Abstract
Although tricyclic antidepressant (TCA) blockade of cardiac Na+ channels is appreciated, actions on neuronal Na+ channels are less clear. Therefore, the effects of TCAs (amitriptyline, doxepin and desipramine) as well as trazadone and fluoxetine on voltage-gated Na+ current (INa) were examined in bovine adrenal chromaffin cells using the whole-cell patch-clamp method. Amitriptyline produced concentration-dependent depression of peak INa evoked from a holding potential of −80 mV with KD value of 20.2 μM and a Hill coefficient of 1.2. Although 20 μM amitriptyline induced no change in the rate or voltage dependence of INaactivation, steady-state inactivation demonstrated a 15-mV hyperpolarizing shift. Similar results were observed for doxepin and desipramine. This shift in steady-state inactivation was associated with a slowed rate of recovery from the inactivated state. Contrasting results were observed with the atypical antidepressants: while 20 μM fluoxetine depressed peak INa by 61% and caused a 7-mV hyperpolarizing shift in steady-state inactivation, 100 μM trazodone decreased peak INa by only 19% and caused only a 3-mV shift. Although the magnitude of fluoxetine effects was similar to those of the TCAs, the onset of fluoxetine effects was substantially slower than for amitriptyline. In voltage-clamp and current-clamp measurements from neonatal rat dorsal root ganglion neurons, 20 μM amitriptyline decreased INa by 52% and depressed action potential dynamics consistent with enhanced Na+ channel inactivation. The effects of the TCAs on INa are similar to local anesthetic behavior and could contribute to certain analgesic actions.
Although the primary mechanism by which TCAs relieve the symptoms of depressive illness appears to involve the inhibition of neurotransmitter reuptake, TCAs have additional effects on a variety of ion channels (Choiet al., 1992; Ogata et al., 1989; Schauf et al., 1975). For example, TCAs block Na+channels in cardiac tissues in a fashion similar to the action of local anesthetics (Barber et al., 1991; Nattel, 1987), an effect that may contribute to their alteration in cardiac conduction and dysrhythmogenic actions. In addition to their antidepressant actions, TCAs have been known to exert an analgesic effect (Paoli et al., 1960) that is independent of the antidepressant action (Panerai et al., 1990). Acute administration of TCAs can exert an analgesic effect (Bromm et al., 1986; Coquozet al., 1991; Poulsen et al., 1995), which may be useful in postoperative pain management (Tiengo et al., 1987). Long-term administration of the TCAs amitriptyline and desipramine diminishes pain in patients with diabetic neuropathy (Maxet al., 1987, 1992) and other chronic pain conditions (Magni, 1991; Watson et al., 1982). Because the effects of TCAs on neuronal Na+ currents are less well described, we conducted a series of experiments to determine the sensitivity of neuronal Na+ channels to TCAs, and we compared their potency with that of the two atypical antidepressants, trazodone and fluoxetine, which differ greatly in analgesic potency. Bovine adrenal chromaffin cells, a well-characterized model for neuronal electrophysiology and secretion, were used to study the alteration of voltage-gated Na+ current (INa) and modulation of Na+ channel inactivation by antidepressants. In addition, the use-dependent Na+ channel blockade by amitriptyline was characterized in isolated DRG neurons. A preliminary account of this work has been reported (Pancrazio et al., 1996).
Methods
Bovine adrenal chromaffin cells were generously supplied by Dr. Y. I. Kim (Department of Biomedical Engineering, University of Virginia, Charlottesville, VA) and were isolated according to a previously reported method (Greenberg and Zinder, 1982) with modifications (Creutz et al., 1987). Neonatal (P8-P12) rat DRG neurons were isolated by a method modified from McLean et al. (1988). Briefly, after dissection from the perispinal tissue, ganglia were minced and incubated in 2.5 mg/ml trypsin (Sigma Chemical, St. Louis, MO) supplemented with 2 mg/ml type I collagenase (Sigma Chemical) for 30 min at 37°C. Cells were triturated five to eight times and resuspended in Dulbecco’s modified Eagle’s medium (GIBCO BRL, Gaithersburg, MD) with 10% fetal calf serum (Hyclone Laboratories, Logan, UT), 50 U/ml penicillin and 50 μg/ml streptomycin (Sigma Chemical). Cells were plated onto poly-l-lysine-coated glass coverslips and maintained in an incubator at 37°C in 5% CO2/95% air. Best results for voltage-clamp experiments were obtained from cells within 2 days of isolation, before extensive processes developed.
For measurement of INa in bovine adrenal chromaffin cells, the external bathing solution contained (in mM) 141 NaCl, 5 KCl, 0.2 CaCl2, 1 CoCl2, 1 MgCl2 and 10 HEPES, adjusted to pH 7.4 with 1 M NaOH. The patch pipette solution contained (in mM) 120 CsCl, 20 tetraethylammonium chloride, 1 CaCl2, 11 EGTA-CsOH, 11 HEPES and 5 MgATP, adjusted to pH 7.3 with 1 M HCl. For voltage-clamp measurements from DRG neurons, the external recording solution contained (in mM) 10 or 30 NaCl, 130 or 150 tetraethylammonium chloride, 1 MgCl2, 0.2 CaCl2, 1 CoCl2 and 10 HEPES, adjusted to pH 7.4 with 1 M CsOH. The patch pipette solution for these experiments contained (in mM) 100 CsCl, 2.5 MgCl2, 10 EGTA, 30 CsOH, 40 HEPES, 2 MgATP and 0.3 NaGTP, adjusted to pH 7.3 with 1 M CsOH. For current clamp measurements from DRG neurons, the external recording solution contained (in mM) 140 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2 and 10 HEPES, adjusted to pH 7.4 with 1 M NaOH. The pipette solution contained (in mM) 140 KCl, 0.5 EGTA-KOH, 5 MgATP and 5 HEPES, adjusted to pH 7.3 with 1 M HCl. Using a gravity-feed perfusion system, complete solution exchange was accomplished within 2 sec, and measurements were typically made 120 sec after solution application. Amitriptyline was obtained from Sigma Chemical. Desipramine, doxepin and trazadone were obtained from Research Biochemicals (Natick, MA). Fluoxetine was provided by Eli Lilly (Indianapolis, IN). For our detailed mechanistic studies, we focused on amitriptyline as a model TCA.
Standard patch-clamp methods were used as described by Hamill et al. (1981). Voltage-clamp or current-clamp measurements were taken 4 to 6 min after initiation of the whole-cell recording configuration using the Axopatch 200 (Axon Instruments, Foster City, CA) patch-clamp amplifier. Patch electrodes were fabricated from KIMAX-51 borosilicate glass (American Scientific, Charlotte, NC) with a two-stage micropipette puller and heat-polished with a microforge and had resistances of ≤1.5 MΩ when filled with internal solution. Unless otherwise noted, cells were voltage-clamped at −80 mV and the commonly used P/n approach (n = −4) was used to estimate leakage and capacitive current. Whole-cell currents were filtered at 5 kHz with a four-pole Bessel low-pass filter and digitized at 20 kHz. Current records were analyzed offline using a custom program capable of preparing I-V relations and nonlinear curve fitting (Pancrazio, 1993) or with Sigmaplot for Windows version 2 (Jandel, San Rafael, CA). For steady-state activation and inactivation estimates, data were fit to the Boltzmann function, f(V):
Where appropriate, data are presented as mean percentage of control ± S.E.M. and the number of cells tested (n). The rates of recovery from inactivation and the onset of use-dependent blockade were fit to exponential equations using Sigmaplot. For the analysis of concentration-dependence and rate of recovery from inactivation, standard errors derived from the fitted data were used to test for significant differences. Statistical significance of a drug effect was determined using paired Student’s t test with P < .05 considered significant.
Results
As initially reported by Fenwick et al. (1982), bovine adrenal chromaffin cells express a rapidly activating and inactivating INa, which in our experiments reached a peak of 845 ± 450 pA (mean ± S.D., n = 54) in response to a step depolarization of −10 to +10 mV from a holding potential of −80 mV. The addition of amitriptyline to the bathing solution induced a marked decrease in INa over a range of test potentials (fig. 1A) within 30 sec of drug application. In the presence of 20 μM amitriptyline, peak INa fell by 51 ± 2% (mean ± S.E.M., n = 13 cells). The time to peak (tP) of INa was not affected by antidepressant treatment; for example, tP at 0 mV under control conditions and in the presence of amitriptyline was 1.3 ± 0.1 and 1.2 ± 0.1 msec (n = 5), respectively. Figure 1B summarizes the peak INa triggered by test potentials ranging from −50 to +60 mV from a holding potential of −80 mV under control, 20 μM amitriptyline-treated and recovery conditions. Although there appeared to be no obvious shift in the voltage dependence of activation based on the I/V relation, steady-state activation was estimated by calculation of the peak conductance, GNa(V), at each test potential, V, according to Ohm’s law:
The suppression of INa by amitriptyline was concentration dependent and also characteristic of the TCAs doxepin, desipramine and nortriptyline (fig. 2). The concentration dependence of amitriptyline-induced blockade was fitted to a logistic equation:
Time-dependent Na+ channel inactivation was assessed by fitting the inactivating phase of INato a single exponential decay function of time, INa = A·exp(−t/τH), where A is the current amplitude and τH is the time constant of inactivation. Amitriptyline (20 μM) had no effect on the time course of inactivation for test potentials ranging from −10 to +40 mV. For example, τH was 1.5 ± 0.2 and 1.4 ± 0.2 msec (n = 5) under control and amitriptyline-treated conditions, respectively, in response to a voltage step to 0 mV. Although the TCAs did not change the inactivation rate, steady-state inactivation was markedly altered (fig.5A). To estimate steady-state inactivation, peak INa was triggered by a test pulse to 0 mV from prepulse potentials, 5 sec in duration, from −95 to +40 mV. For one set of experiments, data were normalized and fit to the Boltzmann function, yielding control values forkn
of −6 ± 1 mV (n = 5) and for Vn of −58 ± 2 mV. Amitriptyline (20 μM) shifted Vntoward hyperpolarized potentials by 15 ± 1 mV (n= 5), while it exerted no observable effect onkn
. This effect was readily reversible and concentration dependent (fig. 5B). To gain further insight into the mechanisms of amitriptyline-induced blockade of INa inactivated state, the rate of recovery from inactivation was examined. This was accomplished using two voltage steps to +10 mV separated by a repolarizing interpulse to −80 mV for durations ranging from 120 msec to 6 sec. Recovery from inactivation was estimated by the ratio of peak INa evoked by the second pulse (INa,2) to peak INa triggered by the initial pulse (INa,1). The results plotted in figure6 show that amitriptyline delayed recovery of a major fraction of INa from inactivation. Under both control and amitriptyline-treated conditions, INa,2/INa,1 was well fit as a biexponential process:
To verify Na+ channel inhibition in the DRG neurons by amitriptyline, voltage-clamp measurements were taken using an external recording solution with the Na+concentration decreased from 141 mM to 10 or 30 mM to ensure suitable clamp conditions. DRG neurons can express two types of Na+ channels (Roy and Narahashi, 1992), which may account for the differential response of neurons to sustained depolarizing current injections under current-clamp conditions (Elliott and Elliott, 1993). As shown in figure 7, both slow and fast INa, measured from two different neurons, appeared to exhibit a similar range of sensitivity to amitriptyline. Overall, DRG peak INa fell to 48 ± 4% of the control amplitude with application of 20 μM amitriptyline, a level similar to that of chromaffin cell INa.
To assess the effect of the slowed recovery from inactivation in generating use-dependent blockade, DRG neurons were voltage-clamped and depolarized from −80 mV to 10 mV for 20 msec at a rate of 10 Hz. The results of a 10-Hz stimulus for 14 pulses is shown in figure8, and the greater use-dependence in 10 μM amitriptyline is evident from the tracings of INa with subsequent depolarizations. The peak current of the nth depolarization (INa,n) could be described as an exponential function of the number of depolarizations following the initial peak current after rest (INa,1):
To determine whether amitriptyline inhibition of neuronal Na+ channels would alter physiological processes, we examined AP behavior and INa in neonatal rat DRG neurons. Current-clamped neurons exhibited resting potentials ranging from −55 to −65 mV, consistent with previous work (Caviedeset al., 1990; Elliott and Elliott, 1993), and all-or-nothing APs in response to depolarizing current injections of +200 to +800 pA. In 4 of 7 neurons, sustained depolarizing current injection triggered bursts of 5.6 ± 0.8 Aps/stimulus of 200 msec in duration, as shown in figure 9A. The application of 20 μM amitriptyline reversibly decreased the number of APs evoked by the same depolarizing stimulus to 55 ± 4% of control. The inhibition followed the pattern illustrated in figure 9A, which is consistent with a use-dependent inhibition induced by amitriptyline. In the remaining neurons, that displayed only a single AP with current injection, a series of current steps were applied to assess changes in excitability. Although amitriptyline failed to alter the voltage response to a hyperpolarizing current injection of −30 pA (fig. 9B), effects consistent with decreased membrane excitability were observed with depolarizing current pulses. Larger depolarizing current steps than those shown (typically >400 pA) did trigger APs, which were virtually indistinguishable between control and amitriptyline-treated conditions.
Discussion
In addition to the blockade of the cardiac Na+ channel, the TCAs inhibit the neuronal Na+ channel but with a far lower potency. The calculated neuronal Na+ channelKD value of ≈20 μM for amitriptyline is similar to that for desipramine, 9 μM, previously reported in Myxicola giant axon (Schauf et al., 1975). Amitriptyline (10–30 μM) reduced AP amplitude and maximum rate of depolarization (dV/dtmax) in crayfish giant axon (Wang et al., 1981), while TCAs inhibited22Na+ influx in bovine chromaffin cells with IC50 values of 10 to 17 μM, levels similar to the present study (Arita et al., 1987). Amitriptyline also has been shown to inhibit neurotransmitter release from rat striatal brain slices via blockade of Na+ channels (Ishii and Sumi, 1992). Ogataet al. (1989) directly demonstrated a voltage-dependent reduction of INa from neuroblastoma cells by 3 μM imipramine, another TCA. In these neurally derived cells, the concentrations of amitriptyline and other TCAs required to cause >50% depression of INa is ∼10 to 50 times higher than the 0.4 to 3.2 μM concentration of amitriptyline that is required for an equivalent effect in cardiac myocytes (Barber et al., 1991) and Purkinje fibers (Nattel, 1987) when stimulation rates are increased above 3 Hz.
Our results provide a more complete description of TCA action on steady-state inactivation and demonstrate the apparent stabilization by amitriptyline of an inactivated state that requires a more sustained repolarization to revert to an available closed state. Figure 6predicts that after a repolarization for 80 to 100 msec, approximately twice as many channels will be inactivated in the presence of amitriptyline as in the control setting. Stabilization of the inactivated state by the antidepressants results in use-dependent blockade of Na+ channels, as evidenced in the DRG experiments with 10-Hz stimulation. Such rapid stimulation increased the depression of INa by 10 μM amitriptyline from 27% for the initial depolarization to 44% for the steady-state 10 Hz INa.
Based on competition experiments in myocytes, Barber et al.(1991) suggested that amitriptyline and lidocaine compete for the same LA binding site, recently shown to reside in amino acids present on the transmembrane segment S6 of domain IV of the alpha subunit (Ragsdale et al., 1994). It is noteworthy that the action of amitriptyline and the other TCAs is similar to the LAs in a number of regards. First, amitriptyline shifts the steady-state inactivation curve in the hyperpolarizing direction, just as the LAs. Second, amitriptyline depresses INa in a frequency-dependent manner, while the degree of use-dependent inhibition may be somewhat less for the neuronal than the myocardial Na+ channels. Third, the cardiac Na+ channel is more sensitive than the neuronal type to blockade by amitriptyline. Likewise, the LA concentration required for equivalent inhibition in nerve is typically ≈10-fold greater than that for myocardium if one compares studies of LA depression of INa or action potential rate of depolarization (dV/dt) in myocardium vs. nerve. However, amitriptyline must have a 20- to 200-fold greater affinity for the LA binding site than do the LAs themselves if one compares the effective concentrations for Na+ channel or conduction blockade. It is noteworthy that TCAs have a tertiary amine group connected by a three carbon chain to a large aromatic moiety, while LAs have a tertiary amine connected to an aromatic group by a chain of two carbons and an amide (or ester) linkage.
The clinical relevance of such neuronal Na+channel inhibition is unclear. There is considerable evidence demonstrating the analgesic effects of tricyclic (Hameroff et al., 1984; Max et al., 1992; Valverde et al., 1994) and certain atypical (Max et al., 1992; Raniet al., 1996) antidepressants under differing clinical conditions of chronic pain. For example, the TCA doxepin has been shown to effectively treat migraine headache (Mørland et al., 1979) and lower back pain (Hameroff et al., 1982, 1984), and fluoxetine is a useful analgesic against rheumatic pain (Rani et al., 1996). Although the ability of all antidepressants to inhibit the reuptake of neurotransmitters may underlie their effectiveness against chronic pain, the usefulness of antidepressants in the treatment of acute pain appears to be primarily limited to the TCAs, suggesting a mechanism of acute analgesic activity independent from the antidepressant mechanism of action. The demonstrated voltage- and frequency-dependent inhibition of Na+ channels by TCAs, particularly amitriptyline, are properties that appear to be necessary properties of Na+ channel modulators with analgesic efficacy (Tanelian and Brose, 1991). In view of the use-dependent effect on Na+ channels, sensitized neurons that fire repetitive APs or are partially depolarized would be predicted to be more susceptible.
Several observations raise the possibility that Na+ channels may be involved in the usefulness of TCAs in the treatment of pain. All TCAs have the same effect on neuronal Na+ channels, whereas the effects of the atypical antidepressants on Na+ channels varies. A high concentration of trazodone, with little apparent clinical efficacy in the treatment of acute pain (Davidoff et al., 1987), had negligible effect on INa. Conversely, fluoxetine has questionable usefulness in pain management (Messiha, 1993) but can inhibit INa at concentrations comparable to the TCAs. Important differences between the blockade by fluoxetine and the TCAs is the use-dependence of this activity. Unlike amitriptyline, fluoxetine appears to cause less use-dependent inhibition of current, which might make its physiological effects less prominent.
Amitriptyline has also been administered intrathecally and found to augment the action of narcotics, an effect attributed to alteration in monamine reuptake (Eisenach and Gebhart, 1995). However, an analgesic dose of drug may achieve concentrations sufficient to mediate some effects via Na+ channel blockade. In sheep, 5 mg of amitriptyline injected into the CSF (20–25 ml) gives an estimated concentration of >500 μM and has been found to yield a CSF concentration of 1.3 ± 0.6 μg/ml (4.8 μM) after 2 hr, with higher concentrations present in the gray (5.9 ± 2.7 μg/g of tissue) and white (2.9 ± 1.5 μg/g of tissue) matter near the site of injection (Cerda et al., 1997). A decrease in blood pressure in the animals with acute intrathecal amitriptyline is similar to changes observed with local anesthetics in blocking sympathetic outflow from the spinal cord.
The concentrations of antidepressants used in this study are far higher than clinically relevant plasma concentrations. In patients receiving daily doses of 75 to 300 mg of amitriptyline, plasma steady-state concentrations range from 0.3 to 0.9 μM (Baldessarini, 1985). However, the plasma binding (>90%) is such that the free drug concentration in plasma is much lower. On the other hand, antidepressants do accumulate in the brain in concentrations 20-fold greater than in plasma (Karson et al., 1993), but the free drug concentration available for action at central neuronal Na+ channels is unclear.
Antidepressants block a variety of other ion channels, including the Ca2+-activated K+ channels, which are inhibited by amitriptyline with aKD value of ≈40 μM (Kamatchi and Ticku, 1991). Amitriptyline also inhibits voltage-gated K+ channels in rat superior cervical ganglion cells with a KD value of 12 μM (Wooltorton and Mathie, 1995). Studies have reported the sensitivity of L-type Ca++ channels to antidepressant inhibition (Antkiewicz-Michaluk et al., 1991; Choi et al., 1992; Schwaninger et al., 1995). In mouse DRG neurons, theKD value of imipramine blockade of L-type calcium channel is 30 μM (Choi et al., 1992). It is striking that the concentration of TCAs that inhibits Ca++ and K+ channels is only slightly greater than that for neuronal Na+channels. The significance of antidepressant effects of these ion channels to clinical effects is presently unclear; however, the potential exists for synergistic action due to combined inhibition of Na+ and Ca++ channels.
Acknowledgments
The authors thank Dr. Donald Manning for useful and stimulating discussions and Ms. Jill Tatum for her excellent editorial assistance.
Footnotes
-
Send reprint requests to: Carl Lynch III, M.D., Ph.D., Department of Anesthesiology, Box 10010, University of Virginia, Health Sciences Center, Charlottesville, VA 22906-0010. E-mail:cl7y{at}virginia.edu
-
↵1 This study was supported in part by National Institutes of Health Research Grant GM31144 (C.L.).
-
↵2 Present address: Code 6900, Naval Research Laboratory, Washington, DC 20375.
- Abbreviations:
- AP
- action potential
- CSF
- cerebrospinal fluid
- DRG
- dorsal root ganglion
- dV/dtmax
- maximum rate of depolarization
- INa
- sodium current
- I-V
- current voltage
- VH
- holding potential
- LA
- local anesthetic
- EGTA
- ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid
- HEPES
- 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
- TCA
- tricyclic antidepressant
- Received April 29, 1997.
- Accepted September 15, 1997.
- The American Society for Pharmacology and Experimental Therapeutics