Treatments of chronic pain may cause severe side effects, among which motor dysfunction is most prominent. A relatively new and promising approach to chronic pain treatment is the epidural administration of an aqueous suspension of the local anesthetic n-butyl-p-aminobenzoate (BAB), also known as butamben (1–3). Application of the BAB suspension to the spinal dura results in long-lasting (median, 29 days) pain relief without impairment of motor function. This indicates that the BAB suspension selectively inhibits pain-signaling sensory neurons, but the mechanism of its specific analgesic action is unknown.
The BAB molecule is an aminobenzoate ester-linked to a butyl group. The structure is similar to that of other ester-linked local anesthetics, such as benzocaine and procaine, which profoundly affect sodium channels involved in impulse generation and transmission in neurons. The effects of BAB on sodium currents have previously been studied in small dorsal root ganglion (DRG) neurons (4,5). However, the widespread opinion that the action mechanism of local anesthetics is mediated by sodium channels alone is, particularly for epidural anesthesia, unproven (6).
Recently, we have shown in DRG neurons an effect of BAB on potassium channels, and Kv1.1 channels in particular, which could contribute to the analgesia caused by the BAB suspension (7). Calcium channels also play an important role in action potential firing of sensory neurons. A variety of calcium channel subtypes are expressed in sensory neurons of rodents (8,9). In the rat, the N-type calcium current comprises ∼50% of the total calcium current and is involved in calcium entry during action potentials in small-diameter DRG neurons (10–12), which include the pain-sensing neurons (13).
In this study, we addressed the question of whether voltage-activated calcium channels are affected by BAB. To this end, the patch-clamp technique in a whole-cell voltage-clamp configuration was applied to acutely isolated small-size DRG neurons from neonatal mice. We did find inhibitory effects of BAB on the whole-cell current through calcium channels, including the N-type component. The physiological significance of these findings is considered in the Discussion.
Neonatal mice were killed by decapitation, and DRGs from all accessible levels of the spinal cord were rapidly collected (approved by the Animal Ethics Committee at the Leiden University Medical Center). Cells were mechanically dissociated from two or three ganglia and cultured on a circular glass coverslip as previously described (7). Within 8 h of culture, spherical neurons with a diameter of ∼20 μm were selected for patch-clamp measurements. At this stage, neurite outgrowth was still negligible.
For voltage-clamp experiments, a coverslip with DRG cell culture was mounted in a chamber on the stage of an inverted microscope. Patch pipettes were pulled from borosilicate glass (Clark GC-150 TF-15) and had resistances of 2.0 to 2.5 MΩ measured in the standard bath solution. Sintered Ag/AgCl electrodes coupled the amplifier input leads to the solutions. To minimize offset caused by low-Cl− pipette solutions, the pipette holder (14) contained a Cl−-rich solution at the Ag/AgCl electrode.
Gigaseals were made in a microbath of ∼75 μL and continuously perfused (∼300 μL/min) with the standard bath solution (mM): NaCl 145, KCl 5, CaCl2 2, MgCl2 1, and HEPES 10 (pH 7.4; NaOH). The pipette solution contained (mM) Cs-methanesulfonate 103, MgCl2 4, HEPES 9, EGTA 9, (Mg)adenosine triphosphate 4, (tris)guanosine triphosphate 1, and (tris)phosphocreatine 14 (pH 7.4; CsOH).
After establishment of the whole-cell configuration, the barium current was recorded with voltage-step protocols during extracellular perfusion with (mM) Tetra-Ethyl-Ammonium (TEA)-chloride 160, HEPES 10, EGTA 0.1, and BaCl2 5 (pH 7.4; TEA-OH). A personal computer running Clampex 7 (Axon Instruments, Foster City, CA) and a List EPC 7 amplifier provided voltage protocols. Up to 80%–90% of the series resistance was compensated. The membrane currents were filtered at 3 kHz in general and at 10 kHz for tail current measurements. Control experiments with an equivalent RC circuit (R = resistance, C = capacitance) of the whole cell showed that current transients with time constants of >100 μs can be reliably measured at the 10-kHz filter setting (three-pole Bessel) under our conditions. All currents were leak-subtracted by using the P/4 method. Membrane capacitance of the selected DRG neurons was 14 ± 3 pF (n = 55). Calcium currents during action-potential clamp were measured under constant perfusion with (mM) TEA-Cl 160, HEPES 10, and CaCl2 2 (pH 7.4; TEA-OH). The pipette solution was the same as above in the step voltage-clamp conditions.
BAB (OPG Farma, Utrecht, The Netherlands) was added to the extracellular solution from a stock of BAB in ethanol (1–500 mM). The final ethanol concentration never exceeded 0.1%. Because BAB has low water solubility (<700 μM at room temperature; Merck Index 1989) and easily binds to the plastic surfaces of the perfusion system, final BAB concentrations up to 500 μM were verified by using absorption spectrophotometry (290 nm). ω-Conotoxin-GVIA (CnTx; Peptide Institute Inc., Osaka, Japan) was dissolved in distilled water and added with a final fully blocking concentration of 3.3 or 5 μM (8,10,11,13).
Normalized data were corrected for rundown in the presence of vehicle (0.1% ethanol) at all potentials measured in control experiments (n = 8). For example, at test pulses of 0 mV, an apparent linear barium current (I) decline (rundown) of ∼6% in 5 min was observed. The concentration-inhibition data were fitted by using the Hill equation:
where Io is the current at [BAB] = 0 and the IC50 is the concentration at which the current is reduced by 50% and n is the Hill coefficient. Results are presented as mean ± sd for n cells, unless stated otherwise, and were compared by using paired or independent Student’s t-tests with the level of significance (P) chosen as 0.05.
To assess the effect of BAB on calcium channels in neonatal mouse DRG neurons, barium currents through these channels were elicited by a square test pulse to 0 mV from −80 mV. Application of 200 μM BAB resulted in a decline of the current amplitude, which reached a steady-state value after 2–3 min. In Figure 1A, the steady-state effect of 200 μM BAB is shown for a representative cell. At this BAB concentration, the reduction of the peak whole-cell barium current amounted to 49% ± 7% (n = 8). During washing out of the drug, the inhibiting effect of BAB proved to be partly reversible, reaching 86% of the amplitude of the control situation (5).
The concentration dependency of the current reduction by BAB was determined by measuring steady-state barium currents at different BAB concentrations. A prepulse to −120 mV was applied to remove possible inactivation at −80 mV (<10%). The peak currents in the presence of BAB were normalized to the corresponding whole-cell peak currents in the absence of BAB and were plotted as a function of concentration in the range from 1 to 500 μM (Fig. 1B). A single Hill function could be fitted to the data, and this yielded an IC50 of 207 ± 14 μM and a Hill coefficient of 1.7 ± 0.2 (n = 40).
N-type calcium channels have a selective sensitivity to CnTx. To isolate the N-type current component, we used the procedure shown in Figure 2, A and B. Whole-cell barium currents were measured in the absence and presence of 3.3 μM CnTx, and the CnTx-insensitive currents were subtracted from the control currents (Fig. 2B, left). CnTx caused an inhibition in peak current of 58% ± 5% (n = 6). To investigate whether N-type currents are affected by BAB, currents were measured after preincubation with, e.g., 200 μM BAB and after subsequent perfusion with the CnTx solution still containing 200 μM BAB (Fig. 2A, right). The resulting difference current represents the current through N-type channels in the presence of 200 μM BAB (Fig. 2B, right). By repeating this procedure at different BAB concentrations and by plotting the normalized current density (pA/pF; pA = picoAmpere, pF = picoFarad cell membrane capacitance) as a function of BAB concentration, the N-type concentration-response curve was obtained (Figure 2C, solid curve). Fitting the Hill equation to this relation yielded an IC50 of 220 ± 35 μM and a Hill coefficient of 1.4 ± 0.3 (n = 35).
The current decay of the N-type component during maintained depolarization (500 ms) was fitted with a single exponential function. The mean time constant in control solution was 78 ± 12 ms (n = 8), whereas in the presence of 200 μM BAB, the N-type current inactivated significantly faster, with a time constant of 64 ± 8 ms (n = 7; P = 0.024).
The tail current (Fig. 2B), representing the deactivation of the N-type channels, was elicited by stepping back from 0 to −80 mV and could also be fitted by an exponential function (fits not shown), yielding a time constant in control conditions of 167 ± 24 μs (n = 6). In the presence of 200 μM BAB, the time constant was 136 ± 20 μs (n = 6)—significantly less (P = 0.043) than that obtained under control conditions.
The residual current in the presence of CnTx represents the non-N-type current through calcium channels. Its BAB concentration-response curve is also given in Figure 2C (dotted curve) and is characterized by an IC50 of 189 ± 28 μM and a Hill coefficient of 1.1 ± 0.2. Because of its heterogeneity, the non-N-type current was not further investigated.
We previously used voltage steps to elicit barium currents. However, under physiological conditions, calcium ions are the charge carriers, and the calcium channels are activated by naturally occurring changes in the transmembrane potential, e.g., during an action potential. Therefore, in this study, we also measured effects of BAB on calcium channels in the presence of a physiological concentration of calcium ions in the extracellular solution and used a previously recorded action potential as command voltage.
Figure 3A shows a representative action potential from a DRG neuron with a resting potential of approximately −75 mV. Although there is a marked neuron-to-neuron variability in action potential shapes, this action potential was applied as a standard voltage profile to get insight into the participation of the different calcium channels in the generation of the action potential. The 20-ms digitized standard action-potential record was applied from the holding voltage of −80 mV. The resulting control ion current is depicted in Figure 3B and shows an initial outward current followed by an inward calcium current. This short outward current is resistant to application of 600 μM cadmium (blocking all inward current; n = 5) and does not interfere with the measurement of the subsequent inward calcium current. This current is likely carried by cesium ions flowing through unblocked (fast) sodium and potassium channels (11). The peak of the inward current coincided with the shoulder in the repolarizing phase of the action potential. The inward current decayed in two phases: an initial fast and subsequent slow one. The slower current decay occurred after nearly complete repolarization of the action potential, i.e., during the after-depolarization.
BAB caused an overall decrease of the action potential clamp-evoked calcium current (Fig. 3C). The initial positive current, the large negative peak, and the fast- and slow-decay components were all affected. Figure 3D gives the concentration-response curve (dotted line) for the effect of BAB on the normalized integral (to reduce variability) of the total inward current. The parameters of the fitted Hill curve were an IC50 of 206 ± 8 μM and a Hill coefficient of 1.3 ± 0.1 (n = 47).
During voltage clamping with the standard action potential, CnTx was perfused over the cell to specifically block the N-type current. CnTx did not affect the initial outward current or the slower component of the current decay, whereas the faster decaying current was removed to a great extent (Fig. 3B, top). The difference between the total current and the current in the presence of CnTx is the N-type calcium current (Fig. 3B, bottom). CnTx blocked 48% ± 10% (n = 8) of the integrated current corresponding to the total calcium inflow. To determine the effect of BAB on the N-type component of the calcium current, action potential-clamped currents were measured after preincubation with, e.g., 100 μM BAB and after the subsequent perfusion with the CnTx solution, still containing 100 μM BAB (Fig. 3C, middle). The resulting difference current represents the current through N-type channels in the presence of 100 μM BAB (Fig. 3C, bottom). By repeating this procedure at different BAB concentrations, the concentration dependency of BAB on the CnTx-sensitive current was determined. The relation of BAB and the normalized integral of the CnTx-sensitive current was described with a Hill equation (Fig. 3D, solid curve), yielding an IC50 of 177 ± 47 μM and a Hill coefficient of 1.4 ± 0.5 (n = 47), similar to the parameters of the total calcium current (see above). For non-N-type calcium currents, a similar concentration-response curve is implicated by the very similar curves in Figure 3D.
This study shows that the local anesthetic, butamben (BAB), inhibits voltage clamp-evoked barium and calcium currents, including their N-type components. Unlike sodium channels, in which the effects of 100 μM BAB ranged from a nearly complete block to insensitivity (4,5), calcium and potassium channels show similarities in the measured effects of BAB. On both calcium (native) and Kv1.1 channels (native and cloned) (7), BAB caused an inhibition of the current with an IC50 of ∼200 μM, a Hill coefficient of 1–2, and accelerated deactivation. These data allow the possibility of two BAB binding sites per channel and suggest an allosteric mechanism of BAB action, by which the channel is biased toward the closed state. However, more experiments are needed to come to more definite conclusions about the mechanism of current reduction by BAB.
By using the action potential clamp and minimizing sodium and potassium currents, the total and N-type calcium currents flowing during a standard action potential could be measured. The shoulder of the action potential coincided with the peaks of the inward currents, as in rat sensory neurons (10,11). The finding that CnTx eliminated approximately half of the total charge displaced through the calcium channels during the applied action potential suggests a significant role for N-type calcium currents in nociceptive neurons. To illustrate the contribution of the N-type calcium current to the action potential wave form, Figure 4 shows the effect of CnTx on an action potential evoked in current clamp. Upon perfusion with CnTx, the shoulder of the action potential was partially removed, consistent with the results shown in Figure 3 and results of others (10,11). The importance of the N-type calcium channels in pain signaling is emphasized by findings that nociceptive neurons abundantly express a unique splice variant of the N-type channel (12) and that mice lacking the Cav2.2 gene, encoding for N-type calcium channels, show altered nociceptive responses (15–17). Our results indicate that non-N-type calcium currents are also inhibited by BAB. Inhibition of the low-voltage-activated T-type calcium channels by BAB was confirmed in unpublished experiments (IC50 ∼180). Thus, inhibition of both N-type and non-N-type calcium channels may contribute to BAB’s epidural analgesic action.
The just-below-maximal-water-solubility BAB concentration of 500 μM (∼2.5× IC50) largely inhibited the total calcium current (Fig. 3D). The maximal-solubility concentration is the upper value in the clinical situation in which BAB is applied as an aqueous suspension on the spinal dura. In the epidural space, BAB diffuses from its depot and will affect the spinal nerves passing that space. Apparently, a local long-lasting BAB concentration gradient is established as a result of the local balance among release, diffusion, and degradation of BAB, including pharmacologically effective concentrations (18).
The interesting question remains why epidural BAB suspensions selectively affect the small-diameter pain-transmitting nerve fibers, whereas the thick motor and sensory fibers are not influenced. For the explanation of this differential blockade, three mechanisms should be considered, all of which may contribute. The first one explains differential nerve block with the classic observation that thinner axons cease firing with shorter segmental exposure to impulse-blocking drugs than thicker axons (19). Korsten et al. (3) explained in this way the selective action of BAB from differences in the critical length of axons traversing the epidural space. Grouls et al. (18) suggested that selective pain suppression by BAB was the result of a stable establishment of relatively small epidural concentrations due to the low solubility of BAB, which would favor inhibition of the thinner pain fibers. Finally, there are possible differences in BAB-sensitive ion channel expression in axonal membranes of myelinated and unmyelinated fibers. A rich supply of sodium and potassium channels is present in Ranvier nodes, but calcium channels are lacking (20). For the unmyelinated sensory fibers, the spectrum of ion channels is not well studied, but apart from tetrodotoxin-sensitive and -resistant sodium channels, calcium channels belong to their ion channel palette (21). Calcium spikes have been recorded from human nociceptive C-fibers of the sural nerve (22) and could be evoked by capsaicin (23). It is thus tempting to speculate that calcium channels play a key role in selective analgesia by BAB by serving as targets for blocking the calcium spikes in pain-transmitting fibers. In this respect, it is of interest to mention that others have shown that some other, more hydrophilic, local anesthetics (e.g., bupivacaine) also inhibit calcium currents in mammalian sensory neurons (24) and dorsal horn neurons (e.g., ropivacaine) (25). Ropivacaine, which also has motor-sparing properties, seems to act by a mechanism other than that of BAB: it must have a less localized epidural distribution because of its larger water solubility and insensibility to esterases (18) and because of its property to increase calcium currents at smaller concentrations (25).
In conclusion, submaximal water-solubility BAB concentrations inhibit the calcium channels of sensory neurons. This inhibition is likely to contribute, in addition to the inhibition of sodium and potassium channels, to the long-duration selective analgesia after epidural application of BAB suspensions.
1. Shulman M. Treatment of cancer pain with epidural butyl-aminobenzoate suspension. Reg Anesth 1987;12:1–4.
2. Shulman M, Lubenow TR, Nath HA, et al. Nerve blocks with 5% butamben suspension for the treatment of chronic pain syndromes. Reg Anesth Pain Med 1998;23:395–401.
3. Korsten HH, Ackerman EW, Grouls RJ, et al. Long-lasting epidural sensory blockade by n-butyl-p-aminobenzoate in the terminally ill intractable cancer pain patient. Anesthesiology 1991;75:950–60.
4. Van den Berg RJ, Van Soest PF, Wang Z, et al. The local anesthetic n-butyl-p-aminobenzoate selectively affects inactivation of fast sodium currents in cultured rat sensory neurons. Anesthesiology 1995;82:1463–73.
5. Van den Berg RJ, Wang Z, Grouls RJ, Korsten HH. The local anesthetic, n-butyl-p-aminobenzoate, reduces rat sensory neuron excitability by differential actions on fast and slow Na+ current components. Eur J Pharmacol 1996;316:87–95.
6. Butterworth JF, Strichartz GR. Molecular mechanisms of local anesthesia: a review. Anesthesiology 1990;72:711–34.
7. Beekwilder JP, O’Leary ME, van den Broek LP, et al. Kv1.1 channels of dorsal root ganglion neurons are inhibited by n-butyl-p-aminobenzoate, a promising anesthetic for the treatment of chronic pain. J Pharmacol Exp Ther 2003;304:531–8.
8. Diochot S, Richard S, Valmier J. Diversity of voltage-gated calcium currents in large diameter embryonic mouse sensory neurons. Neuroscience 1995;69:627–41.
9. Mintz IM, Adams ME, Bean BP. P-type calcium channels in rat central and peripheral neurons. Neuron 1992;9:85–95.
10. Scroggs RS, Fox AP. Multiple Ca2+
currents elicited by action potential waveforms in acutely isolated adult rat dorsal root ganglion neurons. J Neurosci 1992;12:1789–801.
11. Blair NT, Bean BP. Roles of tetrodotoxin (TTX)-sensitive Na+ current, TTX-resistant Na+
current, and Ca2+
current in the action potentials of nociceptive sensory neurons. J Neurosci 2002;22:10277–90.
12. Bell TJ, Thaler C, Castiglioni AJ, et al. Cell-specific alternative splicing increases calcium channel current density in the pain pathway. Neuron 2004;41:127–38.
13. Scroggs RS, Fox AP. Calcium current variation between acutely isolated adult rat dorsal root ganglion neurons of different size. J Neurosci 1992;445:639–58.
14. Buisman HP, De Vos A, Ypey DL. A pipette holder for use in patch-clamp measurements. J Neurosci Methods 1990;31:89–91.
15. Hatakeyama S, Wakamori M, Ino M, et al. Differential nociceptive responses in mice lacking the alpha(1B) subunit of N-type Ca2+
channels. Neuroreport 2001;12:2423–7.
16. Kim C, Jun K, Lee T, et al. Altered nociceptive response in mice deficient in the alpha(1B) subunit of the voltage-dependent calcium channel. Mol Cell Neurosci 2001;18:235–45.
17. Saegusa H, Kurihara T, Zong S, et al. Suppression of inflammatory and neuropathic pain symptoms in mice lacking the N-type Ca2+
channel. EMBO J 2001;20:2349–56.
18. Grouls RJ, Meert TF, Korsten HH, et al. Epidural and intrathecal n-butyl-p-aminobenzoate solution in the rat: comparison with bupivacaine. Anesthesiology 1997;86:181–7.
19. Franz DN, Perry RS. Mechanisms for differential block among single myelinated and non-myelinated axons by procaine. J Physiol 1974;236:193–210.
20. Waxman SG, Ritchie JM. Molecular dissection of the myelinated axon. Ann Neurol 1993;33:121–36.
21. Elliott P. Action of antiepileptic and anaesthetic drugs on Na- and Ca-spikes in mammalian non-myelinated axons. Eur J Pharmacol 1990;175:155–63.
22. Quasthoff S, Grosskreutz J, Schroder JM, et al. Calcium potentials and tetrodotoxin-resistant sodium potentials in unmyelinated C fibres of biopsied human sural nerve. Neuroscience 1995;69:955–65.
23. Mayer C, Quasthoff S, Grafe P. Confocal imaging reveals activity-dependent intracellular Ca2+
transients in nociceptive human C fibres. Pain 1999;81:317–22.
24. Sugiyama K, Muteki T. Local anesthetics depress the calcium current of rat sensory neurons in culture. Anesthesiology 1994;80:1369–78.
25. Liu B-G, Zhuang X-L, Li S-T, et al. Effects of bupivacaine and ropivacaine on high-voltage-activated calcium currents of the dorsal horn neurons in newborn rats. Anesthesiology 2001;95:139–43.