Differential Effects of Intrathecally Administered Morphine and Its Interaction with Cholecystokinin-B Antagonist on Thermal Hyperalgesia following Two Models of Experimental Mononeuropathy in the Rat 

The following investigations were performed according to a protocol approved by the Institutional Animal Care Committee of Chiba University, Chiba, Japan. Male Sprague-Dawley rats weighing 250–300 g were fitted with long-term intrathecal catheters and observed to monitor the effect of the drugs.

Long-term intrathecal catheters were inserted 3 days before the nerve injury, during halothane anesthesia, by passing a PE-10 catheter through an incision in the atlantooccipital membrane to a position 8 cm caudal to the cisterna at the level of the lumbar enlargement. [14] The external opening of the catheter was fastened to the top of the skull and sealed with a piece of steel wire. Rats that showed neurologic deficits were not studied.

Anesthesia was induced by having the rats inhale 5% halothane, which was maintained at a concentration of 2% to 3% as needed. After a local incision was made, the biceps femoralis of each leg was bluntly dissected at the mid thigh to expose the sciatic nerve. Each nerve was carefully mobilized, with care taken to avoid undue stretching. After the sciatic nerve was mobilized, a chronic constriction injury or a partial sciatic nerve injury was created on the right sciatic nerve. The left sciatic nerve was only mobilized. After surgery, the animals were maintained individually in plastic cages with solid floors covered with 3 to 6 cm of sawdust. After these procedures, all animals ate and drank normally.

Chronic Constriction Injury. The chronic constriction injury was created by placing four loose ligations around the right sciatic nerve. [15] Each of four 4–0 chromic gut sutures was tied loosely with a square knot around the sciatic nerve. We used a brief twitch in the muscle surrounding the exposure as an indicator of the desired degree of constriction. Incisions were closed, layer to layer, with 3–0 silk sutures, and the rats were allowed to recover from the anesthetic.

Partial Sciatic Nerve Injury. A partial sciatic nerve injury was created by tight ligation of one third to one half of the right sciatic nerve. [16] An 8–0 silicon-treated silk suture was inserted into the sciatic nerve just proximal to the sciatic trifurcation using a 3/8 curved, reversed-cutting miniature needle and was tightly ligated so the dorsal one third to one half of the nerve thickness was trapped in the ligature. Incisions were closed, layer to layer, with 3–0 silk sutures, and the rats were allowed to recover from the anesthetic.

Sham Sciatic Nerve Injury. To obtain control data, a sham sciatic nerve injury group (sham injury group) was added. In the sham injury group, right and left sciatic nerves were only mobilized, and incisions were closed layer to layer with 3–0 silk sutures.

Paw withdrawal latency (PWL) in response to thermal stimulation was measured using a device similar to that described previously. [17] The rats were placed in a clear plastic cage (10 x 20 x 24 cm) on an elevated floor of clear glass (2 mm thick). A radiant heat source (Eye Projector Halogen Lamp JRC-12V-100W; Iwasaki Electric, Tokyo, Japan) with a 5-mm aperture was contained in a movable holder placed beneath the glass floor. The voltage to the thermal source was controlled by a constant voltage supply. To reduce the variability in plate surface temperature resulting from minor changes in room temperature, the interior of the box under the animal was prepared with a heat source such that the glass temperature was regulated at 30 [degree sign]C. The calibration of the thermal test system was such that the average response latency in 10 healthy, untreated rats was maintained at 10 s before an experimental series was begun.

To initiate a test, a rat was placed in the box and allowed 5–10 min to habituate. The halogen lamp beneath the floor was positioned so it focused on the plantar surface of one hind paw that was resting on the glass. Care was taken not to focus the lamp on skin that was not in contact with the glass floor. The light was activated, initiating a timing circuit. The interval between the application of the light beam and the brisk hind paw withdrawal response was measured to the nearest 0.1 s. The time value was assigned as the response latency. The trial was discontinued and the lamp was removed in the absence of a response within 20 s.

A preliminary study performed in the authors' laboratory revealed that maximum thermal hyperalgesia occurred between 7 and 14 days after the chronic constriction injury or the partial sciatic nerve injury. Thus, each hyperalgesic animal and the sham injury rats received one medication administered intrathecally at each of two time points: 7 and 11 days after the creation of the nerve injury. Before the drug injection, the hind paws were tested three times, with 5-min intervals between the subsequent testing of one paw as the baseline data. Both paws were tested alternately once at 5, 15, 30, and 60 min after drug injection. We compared the effects of drugs in rats with chronic constriction injuries or partial sciatic nerve injuries with the effects in the rats with sham injuries.

The intrathecally administered drugs were delivered in a total volume of 10 [micro sign]l, followed by 10 [micro sign]l saline to flush the catheter. All agents were dissolved in saline such that the final dose was delivered in 10 [micro sign]l. The agents used in this study were morphine hydrochloride (molecular weight = 322 g; Takeda, Osaka, Japan), naloxone hydrochloride (molecular weight = 364 g; Sigma Chemical Co., St. Louis, MO), PD135158 N-methyl-D-glucamine ([1S-[1 [Greek small letter alpha],2 [Greek small letter beta][S*(S*),4 [Greek small letter alpha]]-4-[[2-[[3-(1H-Indol-3-yl)-2-methyl-1-oxo -2[[[(1,7,7,-trimethylbicyclo[2.2.1]hept-2-yl)oxy-]carbonyl]amino] propyl]amino]-1-phenylethyl]amino]-4-oxo-butanoic acid N-methyl-D-glucamine salt, molecular weight = 812 g; Research Biochemicals International, Nattick, MA).

To compare the predrug right and left PWLs of each group, we used one-way analysis of variance (ANOVA) with the Tukey multiple comparison test. It has been reported that, after nerve injury, the spinal dorsal horn reorganizes and that this reorganization depends on the type of nerve injury. For example, chronic constriction injury increases the number of [micro sign]-opioid receptors and tight ligation of the sciatic nerve decreases the number of [micro sign]-opioid receptors. [18,19] It is reasonable to assume that the time course of the effect of intrathecally administered drugs on the PWL depends on the nerve injury. Therefore, to analyze the effects of the drugs on the PWL, we chose maximum PWL, which was defined as the single longest PWL value during the first 30 min after drug injection. We did not use area under the curve of the time-effect curves. To obtain a dose-response curve, the dose was plotted against the Delta PWL. The Delta PWL was calculated by subtracting the predrug PWL value from the maximum PWL value. Dose dependency was evaluated using ANOVA with the Dunnett multiple comparison test. In this case, Delta PWLs at each dose were compared with each other. When dose dependency existed, the dose-response curves were fitted using a least-squares linear regression analysis, and the slopes of each regression line and the 95% confidence intervals were calculated to compare each regression line. The doses that resulted in Delta PWL = 5.5 s (ED⁵.5s) were calculated with their 95% confidence intervals. We chose 5.5 s because the Delta PWL of the uninjured paw in the saline-treated group was approximately 1.5 s and the Delta PWL in the highest morphine dose group was approximately 9 s. Wherever appropriate, results are expressed as the mean +/- SD. Critical values that reached a P < 0.05 level were considered significant.

No difference was apparent between predrug right and left PLWs in each group (Table 1, right PWL: P > 0.8; left PWL: P > 0.6 by ANOVA). Intrathecally administered morphine increased the Delta PWL of either the right or the left hind paw in a dose-dependent manner at a dose between 1 and 10 [micro sign]g (P < 0.005 by ANOVA;Figure 1 and Figure 2), and the morphine dose-response curve of the right hind paw overlapped that of the left hind paw (Table 2, Figure 2). When administered with morphine, 80 [micro sign]g PD135158 significantly shifted the morphine dose-response curve to the left (Table 2, Figure 2). Conversely, 20 [micro sign]g PD135158 administered with morphine had no effect on the morphine dose-response curve (Table 2, Figure 2).

Before the drugs were injected, the values of right and left PWLs were 7 +/- 1.1 and 10.5 +/- 1 s (n = 57) in the rats with chronic constriction injuries and 7.1 +/- 1.1 and 10.5 +/- 1.4 s (n = 75) in the rats with partial sciatic nerve injuries, respectively, and the right PWL was significantly shorter than the left PWL (P < 0.001 by Student t test). No difference was apparent between predrug right and left PWLs of each group (chronic constriction injury model: right PWL = P > 0.8, left PWL = P > 0.8; partial sciatic nerve injury model: right PWL = P > 0.8, left PWL = P > 0.8 by ANOVA;Table 1).

Chronic Constriction Injury Model. Intrathecally administered morphine produced equal increases in the Delta PWL of the injured and uninjured paws in a dose-dependent manner at a dose between 0.01 and 1 [micro sign]g (injured paw, P < 0.01; uninjured paw, P < 0.05 by ANOVA;Figure 3 and Figure 4, Table 2), and the morphine dose-response curve of the injured paw overlapped that of the uninjured paw (Figure 4, Table 2). The value of ED⁵.5s and the slope of the injured (right) and uninjured (left) paws in the chronic constriction injury model were significantly less than those of the right and left paws in the sham injury model, respectively (Table 2).

Intrathecal injection of 80 [micro sign]g PD135158 had no effect on the maximum PWL of the injured and uninjured paws compared with that in the saline-treated rats (injured paw, P > 0.5; uninjured paw, P > 0.8 by t test;Figure 5). When 80 [micro sign]g PD135158 was administered with morphine, the morphine also produced proportionally equal increases in the Delta PWLs of the injured paw and the uninjured paw in a dose-dependent manner at a dose between 0.001 and 0.1 [micro sign]g (injured paw, P < 0.005; uninjured paw, P < 0.001 by ANOVA;Figure 3 and Figure 4), and the morphine dose-response curve for the Delta PWL of the injured paw overlapped that of the uninjured paw (Table 2, Figure 4). When 80 [micro sign]g PD135158 was administered with morphine, the morphine dose-response curve of the injured and uninjured paws shifted to the left significantly (Table 2, Figure 4). After intrathecal administration of 80 [micro sign]g PD135158 with morphine, the values of ED⁵.5s of the injured (right) and uninjured (left) paws in the chronic constriction injury model were significantly less than those of the right and left paws in the sham injury model (Table 2).

When 10 [micro sign]g naloxone was administered with 1 [micro sign]g morphine or 0.1 [micro sign]g morphine plus 80 [micro sign]g PD135158, the morphine antinociceptive effect was antagonized significantly (maximum PWLs: 1 [micro sign]g morphine to the injured paw = 16.3 +/- 3.1 s and to the uninjured paw = 18.9 +/- 1.5; 0.1 [micro sign]g morphine plus PD135158 80 [micro sign]g to the injured paw = 15.4 +/- 3 s and to the uninjured paw = 19.5 +/- 0.7; naloxone plus 1 [micro sign]g morphine to the injured paw = 9.2 +/- 1.3 s and to the uninjured paw = 13.6 +/- 0.7; naloxone plus 0.1 [micro sign]g morphine plus PD135158 80 [micro sign]g to the injured paw = 8.5 +/- 2.3 s and to the uninjured paw = 12.5 +/- 1.5; P < 0.005 by t test). Intrathecal injection of 10 [micro sign]g naloxone had no effect on the maximum PWLs of the injured and the uninjured paws compared with that in the saline-treated rats (naloxone-treated rats: injured paw = 9.4 +/- 0.9 s, P > 0.1; uninjured paw = 13.2 +/- 2 s; P > 0.3 by t test).

Partial Sciatic Nerve Injury Model. Intrathecal injection of morphine increased the Delta PWL of the uninjured paw, but not the injured paw, in a dose-dependent manner at a dose between 0.1 and 10 [micro sign]g (injured paw: P > 0.05; uninjured paw: P < 0.001 by ANOVA;Figure 3 and Figure 4, Table 2). The value of ED⁵.5s of the uninjured paw in the partial sciatic nerve injury model is significantly greater than that of the uninjured paw in the chronic constriction injury model and is the same as that of the left paw in the sham injury model (Table 2).

The intrathecal injection of 80 [micro sign]g PD135158 had no effect on the maximum PWL of the injured and uninjured paws compared with that in the saline-treated rats (injured paw, P > 0.4; uninjured paw, P > 0.1 by t test, Figure 3 and Figure 5). When 80 [micro sign]g PD135158 was administered with morphine, the morphine produced increases in the Delta PWL of the uninjured paw in a dose-dependent manner at a dose between 0.001 and 0.1 [micro sign]g (P < 0.001), and, after the administration of 20 [micro sign]g PD135158 with morphine, morphine increased the Delta PWL of the uninjured paw in a dose-dependent manner at a dose between 0.01 and 1 [micro sign]g (P < 0.001, by ANOVA). Administration of morphine with 80 [micro sign]g or 20 [micro sign]g PD135158 did not increase the Delta PWL of the injured paw in a dose-dependent manner (80 [micro sign]g PD135158, P > 0.1; 20 [micro sign]g PD135158, P > 0.1 by ANOVA;Figure 3 and Figure 4), but the Delta PWL of the injured paw in the rats treated with 0.1 [micro sign]g morphine plus 80 [micro sign]g PD135158 was significantly greater than that in the rats treated with 0.1 [micro sign]g morphine (P < 0.05 by t test, Figure 4). The value of ED⁵.5s of the uninjured paw in the rats given 20 [micro sign]g PD135158 plus morphine was significantly greater than that in the rats given 80 [micro sign]g PD135158 plus morphine and was significantly less than that in the rats given morphine (Table 2).

When 10 [micro sign]g naloxone was administered with either 10 [micro sign]g morphine or 0.1 [micro sign]g morphine plus 80 [micro sign]g PD135158, the morphine antinociceptive effect was antagonized significantly (maximum PWLs in rats given 10 [micro sign]g morphine: injured paw = 12.1 +/- 2.9 s, uninjured paw = 19.1 +/- 1.6 s; in rats given 0.1 [micro sign]g morphine plus 80 [micro sign]g PD135158: injured paw = 11.4 +/- 1.3 s, uninjured paw = 18.9 +/- 1.6 s; in rats given naloxone plus 10 [micro sign]g morphine: injured paw = 9.1 +/- 1.1 s, uninjured paw = 1.36 +/- 1.9 s; in rats given naloxone plus 0.1 [micro sign]g morphine plus 80 [micro sign]g PD135158: injured paw = 8.1 +/- 1.9 s, uninjured paw = 12.0 +/- 0.9; P < 0.005 by t test). Ten micrograms naloxone had no effect on the maximum PWLs of the injured and uninjured paws compared with the PWLs of the saline-treated rats (naloxone-treated rats: injured paw = 8.8 +/- 1 s, P > 0.9; in the uninjured paw = 12.6 +/- 2.8 s, P > 0.9 by t test).

The current study clearly shows that intrathecally administered morphine increased the maximum PWL of the uninjured paw, but not the maximum PWL of the injured paw in a dose-dependent manner in the partial sciatic nerve injury model. Conversely, intrathecally administered morphine increased the maximum PWL of both the injured and the uninjured paw in a dose-dependent manner in the chronic constriction injury model. These data suggest that the effect of morphine on thermal hyperalgesia induced by a chronic constriction injury is more prominent than that induced by a partial sciatic nerve injury. The morphine effect on thermal hyperalgesia induced by a chronic constriction injury is consistent with the findings of another report. [20] These results indicate that the effect of morphine on thermal hyperalgesia depends on the type of nerve injury.

Few patients with neuropathic pain report thermal hyperalgesia, and tactile allodynia is a more common symptom than thermal hyperalgesia in patients with neuropathic pain. [21] However, Rowbotham and Fields [22] reported that, in the syndrome of postherapeutic neuralgia, the patients with heat hyperalgesia had significantly higher pain ratings and more severe allodynia than did patients with heat hypoalgesia. This indicated that thermal hyperalgesia is an important clinical feature in patients with neuropathic pain.

Weight bearing on the hind paw is an important factor to assess the level of hyperalgesia after nerve injury or inflammation. [23] For example, after a tibial nerve was severed, there was a marked threshold decrease in the nerve-injured paw in a standing rat, but when the hind limbs were not bearing weight, the unilateral threshold decrease was abolished completely. [23] Thus, we compared the effect of drugs on the Delta PWLs of the injured (right) and uninjured (left) paws in the rats with nerve injuries with those of the right and left paws in the rats with sham injuries, respectively.

In the current study, we analyzed the effect of morphine with Delta PWL, which was calculated by substracting the predrug PWL from the maximum PWL. Thus, Delta PWL indicated the amount of the increase in PWL after the drug administration, regardless of the level of the pre-drug PWL. In the chronic constriction injury model, the dose-response curve of morphine in the injured paw overlapped with that in the uninjured paw, but, as previously reported, [20] intrathecal morphine did not influence the difference between the PWL of the uninjured paw and that of the injured paw.

Thermal hyperalgesia induced by a chronic constriction injury is not mediated by capsaicin-sensitive C fibers, and thermal hyperalgesia induced by partial sciatic nerve injury is mediated by capsaicin-sensitive C fibers. [24,25] This suggests that thermal hyperalgesia after chronic constriction injury is mediated by mechanisms different than those after partial sciatic nerve injury. In the current study, C-fiber-mediated thermal hyperalgesia is morphine resistant and C-fiber-independent thermal hyperalgesia is morphine sensitive. Morphine inhibits C-fiber-mediated neurotransmission in the spinal dorsal horn, [26] but our data suggested that the mechanisms of the analgesic effect of morphine on thermal hyperalgesia induced by nerve injury is not directly related to the inhibition of the C-fiber-mediated neurotransmission in the spinal cord. Gastric acid challenge induced hydrochloride concentration-dependent excitation of medullary neurons, and this excitation was thought to be associated in part with behavioral manifestations of pain. [27] It has been reported that gastric acid challenge caused many neurons in the nucleus tractus solitarii and area postrema to express C-fos mRNA, that morphine inhibited this C-fos mRNA expression in a naloxone-reversible manner, and that pretreatment of rats with capsaicin had no effect. [27] These data also suggested that some types of pain, which were not mediated by capsaicin-sensitive C fibers, were effectively treated with morphine.

The precise mechanisms that cause thermal hyperalgesia induced by chronic constriction injury to be more sensitive to the morphine analgesic effect than that by the partial sciatic nerve injury model are unknown. In the chronic constriction injury model, it has been reported that an approximately 50% increase in the number of [micro sign]-opioid binding sites in the lumbar spinal cord was found ipsilateral and contralateral to the nerve injury compared with intact rats. [19] Therefore, it is reasonable that the values of ED⁵.5s in injured (right) and uninjured (left) paws in the chronic constriction injury model are significantly smaller than those in the sham injury model, respectively. Conversely, the values of ED⁵.5s of the uninjured (left) paw in the partial sciatic nerve injury model are the same as those of the left paw in the sham injury model, and morphine did not increase the Delta PWL of the injured (right) paw in a dose-dependent manner in the partial sciatic nerve injury model. Unfortunately, the time course of the density of the [micro sign]-opioid receptor in the spinal cord has not been reported in the partial sciatic nerve injury model. Goff et al. [18] reported that the tight ligation of the sciatic nerve decreased the [micro sign]-opioid receptor immunoreactivity in the dorsal horn of the side ipsilateral to the nerve injury. This suggested that the effect of the nerve injury on the [micro sign]-opioid receptor immunoreactivity depends on the type of nerve injury. It has been reported that, in the chronic constriction injury model, large myelinated fibers are more severely damaged than small myelinated and unmyelinated fibers, and changes in the unmyelinated fibers are probably involved in the production of thermal hyperalgesia. [28] Conversely, axotomy damaged all myelinated and unmyelinated fibers. We think that different degrees of nerve degeneration may contribute to the different levels of reorganization of opioid receptors and to the different morphine sensitivity.

As noted before, it has been assumed that one possible mechanism for the reduced effect of morphine on neuropathic pain is the antimorphine effect of cholecystokinin. [9] Nerve section of sciatic nerve reportedly causes a marked increase in the number of cholecystokinin-B receptor-positive neurons in the lumbar dorsal root ganglia ipsilateral, but not contralateral, to the nerve injury. [12] Furthermore, upward regulation of the mRNA for cholecystokinin in the ipsilateral dorsal root ganglion to the nerve injury was also observed after sciatic nerve transection. [13] No data exist concerning the number of cholecystokinin-B receptors or the content of cholecystokinin after either chronic constriction injury or partial sciatic nerve injury. If both the chronic constriction injury and the partial sciatic nerve injury increase the number of cholecystokinin-B receptor-positive neurons and the content of cholecystokinin in the dorsal root ganglia ipsilateral, but not contralateral, to the nerve injury, the effect of the cholecystokinin-B receptor antagonist on morphine analgesia should be more prominent in the injured paw than in the uninjured paw. In the chronic constriction injury model, the ED⁵.5s value of the injured paw is the same as that of the uninjured paw after administration of PD135158 with morphine. In the partial sciatic nerve injury model, an intrathecal injection of PD135158 potentiated the effect of intrathecal morphine on the PWL of the uninjured paw significantly, but administration of morphine with PD135158 had only a small effect on the PWL of the injured paw. Therefore, it is unlikely that the upward regulation of cholecystokinin-B receptors and the increase in the content of cholecystokinin in the lumbar dorsal root ganglia ipsilateral, but not contralateral, to the nerve injury occurs after chronic constriction injury or partial sciatic nerve injury. These data also suggest that, in the partial sciatic nerve injury model, cholecystokinin and cholecystokinin-B receptors do not play important roles in the reduction of analgesic effect of morphine in the injured paw. After administration of 80 [micro sign]g PD135158 with morphine, the ED⁵.5s values of the uninjured paw in the chronic constriction injury and the partial sciatic nerve injury models were significantly smaller than that of the left paw in the sham injury model. We do not know why the effect of cholecystokinin-B receptor antagonist on morphine analgesia in the uninjured paw in both the chronic constriction in the injury and the partial sciatic nerve injury models is more prominent than that in the left paw in the sham injury model. No data exist, but the reorganization of the mechnanisms of nociceptive information transmission in the spinal cord could have occurred after the nerve lesion was induced, and this reorganization may contribute to the prominent effects of the cholecystokinin-B antagonist on morphine analgesia in an uninjured paw in the chronic constriction injury and the sciatic nerve injury models.