Gabapentin Activates Spinal Noradrenergic Activity in Rats and Humans and Reduces Hypersensitivity after Surgery

Male Sprague-Dawley rats (Harlan Industries, Indianapolis, IN), weighing 200–250 g, were used in this study. Animal surgery conformed to the Wake Forest University Guidelines on the ethical use of animals, and studies were performed under Animal Care and Use Committee approval from Wake Forest University School of Medicine, Winston-Salem, North Carolina. Animals were housed under a 12-h light–dark cycle, with food and water available ad libitum .

Paw incision was performed as previously described.14Animals were anesthetized with halothane, the plantar surface of the right hind paw was prepared with 70% ethanol, and a 1-cm longitudinal incision was made through the skin and fascia, starting 0.5 cm from the edge of the heel and extending toward the toes. The plantaris muscle was elevated and incised longitudinally. The wound was then closed with two silk sutures. Behavior tests were performed 24 h after paw incision.

Animals were anesthetized with halothane, and intrathecal catheters were implanted as previously described.17Animals were placed prone in a stereotaxic frame, and a small incision was made at the back of the neck. A small puncture was made in the atlanto-occipital membrane of the cisterna magnum, and a polyethylene catheter, 8.5 cm, was inserted so that the caudal tip reached the lumbar enlargement of the spinal cord. The rostral end of the catheter was exteriorized at the top of the head, and the wound was closed with sutures.

During anesthesia with an intraperitoneal injection of sodium pentobarbital (50 mg/kg) and atropine (0.1 mg/kg), animals were placed securely into a stereotaxic device (KOPF, Tujunga, CA), and a 1-cm midline incision was made to expose the skull surface. A sterile stainless steel guide cannula (22-gauge needle shaft; Plastics One, Roanoke, VA) was implanted into the left lateral cerebral ventricle. The coordinates for placement of the tip of the guide cannula were 0.80 mm posterior and 1.5 mm lateral to the bregma, and 3.5 mm ventral from the surface of the dura mater, according to the rat brain atlas.18Each cannula was kept in place with acrylic dental cement secured by two stainless steel skull screws and was fitted with a dummy cannula to maintain patency, and then the wound was closed with sutures.

After implantation of the intracerebroventricular and/or intrathecal catheters, rats were housed individually with free access to food and water. Animals were allowed at least 5 days to recover from the surgery, and those displaying signs of motor dysfunction (forelimb or hind limb paralysis) were to be excluded from the study and killed immediately. No animal in the current study was excluded for these reasons.

Nociceptive mechanical thresholds, expressed in grams, were measured with the Randall-Selitto test using an analgesimeter (Ugo Basile, Comerio, Italy). The test was performed by applying pressure to the hind paw. By pressing a pedal that activated a motor, the force increased at a constant rate on a linear scale. When the animal withdrew the paw or vocalized, the pedal was immediately released and the nociceptive threshold was read on a scale. A cutoff of 250 g was used to avoid potential tissue injury.

All drugs were purchased from Sigma (St. Louis, MO) except gabapentin solution (Neurontin, 50 mg/ml; Park-Davis, New York, NY). For oral administration, gabapentin solution was diluted with sterilized water and given to the animals with a feeding tube (30–300 mg/6 ml/kg). For intracerebroventricular administration, gabapentin hydrochloride was dissolved in saline and injected (30–100 μg/5 μl/rat). For intrathecal administration, atropine sulfate, idazoxan hydrochloride, and tertiapin-Q trifluoroacetate were dissolved in saline and were injected in a volume of 10 μl followed by 10 μl saline at 1.5 h after oral gabapentin or at the same time as intracerebroventricular gabapentin administration. Drug doses and timing of administration were determined in preliminary experiments or from previously published ranges.

After approval was obtained from the Institutional Review Board of Wake Forest University School of Medicine, 44 patients with American Society of Anesthesiologists physical status I–III who were scheduled to undergo orthopedic or urogenital surgery during spinal or combined spinal–epidural anesthesia were recruited. All patients provided written informed consent. Patients taking or allergic to gabapentin were excluded, as were those with leukopenia or renal insufficiency or failure. On the day of surgery, patients were randomly assigned to receive, approximately 90 min before surgery, 1,200 mg gabapentin (n = 21) or placebo (n = 23) using pills prepared by our research pharmacy. The study was double blind. At the time of anesthesia initiation, a 25- or 27-gauge Whitacre needle was inserted in a mid or lower lumbar interspace, and 3 ml clear CSF was aspirated for norepinephrine assay. Intraoperative anesthesia was provided with spinal or spinal epidural anesthesia with sedation or general anesthesia. CSF samples were stored at −80°C until norepinephrine assay. Patients were visited on the first postoperative day and asked to provide a global pain score since surgery, using a 10-cm visual analog scale. We did not control the method of postoperative analgesia, but recorded morphine use during the first 24 h.

Norepinephrine content of CSF was determined as previously described.19After treatment with 0.1 m perchloric acid, the CSF samples were centrifuged. Supernatants were collected, and norepinephrine content was measured by high-pressure liquid chromatography with electrochemical detection.

Unless otherwise indicated, data are presented as mean ± SEM. Differences among groups for withdrawal threshold were determined using an unpaired t  test, one-way analysis of variance for repeated measures followed by the Dunnett post hoc  test within groups, or two-way analysis of variance for repeated measures followed by the Holm-Sidek post hoc  test among groups. CSF concentrations of norepinephrine were not normally distributed, are presented as median ± 25th and 75th percentiles, and were compared using the Mann–Whitney rank sum test. The significance level was set at P < 0.05.

Paw incision decreased mechanical withdrawal threshold of the incised paw from 191 ± 25 g (mean ± SD) to 81 ± 17 g, accompanied by swelling of the paw 24 h after surgery. Withdrawal threshold of the contralateral paw did not change after surgery (data not shown).

Oral (30–300 mg/kg) and intracerebroventricular gabapentin produced a dose-dependent increase in withdrawal threshold (fig. 1). Oral gabapentin increased withdrawal threshold compared with vehicle 1–4 h after administration (fig. 1A), whereas intracerebroventricular administration of 30 and 100 μg/rat of gabapentin increased withdrawal threshold compared with vehicle only 15–60 min after administration (fig. 1B). Based on these data, we used 100 mg/kg gabapentin orally and 100 μg/rat intracerebroventricularly in subsequent experiments.

Intrathecal administration of the α²-adrenergic receptor antagonist idazoxan (30 μg/rat) alone did not change withdrawal threshold of the incised paw but significantly blocked oral gabapentin analgesia (fig. 2A). Similarly, intrathecal idazoxan (3 and 30 μg/rat) reduced analgesia from intracerebroventricular gabapentin (fig. 2B).

Intrathecal administration of the non–subtype-selective muscarinic receptor antagonist atropine (30 μg/rat) was used to test whether a cholinergic component is involved in gabapentin analgesia. This dose of atropine, which diminished the effect of intrathecal α²-adrenergic receptor agonists on hypersensitivity after surgery,20did not change withdrawal threshold of the incised paw when administered alone and did not reverse the effect of gabapentin (fig. 3).

Because α²-adrenoceptor activation stimulates GIRKs,15we used the GIRK blocker tertiapin-Q21to test this potential downstream mediator of gabapentin analgesia. Intrathecally administered tertiapin-Q (1.0 μg/rat) alone did not change withdrawal threshold of the incised paw (fig. 4A) but dose-dependently inhibited oral gabapentin analgesia (fig. 4B). Higher doses than 1.0 μg/rat were not tested because they resulted in continual tail-flick behavior (data not shown).

The group did not differ in demographic variables, with the study population as a whole 57 ± 2.4 yr in age, 92 ± 3.6 kg in weight, 171 ± 1.3 cm in height, and 43% female. The group also did not differ in surgery type, with 86% orthopedic and 14% urogenital procedures in the study population as a whole. Of the orthopedic procedures, 24% overall were hip replacement and 76% were open knee procedures. Groups did not differ in preoperative pain score (overall 4.3 ± 0.4 cm), and only two patients, one in each group, had no pain before surgery.

The CSF concentration of norepinephrine was greater in patients treated with gabapentin (median [25th–75th percentile], 461 [400–864]) than in patients who received placebo (329 [238–432]; P < 0.005). The groups did not differ in time from oral intake of study drug until CSF sampling (105 ± 8 min in the gabapentin group vs.  143 ± 13 min in the placebo group), and there was no significant correlation between this time and CSF norepinephrine concentration in either group or in the population as a whole.

The purpose of this study was not to determine the effect of oral gabapentin on postoperative pain, and we did not control the method of postoperative pain treatment. As such, postoperative analgesia was provided using a variety of methods, with sustained-release epidural morphine (n = 7), peripheral nerve catheter infusion of local anesthetic (n = 21), epidural catheter infusion of local anesthetic with opioid (n = 3), or systemic analgesic (n = 11), with a similar distribution between gabapentin and placebo groups. Despite this variability, the amount of opioid administered in the first 24 h was lower in gabapentin-treated patients (32 ± 5.2 mg morphine equivalents) than in patients who received placebo (50 ± 9.3 mg; P = 0.05). Postoperative pain score, however, did not differ between groups (4.1 ± 0.5 vs.  4.4 ± 0.6 in the gabapentin and placebo groups, respectively).

Our understanding of postoperative pain has been considerably advanced by the recent development of animal models of incisional14and deep tissue22surgical injury. Studies in these models have shown that postoperative pain, once considered akin to acute inflammation, rather reflects mechanisms that overlap with, but also differ from those of inflammation and nerve injury. It was therefore only slightly surprising that gabapentin, a drug used to treat chronic neuropathic pain, also reduced hypersensitivity from incisional surgery in rats9,11and postoperative analgesia in humans.10The current study adds significantly to our understanding of the mechanisms by which gabapentin is effective in the postoperative setting.

Gabapentin reduces hypersensitivity from nerve injury in rodents3–5and patients,2and its mechanisms of action in this setting have been extensively studied. Because spinal plasticity and sensitization are recognized as critical components in pain amplification after peripheral nerve injury and inflammation, most studies have focused on peripheral afferents and spinal cord neurons. Recently, Tanabe et al.  5demonstrated that gabapentin acts on supraspinal structures to stimulate the descending noradrenergic system after nerve injury in mice. Norepinephrine is released in the spinal dorsal horn by descending noradrenergic axons, which mainly originate from the locus ceruleus and adjacent nuclei in the brainstem, and produces analgesia by stimulating α²adrenoceptors.12Peripheral inflammation enhances descending noradrenergic inhibition and increases sensitivity of spinal neurons to α²-adrenoceptor agonists such as clonidine.13Incision of the paw also activates descending noradrenergic inhibition, as evidenced by efficacy of intrathecal administration of norepinephrine transporter inhibitor Xen2714 after incisional surgery.23 

The current study indicates that incisional surgery also alters the ability of gabapentin to activate descending noradrenergic inhibition, because systemic or intracerebroventricular gabapentin did not produce antinociception in normal animals,5,24but reduced hypersensitivity after surgery. The rapid onset of effect from intracerebroventricular gabapentin (15 min) is consistent with a supraspinal site of action, although we cannot absolutely exclude the possibility that gabapentin was distributed to the spinal cord during this period of time. Consistent with studies in a neuropathic pain model,5the antihypersensitivity effect of both systemic and intracerebroventricular gabapentin was significantly blocked by intrathecal injection of the selective α²-adrenoceptor antagonist idazoxan. Intrathecal idazoxan alone did not affect hypersensitivity in this model, consistent with a lack of tonic descending inhibition, or with a floor effect of hypersensitivity after surgery. Blockade of intracerebroventricular gabapentin's effect by intrathecal idazoxan in the current study, and its blockade by depletion of spinal norepinephrine after 6-hydroxydopamine after nerve injury,5argues strongly that gabapentin acts at supraspinal sites primarily to stimulate the descending noradrenergic system. Although we could not administer intrathecal idazoxan to humans, the current study, which showed increased norepinephrine concentrations in CSF after gabapentin compared with placebo, supports the clinical relevance of these observations in rats.

The mechanisms by which gabapentin activates noradrenergic neuron in the brainstem are still unknown. A direct effect seems unlikely, both because gabapentin does not exert direct excitatory effects on neurons and on previous observations that gabapentin inhibits rather than excites norepinephrine release in other sites in the brain.25Gabapentin binds with high affinity to α²δ subunits of voltage-gated calcium channels,6and its efficacy depends on the presence of α²δ subunits.7,8Because gabapentin has no antinociceptive effects against acute pain,5,24α²δ subunits or other heretofore uncharacterized gabapentin binding sites in supraspinal structures may be up-regulated or novelly expressed after incisional surgery. Further studies are ongoing in our laboratory to clarify the supraspinal mechanisms of gabapentin to treat postoperative hypersensitivity.

Stimulation of spinal cholinergic circuits by activation of spinal α²adrenoceptors is widely documented in humans and animals.26We hypothesized that gabapentin analgesia in the postoperative model might rely on this circuit, because intracerebroventricular gabapentin analgesia is completely blocked by systemic or intrathecal pretreatment with atropine after peripheral nerve injury.27This hypothesis was not supported, because atropine did not block the effect of gabapentin after paw incision. Sensitivity to atropine inhibition of spinal α²-adrenoceptor activation differs between neuropathic and postoperative pain models. We previously reported that clonidine increased acetylcholine release in spinal cord slices from nerve-injured but not normal rats.28This agrees with the observation that the antihypersensitivity effect of intrathecal clonidine to mechanical stimuli is abolished by intrathecal atropine in nerve injured rats, but its antinociceptive effect is not inhibited by atropine in normal rats.16,29 

G protein–coupled inwardly rectifying potassium channels have been identified in the spinal dorsal horn and play an important role in analgesia induced by various drugs, including α²-adrenoceptor, muscarinic, and opioid receptor agonists.15,21,30We observed in the current study that oral gabapentin analgesia was strongly blocked by intrathecal administration of the GIRK blocker tertiapin-Q, consistent with gabapentin-induced activation of G protein–coupled receptors in the spinal cord which interact with GIRKs. Because gabapentin does not directly affect GIRK activity,31these data suggest that gabapentin-induced norepinephrine and acetylcholine release in the spinal cord causes GIRK activation, which participates in its antihypersensitivity effect.

To our knowledge, this is the first report to test a mechanism of gabapentin action in humans. Oral gabapentin, in a dose previously demonstrated to produce postoperative pain relief,10significantly increased norepinephrine concentration in CSF compared with placebo group, consistent with our animal data of activation of the descending noradrenergic pathway by gabapentin. Nearly all of the patients in the current study had chronic pain, primarily from orthopedic causes, so we do not know whether similar results would be obtained in patients without pain before surgery. Oral gabapentin efficacy in a variety of surgical populations, however, argues that its action does not depend on the presence of pain before surgery. Gabapentin reduces pain intensity and opioid consumption after surgery,10whereas the current study only observed a reduction in opioid consumption in patients treated with gabapentin. This slight discrepancy likely reflects the fact that we did not control the method of postoperative analgesia. In addition, although oral gabapentin was associated with higher concentrations of norepinephrine in CSF at the time of surgery than placebo, whether this was sustained during the postoperative period and accounted for or contributed to analgesia from gabapentin is not determined by the current study.

In summary, oral and intracerebroventricular gabapentin reduces hypersensitivity in a rat model of postoperative pain. This effect is completely blocked by intrathecal idazoxan, suggesting that gabapentin analgesia is mainly mediated by spinal α²adrenoceptors as has been previously reported in animals after nerve injury. Gabapentin efficacy after incisional surgery is also reduced by intrathecally administered muscarinic antagonist and a GIRK inhibitor, consistent with spinal activation of α²adrenoceptors and a cholinergic circuit. The clinical relevance of these observations is supported by increased CSF concentration of norepinephrine in patients receiving oral gabapentin before surgery. These data suggest that gabapentin produces postoperative analgesia in part by activating descending noradrenergic inhibition, and that its effect may therefore be augmented by drugs that further amplify the consequence of norepinephrine release.

The authors thank Xinhui Li, Ph.D. (Assistant Professor of Anesthesiology, Wake Forest University School of Medicine, Winston-Salem, North Carolina), for performing the norepinephrine assays.