A Combination of Gabapentin and Morphine Mediates Enhanced Inhibitory Effects on Dorsal Horn Neuronal Responses in a Rat Model of Neuropathy

Male Sprague-Dawley rats, initially weighing 130– 150 g, were used in this study. All experimental procedures were approved by the Home Office (animal procedures section, London, United Kingdom) and follow the guidelines under the International Association for the Study of Pain. Selective tight ligation of spinal nerves L5 and L6 and a sham procedure were performed as first described by Kim and Chung (1992). 10Briefly, during gaseous anesthesia, the rat was placed in a prone position, a midline incision was made from L4–S2, and the left paraspinal muscles were separated from the spinous processes. Part of the L6 transverse process was removed to expose the L4 and L5 spinal nerves, and L6 was identified lying just under the sacrum. Using 6-0 silk thread, the left spinal nerves L5 and L6 were tightly ligated distal to their dorsal root ganglion and proximal to their conjunction to form the sciatic nerve. Hemostasis was confirmed, the wound was sutured, and the animal recovered from anesthesia. A sham operation was performed to produce a control group, whereby the surgical procedure was identical to that of the experimental group, but spinal nerve ligation was omitted.

For 2 weeks after surgery, the rats were housed in groups of four in plastic cages under a 12/12 h day/night cycle, and their general health was monitored. Successful reproduction of the neuropathic model was confirmed by behavioral testing at postoperative day 14, assessing the sensitivity of both the ipsilateral and contralateral hind paws to normally nonnoxious punctate mechanical (von Frey filaments) and cooling (acetone) stimuli. Rats were placed in transparent plastic cubicles on a mesh floor and were allowed to acclimatize before tests were initiated. Foot withdrawals to trials of ascending von Frey filaments (bending forces of 1, 5, and 9 g: 9.9, 49.5, and 89.1 mN, respectively), considered nonnoxious under normal circumstances, were quantified. In a trial, a single filament was applied 10 times to the plantar surface of the foot, through the mesh floor, for 2–3 s each time. A period of 3–4 min was left before commencing with the next filament. Foot withdrawals to the application of a drop of acetone to the plantar region of the foot were quantified. In a trial, a drop of acetone was gently squirted through the mesh floor via  a syringe, five times, at 5-min intervals. The number of foot withdrawals were measured on both the ipsilateral and contralateral hind paws for each stimulus modality and were expressed as Difference scores = ipsilateral response − contralateral response, as in the original article 10and as previously described in Chapman et al.  (1998). 11 

Subsequent to behavioral testing, the operated rats were used for electrophysiologic studies at postoperative days 14–17. Briefly, anesthesia was induced with 3% halothane in a mixture of 66% N²O and 33% O², and a cannula was inserted into the trachea. A laminectomy was performed (vertebrae L1–L3) to expose segments L4–L5 of the spinal cord, and the level of halothane was reduced to 1.8%. Extracellular recordings of single convergent neurons, located deep within the dorsal horn (> 500 μm), receiving input from the toe region ipsilateral to the spinal nerve ligation or sham procedure, were made using a parylene-coated tungsten electrode (2 MΩ impedance, 0.005-in uninsulated tip). Neurons selected responded to both noxious (pinch) and nonnoxious (touch) stimuli.

Any spontaneous activity exhibited by a neuron was recorded over 10 min. Action potentials evoked by natural stimuli applied constantly over 10 s were quantified by the application of both punctate mechanical (von Frey filaments, 9 and 75 g) and thermal (constant water jet at 45°C) stimuli applied to the center of the neuron's receptive field. On each separate application, the von Frey filaments were applied to the exact same location, and pressure was applied so as to bend the filament to the same degree. This mediated an accurate and consistent bending force as indicated by the acquisition of stable predrug control responses. The thermal response to 45°C was determined by subtracting the response to 32°C (a nonnoxious temperature so as to ascertain any mechanical response evoked by the water jet) from the response to 45°C. All responses to natural stimuli were normalized by the subtraction of any spontaneous activity over 10 s measured before the application of each stimulus. A representative response of a neuron to punctate mechanical stimuli (von Frey filaments, 9 and 75 g) and noxious heat (constant water jet at 45°C) is shown in figure 1. Response of the neuron to transcutaneous electrical stimulation was established by insertion of two fine needles into the center of its peripheral receptive field and, for the majority of cells involved, adjacent toes. A test consisted of a train of 16 stimuli (2-ms-wide pulse at 0.5 Hz at 3 times the threshold required to evoke a C-fiber response), and a poststimulus histogram was constructed. The thresholds were determined by increasing the electrical stimulus from 0 mA until an action potential was evoked in the corresponding latency band. Electrically evoked action potentials were separated on a latency basis into Aβ fibers (0–20 ms), Aδ fibers (20–90 ms), C fibers (90–300 ms), and postdischarge (300–800 ms). Examples of a representative response of a neuron to a single electrical pulse at 3 times the C-fiber threshold (as seen on the oscilloscope) and a poststimulus histogram after a train of 16 electrical stimuli is shown in figure 1. The “input” is the number of action potentials (90–800 ms) evoked by the first stimulus of the train. “Excess spikes” is a measure of “wind-up,” which is increased neuronal excitability to repeated constant stimulation. Excess spikes was calculated as the total action potentials (90–800 ms) after the 16-stimulus train minus the input times 16.

The testing protocol, initiated every 10 min, consisted of an electrical test followed by the natural stimuli, as described. Stabilization of the neuronal responses was confirmed with at least three consistent predrug responses (< 10% variation) for all measures. These values were then averaged to generate predrug control values with which to compare the effect of drug administration on the subsequently evoked responses. Morphine sulfate and gabapentin (a gift from Parke Davis, Cambridge, United Kingdom) each were dissolved in saline and were administered subcutaneously into the scruff of the neck in volumes of 250 μl. The effect of morphine alone (1 and 4 mg/kg applied cumulatively) was monitored until exertion of maximum effect (a minimum of 50 min). The drug combination protocol consisted of a single dose of gabapentin (either 10 or 20 mg/kg), monitored for 60 min, preceded by two doses of morphine (1 and 4 mg/kg). Reversal of opiate-mediated effects was assessed by spinal application of 5 μg naloxone.

The results were calculated as maximum percentage inhibition from the averaged predrug value for each neuron, and the overall results for each dose were expressed as mean ± standard error of the mean of the normalized data. Statistical analysis of maximal drug effect for each measurement at each dose compared with its averaged predrug value was determined by paired t  test on raw data. An unpaired t  test on the normalized data was used for the comparison of drug effects between different experimental groups and for comparison of the inhibitory effect of morphine in the presence and absence of gabapentin. The level of significance was P ≥ 0.05.

During the postoperative period, the animals showed normal weight gain and maintained good general health. Rats subjected to spinal nerve ligation showed abnormal foot posture ipsilateral to nerve injury whereby toes were held together in a “guarding” behavior. This did not occur either in the contralateral hind paw or in the sham-operated rats. Successful replication of the nerve injury model was confirmed by behavioral testing at postoperative day 14 (fig. 2), which showed the presence of mechanical and cooling allodynia in the injured hind paw of spinal nerve–ligated rats. Evoked allodynia, in response to innocuous mechanical (von Frey filaments, bending force 1–9 g) and cooling (acetone) stimuli, was displayed as a brisk withdrawal, accompanied in some cases by shaking and licking of the foot ipsilateral to spinal nerve ligation. Consistent withdrawal responses were not exhibited by the control group or by the contralateral hind paw of the experimental group, and when present, were never accompanied by the pain-like behaviors displayed by the lesioned hind paw of spinal nerve–ligated rats. The development of neuropathic behaviors has previously been shown to be evident at postoperative day 2, maximal at days 7–12, and still maintained at day 14. 7 

The numbers of ipsilateral dorsal horn neurons characterized in each group were 24 in spinal nerve–ligated rats, 19 in sham-operated rats, and 19 in naive rats. All neurons had a receptive field over the left ipsilateral hind paw. No significant differences were found between experimental groups in the mean values of recorded neuron depth or responses evoked by electrical and natural stimulation. However, the mean level of ongoing spontaneous activity recorded from neurons in nerve-injured rats was 1.73 ± 2.53 Hz, which was significantly higher (P < 0.05) than that observed in sham and naive rats (0.39 ± .89 and 0.42 ± 0.71 Hz, respectively). It is also worth noting that 71% of neurons characterized in spinal nerve–ligated rats showed spontaneous activity at a rate greater than 0.1 Hz in comparison with only 37% of characterized neurons in sham-operated rats and 39% of naive rats.

The effect of subcutaneously administered morphine (1 and 4 mg/kg applied cumulatively) on the electrically and naturally evoked dorsal horn neuronal responses was tested in spinal nerve–ligated, sham-operated, and naive animals. Morphine produced a dose-related inhibition of the evoked responses in neurons recorded from naive and sham-operated animals, with clear effects seen at approximately 40–50 min. The greatest effect was seen in the sham group (figs. 3A–D). In addition to the significances indicated on the measurements displayed in figure 3, in the sham group, both doses of morphine significantly inhibited the excess spike and heat response in comparison to predrug control values, as did 4 mg/kg for the von Frey 9-g evoked response (n = 6, P < 0.05). In addition, for the naive group, the top dose of morphine also significantly inhibited the excess spike measurement (n = 6, P < 0.05). Morphine was noticeably less effective at inhibiting the evoked responses in nerve-injured rats in comparison with naive and sham-operated rats (figs. 3A–D). Two out of 7 cells showed an increase in their response, resulting in an average excitation for the input (fig. 3C) and excess spike measurements at the low dose. The von Frey–evoked responses were significantly inhibited. No statistical significance was determined for the direct comparison of the effects of morphine between experimental groups. Morphine-mediated inhibitions were reversed with spinally applied naloxone, and for the majority of neuronal responses, this often exceeded control values, indicative of some activation of endogenous opioid systems.

The effect of subcutaneously administered gabapentin (10 and 20 mg/kg), on the electrically and naturally evoked dorsal horn neuronal responses, was tested in spinal nerve–ligated, sham-operated, and naive animals. Gabapentin produced a dose-related inhibition of the evoked responses in neurons in spinal nerve–ligated and sham-operated animals, with clear effects seen at approximately 40–50 min. The greatest effect was observed in the neuropathic animals for which statistically significant inhibitions (n = 10, P < 0.05) compared with predrug control values were achieved with 20 mg/kg for all measurements (excluding A fibers, which were relatively spared) (figs. 3A–D). In contrast, in the naive group, gabapentin did not have a constant inhibitory effect on all neuronal measurements. With the exception of the heat response, either or both doses of gabapentin produced an overall excitation of the responses recorded from four out of seven neurons. This was most marked for postdischarge (fig. 3B) and excess spikes (data not shown). No statistical significance was determined for direct comparison of the effects of gabapentin between experimental groups.

The effect of subcutaneously administered morphine (1 and 4 mg/kg applied cumulatively), 60 min after the subcutaneous administration of either 10 or 20 mg/kg gabapentin, on the electrically and naturally evoked dorsal horn neuronal responses was tested in spinal nerve–ligated, sham-operated, and naive animals. In combination with gabapentin, morphine produced a dose-related inhibition of the evoked responses in neurons in naive, sham-operated, and nerve-injured animals, with clear effects seen at approximately 40–50 min, and these were reversed by spinally applied naloxone (fig. 4). Maximal inhibitions within each experimental group were achieved with a combination of 10 mg/kg gabapentin and 4 mg/kg morphine (figs. 3A–D). In each experimental group, these inhibitions were greater than those observed with morphine alone, with the greatest effect observed in the nerve injury group. Further to the measurements displayed in figure 3, in the presence of 10 mg/kg, both doses of morphine significantly inhibited the von Frey 9-g and heat-evoked responses in all three experimental groups. When morphine was used in combination with 20 mg/kg gabapentin, the levels of inhibition reached were no greater than those achieved in the presence of 10 mg/kg gabapentin (figs. 3A–D). The effects of the gabapentin–morphine combination were fairly comparable between sham and spinal nerve ligation groups. However, in the neuropathic group, there was a more marked increase in the inhibitions achieved with morphine in the presence of gabapentin compared with the effects of morphine alone (fig. 3). Because in the sham group the levels of inhibition mediated by morphine alone were greater than compared with that observed after nerve ligation, the increase observed in the presence of gabapentin in the sham group was not as marked. However, statistical significance was found for input (fig. 3C) using 4 mg/kg morphine in conjunction with 20 mg/kg gabapentin and for heat (data not shown) using 1 mg/kg morphine with 20 mg/kg gabapentin. The neuronal responses of the naive group were least inhibited by the drug combination (figs. 3A–D) for which no statistically significant difference was found in the comparison of the effects of morphine in the presence of gabapentin to morphine alone.

Unilateral tight ligation of L5 and L6 spinal nerves produced reproducible nociceptive syndromes in the lesioned hind paw. A clear withdrawal reflex with associated aversive behaviors, indicative of the development of mechanical and cooling allodynia, was produced as previously described. 10,11Thus, all spinal nerve–ligated animals used for the electrophysiology and subsequent pharmacology exhibited neuropathic signs; sham-operated animals did not. This is the first electrophysiologic study examining the effect of a combination of gaba-pentin and morphine on spinal processing of sensory information after nerve injury. In naive, sham-operated, and spinal nerve ligation groups, significant inhibitions of the evoked dorsal horn neuronal responses were achieved using a combination of the two drugs at doses that, when administered alone, lacked any effect. The greatest improvement in the effectiveness of morphine after gabapentin treatment occurred after neuropathy, in which systemic morphine was almost ineffective.

The current study has shown that subcutaneously administered morphine had a reduced inhibitory effect on the electrically evoked dorsal horn neuronal responses in rats subject to spinal nerve ligation in comparison with sham-operated and naive animals. In addition, morphine dose escalation did not improve the inhibition mediated in nerve-injured rats, whereas its effect was dose-related in sham and naive groups. Doses used here were identical to those shown to have effects in behavioral studies of neuropathy and inflammation. 12A reduced sensitivity of evoked neuronal responses in spinal nerve–ligated rats to systemic morphine has previously been demonstrated in a similar electrophysiologic study. 13However, the antiallodynic and antinociceptive abilities of morphine in behavioral studies involving neuropathy are somewhat variable and seem to be dependent on the model of neuropathy used, behavioral assessment, and nature of stimuli used, alongside the route of morphine administration. In accordance with the current study, a reduced effectiveness of intrathecally administered morphine after neuropathy has been observed 14–17and may relate to the observed spinal reduction of opioid receptors ipsilateral to spinal nerve ligation. 18Spinal morphine activates spinal μ receptors, the majority of which are located on the presynaptic terminals of nociceptive C-fiber primary afferent fibers to suppress neurotransmitter release and, to a lesser extent, postsynaptic μ receptors, which result in a hyperpolarization of dorsal horn neurons. A lack of opioid receptors on the terminals of large-diameter, low-threshold Aβ fibers that convey nonnoxious information explains the nociception-specific actions of morphine, as seen in the current study, whereby in all three experimental groups, Aβ fiber–evoked responses were relatively spared in comparison with the C fiber–mediated responses (see also Dickenson and Suzuki 19).

It has been demonstrated here that subcutaneously administered gabapentin, at the highest dose used (20 mg/kg, which was intentionally low for the main focus of the study but interestingly within the range used clinically), had a greater inhibitory effect on the evoked neuronal responses in rats subject to spinal nerve ligation in comparison with sham-operated and naive animals. Effects of gabapentin were established at approximately 40–50 min. The systemic doses used here and the time course of effects observed electrophysiologically are in accordance with those seen behaviorally. 12Behavioral studies show gabapentin has a negligible effect against physiologic sensory nociception, 20–23but is effective in pathophysiologic situations in which central sensitization is present. In various models of nerve injury, gabapentin has been shown to reduce any resultant mechanical or thermal allodynias and hyperalgesias. Systemically, gabapentin is antinociceptive and antiallodynic in nerve injury models 12,20,24–26and reduces evoked dorsal horn neuronal responses after spinal nerve ligation. 27Here, gabapentin was not without inhibitory effect in the experimentally appropriate sham-operated control group, but there was a tendency for the neuronal responses of the unoperated naive group to be facilitated. This is in accordance with other similar studies and may result from the ability of the drug to act in inflammatory states 27,28because for sham surgery, the initial invasive injury may be sufficient to induce some inflammatory response.

Effective systemic doses of gabapentin used in preclinical studies are within a similar range (10–300 mg/kg), which fits well with the doses used in this investigation. Using the spinal route of administration, therapeutic doses of gabapentin required are much lower (10– 100 μg per animal), which indicates a spinal site of action, but the mechanisms of its anticonvulsant–analgesic effects remain elusive. Gabapentin is an analog of GABA designed to cross the blood–brain barrier, but it does not seem to modulate GABA receptor function, and there are no other known neurotransmitter receptor targets (see Taylor et al.  4). A gabapentin binding site has been identified as the modulatory α²δ subunit of voltage-dependent calcium channels, 5and there is an established role for these channels in nociception, the contribution of the different subtypes dependent on the nature of the pain. 6–8Thus, interaction of gabapentin with α²δ may influence neuronal excitability because voltage-dependent calcium channels mediate the release of excitatory neurotransmitters critical for wind-up and central sensitization (see Matthews and Dickenson 7).

The current investigation has convincingly shown that after nerve injury, the evoked dorsal horn neuronal responses, shown to be refractory to systemic morphine, become susceptible to the inhibitory actions of morphine in a dose-related manner when in the presence of systemic gabapentin (itself at a dose that alone mediated negligible neuronal inhibitions). Morphine- and gabapentin-mediated inhibitions were reversed with spinally applied naloxone, and for the majority of neuronal responses, this often exceeded control values, indicative of some activation of endogenous opioid systems. Furthermore, the ability of spinal naloxone to reverse the combined effects of the drugs provides a means to nullify the effects of an overdose but, importantly, reveals that the site of interaction of the systemically applied drugs ultimately resides within the spinal cord. In behavioral studies, systemic gabapentin has been shown to have effects that peak 1 h after administration and that are still apparent at 5 h. 20This indicates that gabapentin would still be active throughout the entire time course of an experiment in the current study in which morphine was administered 1 h after gabapentin and then continued for a further 2 h. Thus, there is great clinical relevance for the treatment of poorly opioid-responsive neuropathic pain for which many cases morphine dose escalation becomes limited by adverse side effects. The majority of the evoked neuronal responses were significantly inhibited by a combination of gabapentin and morphine, with the greatest effects seen on the C-fiber, postdischarge, input, and von Frey–evoked responses. These observations are important because high-threshold C fibers specifically relay nociceptive information, postdischarge is a measure of spinal cord hyperexcitability, and von Frey–evoked responses may relate to behaviorally observed mechanical allodynia or hyperalgesia, all important features of neuropathy. In contrast, the Aβ-fiber response was relatively spared, which may indicate that physiologic touch sensations would be left intact.

There are many benefits to a multidrug regimen. Because neuropathic pain has many causes and resultant symptoms, the underlying dysfunctional mechanisms are likely to be diverse—hence its problematic clinical management. The use of gabapentin to target the excitatory system and morphine to target the inhibitory system tackles a sensitized spinal cord from two angles. It is also likely to offer a better side effect profile because as shown in this study, only low doses of both gabapentin and morphine, each alone being ineffective, were required to mediate significant effects. The use of gabapentin in conjunction with morphine may also impart important characteristics of gabapentin-mediated antinociception seen in preclinical studies, which morphine lacks—in particular, its ability to block both static and dynamic components of mechanical allodynia 12and the lack of tolerance development with chronic use. 21 

There is an increasing amount of clinical data in support of the control of various neuropathic pain with gabapentin, such as postherpetic neuralgia, painful diabetic neuropathy, multiple sclerosis, and trigeminal neuralgia, to name a few. 2Two separate randomized, double-blind studies in postherpetic neuralgia patients 29and patients with diabetic peripheral neuropathy 30reported significant reductions in pain scores compared with placebo over a period of 8 weeks. Furthermore, improvements in sleep and mood were noted for the gabapentin-treated patients, and tolerance did not seem to develop. Reported adverse effects included somnolence and dizziness; however, occurrence of these was minimal and, when experienced, only mild to moderate. Alongside other studies and case reports, it has been proposed that gabapentin is a useful first-line treatment for chronic neuropathic pain, especially when other therapies fail.

The clinical use of opioids in neuropathic pain is more complex, and varied responsiveness is reported, which may be due to the different types of pain, with differing causal dysfunctions. There have been many randomized, double-blind controlled trials conducted in this area, 31and morphine effectiveness ranges from opioid resistant 32to modest pain relief 33to good pain relief. 34It is widely believed that neuropathic pain is less susceptible but not resistant to systemic morphine, 35,36as observed in the current study. Documented morphine side effects include sedation, nausea, vomiting, constipation, respiratory depression, and tolerance, which may prevent dose escalation and thus provide inadequate analgesia. 37Interestingly, the results obtained from this study are generally in accordance with the observed clinical data surrounding the use of morphine and gabapentin alone in neuropathic pain. The limitations of opioids have led to the concept of combination therapy in the hope of attaining a favorable balance of analgesia and side effects with a reduction in the dose. We have demonstrated here the potential for the use of low dose morphine and gabapentin combinational treatment after neuropathy.