Intracerebroventricular Morphine Produces Antinociception by Evoking γ-Aminobutyric Acid Release through Activation of 5-Hydroxytryptamine 3 Receptors in the Spinal Cord

The protocol for this study was approved by Sapporo Medical University Animal Care and Use Committee (Sapporo, Japan). Experiments were conducted on male Sprague-Dawley rats (weight, 250–350 g; Japan SLC, Hamamatsu, Japan) housed individually in a temperature-controlled (21 ± 1°C) room with a 12-h light–dark cycle and given free access to food and water. Each animal was used in only one experiment.

Rats were anesthetized with pentobarbital (50 mg/kg administered intraperitoneally) and were placed in a stereotaxic head holder to implant an intracerebroventricular guide cannula. The skull was exposed, and a guide cannula equipped with a 32-guage stylet was directed into the left lateral ventricle according to Paxinos and Watson 7(anterior-posterior +0.8 mm, midline +1.4 mm, dorsal-ventral 3.8 mm with respect to bregma) and was attached to the skull with three anchoring screws and cemented in place. For intracerebroventricular injection, the stylet was removed and the injection cannula was inserted.

For the behavioral study, a lumbar intrathecal catheter was implanted in the rats with an intracerebroventricular guide cannula. A polyethylene intrathecal catheter (PE-10; Becton Dickenson, Franklin Lakes, NJ) was inserted 15 mm cephalad into the lumbar subarachnoid space at the L4–L5 intervertebral level, with the tip of the catheter located near the lumbar enlargement of the spinal cord to administer the drugs intrathecally. The catheter was tunneled subcutaneously and externalized through the skin in the neck region. At least 6 days of postsurgical recovery were allowed before animals were used in the behavioral study.

The spinal cord dialysis probe was prepared according to our modification of the method described by Skilling et al.  8The dialysis probe was constructed from a 1-cm length of dialysis fiber (an ID of 200 μm, an OD of 220 μm, and 50-kd molecular weight cutoff; DM-22, Eicom Co., Kyoto, Japan) which was coated with a thin layer of epoxy glue (Devcon Co., Danvers, MA) along the whole length, except for a 2-mm region in the middle. To make the fiber firm enough for implantation, a Nichrome-Formvar wire with a 78-μm ID (A-M Systems, Inc., Everret, WA) was passed through the fiber. Each end of the fiber was attached to 2-cm polyethylene catheters (PE-10; Becton Dickinson, Franklin Lakes, NJ), and each end of the polyethylene catheters was then attached to an 8-cm teflon tube with an ID of 100 μm and an OD of 400 μm (JT-10; Eicom). Seven days after an intracerebroventricular guide cannula implantation, rats were anesthetized with pentobarbital (50 mg/kg administered intraperitoneally) and incised along the dorsal midline from T2 to L2. The lateral surfaces of vertebra L1 were exposed, and bilateral holes were carefully made through the bone, exposing the spinal cord laterally at the level of the dorsal horn. A dialysis tube was placed through the holes by hand, passing transversely through the dorsal spinal cord. The two distal ends of the probe were tunneled subcutaneously and externalized through the skin in the neck region. The experiments were performed 18–24 h after the implantation of the dialysis probe. After a recovery period, the animals showing any signs of limb paralysis or impaired movement were excluded in this study. After each experiment, we perfused methylene blue dye through the dialysis probe to verify the position of the dialysis fiber, and then rats were killed with an overdose of pentobarbital. The data to be reported are from rats in which methylene blue dye remained in the area of the dorsal half of the dorsal horn.

The animals freely moved in a plastic cage with dimensions of 30 × 30 × 35 cm during the dialysis experiments. A liquid switch (SI-50; Eicom) was placed between the syringe pump and the dialysis probe to enable a different drug to be administered locally via  the probe. The dialysis probe was perfused with artificial cerebrospinal fluid (ACSF; 140 mm NaCl, 4.0 mm KCl, 1.26 mm CaCl², 1.15 mm MgCl², 2.0 mm Na²HPO⁴, 0.5 mm NaH²PO⁴, pH 7.4) at a constant flow rate of 3 μl/min. The samples were collected as 20-min fractions and divided into two samples: 20 μl dialysate collected for analysis of GABA and 20 μl for 5-HT. The samples were frozen at −80°C until analysis. Three consecutive samples were collected for determination of basal concentrations 180 min after starting perfusion of ACSF. After obtaining three consecutive samples, 40 nmol/3 μl of intracerebroventricular morphine (Sankyo Co., Tokyo, Japan) was administered over a 120-s period. The samples were collected up to 360 min after morphine administration. To examine the naloxone sensitivity and the involvement of spinal 5-HT³receptor activation in morphine-induced spinal GABA release, either intracerebroventricular naloxone (Sigma, St.Louis, MO) or the selective 5-HT³receptor antagonist, 3-tropanyl-indole-3-carboxylate methiodide (TICM; Research Biochemical International, Natick, MA) was administered 180 min after morphine administration. Naloxone 40 nmol was injected in a volume of 3 μl over a 120-s period. TICM (1 mm) was perfused through the dialysis catheter for 60 min. All drugs used in microdialysis study were prepared in ACSF.

Serotonin in the dialysate was analyzed using high-performance liquid chromatography with electrochemical detection (Degasser, DG-100; liquid chromatograph, EP-100; electrochemical detector, ECD-100; Eicom). The chromatographic conditions were as follows: column, Eicompak (CA-5ODS 2.1 × 150 mm; Eicom); mobile phase, 0.1 m phosphate buffer (pH 6.0) containing 20.0% methanol, 0.02 mm EDTA, and 0.72 mm sodium 1-octanesulfonate for the analysis of 5-HT; working electrode, glassy carbon (WE-3G, Eicom); flow rate, 0.23 ml/min. Detector voltage and detector temperature were set at 0.45 V and at 25.0°C, respectively. Retention time for 5-HT was 14.85 min. The detection limit for 5-HT analysis is 0.05 fg/20 μl.

γ-Aminobutyric acid in the dialysate was analyzed using high-performance liquid chromatography with fluorescence detection (Degasser, DG-100; liquid chromatograph, EP-100; fluorescence detector, FLD-370; Eicom) after derivatization with o-phthaldialdehyde (Sigma). The o-phthaldialdehyde derivatizing reagent was prepared by dissolving o-phthaldialdehyde (54 mg) in absolute methanol (1 ml) and adding 2-mercaptoethanol (40 μl) and sodium carbonate (0.1 m, pH 9.5, 99 ml). Ten microliters of this derivatizing reagent was mixed with 20 μl of the dialysate and allowed to react for 2.5 min. Twenty microliters of this solution was analyzed. The chromatographic conditions were as follows: column, Eicompak (SC-5ODS 2.1 × 150 mm; Eicom); mobile phase, 0.1 m sodium dihydrogenphosphate and 0.1 m disodium hydrogenphosphate (pH 3.5) containing 50.0% methanol and 0.1 mm disodium EDTA; flow rate, 0.23 ml/min. Detector temperature was set at 30°C. Retention time for GABA was 14.83 min. The detection limit for GABA analysis is 0.2 pg/20 μl.

A behavioral study was performed separately from the microdialysis study. The tail-flick test was used to assess thermal nociceptive threshold. Tail-flick testing was performed by monitoring latency to withdrawal from the heat source (a 50-W projection lamp bulb) focused on a distal segment of the tail, using a thermal analgesimeter (KN-205E; Natsume, Tokyo, Japan). A cutoff time of 15.0 s was used to minimize damage to the skin of the tail. After determination of baseline tail-flick latencies, 40 nmol/3 μl of intracerebroventricular morphine was administered. The tail-flick latencies were measured at 30, 60, 120, 180, 240, 300, and 360 min after the intracerebroventricular injection. To examine the involvement of spinal 5-HT³receptors and GABA^Areceptors in morphine analgesia, either TICM or a GABA^Areceptor antagonist bicuculline (Sigma) was intrathecally administered 170 min after morphine administration, and the effect of the antagonist on tail-flick latencies was evaluated 10 min later. TICM and bicuculline were freshly dissolved in ACSF in concentrations that allowed intrathecal injections in 10-μl volumes. Intrathecal drug administration was accomplished by using a microinjection syringe (Hamilton, Reno, NV) connected to the intrathecal catheter in awake, briefly restrained rats. All intrathecal drugs were administered manually over 10 s.

In the microdialysis study, the changes of GABA and 5-HT concentrations are presented as mean ± SD of percentage of basal concentrations. Data were compared with basal concentration using a one-way analysis of variance for repeated measures followed by Dunnett test within a single group, and were analyzed using a two-way analysis of variance for repeated measures followed by the Tukey Kramer test for between-group comparisons. In the behavioral study on morphine antinociception, the values of withdrawal latency were converted to percent maximum possible effect:

Changes in percent maximum possible effect were compared with baseline using a one-way analysis of variance for repeated measures followed by Dunnett test within a single group, and analyzed using a paired t  test for the comparison between pretreatment and posttreatment in the behavioral study. P < 0.05 was considered statistically significant.

The basal concentration of 5-HT was 2.57 ± 1.18 pg/20 μl. The coefficient of variation among the three consecutive samples for determination of basal concentrations was less than 7%. Intracerebroventricular morphine (40 nmol) evoked a significant increase in spinal 5-HT concentration (P < 0.01;fig. 1). Significant increases in 5-HT concentration were observed 40–240 min after morphine administration.

The basal concentration of GABA was 0.27 ± 0.14 pg/20 μl. The coefficient of variation among the three consecutive samples for determination of basal concentrations was less than 5%. Figure 2shows the effect of 40 nmol intracerebroventricular morphine on spinal GABA concentration. GABA concentration was gradually increased after intracerebroventricular morphine administration and reached a plateau from 180 to 260 min after morphine administration. Significant increases in GABA concentration were observed from 80 up to 360 min after administration of morphine compared with the basal concentration (P < 0.01). Intracerebroventricular naloxone (40 nmol) administered at 180 min after morphine significantly reversed the morphine-induced increase in GABA concentration (P < 0.01). Spinal perfusion of 1 mm TICM for 60 min (from 180 to 240 min after morphine) also reversed morphine-induced increase in GABA concentration (P < 0.01). GABA concentrations 180 min after morphine administration were comparable among the three different treatments. TICM (1 mm) alone did not significantly affect the basal spinal GABA concentration for at least 240 min (fig. 3).

Intrathecal administration of bicuculline at doses of greater than 0.02 μg shortened the tail-flick latency in the tail-flick test. For example, 0.2 μg bicuculline shortened the tail-flick latency, which was observed within 5 min after the administration and continued for 30 min (table 1). Doses of intrathecal TICM greater than 1.0 μg also shortened the tail-flick latency. For example, 10.0 μg TICM shortened the tail-flick latency that was observed within 5 min and continued for 30 min (table 2). Neither 0.02 μg intrathecal bicuculline nor 1.0 μg TICM affected the basal tail-flick latency (tables 1 and 2). Therefore, these doses were used in the behavioral study, and the antagonistic effects of these drugs were assessed at 10 min after intrathecal administration.

The mean baseline tail-flick latency in this experiment was 5.7 s (range, 5.1–6.2 s). Figure 4shows the time course of the change in percent maximum possible effect after intracerebroventricular morphine administration. Intracerebroventricular morphine (40 nmol) produced significant antinociceptive effects from 30 min up to 360 min after its administration. The maximal antinociceptive effect was observed 30–180 min after morphine administration. Figure 5shows the effects of intrathecal bicuculline and TICM on the intracerebroventricular morphine–induced antinociceptive effect. We evaluated the antagonistic effects of bicuculline and TICM 180 min after morphine administration, because at this time point, the maximal effect of morphine in GABA analysis was observed (fig. 2). Intrathecal bicuculline significantly reversed the morphine-induced antinociception (P < 0.05). Intrathecal TICM also significantly reversed the morphine-induced antinociception (P < 0.05). On the other hand, intrathecal ACSF alone did not produce any significant changes in the tail-flick latency. There were no significant differences in percent maximum possible effect before intrathecal administration among the three different treatments.

The current study showed that intracerebroventricular morphine evoked spinal 5-HT release. Previous reports also found morphine–induced spinal 5-HT release. 9–11The major source of serotonergic projections to the dorsal horn are neurons in the rostral ventromedial medulla, which contains both opioid peptides and opioid receptors. 12,13Several investigators found that electrical or chemical stimulation of the dorsal lateral funiculus, bulbospinal neurons, or the nucleus raphe magnus evoked 5-HT release in the dorsal horn of the spinal cord. 14–16Therefore, it would be considered that morphine evokes spinal 5-HT release through activation of rostral ventromedial medulla serotonergic neurons by the inhibition of inhibitory interneurons. However, this may not be always consistent. First, Matos et al.  17concluded that intracerebroventricular morphine at an analgesic dose did not necessarily increase the extracellular concentration of 5-HT in the spinal dorsal horn using high-performance liquid chromatography with electrochemical detection, because morphine increased 5-HT release in only two of four rats. The discrepancy between their results and ours may be caused by the sensitivity of assay system for 5-HT. Baseline 5-HT concentrations were close to the sensitivity limit in the high-performance liquid chromatography with electrochemical detection, and hence it is difficult to measure small changes reliably. 18We used a smaller diameter (2.1 mm) detector column in the current study, compared with that used in the study by Matos et al.  17(4.6 mm). This column enabled us to analyze 5-HT with high sensitivity. Using this assay system, morphine at the dose we used (40 nmol = 11.4 μg), which was almost comparable to the dose used by Matos et al.  (10 μg), consistently increased spinal 5-HT concentration in all eight rats in the current study, and the degree of increase in 5-HT concentration was approximately 150% of the basal concentration. We performed the experiments 18–24 h after the implantation of the dialysis probe because of prevention of the decrease in recovery rate caused by gliosis. Therefore, there may be the possibility that the acute trauma of probe implantation modified the effect of morphine on 5-HT release. Second, electrophysiologic studies have demonstrated that most serotonergic rostral ventromedial medulla neurons are not affected by morphine. 19–21However, those studies used the anesthetized animal. There is a possibility that anesthesia reduced serotonergic neuronal activity, 18,22and serotonergic neuronal activity is lower during sleep than wakefulness. 23We used the awake free-moving rats implanted with microdialysis probe in this study.

Several lines of study have suggested that supraspinal morphine or PAG stimulation activates a spinal GABAergic system. 1,3,5,24–26However, there has not been direct evidence that supraspinal morphine evokes spinal GABA release. In the current study, we clearly demonstrated that intracerebroventricular morphine evokes spinal GABA release concomitant with 5-HT release. There was the discrepancy in time course between intracerebroventricular morphine–induced increases in 5-HT and GABA concentrations in our results. Spinal microdialysis does not directly measure neurotransmitter concentrations at synaptic sites, but rather neurotransmitter concentrations that have diffused into the extracellular space. Diffusion into the extracellular space is likely to involve processes of uptake and metabolism, which can modify the amount of neurotransmitter in the dialysate. Therefore, it is possible that diffusion into the extracellular space would contribute to the discrepancy in time course between 5-HT and GABA concentration.

There are at least four subtypes of 5-HT receptors (5-HT¹, 5-HT², 5-HT³, and 5-HT⁴) in the spinal cord. 27,28Among the subtypes of 5-HT receptors, 5-HT³receptors are unique because G proteins are not affected by agonist binding to 5-HT³receptors. 5-HT³receptors are directly linked to the opening of nonselective monovalent cation channels 4and should mediate neuronal excitation. Alhaider et al.  3demonstrated that the activation of the GABAergic system contributed to spinal 5-HT³receptor–mediated antinociception. Therefore, we focused on 5-HT³receptors. TICM used in the current study is a potent and selective 5-HT³receptor antagonist and has been commonly known as IC 205–930. Radioligands of IC 205–930 have been used to detect 5-HT³receptor binding site in the central nervous system. 29In the current study, blockade of spinal 5-HT³receptors inhibited intracerebroventricular morphine–induced spinal GABA release. This indicates that morphine activates supraspinal serotonergic neurons projecting to the spinal cord, and then released 5-HT excites spinal GABAergic interneuron through 5-HT³receptors. Immunohistochemical and autoradiographic studies revealed that 5-HT³receptors are restricted to the superficial layers of the dorsal horn in the spinal cord. 29–31Intense GABA immunoreactivities are also found within laminae I–III of the spinal cord. 32,33It would be supposed that the GABAergic neuron with 5-HT³receptors are directly activated or activation of presynaptic 5-HT³receptors indirectly induce the activation of GABAergic neuron. There is a report that the 5-HT³receptors mainly were located in the primary afferent terminal in the dorsal horn. 34In addition, Peng et al.  35recently suggested that 5-HT³receptors located in central terminals of primary afferents might contribute to the generation of dorsal root reflex, leading to presynaptic inhibition. Therefore, further study is necessary to clarify the mechanism of 5-HT³receptor activation-induced GABA release.

In a behavioral study, we examined whether spinal GABA release through the activation of 5-HT³receptors observed in our microdialysis study would be a component of supraspinal morphine–induced antinociception. GABA and 5-HT concentrations were measured within dorsal horn at the level of the L1 vertebra, and the tail-flick test was used in the behavioral study. The primary afferent neurons responsible for tail-flick response largely innervate in the dorsal horn of the sacrococcygeal spinal cord. In this study, GABA and 5-HT concentrations were measured within dorsal horn at the level of the L1 vertebra at which L4 nerve root mainly terminated. Therefore, results of the current microdialysis study may not directly reflect the changes of GABA and 5-HT in the sacral spinal cord. However, neurons originating in the nucleus raphe magnus, on which intracerebroventricular morphine acts, project to all levels of the spinal cord. 36–38Therefore, intracerebroventricular morphine–induced GABA and 5-HT release we observed should contribute to antinociception in the tail-flick test.

It has been generally considered that supraspinal morphine produces antinociceptive effects through the serotonergic and noradrenergic descending inhibitory systems. In respect to the serotonergic mechanism, the analgesic action of morphine is reduced by destruction of the serotonergic innervation of the spinal cord with selective neurotoxins 39or lesions of the medullary raphe, 40,41which contains the cell of the origin of the bulbospinal 5-HT pathway innervating the spinal dorsal horn. Intrathecal injection of a 5-HT receptor antagonist also attenuates the analgesia produced by microinjection of morphine into the PAG. 42In addition, intrathecal 5-HT produces antinociceptive effects. However, the nature of the receptors involved in the 5-HT–induced modulation of pain in the spinal cord remains to be elucidated. To the best of our knowledge, there has been no report that examined the contribution of spinal 5-HT³receptors to intracerebroventricular morphine–induced antinociception in behavioral tests. The results of the current study have shown that spinal 5-HT³receptors are involved in supraspinal morphine–induced antinociceptive effects. In contrast, several studies found only a minor role of the serotonergic system in supraspinal morphine–induced antinociceptive effect. 43–45However, because these studies used methysergide as the 5-HT receptor antagonist, which blocks 5-HT^1/2, not 5-HT³receptors, the role of a serotonergic mechanism may be underestimated. The current study showed that TICM did not completely reverse the morphine-induced antinociceptive effect. This suggests the involvement of the other subtypes of 5-HT receptors or noradrenergic system in intracerebroventricular morphine–induced analgesia. Indeed, it has been reported that GABA^Areceptor response is potentiated by the activation of 5-HT²receptor, 46and that 5-HT¹receptor activation inhibits the nociceptive transmission. 47The contribution of noradrenergic system in intracerebroventricular morphine–induced analgesia was also well demonstrated.

Several investigators reported that the spinal GABA system contributes to nucleus raphe magnus or PAG stimulation– and supraspinal morphine–induced antinociceptive effects. 1,3,5,25,26,48,49Although our microdialysis study demonstrated that intracerebroventricular morphine gradually increased spinal GABA release with a maximal effect at 180–240 min after administration, an antinociceptive effect was observed 30 min after the administration and thereafter. Because we did not assess the antagonistic effect of bicuculline at the early phase of morphine-induced antinociception, the involvement of GABA in morphine analgesia at the early phase is unclear. We speculate that another mechanism such as noradrenergic or direct serotonergic inhibitions, in addition to GABAergic inhibition, might be involved in the antinociceptive effect observed in the early phase.

In conclusion, the current study showed that intracerebroventricular morphine produced antinociceptive effect, in part through GABA release by the activation of 5-HT³receptors in the spinal cord.