Synergistic Effect between Intrathecal Non-NMDA Antagonist and Gabapentin on Allodynia Induced by Spinal Nerve Ligation in Rats

Male rats (Harlan Sprague-Dawley) weighing 150–180 g at the time of acquisition were used in this study. Ligation of the L5 and L6 spinal nerves in rats was used in this study as an experimental model of neuropathic pain because it produces profound and sustained tactile allodynia, which resembles the condition observed in patients with neuropathic pain. 19,20During halothane anesthesia, the left L5 and L6 spinal nerves were isolated under a surgical microscope and ligated tightly with 5-0 silk suture, according to the method described by Kim and Chung. 19The animals were allowed to recover for 5–7 days before intrathecal cannulation. Intrathecal catheters (PE-10 tubing) were inserted in rats during halothane anesthesia, as previously described by Yaksh and Rudy. 21Intrathecal catheters were advanced 8 cm caudal through an incision in the cisternal membrane and secured to the musculature at the incision site. Only animals with no evidence of neurologic deficits after catheter insertion were studied. All pharmacologic experiments were conducted between 3 and 4 weeks after spinal nerve ligation because tactile allodynia develops within 1 week after surgery and lasts for at least 6–8 weeks. Each rat received only a single intrathecal injection of drugs. The surgical preparations and experimental protocols were approved by the Animal Care and Use Committee at the Wake Forest University School of Medicine.

The mechanical threshold was determined before and after spinal nerve ligation in all animals. To quantify mechanical sensitivity of the paw, rats were placed in individual plastic boxes on a mesh floor, which allowed access to their hind paw. A series of calibrated von Frey filaments (Stoelting Co., Wood Dale, IL) were applied perpendicularly to the plantar surface of the left paw with sufficient force to bend the filaments for 6 s. Brisk withdrawal or paw flinching were considered as positive responses. In the absence of a response, the filament of next greater force was applied. In the presence of a response, the filament of next lower force was applied. The tactile stimulus producing a 50% likelihood of withdrawal was determined using the “up–down” calculating method, as described in detail by Chaplan et al.  20,22Each trial was repeated 2 or 3 times at approximately 2-min intervals, and the mean value was used as the force to produce withdrawal responses. Motor dysfunction was evaluated by testing the animals’ ability to stand and ambulate in a normal posture and to place and step with the hind paws. We assessed the motor function in a simple manner by grading the ambulation behavior of rats as the following: 2 = normal; 1 = limping; 0 = paralyzed. The rats were tested for 2 min once every test period.

This protocol was used to determine the dose–response and time course of the analgesic action of intrathecal injection of a non-NMDA receptor antagonist, CNQX. Other non-NMDA receptor antagonists, such as 2,3-dihydro-6-nitro-7-sulfamoyl-benzo(f)quinoxaline (NBQX) and DNQX, were not chosen for this study because of their effect of impairment on motor function after intrathecal injection or their poor solubility in saline. 10,12Rats were acclimated for 30 min before testing. Then, baseline withdrawal thresholds to von Frey filament stimulation were determined. Rats were assigned randomly to three groups receiving intrathecal injection of three different doses of CNQX: 0.5 μg (n = 6); 2 μg (n = 7); and 8 μg (n = 7). These doses were selected based on our pilot study in which we found that 0.2 μg CNQX had no effect, whereas 10 μg CNQX produced a maximal effect on the paw withdrawal responses to mechanical stimulation. The mechanical thresholds were determined 15, 30, 45, 60, 90, and 120 min after treatment.

This protocol was used to determine the dose-dependency and time course of the analgesic action of intrathecal injection of gaba-pentin. After the baseline withdrawal threshold was measured, separate animals were randomly organized in the following three groups for intrathecal administration of three doses of gabapentin: 10 μg (n = 7); 30 μg (n = 7); and 100 μg (n = 7). These doses were chosen based on preliminary experiments and a previous study. 17A previous study by Hwang and Yaksh 17showed that the antiallodynic effect of intrathecal gabapentin in this animal model lasts about 4 h.

To study the interaction between CNQX and gabapentin, an isobolographic analysis was performed. 23The respective ED⁵⁰values were calculated from the dose–response curve of CNQX or gabapentin alone. Then, a dose–response curve was obtained by administration of both agents in a constant dose ratio (gabapentin:CNQX, 15:1, see Results) based on the ED⁵⁰values of the single drug. Twenty additional rats randomly received one of the following combinations of gabapentin and CNQX: 4.5:0.3 μg (n = 7); 18:1.2 μg (n = 7); or 72:4.8 μg (n = 6). The mechanical thresholds were determined 15, 30, 45, 60, 90, and 120 min after intrathecal treatment. Subsequently, from the dose–response curve of the combined drugs, the ED⁵⁰value of the total dose of the mixture was determined using linear regression. In addition, we performed experiments using separate rats (n = 6) to show that the effect of CNQX plus gabapentin was reversible. We evaluated the effect of a combination of CNQX and gabapentin (4.8:72 μg) for the first 5 h and 20 h after intrathecal injection.

Both drugs for intrathecal injection were dissolved in normal saline and administered in a volume of 5 μl, followed by a 10-μl flush with normal saline. Previous studies have shown that intrathecal injection of saline has no effect on the allodynic condition in spinal nerve–ligated rats. 10,22Gabapentin was supplied by Parke-Davis Pharmaceutical Research (Ann Arbor, MI) and CNQX disodium, a water-soluble formulation of CNQX, was purchased from Research Biochemical International (Natick, MA).

To calculate the ED⁵⁰value of each drug, the response threshold data were converted to a percentage of the maximum possible effect (%MPE) according to the following formula: The baseline threshold was the withdrawal threshold after nerve ligation and recovery. ED⁵⁰was determined using linear regression for each drug. Isobolographic analysis using the ED⁵⁰for maximal effect was performed as described previously. 23Data are presented as the mean ± SD. The paw withdrawal thresholds in response to mechanical stimulation before and after nerve ligation were compared using a paired Student t  test. Effects of drugs on allodynia were determined by repeated-measures analysis of variance, followed by the Tukey post hoc  test. P < 0.05 was considered to be statistically significant.

The tight ligation of spinal nerves reliably produced allodynia, as defined by a marked reduction in the tactile stimulus necessary to evoke withdrawal of the paw ipsilateral to the nerve lesion. The paw withdrawal threshold before spinal nerve ligation was 28.7 ± 2.4 g. The mechanical threshold decreased significantly to 2.4 ± 0.5 g (P < 0.05) within 7 days after nerve ligation and remained stable for at least 6 weeks in all animals studied.

Intrathecal injection of CNQX increased significantly the withdrawal threshold in a dose-dependent manner (fig. 1). Blockade of non-NMDA receptors with CNQX produced a profound but less-sustained reduction in allodynia. The antiallodynic effect of intrathecal injection of three different doses of CNQX quickly reached a maximum within 30 min, then gradually decreased and lasted approximately 60 min (fig. 1). The ED⁵⁰of CNQX was 3.4 ± 0.22 μg.

Intrathecal administration of gabapentin also produced a dose-dependent antiallodynic effect (fig. 2). The ED⁵⁰of gabapentin was 45.9 ± 4.65 μg. In contrast to the time course of the effect of CNQX, intrathecal injection of gabapentin gradually increased the baseline withdrawal thresholds, with a peak effect at 45 min (fig. 2). The antiallodynic effect of gabapentin lasted for at least 2 h (fig. 2).

Intrathecal coadministration of gabapentin and CNQX in a constant ratio of the ED⁵⁰s showed a significant increase in the withdrawal threshold to mechanical stimulation in a dose-dependent fashion (fig. 3). Isobolographic analysis showed that this interaction was synergistic because the dose of CNQX plus gabapentin necessary to produce a 50% return to presurgery withdrawal threshold was significantly less than that calculated as necessary for an additive interaction (fig. 4). The antiallodynic effect of CNQX plus gabapentin decreased gradually toward the control level over a 5-h period, and the tactile threshold returned completely to the baseline level within 20 h (fig. 5).

Animals receiving intrathecal injection of CNQX, gabapentin, or coadministration of gabapentin plus CNQX did not exhibit motor dysfunction, as judged by the placing–stepping reflex and ambulation behavior or other visible behavioral changes, such as sedation or agitation. The ambulation score was not affected by CNQX, gabapentin, and their combination in the doses used. All rats tested reached the maximum score.

There are two new observations from the current study. First, intrathecal administration of a non-NMDA receptor antagonist, CNQX, attenuated significantly the allodynic condition in a rat model of neuropathic pain without evident effect on motor function. These results support the notion that spinal non-NMDA receptors are important for the maintenance of pain behaviors in this rat model of neuropathic pain. Second, intrathecal CNQX and gabapentin produced a potent synergistic analgesic effect in spinal nerve–injured rats. Therefore, data from the current study suggest that intrathecal coadministration of non-NMDA receptor antagonists and gabapentin may be more effective for treatment of chronic neuropathic pain.

Excitatory amino acids, such as glutamate and aspartate, are important in the transmission of nociceptive inputs in the dorsal horn of the spinal cord. 3,5,6These excitatory amino acids are known to facilitate spinal sensory transmission and contribute to the enhanced excitability of dorsal horn neurons in the chronic pain state through activation of NMDA and non-NMDA receptors. 3,5,24Non-NMDA receptors are important for the processing of noxious and innocuous mechanical stimuli in the dorsal horn of the spinal cord. 5,6The non-NMDA receptors are ligand-gated ion channels and are permeable to sodium and calcium ions. These receptors can be activated by kainate or α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA), from which these pharmacologic names are derived. Non-NMDA receptors mediate fast excitatory neurotransmission in the central nervous system, including the spinal cord dorsal horn neurons. 5,25Receptors for AMPA and kainate both are expressed in the dorsal horn of the spinal cord. 4,26Non-NMDA receptors are present in the superficial lamina of the dorsal horn of the spinal cord and are activated by stimulation of nociceptors and mechanoreceptors. 26,27It has been reported that spinal non-NMDA receptors are upregulated after peripheral nerve injury, 24and non-NMDA receptor antagonists suppress dorsal horn neuronal responses to mechanical stimuli in rats subjected to spinal nerve ligation. 3The antiallodynic effects of non-NMDA antagonists on neuropathic pain remain controversial because one previous study failed to show such an analgesic effect in a neuropathic pain model induced by sciatic nerve ligation. 12Although DNQX has been used intrathecally in the same neuropathic pain model, 10the focus of that study was on NMDA, not on non-NMDA receptors. Furthermore, the analgesic effect of DNQX is largely limited by its effect (or the vehicle used;i.e. , NaOH) on the motor function. 12In the current study, to avoid the undesirable effects of some non-NMDA antagonists (i.e. , NBQX and DNQX) and their vehicles on the motor function, we evaluated the antiallodynic action of a preformed water-soluble formulation of CNQX in a well-established rat model of neuropathic pain. We observed that intrathecal injection of CNQX up to 8 μg produced no detectable motor dysfunction, which is in agreement with the findings of others using this compound. 7,28Our results indicate that intrathecal CNQX possesses an antiallodynic effect on this neuropathic pain model without affecting the motor function. This information is important in establishing the role of spinal non-NMDA receptors in the maintenance of allodynia in this rat model of neuropathic pain. Leem et al.  3demonstrated that NMDA and non-NMDA receptor antagonists differentially suppress the responses of spinal dorsal horn neurons to noxious and non-noxious mechanical stimuli, respectively. Thus, the preferential effect of CNQX on innocuous peripheral input may be the basis of its antiallodynic effect on this rat model of neuropathic pain. Recently, the effect of another non-NMDA antagonist, LY29358, on capsaicin-induced allodynia has been shown in humans. 29We found that intrathecal administration of CNQX produced a significant antiallodynic effect, which lasted approximately 60 min. Therefore, blockade of non-NMDA receptors alone may be insufficient to produce a sustained antiallodynic action in this model. For this reason, it would be logical to combine this class of drugs with other agents for a prolonged effect.

The antiallodynic effect of intrathecal gabapentin observed in the current study is similar to what has been reported by others. 17Intrathecal injection of gabapentin had a slow onset of action but was long-lasting (> 2 h). Systemically administered gabapentin reduces the responses of spinal dorsal horn neurons to mechanical stimuli, indicating that the spinal cord is one of the sites of action of gabapentin. 18Although the mechanisms of the antinociceptive action of gabapentin in the spinal cord remain unclear, previous studies have suggested that gabapentin treatment elevates GABA concentration and decreases glutamate concentration in rat brain tissues. 30,31However, gabapentin may have effects through other receptors or ion channels. For example, recent studies have shown that gabapentin has a high affinity for the α²δ-subunit of voltage-sensitive calcium channels in the brain tissue, 32an action that may reduce intracellular calcium accumulation and attenuate neuronal excitability after nerve injury. Calcium channel activation is know to play a role in allodynia development because intrathecal injection of N-type calcium channel blockers alleviates allodynia induced by spinal nerve ligation in rats. 33Conversely, in vitro  electrophysiologic experiments have failed to show any effect of gabapentin on voltage-sensitive calcium channels. 15Further study is necessary to elucidate the pharmacologic action of gabapentin in the spinal cord.

The interaction between intrathecal gabapentin and non-NMDA receptor antagonists on chronic neuropathic pain has not been studied. The current study indicates that coadministration of gabapentin and CNQX resulted in a greater reduction of allodynia induced by ligation of spinal nerves in rats. This functional synergy may result from a distinct drug interaction that acts independently to inhibit spinal nociceptive processing. Because activation of non-NMDA receptors, N-type calcium channels, and a decrease in GABA availability in the spinal cord have been considered to play a role in the development of neuropathic pain, 3,5,33the synergistic action between intrathecal gabapentin and CNQX may be achieved by blockade of non-NMDA receptors, inhibition of voltage-gated calcium channels, and an increase in GABA content in the spinal cord. As shown in this study, the benefits of these two agents, when combined, are not only reflected by the increased potency (synergistic interaction), but also characterized by a rapid and prolonged duration of analgesia in this rodent model of neuropathic pain. It should be acknowledged that the threshold after intrathecal gabapentin plus CNQX (72:4.8 μg;fig. 3) was significantly higher than the preligation threshold. It appears that the combination of these two drugs produces an antinociceptive effect in addition to a reversal of the allodynic condition in this neuropathic pain model. However, their interaction in inhibiting noxious sensory inputs in the spinal cord was not studied in the current study. Alternatively, because non-NMDA antagonists attenuate painful responses to the paw formalin injection in rats, 34,35it is possible that a large dose of CNQX alone could produce an antinociceptive effect. We avoided using a high dose of CNQX in this study because it may cause unspecific effects, as reported by Ren et al.  7 

In summary, this study shows that spinal non-NMDA receptors play an important role in the maintenance of neuropathic pain. Furthermore, intrathecal administration of a non-NMDA receptor antagonist plus gabapentin has a synergistic antiallodynic effect in this model of neuropathic pain. The synergy may arise from a decrease in excitatory neurotransmission (CNQX) accompanied by inhibiting calcium channels (gabapentin) in the spinal cord. Therefore, intrathecal administration of non-NMDA receptor antagonists and gabapentin may represent an alternative therapy for patients with neuropathic pain conditions.