Amitriptyline (AMI), gabapentin (GBP), and lidocaine (LDC) are widely used for the treatment of neuropathic pain syndrome but produce variable results. AMI is a tricyclic antidepressant that non-selectively inhibits the reuptake of serotonin and norepinephrine. GBP (1-aminomethyl cyclohexane acetic acid), a new generation anticonvulsant with relatively low toxicity and few side effects, has been the drug of choice for refractory epilepsy . Structurally, it is a gamma-aminobutyric acid (GABA) analog, with an unknown mechanism of antiseizure action , LDC is an amide local anesthetic with class I antiarrhythmic action. This drug is well recognized as both a diagnostic and therapeutic drug for neuropathic pain . Its mechanisms of analgesic action are depressing conduction in nociceptive afferent nerves , inhibiting dorsal horn neural transmission , and possibly modifying cerebral pain perception . Although there are several studies of the efficacy of antidepressants, GBP, and LDC in the management of the neuropathic pain syndrome in animal models and in humans, the mechanism and site of action of these drugs are still unclear. Therefore, the purpose of this study was to systematically compare the effects of AMI, GBP, and LDC on pain behavior (mechanical allodynia) using an animal model of neuropathic pain. Further, we studied the site of action (peripheral versus central) of these drugs by using electrophysiological recordings of ectopic discharges in the injured peripheral nerve fibers (L5 spinal segment) in these experimental rats before and after treatment. We used the segmental spinal nerve injury (SSI) model of neuropathic pain in rats , which mimics some characteristics of human neuropathic pain.
The experimental protocols were approved by our animal care and use committee and were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
Young adult Sprague-Dawley male rats weighing 150-200 g were used for the experiments. The animals were housed in groups of three in plastic cages with soft bedding under a 12-h reversed light-dark cycle (light cycle 9:00 PM to 9:00 AM; dark cycle from 9:00 AM to 9:00 PM). They were kept in the same room with a constant ambient temperature and had free access to food and water. The entire study was performed during the dark cycle, which is the active period for the rats.
After 1 wk of acclimatization to the laboratory conditions, surgery was performed in all animals, as previously described . Briefly, the rats were anesthetized with halothane in oxygen (halothane concentration of 2% for induction and 0.8% for maintenance). Thereafter, they were placed in a prone position, a midline skin incision at L4-S2 was made, and paraspinal muscles were separated from the spinous processes at these levels. The transverse process of L6 on the left side was then exposed and carefully removed with a small rongeur. Thereafter, the left L5 and L6 spinal nerves were identified, each was tightly ligated with 6-0 silk thread, and the wound was closed. At the end of surgery, anesthesia was discontinued, and the animals were returned to their cages with food pellets and water ad libitum under the same environmental conditions as before. The animals recovered from anesthesia within approximately 10 min. They were allowed a 1-wk recovery period before the experiments were performed. Three rats with a damaged L4 spinal nerve, indicated by a dragging of the left hind limb, were excluded from the study.
All animals were tested for quantitative changes in mechanical sensitivity of the plantar surface of the foot. The behavioral tests were conducted in a blinded manner with the experimenter unaware of which method of treatment was applied to the animals. To quantify mechanical sensitivity of the foot, the median 50% foot withdrawal (paw flinching) threshold was determined with the up-down method using von Frey filaments (bending force of 4.4-142.1 mN) . Based on electrophysiological recordings, we found that thresholds for mechanoreceptors in the rat foot did not exceed 25 mN, whereas those of nociceptors were rarely lower than that level . However, the threshold for paw flinching after neuropathic surgery often became lower than 10 mN, which suggests that a nociceptive reflex was elicited by the activation of mechanoreceptors in this situation. Therefore, a significant reduction in the mechanical threshold for paw flinching was interpreted as a sign of mechanical allodynia.
AMI was obtained from Stuart Pharmaceuticals (Wilmington, DE), diluted with saline, and given intraperitoneally (IP) at a dose of 1.5 mg/kg. GBP was kindly donated in purified powder form by Parke-Davis (Ann Arbor, MI). GBP was dissolved in saline and given IP at a dose of 50 mg/kg. LDC was obtained from Abbott Laboratories (Abbott Park, IL), diluted with saline, and given at a dose of 10 mg/kg as an IV infusion over a period of 10 min, followed by a 0.3-mL saline flush. Choices of doses of these drugs were based on the results of previous studies and our own preliminary study [10,11].
Measurements of the mechanical threshold for paw flinching were made using von Frey filaments, as described above, before and 7 days after surgery. The animals were then randomly assigned to one of five groups. Group 1 (n = 8) received isotonic sodium chloride solution IP and served as control animals, Group 2 (n = 8) received AMI IP, Group 3 (n = 8) received GBP IP, Group 4 (n = 6) received an IV infusion of saline for 10 min, and Group 5 (n = 6) received an IV infusion of LDC for 10 min. Thereafter, mechanical thresholds for paw flinching were measured 0.5, 1, 2, and 4 h and 1, 3, and 7 days after the drug treatment. To avoid the effect of the circadian cycle, all studies were performed during the same time period (between 9 AM and noon) on each of the experimental days.
We studied the site of action of the tested drugs by examining the effects of these drugs on the rate of ectopic discharges of injured afferents in the left L5 dorsal root in 11 additional male Sprague-Dawley rats. These animals were surgically prepared in the same manner as the animals in the first experiment.
The electrophysiological experiment was conducted 7 days after nerve ligation. The rats were anesthetized with a mixture of halothane (2 vol% for induction and 0.8 vol% for maintenance) and 1:2 flow ratio of N2 O and O2. Polyethylene cannulas (PE-20) were inserted into the right carotid artery and the right jugular vein for monitoring the blood pressure and for drug administration, respectively. Tracheotomy was performed, and all animals were thereafter paralyzed with pancuronium (1 mg/kg, IV bolus followed by a continuous infusion 0.4 mg [center dot] kg-1 [center dot] h-1) and mechanically ventilated throughout the rest of the experimental period. The settings on the ventilator were adjusted to maintain the end-tidal CO (2) at 4.0% +/- 0.5%.
The spinal cord was exposed by a laminectomy at the level of L1-6, and the animals were mounted on a spinal investigation frame. We put a few drops of warm (36[degree sign]C) mineral oil on the exposed tissue to prevent it from drying. The L5 dorsal root was cut near the spinal cord, and the distal stump was placed on a mirror-based plate. The distal stump of the cut dorsal root was teased into small fascicles until spontaneously active single units could be recorded. Single unit activity was assumed when individual action potentials were monitored with a digital oscilloscope and amplified with an alternating current-coupled amplifier. The amplified signals were led to a window discriminator whose output was used to drive an audio monitor and then fed to a data acquisition system to construct peristimulus time histograms. The activity was also recorded on a digital recording device for data analysis off-line. During the recording period, the concentration of halothane was maintained between 0.3% and 0.6% (the level was adjusted to maintain stable blood pressure), and the experiment was discontinued if mean arterial blood pressure dropped below 70 mm Hg.
Tight ligation of the spinal nerve essentially transects the nerve. Therefore, most of the sensory neurons in the L5 dorsal root are disconnected from their sensory receptors, which are the normal sites of action potential generation. Therefore, we assumed continuing discharges in the L5 dorsal root filaments to be ectopic discharges. In addition, once a spontaneously active unit was found, we applied mechanical explorative stimuli all over the hind limb to ensure that there were no receptive fields.
Baseline activity of continuing ectopic discharges was recorded for 10 min before drug injection as the control. A bolus of AMI (1.5 mg/kg IP) or GBP (50 mg/kg IP), or a LDC (10 mg/kg) infusion over a period of 10 min via a cannula placed in the jugular vein was administered. Ectopic discharges were recorded for the next 60 min.
All values are expressed as means +/- SEM. Data were compared among groups by using the Kruskal-Wallis one-way analysis of variance, a nonparametric statistical test, followed by the Dunnett's or Dunn's test for multiple comparisons. Statistical significance was accepted if P <or=to 0.05.
Before preparatory surgery, all animals had normal mechanical threshold (>100 mN) for paw flinching, and there was no sign of mechanical allodynia. However, all rats exhibited characteristic neuropathic pain behavior and, as shown in Figure 1, a marked decrease in mechanical threshold for paw flinching (<10 mN) within 7 days after the spinal nerve ligation. There were no statistically significant differences in threshold values among groups before or after surgery. On the day of drug treatment (7 days after surgery), animals typically showed guarding behavior of the left hind paw, which was abnormally everted with slightly flexed toes.
In a previous study , doses of 0.5 and 2.0 mg/kg of AMI were effective in reducing pain behaviors in the rat. One week after the spinal nerve ligation, IP injections of four doses of AMI (0.1, 0.5, 1.5, and 3.0 mg/kg) were made in four separate groups of rats (two rats in each group). There was a graded increase in mechanical threshold for paw flinching with graded doses <or=to1.5 mg/kg; however, 3.0 mg/kg produced a smaller increase than 1.5 mg/kg. Therefore, 1.5 mg/kg was chosen as the optimal dose used for AMI in this study.
IP injections of three doses of GBP (10, 50, and 100 mg/kg) were made in three separate groups of rats (two rats in each group). The increase in mechanical threshold for paw flinching was largest using a dose of 50 mg/kg. Therefore, 50 mg/kg was chosen as the optimal dose used for GBP in this study.
Chaplan et al.  reported that 30 mg/kg LDC infused IV over 60 min was effective in reducing neuropathic behavior in rats. When we infused 15 mg/kg LDC over 10 min into four rats, two died. Therefore, we decided that 10 mg/kg LDC is the maximal dose that could be safely infused over 10 min in the rat.
In Experiment 1, the IP injection of AMI (1.5 mg/kg) or GBP (50 mg/kg) produced a significant increase in the mechanical threshold for paw flinching compared with the saline control (Figure 1A). The level of increase was significant 1-4 h postinjection. The IV infusion of LDC (10-mg/kg infusion for 10 min) was also effective in elevating the mechanical threshold for paw flinching (Figure 1B). LDC had a quick onset and produced a long-lasting effect so that the threshold was significantly increased up to 1 day after the infusion. Using the given doses, GBP was the most effective in producing an anti-allodynic effect in this model of neuropathic pain.
In Experiment 2, the effects of AMI, GBP, and LDC on continuing discharges of afferent fibers were studied on single units recorded from the distal stump of the cut L5 dorsal root. Because the L5 spinal nerve had been tightly ligated at the time of neuropathic surgery, recorded afferents were disconnected from the sensory receptors in which normal impulse generation occurs. In addition, the absence of receptive fields on each unit was confirmed at the time of recording from each unit. Therefore, recorded continuous discharges of afferent fibers must have originated from a site other than a normal impulse generation site; that is, these were ectopic discharges.
GBP had no effect on spontaneous ectopic discharges of three separate units, whereas LDC produced a marked reduction of ectopic discharges (Figure 2). AMI, however, produced a delayed weak reduction of discharges.
The effects of these drugs on ectopic discharges were studied on a total of 36 afferent fibers: 10 for AMI, 8 for GBP, 7 for control, and 11 for LDC. The average responses of these units are shown in Figure 3. LDC reduced ectopic discharges significantly during the drug infusion, and the effect lasted for at least 30 min longer. However, AMI reduced the response of ectopic discharges at a delayed time period; therefore, the reduction was only significant 20-30 min after the drug injection. GBP did not produce any change in ectopic discharge rate. None of the drugs produced changes in the arterial blood pressure during the entire recording session, which suggests that the changes in ectopic discharges are not secondary to blood pressure changes.
Peripheral nerve injury can result in mechanical allodynia; thus, innocuous stimuli can evoke significant pain in humans and pain behavior in experimental animals . In the present study, we compared the effects of three widely used treatments of neuropathic pain in an animal model. Saline treatment did not have a statistically significant effect on pain threshold throughout the experimental period; however, a slight increase (although not statistically significant) was evident 1 and 2 h after the saline infusion (Figure 1B). This effect may be due to the rats acclimating to the test behavior. However, all three treatment drugs were effective in reducing behavioral signs of mechanical allodynia, although they seem to exert their effect in different sites and, apparently, through different mechanisms.
Although the exact mechanisms for the generation of neuropathic pain are not clear, it is obvious that multiple central, as well as peripheral, mechanisms are involved. Ectopic discharges are produced by injured nerve fibers and their dorsal root ganglion cells . These ectopic discharges then enter the spinal cord and sensitize spinal dorsal horn neurons . Signals entering the spinal cord from injured sensory neurons seem to be responsible for the maintenance of neuropathic pain because blocking input from injured sensory neurons to the spinal cord by dorsal rhizotomy abolishes neuropathic pain behaviors in the SSI model , and application of local anesthetics to the painful foci associated with trauma relieves neuropathic pain in humans . Therefore, various treatments for neuropathic pain relief can be targeted to either abnormal sensory processing of the dorsal horn cells or ectopic discharge mechanisms. Apparently, the drugs tested in the present study act on one or more aspects of these mechanisms.
GBP is a structural analog of GABA, which penetrates the blood brain-barrier. Although it is commonly used as an anticonvulsant and antinociceptive drug, its mode of action is not well understood. Taylor et al.  summarized mechanistic hypotheses of GBP pharmacology. They proposed several different modes of action, including its ability to increase the concentration and probably the rate of synthesis of GABA in brain; its ability to bind with high affinity to a novel binding site in brain tissues that is associated with an auxiliary subunit of voltage-sensitive Ca2+ channels; its ability to reduce the release of several monoamine neurotransmitters; its ability to inhibit voltage-activated Na+ channels (although this is controversial); and its ability to increase serotonin concentrations in human whole blood. There have been a few clinical case reports of the effect of GBP on neuropathic pain . Further, Hwang and Yaksh  have shown that the spinal application of this drug alleviated mechanical allodynia without motor dysfunction in a dose-dependent manner in the SSI model. In the present study, we administered GBP systemically via the IP route. The effect of GBP followed a similar time course as that observed by Hwang and Yaksh . The time course of the anti-allodynic effect of GBP in the present study is also consistent with previous studies [19,20] that show that the maximal plasma concentration is reached 1-4 h after a single dose administration in humans, as well as animals.
In the present experiment, GBP had no effect on ectopic discharges of injured afferent fibers. Thus, we propose that the alleviation of mechanical allodynia by GBP is probably entirely mediated by a central mechanism (at the spinal or supraspinal level), such as a reduction in central sensitization. Further, it is possible that GBP stabilizes the sensitized neurons in the spinal cord. Our proposal is supported by a study published by Chapman et al. , who investigated the effects of subcutaneous GBP on spinal neuronal responses in spinal nerve-ligated and sham-operated rats. The authors reported that GBP (10-100 mg/kg) significantly reduced the spontaneous activity of spinal neurones of spinal nerve-ligated and sham-operated rats on Postoperative Days 7-17.
AMI is considered by some as the drug of choice for neuropathic pain. Its effectiveness has been reported in a wide range of pain states, including chronic headaches, migraine, facial pain, diabetic neuropathy, postherpetic neuralgia, and postmastectomy pain . However, the treatment is often compromised by side effects, including sedation, urinary retention, and orthostatic hypotension. Animal studies have shown both direct analgesic effects of the tricyclic antidepressants and potentiation of morphine analgesia . AMI has a wide range of pharmacological actions, including inhibition of norepinephrine and serotonin reuptake and antagonism of muscarinic cholinergic, histamine, and alpha-adrenergic receptors . In previous studies [24,25], AMI was as effective in patients with normal mood as in depressed patients, which suggests that the relief of pain was not mediated solely through a change in mood.
The results of our electrophysiological study show that systemic injection of AMI significantly reduces the rate of ectopic discharges but that the degree of reduction is relatively mild. This suggests that AMI may act on multiple sites: partially and at both central and peripheral sites. The possibility of dual sites of action of AMI is consistent with previous studies because there are reports of both central and peripheral effects of AMI. As for the central effect of AMI, Spiegel et al.  stated that analgesic effects of tricyclic anti-depressants are attributed to an augmentation of descending inhibitory influences on nociceptive pathways via an inhibition of reuptake of catecholamine transmitters at spinal and supraspinal synapses. As an alternative, Eisenach and Gebhart  suggested that AMI produces analgesia by an antagonistic effect on N-methyl-D-aspartate (NMDA) receptors. AMI may produce analgesia by this mechanism rather than by monoamine reuptake inhibition because AMI and other tricyclic antidepressants show high affinity for binding to NMDA receptors, and they function as NMDA antagonists at concentrations similar to those for inhibiting monoamine reuptake. However, Fromm et al.  reported that AMI exerts its therapeutic effect in neuropathic pain by enhancing segmental inhibition of wide dynamic range neurons, thereby preventing their excessive firing, which may be important in neuropathic pain. A peripheral action of AMI has also been suggested. Seltzer et al.  reported that AMI reduces autotomy behavior due to an analgesic effect, not due to sedation. These authors further speculated that the mechanism by which AMI suppresses autotomy could be a peripheral action, either by acting directly on the sensory fibers in the neuroma or via suppression of inflammatory cells. A potential peripheral action of AMI is also supported by the fact that many antidepressants block sodium channels .
The effect of AMI in certain conditions seems to be opiate-mediated because AMI alleviates hyperalgesia, and this alleviation of neuropathic pain behavior can be blocked by naloxone . The effects of AMI on neuropathic pain behavior can be modified through an endogenous opiate system.
LDC is one of the most widely studied and most commonly used local anesthetics. Systemically administered LDC (recommended dose 1-5 mg/kg over 5-35 min) is effective in the control of chronic pain states that present with radicular symptoms of pain arising from a peripheral neuropathy and, possibly, in some sympathetically mediated pain states . The use of a local anesthetic infusion in the management of acute and chronic pain is not new. Boas et al.  reported analgesic responses to an IV LDC infusion and proposed that deafferentation is an important determinant for a positive response of LDC. Mechanisms proposed in experimental models of peripheral nerve injury show that LDC-like compounds are effective because they suppress abnormal ectopic discharge generation from damaged primary afferents or dorsal root ganglion neurons . Chaplan et al.  showed that the alleviation of mechanical allodynia is not mediated via the spinal cord because an intrathecal injection of LDC did not alleviate mechanical allodynia in a spinal nerve ligation model. Therefore, LDC would alleviate mechanical allodynia by blocking ectopic discharge in peripheral sites. This is consistent with our finding of suppression of ectopic discharge in a neuropathic rat model after treatment with LDC. This clearly demonstrates a peripheral mechanism of this local anesthetic.
In the present study, we examined the effectiveness of the three drugs that are most often used clinically for the treatment of neuropathic pain. The single-dose administration of AMI, GBP, and LDC produced anti-allodynic effects in an established model for neuropathic pain in rats. This clearly supports their continued clinical use in managing patients with neuropathic pain syndromes. Further, the present study also suggests that the site of action is central for GBP, peripheral for LDC, and probably both central and peripheral for AMI.
1. Ramsay RE. Clinical efficacy and safety of gabapentin. Neurology 1994;44:23-30.
2. Goldlust A, Su TZ, Welty DF, et al. Effects of anticonvulsant drug gabapentin on the enzymes in metabolic pathways of glutamate and GABA. Epilepsy Res 1995;22:1-11.
3. Rowbotham MC. Chronic pain: from theory to practical management. Neurology 1995;45:5-10.
4. Thoren P, Oberg B. Studies on the endoanesthetic effects of lidocaine and benzonatate on non-medullated nerve endings in the left ventricle. Acta Physiol Scand 1981;111:51-8.
5. Woolf CJ, Wiesenfeld-Hallin Z. The systemic administration of local anaesthetics produces a selective depression of C-afferent fibre evoked activity in the spinal cord. Pain 1985;23:361-74.
6. Garfield JM, Gugino L. Central effects of local anesthetic agents. In: Strichartz GR, ed. Local anesthetics. Berlin: Springer-Verlag, 1987:253-84.
7. Kim SH, Chung JM. An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 1992;50:355-63.
8. Chaplan SR, Bach FW, Pogrel JW, et al. Quantitative assessment of allodynia in the rat paw. J Neurosci Methods 1994;53:55-63.
9. Leem JW, Willis WD, Chung JM. Cutaneous sensory receptors in the rat foot. J Neurophysiol 1993;69:1684-99.
10. Ardid D, Guilbaud G. Antinociceptive effects of acute and "chronic" injections of tricyclic antidepressant drugs in a new model of mononeuropathy in rats. Pain 1992;49:279-87.
11. Chaplan SR, Bach FW, Shafer SL, et al. Prolonged alleviation of tactile allodynia by intravenous lidocaine in neuropathic rats. Anesthesiology 1995;83:775-85.
12. Devor M, Janig W, Michaelis M. Modulation of activity in dorsal root ganglion neurons by sympathetic activation in nerve-injured rats. J Neurophysiol 1994;71:38-47.
13. Woolf CJ, Wall PD. The relative effectiveness of C-primary afferents of different origins in evoking a prolonged facilitation on the flexion reflex in the rat. J Neurosci 1986;6:1433-42.
14. Yoon YW, Na HS, Chung JM. Contributions of injured and intact afferents to neuropathic pain in an experimental rat model. Pain 1996;64:27-36.
15. Gracely RH, Lynch SA, Bennett GJ. Painful neuropathy: altered central processing maintained dynamically by peripheral input. Pain 1992;51:175-94.
16. Taylor CP, Gee NS, Su TZ, et al. A summary of mechanistic hypotheses of gabapentin pharmacology. Epilepsy Res 1998;29:233-49.
17. Rosner H, Rubin L, Kestenbaum A. Gabapentin adjunctive therapy in neuropathic pain states. Clin J Pain 1996;12:56-8.
18. Hwang JH, Yaksh TL. Effect of subarachnoid gabapentin in tactile evoked allodynia in a surgically induced neuropathic pain model in the rat. Reg Anesth 1997;22:249-56.
19. Richens A. Clinical pharmacokinetics of gabapentin. In: Chadwick D, ed. New trends in epilepsy management: the role of gabapentin. London: Royal Society of Medical Services, 1993:41-6.
20. Vollmer KO, von Hodenberg A, Kolle EU. Pharmacokinetics and metabolism of gabapentin in rat, dog and man. Arzneim Forsh Drug Res 1986;36:830-9.
21. Chapman V, Suzuki R, Chamarette HLC, et al. Effects of systemic carbamazepine and gabapentin on spinal neuronal responses in spinal nerve ligated rats. Pain 1998;75:261-72.
22. Bowsher D. Neurogenic pain syndromes and their management. Br Med Bull 1991;47:644-66.
23. Spiegel K, Kalb R, Pasternak GW. Analgesic activity of tricyclic antidepressants. Ann Neurol 1983;13:462-5.
24. Richelson E. Antidepressants: pharmacology and clinical use. In: Treatments of psychiatric disorders. Vol. 3. Washington, DC: American Psychiatric Association, 1989;1773-86.
25. Max MB, Schafer SC, Culnane M, et al. Amitriptyline, but not lorazepam, relieves postherpetic neuralgia. Neurology 1988;38:1427-32.
26. Eisenach JC, Gebhart GF. Intrathecal amitriptyline acts as an N-methyl-D-aspartate receptor antagonist in the presence of inflammatory hyperalgesia in rats. Anesthesiology 1995;83:1046-54.
27. Fromm GH, Nakata M, Kondo T. Differential action of amitriptyline on neurons in the trigeminal nucleus. Neurology 1991;41:1932-6.
28. Seltzer Z, Tal M, Sharav Y. Autotomy behavior in rats following peripheral deafferentation is suppressed by daily injections of amitriptyline, diazepam and saline. Pain 1989;37:245-50.
29. Deffois A, Fage D, Carter C. Inhibition of synaptosomal veratridine-induced sodium influx by antidepressants and neuroleptics used in chronic pain. Neurosci Lett 1996;220:117-20.
30. Edwards WT, Habib F, Burney RG, et al. Intravenous lidocaine in the management of various chronic pain states: a review of 211 cases. Reg Aneth 1985;10:1-6.
31. Boas RA, Covino BG, Shahnarian A. Analgesic responses to i.v. lignocaine. Br J Anaesth 1982;54:501-5.
32. Abram SE, Yaksh TL. Systemic lidocaine blocks nerve injury-induced hyperalgesia and nociceptor-driven spinal sensitization in the rat. Anesthesiology 1994;80:383-91.