were found for the inactivation of the slow Na+ current. Only at short exposure times of 2 min, the effects of BAB could be reversed during a 10-min wash-out.


BAB dramatically increased the firing threshold, and in part of the sensory neurons, it blocked the action potential. The inactivation of the fast Na+ channels, but not of the slow Na+ channels, was increased by BAB. Thus, the block of fast Na+ channels by BAB may contribute to epidural analgesia. At exposure times of 10 min, the effect of BAB was not reversible. This probably originates from its high lipid-solubility, which may be an important factor in determining the duration of the block in vivo.

Key words: Anesthetic, local: n-butyl-p-aminobenzoate. Ions: Sodium sup +. Measurement technique: patch-clamp. Sodium sup + current inactivation. Nerve: dorsal root ganglion.

PAIN treatment in terminal cancer patients often involves application of opioids and local anesthetics or sometimes lesions of sensory nerves by means of alcohol or phenol. These treatments may cause severe side effects, among which motor dysfunction is most prominent. An aqueous suspension of the highly lipid-soluble local anesthetic n-butyl-p-aminobenzoate (BAB) has been injected epidurally on the segmental level of pain perception in terminal cancer patients. The resulting depot of BAB is believed to dissolve slowly in the epidural space, thus causing a gradual and continuous application of BAB. [1]This treatment resulted in a long-lasting (median 29 days) sensory blockade (segmental analgesia), combined with a marked reduction of pain, without a reduced motor function. [1,2]The use of such a suspension is a promising alternative for the classic methods of pain treatment in terminally ill patients.

The preservation of motor functions during and after BAB treatment suggests that the epidural application of the local anesthetic has a selective effect on sensory nerve fibers. It has been proposed that selective block after application of local anesthetics originated from differences in internodal distance of sensory and motor fibers, in relation to the length of nerve exposed to the local anesthetic. [3-5]Korsten et al. [2]postulated that the selective block seen after epidural BAB is caused by exposure of spinal roots crossing the epidural space. This selective effect is not nullified by diffusion to the cerebrospinal fluid, because BAB is hydrophobic. Therefore, the length of nerve exposed to BAB is limited and may explain the selective block of sensory fibers.

To investigate the mechanism of action of BAB on a cellular level, the patch-clamp technique was applied to isolated dorsal root ganglion (DRG) neurons from the rat. These cells are the primary sensory neurons of higher vertebrates, a subpopulation of which mediates pain perception. [6]The DRG neurons are close to the site, where the remnant of the BAB suspension was found in humans at necropsy. [7]In the current study, a primary culture of DRG neurons from newborn rats was used, and attention was focused on the sodium currents in these cells. Sodium currents, in general, are important contributors to excitation in most types of neurons and, therefore, form an excellent target for substances inhibiting signal propagation. Without excluding the possibility that BAB mediates its effects via the modulation of other ion currents, e.g., Potassium sup + or Calcium sup ++, Sodium sup + currents were deemed the most beneficial candidates for a first investigation.

The current study primarily addresses the question of whether the two types of sodium current, present in DRG neurons, are influenced by application of BAB and, if so, in which way. The properties studied are the peak Sodium sup + conductance underlying the fast and slow Sodium sup + currents and the steady-state inactivation of these currents. A part of the results has been described in abstract form. [8].

Neonatal rats 1 day of age were killed by decapitation, and the DRGs from different levels of the spinal cord were rapidly collected under sterile conditions.* With a fine forceps, three or four ganglia were brought into a drop of culture medium on circular cover glasses coated with poly-D-lysine (Sigma) and placed in culture dishes. The ganglia were mechanically dissociated with the fine forceps, and the neurons were allowed to attach to the coated glass for 4 h in a humidified 5% CO2atmosphere at 37 degrees Celsius. Thereafter, we added approximately 1 ml chemically defined culture medium supplemented with nerve growth factor. [9]Within 24 h, the cells were used for patch clamp measurements. [10]To this end, the cover slip was mounted in a chamber [11]on the stage of an inverting microscope (Zeiss IM 35). In the patch pipette holder electrical contact between amplifier and pipette solution was made by an Silver-AgCl electrode. Patch-pipettes, pulled on a vertical two-stage electrode-puller, were from borosilicate glass (Clark GC-150 TF) and had tip diameters of approximately 3 micro meter. Their resistances varied between 1 and 4 M Omega.

Under current clamp conditions, 10% of the DRG neurons, cultured under our circumstances, exhibited Calcium sup ++ -action potentials. [12]To exclude them, we used Calcium sup ++ free extracellular solutions (ECS). The composition of the control ECS (control-ECScc) solution was as follows (in mM): NaCl 130, choline-Cl 10, KCl 5, MgSO4*symbol* 7H2O 2.5, N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulfonic acid] (HEPES) 10, and glucose 10. The second and third type of ECScchad a similar composition but contained, respectively, BAB** 100 nM (BAB-ECScc) of 200 nM of the fast Sodium sup + channel blocker tetrodotoxin (TTX; TTX-ECScc). The pipettes were filled with an intracellular-like solution (ICScc) composed as follows (in mM): NaCl 8, KCl 140, MgSO sub 4 *symbol* 7H2O 2.5, ethylene glycol-bis(beta-amino-ethylether)tetraacetic acid (EGTA) 5, HEPES 10, Na sub 2 ATP 2, and glucose 10.

For measurement of Sodium sup + currents under voltage-clamp (VC) conditions, the pipettes were filled with ICSvc(in mM): CsOH 15, EGTA 5, MgSO4*symbol* 7H2O 2.5, HEPES 10, CsCl 120, NaCl 8, tetraethylammonium-Cl (TEACl) 5, Na2ATP 2, and glucose 10. This type of ICS was used to isolate the Sodium sup + currents, because Caesium sup + and TEA sup + are inhibitors of the Potassium sup + current. Three types of ECSvcwere used. The composition of the first type (control-ECSvc) was in mM: NaCl 130, CsOH 5, HEPES 10, TEACl 10, MgSO4*symbol* 7H2O 2.5, 4-amino-pyridine (4-AP) 1, and glucose 10. Here too Caesium sup +, TEA sup +, and 4-AP were used to reduce the Potassium sup + current as much as possible. The absence of external Calcium sup ++ was necessary to block the Calcium sup ++ current without influencing the sodium current. [13,14]The second and third type of ECSvchad a similar composition as control-ECS sub vc, but included 200 nM TTX (TTX-ECSvc) and 100 micro Meter BAB (BAB-ECSvc).

The ECS was produced in one large batch, whereas the intracellular solutions were prepared for the use of a few weeks at most. ATP and glucose were added to the ICS before the experiment. All solutions were stored at 5 degrees Celsius, and their pH was adjusted to 7.4 immediately before the experiment. The dish containing the cells was continuously perfused with ECS, at a rate of 2 ml/min. Electrical valves in the perfusion system allowed quick changes (plus/minus 20 s) between the different types of ECS. All experiments were performed at room temperature, which was maintained at 19 degrees Celsius.

Current clamp (CC) and voltage clamp (VC) protocols were provided by a custom-built pulse generator. The membrane voltages or currents, recorded by the List EPC-7 amplifier, were filtered at 10 kHz (4-pole Bessel), fed into an A/D-convertor (Biomation), and stored on a PDP 11/73 minicomputer. The tip of the patch-pipette was sealed on the cell surface, and the patch membrane was ruptured by suction (10-40 kPa) to establish the whole cell clamp mode. [10]Under current clamp conditions, we applied the threshold protocol, consisting of 10-20 depolarizing current pulses of 10 ms, with an amplitude increasing from subthreshold to suprathreshold levels at a frequency of 1 Hz. The membrane voltage, at which the action potential appeared for the first time, was taken as the firing threshold of the neuron. [12]Supramaximally evoked action potentials were recorded each 6 s for 2 min or each 30 s for 10 min to follow the onset and recovery from the local anesthetic. Under voltage-clamp conditions, the neurons were clamped at the holding voltage EHof -70 mV. During hyperpolarizing voltage pulses to -140 mV, the capacity currents were compensated by adjusting the capacity cancellation settings of the amplifier. The access resistance to the cell interior was compensated for 50-80%. Readjustments of all compensations were necessary during the experiment. Residual capacity currents Ic(t) were digitally corrected. The ionic leakage current ILof the neuron was measured in the voltage range from -80 to -140 mV. As expected, this current, flowing through the voltage-independent ion channels, was linear and could be described by: Equation 1where gLis the leakage conductance and ELthe leak reversal potential. The parameters in Equation 1were used to subtract ILfrom the membrane currents measured at the different voltages. The measurement of the Sodium sup + current (INa) required a fast spatial control of the membrane voltage. Neurons cultured for 1 day had no or short axons (<< 10 micro meter), and thus a uniform potential could be expected. However, a large Sodium sup + current in conjunction with the residual access resistance could be responsible for inadequate space clamp. Therefore, we selected the smaller neurons, having diameters from 10 to 15 micro meter, to reduce the magnitude of the inward Sodium sup + current, which could be as high as 15 nA in larger cells. This selection limited the magnitude of INato about 5 nA. Only neurons in which the Sodium1+ current activated smoothly and where no delayed action currents appeared were accepted for the analysis of the data.

To understand the voltage protocols and the terminology, following is a simplified scheme of the main states of sodium channels [15]: Equation 2.

The transitions between the different states depend on membrane voltage. At hyperpolarization, Sodium sup + channels are closed; whereas at depolarization, they open (activation) and become inactivated. The transition from the inactivated to the closed state, thus the removal of inactivation, occurs at hyperpolarization. Two types of voltage protocols, under control of the PDP 11/73, were used: (1) the activation protocol, consisting of a 40-ms prepulse to -140 mV to remove inactivation and an 8-ms test pulse to measure the Sodium sup + current. The test pulse was increased from -100 to +100 mV in 10 mV steps at the repetition rate of 1 Hz; and (2) the inactivation protocol, consisting of a 40-ms prepulse, varying from -140 to 0 mV in 5-mV steps, to change the degree of inactivation and a subsequent test pulse to -10 mV. Both protocols were applied on the same neuron in control- and BAB-containing solutions after an incubation period of 10 or 2 min, if quick readjustments of the voltage-clamp allowed such a short measuring time.

Data were retrieved from the minicomputer to a PC-AT (Procom 386) for data analysis. Special customized software (Biomav) was used to correct the measured Sodium sup + currents for the contribution of ILand Ic(t). Curve-fitting was performed using the graphics program FigP6.0 (Biosoft). Statistical testing (Wilcoxon's signed-rank test) was performed with the program StatGraphics. The level of significance was chosen as 0.05. Data are given as mean plus/minus SEM, with n the number of investigated neurons.

Firing Threshold of Sodium Action Potentials

Action potentials were evoked from neurons with a stable resting membrane potential that was more negative than -50 mV. Figure 1, shows the effect of 100 micro Meter BAB on the excitability of 15 of 22 accepted neurons: an increase of the firing level Vthrand the disappearance of the fast rising phase of the action potential. The threshold voltage was raised from -29 plus/minus 2 mV in control-ECSccto -18 plus/minus 2 mV (n = 15) in BAB-ECScc, which were statistically different (P = 0.0001). Provided that the BAB-ECSccwas applied for 2 min or less, these effects were reversible. After washing-out, the threshold voltage was -27 plus/minus 2 mV, not statistically different from the control value (n = 15). The time to reach the peak of the action potential, taupeak, was significantly increased (P = 0.007) from 5.7 plus/minus 0.9 ms to 7.7 plus/minus 1.3 ms (n = 8). Its value became 6.1 plus/minus 1.0 ms after washing-out BAB, which did not differ statistically from the control value. In 7 of 22 neurons, the action potential was blocked by a 2-min application of 100 micro Meter BAB in a reversible way. The threshold voltage of -26 plus/minus 3 mV, determined in control-ECScc, did not differ statistically from the one obtained after washing-out BAB. All investigated neurons exhibited a pronounced effect on their excitability in the presence of 100 micro Meter BAB.

The DRG culture is a heterogeneous population of neurons, which contain varying proportions of fast and slow Sodium sup + channels. [16]To relate the BAB effects to the possible blockade of fast or slow Sodium sup + channels, the toxin TTX was applied before BAB perfusion or directly after washing-out the local anesthetic. At a concentration of 200 nM, TTX selectively blocked the fast Sodium sup + current in our cultured cells. [16]For the first group of neurons (n = 15), the effect of TTX proved similar to that of BAB: The firing threshold and the value of taupeakwere increased significantly, respectively, from -25 plus/minus 2 mV to -18 plus/minus 1 mV (P = 0.0028) and from 6 plus/minus 1 ms to 8 plus/minus 2 ms (P = 0.017) in a reversible way. This strongly suggested that BAB in a concentration of 100 micro Meter inhibited fast Sodium sup + channels and that the BAB-resistant action potential originated from activation of the slow Sodium sup + channels. In the second group of neurons (n = 7), in which BAB blocked the action potential, TTX also reversibly blocked it, supporting the view that fast Sodium sup + channels are blocked by BAB.

During perfusion with BAB or during its washing-out, we measured the supramaximally evoked action potential each 6 s and followed amplitude changes as function of time. The time needed for BAB to depress neuron's excitability, tauonset, amounted to 50 plus/minus 5 s (n = 13). The recovery of the BAB effect in control-ECS was accomplished in a time taurecovof 96 plus/minus 7 s (n = 11), and two neurons needed 2-3 min to recover. This reversible action was only clear when the neurons were incubated in BAB for about 2 min. In eight neurons, incubation with BAB for 10 min did dramatically prolonged the recovery time to 420 plus/minus 92 s (n = 3) or did not recover within 10 min (n = 5). Thus, the washing-out of BAB depends on the exposure time.

Peak Sodium Conductance

Several minutes after the establishment of the whole-cell voltage clamp configuration, outward Potassium sup + currents were blocked by Caesium sup + and TEA sup +, diffusing from the pipette into the cell interior. Under these conditions, inward and outward Sodium sup + currents could be recorded in response to a series of depolarizing test pulses (activation protocol). Figure 2shows typical Sodium sup + currents, measured at four test voltages. At membrane voltages of -30 and -20 mV, the fast Sodium sup + current INa,F dominated, activating within 1 ms and inactivating within a few milliseconds. However, at more depolarized voltages, a second slower activating and inactivating inward Sodium sup + current INa,S started to predominate. The fast currents were blocked by 200 nM TTX within 12 min, leaving the slow current component only. Switching back to control solution, the fast current recovered within 10 min. Most of the investigated neurons exhibited both types of Sodium sup + current, in agreement with published reports. [13,17]Of the 20 neurons, which met our acceptance criteria, 5 neurons showed predominantly the fast Sodium sup + current, whereas in 2, the slow Sodium sup + current was the sole component.

The action of BAB on the Sodium sup + current was measured after incubation with the local anesthetic for 2 or 10 min (Figure 2). No obvious changes in the time course and amplitude of the current were observed under these conditions (activation protocol). After washing-out BAB for 10 min. we recorded the currents again to demonstrate the absence of neuronal degeneration. The peak values of the Sodium sup + current were plotted as function of membrane voltage. An example of the total peak current-voltage relation is shown in Figure 3(A). The voltage at which the Sodium sup + currents reversed, the Sodium sup + reversal potential ENa, was in the vicinity of the calculated Sodium sup + Nernst potential of +60 mV. From the current-voltage relation, we calculated the peak Sodium sup + conductance gp,T (conductance: reciprocal of resistance to Sodium sup + ions) underlying Ip,max according to Hodgkin and Huxley [18]: Equation 3.

The magnitude of gp,T and ENain control solutions was, respectively, 48 nS and 66 mV. In the presence of BAB, for 2 or 10 min, the values of ENa(+68 mV) and gp,T (47 nS) were not significantly different from the values obtained in control solutions (Table 1). After the 10-min wash-out, the values of gp,T and of E sub Na were not statistically different from their values obtained under control conditions (Table 1).

The fact that gp,T was not affected by BAB seems to contradict the TTX-like effect of BAB on the firing threshold and action potential. However, during the standard activation protocol, the hyperpolarizing pre-pulse to -140 mV may remove BAB-induced inactivation of the Sodium sup + current. To investigate the role of the conditioning prepulse, we applied the activation protocol without the hyperpolarizing prepulse. Figure 3(B) shows the current-voltage relation under this condition for a neuron that exhibited predominantly INa,F. The peak currents considerably decreased in the presence of BAB. In five neurons, in which INa,F dominated, the magnitude of gp,T in control was 3.0 plus/minus 1.0 nS, whereas in BAB, it was reduced to 0.64 plus/minus 0.26 nS. The ratio of gp,T in control and in BAB-containing solutions was 0.21 plus/minus 0.01. After applying the hyperpolarizing prepulse, gp,T was 17 plus/minus 7 nS in control and in the presence of BAB, and their ratio was 1.0 plus/minus 0.1 (n = 5). Thus, we can conclude that BAB induced a strong inactivation of the fast Sodium sup + channels, which could be removed by hyperpolarization. After a 2-min exposure with BAB, the reduction of the currents could be reversed by washing-out the anesthetic (Figure 4). The ratio of gp,T after a 10-min wash-out and its control value was 0.70 plus/minus 0.14 (n = 4), thus, for the most part, reversible within this period.

Separation of Fast and Slow Sodium sup + Currents

To separate the Sodium sup + current in its fast and slow components at the test voltage of -10 mV, we considered that the fast current was inactivated at prepulse potentials, whereas the slow current showed no inactivation. [13]Applying our inactivation protocol, the Sodium sup + currents were obtained at the test potential of -10 mV as a function of prepulse potential (Figure 5(A)). When the prepulse was varied from -120 to -50 mV, the amplitude of the fast peak changed from maximal to zero, because of the progressive inactivation of INa,F. In the range of -70 to -40 mV, the slow current hardly changed, indicating the lack of inactivation of INa,S. At prepulse voltages between -40 and 0 mV, the slow current gradually disappeared, indicating that it became inactivated at these voltages. At the prepulse level of about -50 mV, we obtained the noninactivated slow sodium current INasup -50 (t) during the test pulse. By subtracting this current from the Sodium sup + current INa,TEpp(t) at a certain prepulse potential Eppaccording to Equation 4we could obtain the fast Sodium sup + current component as function of Epp, with Eppmore negative than -50 mV. Figure 5(B) shows this procedure with Epp= -140 mV, demonstrating the isolation of the fast activating and inactivating INa,F, which reached its maximum amplitude at this prepulse voltage. From the data, we calculated gp,F according to Equation 3, which was found to be 26 nS (Table 1). After 2- or 10-min BAB perfusion, gp,F was 25 nS, not significantly different from the control value. Similarly, from the peak value of INasup -50 (t) and using Equation 3, we calculated gp,s in control-ECSvcand in BAB-ECSvcas 17 and 13 nS, respectively. This difference was also not significant (Table 1). After the wash-out period, the magnitude of gp,F and gp,s did not differ statistically from the control values.

Inactivation of the Fast and Slow, Sodium sup + Currents

Steady-state inactivation of INa,F and INa,S was studied by varying the level of Epp(inactivation protocol). From Figure 5(A), it is apparent that, at voltages of about -120 mV, the peak of the fast component became maximal, indicating that, at these potentials, its inactivation was removed. Therefore, we limited the amplitude of the prepulse to -140 mV. The fast and slow Sodium sup + -currents were separated according to Equation 4as a function of prepulse voltage. From the peak amplitude of the Sodium sup + current component IpEpp, using a certain prepulse and its maximum amplitude Ipsup -140, obtained with the largest hyperpolarization, we calculated the ratio IpEpp/Ipsup -140. This ratio is equal to the Hodgkin and Huxley inactivation parameter halpha[18]. The relation between halphaof the fast and slow currents was plotted as function of prepulse potential (Figure 6). The halphaversus Epprelation was fitted with a Boltzmann equation of the form [16]: Equation 5where E50is the midpoint potential (50% inactivation) and k the steepness of the curve (slope factor). An example of the steady-state inactivation curves of INa,F and INa,S before and after exposure of 10 min to BAB, is given in Figure 6. The lines were obtained by fitting Equation 5to the experimental data, which yielded the parameters E50and k. Because we will compare these parameters after perfusion with BAB and after washing-out, we determined them in control solution and after a repeated application of control solution or after washing-out TTX. Thus we could follow their values at different times (t), varying from 5 to 30 min after the start of the experiment. The changes Delta E50and Delta k (Delta: value at time t - initial value) were analyzed as function of time. Linear regression of pooled Delta E50,F data on time yielded a regression coefficient (rc) of -0.154 plus/minus 0.043 (n = 10) mV/min, significantly deviating from zero (P = 0.002). Therefore, during a period of 10 min, the value of E50,F may shift spontaneously 1.5 mV to hyperpolarizing voltages. Changes in the value of E50,S (rc = 0.011 mV/min) and in the slopes kF(rc = 0.016 mV/min) and kS(rc = -0.017 mV/min) were not statistically different from zero.

The value of the midpoint potentials of the fast current E sub 50,F obtained in control-ECSvcwas -88 plus/minus 2 mV (n = 19, pooled data). After a 2- or 10-min exposure to 100 micro Meter BAB, the magnitude of E50,F became -100 plus/minus 2 mV (n = 19), significantly different from its value in control solution (P = 0.0001). After the wash-out period, the value of E50,F of neurons exposed to BAB for 10 min remained -100 mV, not statistically different from the value obtained in BAB. However, when BAB was applied for 2 min, it reversed during washout to -90 mV (Table 2(A/B)), statistically different from the value obtained in BAB (P = 0.0001). Therefore, BAB increased the steady-state inactivation of the fast Sodium sup + channels, and this action could be reversed within 10 min only after the short-lasting perfusion. The significant difference (P = 0.0012) between wash-out and control (Table 2(B)) may arise from an incomplete washout or from the spontaneous shift of the inactivation curve. The value of the midpoint potential of the slow current E50,S amounted to -28 plus/minus 2 mV (n = 20) in both control and BAB-containing solutions. In neurons surviving the BAB treatment, its value was -29 plus/minus 2 mV (n = 13) after washing-out BAB (Table 2). The slope factor of the fast current kFwas 11 plus/minus 1 mV (n = 19) in control and 10 plus/minus 1 mV (n = 19) in the presence of BAB. For the slow component, the slope factor kSwas 4.9 plus/minus 0.4 mV (n = 20) in both control and BAB-containing solutions. Neither slope factor changed significantly in the presence of BAB or after washing-out the local anesthetic (Table 2).

The important findings of our study with 100 micro Meter BAB are (1) an 11-mV increase of the firing threshold or action potential blockade; (2) a 12 mV shift of the inactivation curve of the fast Sodium sup + channels to hyperpolarizing membrane voltages; (3) at exposure times of 2 min, but not of 10 min, the shift of the inactivation curve of the fast Sodium sup + channels was reversible within a 10-min wash-out; (4) no shift of the inactivation curve of the slow Sodium sup + channels; and (5) no alteration of the maximum peak conductances underlying the fast and slow Sodium sup + current.

The increase of the firing threshold and the increase of the latency to the peak of the action potential in BAB were similar to the action of TTX. Action potentials blocked by BAB were also abolished by TTX. These findings indicated that, under the current conditions, the fast Sodium sup + channels are the primary target of the local anesthetic. Because the resting membrane potential, determined by Potassium sup + channels, [12]was not altered by BAB (Figure 1), these channels are probably insensitive for the local anesthetic. The action of BAB on the Sodium sup + channels was fast. Considering that the time needed to replace the bath solution is 20 s, the effect of BAB on firing is accomplished within 30 s, whereas it is reversed within 80 s after the short exposure time of 2 min. However, after 10 min of BAB perfusion, its effects on Sodium sup + action potentials and Sodium sup + currents were not reversible within 10 min in most of the investigated neurons. Stable whole-cell patch-clamp recordings can last for about 30 min, after which an unacceptable rundown of the fast and slow currents may take place. This limited the available time to study wash-out effects to about 10 min. During a longer wash-out period, the effects after the 10-min exposure may prove reversible. Selection of neurons, which survive much longer than half an hour, may answer this question, although the variation of E50,F with time will mask reversibility. The dependence of recovery and exposure time is consistent with the extremely lipophilic nature of BAB. The long-lasting action of the drug in DRG neurons may be explained by either the high partition-coefficient of 1,028,*** resulting in a store of BAB in the lipid membrane, or (2) a specific high affinity to binding sites at or near the fast Sodium sup + channel.

The fact that BAB exerts its effect only on the midpoint potential of the fast sodium current is in agreement with the TTX-like effect of BAB on the firing properties. Because of possible spontaneous shifts of the inactivation curve, the mean E50,F change of 12 mV in the hyperpolarizing direction has to be regarded as an upper bound. The shift reflects an increase in the number of inactivated fast Sodium sup + channels by BAB at a given membrane potential. Therefore, at the resting membrane potential, for example, the available number of Sodium sup + channels is considerably reduced, which will cause a substantial reduction of membrane excitability. The rise of the firing threshold and action potential blockade in the investigated neurons are in agreement with this prediction. The relatively large increase of the firing threshold in vitro may result in a complete action potential block in vivo, where small propagated action currents may not be able to cross the raised firing threshold. [19]The clinical observation that epidurally applied BAB has a long-lasting analgesic effect in patients with terminal cancer [1,2]is in accordance with the observed decrease of excitability.

Shifts of Sodium sup + inactivation curves have been found for various anesthetics. In frog myelinated axon, the ethyl-analogs (1 mM) of BAB, benzocaine, metacaine, and ethyl-o-aminobenzoate, shifted the inactivation curves respectively 24, 19, and 17 mV to the hyperpolarizing direction. [20]However, contrary to our findings, the maximum peak Sodium sup + conductances in these studies were considerably reduced. In the same preparation, lidocaine (1 mM) shifted the midpoint potential 30 mV toward hyperpolarizing voltages, while affecting the maximum Sodium sup + conductance. [21]Compared to these anesthetics, BAB only caused a relatively small shift of the inactivation curve, without reducing the maximal Sodium sup + conductance. This difference may originate from the tenfold lower concentration of BAB in our test solution or from a specific property Concentration-dependent shifts of the midpoint potential have been found for carboxylic ester anesthetics in squid axon. [22]Shifting the Sodium sup + inactivation curve is a common feature of local anesthetics and antiepileptic agents. [23]In addition to such shifts, frequency- or use-dependent effects are important. For instance, lidocaine (50-200 micro Meter) reduced the slow Sodium sup + current in a marked use-dependent manner, contrary to its blocking effect on the fast Sodium sup + channels. [16]These effects depend on the recovery time from the blocked or inactivated Sodium sup + channels. In this respect, it is notable that the recovery from the increased inactivation due to BAB, e.g., at the holding voltage, is completed within the duration of the prepulse of 40 ms, whereas lidocaine needs tens of seconds. [21]The fast (partial) recovery from block at hyperpolarization is similar to that of benzocaine, showing a recovery time constant of 27 ms at -120 mV. [24]After the hyperpolarizing prepulse, the Sodium sup + currents, measured at depolarization, seem to be unaltered by the local anesthetic. It is tempting to speculate that, during hyperpolarization, the fast channel is unblocked and the voltage-dependent binding of BAB during depolarization is too slow to affect it, whereas at holding, potential channel block can reach a steady-state level. Such an interpretation favors the modulated receptor hypothesis. [25,26]However, various kinetic analyses should be performed to discriminate between different possibilities. [26].

In the current study, we selected small neurons (< 20 micro meter), which probably correspond to the slow conducting A delta- and C-fibers. [27]Thus, our results seem applicable to pain neurons. Results from C-fibers of the isolated sural nerve indicated that the ED sub 50 value for the magnitude of the compound action potential equals 50-100 micro Meter BAB.**** Therefore, the experiments described here were conducted using a test solution containing 100 micro Meter BAB. The local anesthetic, dissolved from the epidurally injected suspension, will diffuse into the surrounding target tissues, i.e., dorsal root neurons and myelinated and unmyelinated nerve fibers. A concentration of at least 100 micro Meter can be expected at these sites from the maximum solubility of BAB in water, amounting to approximately 1,000 micro Meter/l. Whether concentration gradients along the nerves in conjunction with length-dependent block [4]play a role, is not known. The interesting question why BAB selectively affects sensory neurons remains to be investigated.

In the neonatal DRG neurons, cultured for about 1 day, we found two types of Sodium sup + current. Fast and slow Sodium sup + currents have been found in DRG neurons, [14,28-30]and in newborn rat, a third type of Sodium sup + current the INa,FN may be present. [13,17]The maximum peak inward current of INa,F occurs at potentials varying between -32 mV [30]and -12 mV [13], whereas the maximum inward peak of INa,S occurs at potentials of about -5 mV. Our recordings agree with these findings. The INa,FN, which is faster than the INa,F, should reach its maximum at -20 mV, and its midpoint potential is reported to be -139 mV [13]. However, the constant amplitude of the peak of the fast current at prepulses up to -140 mV excluded significant contributions of INa,FN to the Sodium sup + currents that we were measuring. Under control conditions, our value of E50,F of -88 mV is more negative than reported values of -65 mV [30]to -75 mV [13]. Conversely, the value for E50,S of -28 mV is in the reported range from -23 mV [13]to -43 mV [30]. The difference between our estimate of E50,F and reported values may originate from differences in the composition of the saline solutions and culture media. Under physiologic conditions, the resting membrane potential of the cultured DRG neurons amounts to -64 mV, and the firing threshold of Sodium sup + action potentials was -25 mV. [12]With the current midpoint potential, all fast Sodium sup + channels inactivated at the firing level, indicating that the value of E50,F will be more positive than -88 mV under more physiologic circumstances.

The results reported here concern isolated sensory neurons from neonatal rats at room temperature. Although this preparation differs from the adult humans with whom the clinical study [2]was performed, both isolated neurons and intact humans share a substantial number of properties. The fact that the results from the clinical study and the current one on isolated neurons apparently agree may lead to further understanding of the mechanism of action of long-lasting selective segmental analgesia after epidural administration of a 10% BAB-solution.

The authors thank M. Deenen, for assistance with the culturing of the dorsal root ganglion neurons, and Dr. D. L. Ypey, for support and discussions.

*This procedure had been approved by the Animal Ethical Committee at the Leiden Medical Faculty (approval number FFF-10), as required by the Dutch Law.

**BAB was dissolved in ECS: aqueous solubility 140 mg/l, molecular weight 193.2;pK:2.6.

***Grouls RJE, Ackerman EW, Machielsen EJA. Korsten HHM: Preparation, characterization and quality control of a suspension injection for epidural analgesia. Pharm Wkbld (Sci Edition) 13:13-17, 1991.

****Van den Berg RJ: Unpublished observation. 1994.

Shulman M: Treatment of cancer pain with epidural butyl-amino-benzoate suspension. Reg Anesth 12:1-4, 1987.
Korsten HHM, Ackerman EW, Grouls RJE, van Zundert AAJ, Boon WF, Bal F, Crommelin MA, Ribot JG, Hoefsloot F, Slooff JL: Long-lasting epidural sensory blockade by n-butyl-p-aminobenzoate in the terminally ill intractable cancer pain patient. ANESTHESIOLOGY 75:950960, 1991.
Franz DN, Perry RS: Mechanisms for differential block among single myelinated and non myelinated axons by procaine. J Physiol (Lond) 236:193-210, 1974.
Raymond SA, Steffensen SC, Gugino L, Strichartz GR: The role of length of nerve exposed to local anesthetics in impulse blocking action. Anesth Analg 68:563-570, 1989.
Fink BR: Mechanisms of differential axial block in epidural and subarachnoid anesthesia. ANESTHESIOLOGY 70:851-858, 1989.
Baccaglini PI, Hogan PG: Some rat sensory neurons in culture express characteristics of differentiated pain sensory cells. Proc Natl Acad Sci USA 80:594-598, 1983.
Korsten HHM, Ackerman EW, Grouls RJE: The true epidural space: Possible consequences for administered material-based drug delivery systems (letter). ANESTHESIOLOGY 76:866, 1992.
Van Soest PF, Van den Berg RJ, Grouls RJE, Korsten HHM: Action of n-butyl-p-aminobenzoate on sodium currents in cultured dorsal root ganglion neurons (abstract). Eur J Neurosci 6(suppl):245, 1993.
Romijn HJ, Huizen F, Wolters PF: Towards an improved serum-free, chemically defined medium for long term culturing of the cerebral cortex tissue. Neurosci Biobehav Rev 8:301-334, 1984.
Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ: Improved patch-clamp techniques for high resolution current recording from cells and cell-free membrane patches. Pflugers Arch 391:85-100, 1981.
Ince C, Ypey DL, Diesselhoff-den Dulk MMC, Visser JAM, De Vos A, van Furth R: Micro CO sub 2 incubator for use on a microscope. J Immunol Methods 60:269-275, 1983.
Wang Z, Van den Berg RJ, Ypey DL: Resting membrane potentials and excitability at different regions of rat dorsal root ganglion neurons in culture. Neuroscience 60:245-254, 1994.
Schwartz A, Palti Y, Meiri H: Structural and developmental differences between three types of Sodium channels in dorsal root ganglion cells of newborn rats. J Membr Biol 116:117-128, 1990.
Ikeda SR, Schofield GG: Tetrodotoxin resistant sodium current at nodose neurons: Monovalent cation selectivity and divalent cation block. J Physiol (Lond) 389:255-270, 1987.
Hille B: Ionic Channels of Excitable Membranes. Sunderland Sinauer, 1992, pp 485-503.
Roy ML, Narahashi T: Differential properties of tetrodotoxin-sensitive and tetrodotoxin-resistant sodium channels in rat dorsal root ganglion neurons. J Neurosci 12:2104-2111, 1992.
Omri G, Meiri H: Characterization of sodium currents in mammalian sensory neurons cultured in serum-free defined medium with and without NGF. J Membr Biol 115:13-29, 1990.
Hodgkin AL, Huxley AF: A quantitative description of membrane current and its applications to conduction and excitation in nerve. J Physiol (Lond) 117:500-544, 1952.
Van den Berg RJ, Versluys CAA, de Vos A, Voskuyl RA: Nerve fibre size related block of action currents by phenytoin in mammalian nerve. Epilepsia 35:1279-1288, 1994.
Koppenhofer E, Sommer RG, Froese U: Effects of benzocaine and its isomers on the sodium permeability and steady state sodium inactivation in myelinated nerve, obtained by an improved dissection technique. Gen Physiol Biophys 6:209-222. 1987.
Hille B: Local anesthetics: Hydrophilic and hydrophobic pathways for the drug-receptor interaction. J Gen Physiol 69:497-515, 1977.
Elliott JR, Murrell RD, Haydon DA: Local anesthetic action of carboxylic esters: Evidence for the significance of molecular volume and for the number of sites involved. J Membr Biol 95:143-149, 1987.
Van den Berg RJ, Kok P, Voskuyl RA: Valproate and sodium currents in cultured hippocampal neurons. Exp Brain Res 93:279-287, 1993.
Hondeghem LM, Katzung BG: Antiarrhytmic agents: The modulated receptor mechanism of action of sodium and calcium channel-blocking drugs. Ann Rev Pharmacol Toxicol 24:387-423, 1984.
Courtney KR: Local anesthetics. Int Anesthesiol Clin 26:239-247, 1988.
Chernoff DM, Strichartz GR: Kinetics of local anesthetic inhibition of neuronal sodium currets: pH and hydrophobicity dependence. Biophys J 58:69-81, 1990.
Harper AA, Lawson SN: Conduction velocity is related to morphological cell type in rat dorsal root ganglion neurons. J Physiol (Lond) 359:31-46, 1985.
Fukuda J, Kameyama M: Tetrodotoxin-sensitive and tetrodotoxin-resistant sodium channels in tissue cultured spinal ganglion neurons from adult mammals. Brain Res 182:191 - 197, 1980.
Bossu JL, Feltz A: Patch-clamp study of the tetrodotoxin-resistant sodium current in group C sensory neurones. Neurosci Lett 51: 241-246, 1984.
Ogata N, Tatebayashi H: Comparison of two types of Sodium sup + currents with low-voltage-activated T-type Calcium sup 2+ current in newborn rat dorsal root ganglia. Pflugers Arch 420:590-594, 1992.