Ketamine Attenuates and Reverses Morphine Tolerance in Rodents

All studies were approved by the Institutional Animal Care and Use Committee of Cornell University Medical College.

Adult male CD1 mice (Charles River Laboratories, Kingston, NY) weighing 25 to 30 g were used for studies 1 and 2. Adult male Sprague-Dawley rats (Taconic Farm, Germantown, NY) weighing 350 to 375 g at the time of surgery were used for study 3. The mice were caged in groups of five and the rats were caged individually with free access to food and water; all animals were maintained on a regular light-dark cycle. Treatment groups averaged ten animals.

To assess the analgesic effect of morphine, the tail-flick test was used. The tail-flick apparatus (EMDIE, Richmond, VA) emits radiant heat to the tail at 2 cm from the tip in mice and at 5 to 8 cm in rats. The time from the onset of heat to the withdrawal of the tail (tail-flick latency) was measured. The intensity of the radiant heat was adjusted so that the baseline latencies were between 2.5 and 3.5 s. To avoid causing tissue damage, the heat stimulus was turned off after 10 s (cut-off latency). A mean tail-flick latency was calculated from two repeated measurements. Baseline latency was obtained before opioid or saline administration. Subsequent response latencies were determined at the peak analgesia, which was 30 min after subcutaneous morphine and 10 min after intrathecal morphine were given. The latency data were converted to a quantal form by determining the percentage of analgesic responders in each group from the response latency values compared with baseline latencies.

Morphine sulfate was obtained from Mallinkrodt (St. Louis, MO), and racemic ketamine hydrochloride was obtained from Sigma Chemical Company (St. Louis, MO). The doses of morphine and ketamine are expressed as the free base. Morphine and ketamine were dissolved in normal saline. The pH of the ketamine dosing solutions for both subcutaneous and intraspinal administration were between 5.7 and 6.2. Each subcutaneous injection of morphine, racemic ketamine, and saline was delivered in a volume of 0.1 ml/10 g mouse weight. Each intraspinal injection was delivered in a volume of 5 micro liter, followed by 10 micro liter saline to flush the intrathecal catheter.

For the intraspinal administration of drugs to the rat, a catheter was placed in the intrathecal space. Under halothane anesthesia, a PE-10 tube was inserted through a small hole made in the atlantooccipital membrane and threaded 8.5 cm down the intrathecal space to the lumbosacral level of the spinal cord. [14]The catheterized rats were observed for 24 h after operation, and those with any signs of paralysis were excluded from the study. At the end of the study, 5 micro liter of a 1% methylene blue solution was introduced into the catheter followed by 10 micro liter saline to confirm the position of the catheter and the spread of the dye in the intrathecal space. Animals with a misplaced catheter or inadequate spread of the dye, as confirmed by dissection, were excluded from the study (8.3%).

After measuring the baseline latencies, increasing doses of morphine were administered until each animal responded to the analgesic (cumulative dose-response assessment [2,6]). An analgesic responder was defined as one whose mean response tail-flick latency was two or more times the value of the mean baseline latency. The percentage of analgesic responders in the group for each cumulative dose was calculated, and a cumulative dose-response curve was constructed. The changes in the median effective dose (ED⁵⁰) of morphine determined from the curves were used to express the changes in the relative analgesic potency of morphine.

Tolerance paradigms were based on previous studies of the kinetic-dynamic relationships of NMDA receptor antagonists and morphine tolerance. [2,4-6] 

Study 1 (Figure 1(A)) was designed to estimate the dose-response relationship for subcutaneous ketamine in attenuating the development of morphine tolerance. An escalating schedule of morphine doses administered subcutaneously three times a day for 3 days was used to produce tolerance in mice. On the morning of day 1, each animal received 10 mg/kg morphine given subcutaneously at 08:00 A.M. or underwent a cumulative dose-response assessment using subcutaneous morphine to determine their ED⁵⁰value for morphine. Each analgesic responder was not subjected to further tail-flick assay but received the subsequent dose of morphine so that each animal received the same opioid dose, approximately 10 mg/kg on the morning of day 1. On that same day, 10 mg/kg morphine was administered two more times. On day 2, the animals received 20 mg/kg morphine given subcutaneously three times per day; on day 3, the animals received 40 mg/kg morphine given subcutaneously three times per day. Ketamine at a dose of 0.3, 3, or 10 mg/kg or saline was administered subcutaneously 15 min before and 15 min after each morphine injection on days 1 through 3. On day 1, the first dose of ketamine or saline was administered before the cumulative dose-response assessment. On day 4, all animals underwent the morphine cumulative dose-response assessment without first receiving ketamine or saline. The development of morphine tolerance was determined by comparing the morphine ED⁵⁰values obtained on day 4 to those of day 1.

Study 2 (Figure 1(B)) was designed to determine whether ketamine could reverse established morphine tolerance. Tolerance was produced by implanting two 25-mg morphine pellets on day 1. The animals underwent a cumulative dose-response assessment on day 4. The day 4 morphine ED50 value was compared with the value obtained with morphine-naive animals to determine the degree of tolerance. The pellets were then removed and mice were separated into two groups that received either ketamine (10 mg/kg) or saline subcutaneously once on day 4 and three times per day on days 5 through 7. A repeated cumulative dose-response assessment was done on day 8 without administering ketamine. The morphine ED⁵⁰values of ketamine-treated mice on day 8 was compared with those of morphine-treated mice on day 4 and saline-treated mice on day 8.

Study 3 (Figure 1(C)) was designed to determine whether intraspinal ketamine could attenuate tolerance produced by intraspinal morphine. A paradigm similar to study 1 was used to produce intraspinal morphine tolerance in rats. On the morning of day 1, each animal underwent a cumulative dose-response assessment with intraspinal morphine to determine the morphine ED⁵⁰value. In this study, the cumulative dose-response assessment was performed with each group before administering ketamine to preclude any difference among the groups that might have been caused by the intrathecal catheterization. After this procedure, the animals received additional doses of intraspinal morphine so that each animal received the same dose of morphine, approximately 10 micro gram on the morning of day 1. Ten micrograms of morphine or saline was administered two more times on day 1. Twenty and 40 micro gram morphine or saline were administered intrathecally three times per day at 9:00, 14:00, and 19:00 on days 2 and 3, respectively. Ketamine at 12 micro gram (50 nmol) or saline was administered intraspinally 10 min before each morphine injection on days 1, 2, and 3. Ketamine was administered after the cumulative dose-response assessment on day 1, 10 min before the subsequent dose of morphine was injected. Thus the treatment arms included saline-morphine (SAL + MOR), ketamine-morphine (KET + MOR), ketamine-saline (KET + SAL), and saline-saline (SAL + SAL). Each animal was tested for its analgesic response to morphine, and the morphine ED⁵⁰value was determined again on day 5. This was done on day 5 because the return of tail-flick latencies to baseline values required more than 24 h after the last 40-micro gram intraspinal dose of morphine on day 3. The development of morphine tolerance was determined by comparing the morphine ED⁵⁰values obtained on day 5 to those obtained on day 1.

The acute effect of the intraspinal administration of 12 micro gram ketamine on tail-flick latencies was tested on another group of intrathecally cannulated rats. Twelve micrograms ketamine was given intraspinally followed by 5 micro liter saline. Mean tail-flick latency was obtained before ketamine injection (baseline latency), 10 min after ketamine injection (immediately before saline injection), and 10 min after saline injection.

The quantal dose-response data were analyzed using the BLISS-21 computer program. This program maximized the log-likelihood function to fit a parallel set of Gaussian normal sigmoid curves to the dose-response data and provides ED⁵⁰values, 95% confidence limits (CI), and relative potency estimates. [15] 

Study 1 (Figure 2) shows that the escalating subcutaneous dosing schedule for 3 days results in a shift to the right of the morphine dose-response curve on day 4 (SAL + MOR). However, when 10 mg/kg ketamine is coadministered subcutaneously with morphine, no significant shift of the curve is observed. The rightward shift of the dose-response curve indicates the development of tolerance to morphine, the magnitude of which can be expressed as an increase in the ED⁵⁰value and a decrease in the relative potency of morphine. Table 1shows these values for each group. On day 4, the SAL + MOR group had a morphine ED⁵⁰value of 23.1 mg/kg, more than five times greater than the ED⁵⁰value obtained with saline pretreatment on day 1 (SAL group). Coadministration of ketamine subcutaneously at a dose of 0.3 mg/kg (KET - 0.3 + MOR) did not reduce the day 4 morphine ED⁵⁰value significantly from control. The 3 mg/kg dose of ketamine (KET - 3 + MOR) significantly attenuated tolerance, whereas the 10 mg/kg dose of ketamine (KET - 10 + MOR) completely blocked morphine tolerance. On day 1, ketamine at 0.3, 3, or 10 mg/kg did not significantly alter the morphine ED⁵⁰value (Table 1). These data demonstrate a dose-dependent attenuation of morphine tolerance by ketamine. None of the treatments altered the baseline tail-flick latencies assessed before the morphine cumulative dose-response assessment. The 10 mg/kg dose of ketamine produced behavioral effects characterized by running and posturing. Preliminary dose-ranging studies indicated that these effects lasted approximately 10 min so we elected to use pre- and postdosing of ketamine given the relatively longer duration of morphine's effects on the central nervous system. No central nervous system sedation was observed.

In study 2 (Table 2), mice in which two 25-mg morphine pellets were implanted developed a significant degree of morphine tolerance when tested on day 4. Cumulative dose-response assessment showed a tenfold shift in the morphine ED⁵⁰values (46.3 and 42.3 mg/kg) compared with naive controls (4.3 mg/kg). After the pellets were removed on day 4, the mice received 10 mg/kg ketamine subcutaneously on day 4 after the pellets were removed, the same dose three times per day on days 5 through 7, or saline. A cumulative dose-response study on day 8 revealed that in the ketamine-treated group the morphine ED⁵⁰value decreased significantly to 8.7 mg/kg (not significantly different from the SAL group on day 1), indicating a complete reversal of tolerance. In the control group receiving saline subcutaneously, the morphine ED⁵⁰value on day 8 (46.7 mg/kg) showed no change compared with day 4, demonstrating the persistence of a significant degree of morphine tolerance. None of the treatments altered the baseline tail-flick latencies assessed before the morphine cumulative dose-response assessment.

In study 3, the intrathecal administration of morphine in escalating doses for 3 days (SAL + MOR group) resulted in a large rightward shift of the dose-response curve for morphine on day 5 (Figure 3), indicating a significant degree of analgesic tolerance. The magnitude of this tolerance can be expressed as a 46-fold increase in the ED⁵⁰value for intraspinal morphine from 0.8 micro gram (0.5 to 1.3 micro gram; 95% CI) in morphine-naive rats to 38.5 micro gram (26.9 to 54.9 micro gram; 95% CI) in morphine-treated rats (Table 3). In contrast, the rats that had been given ketamine and morphine (KET + MOR group) concurrently showed only a slight rightward shift on day 5 (Figure 3). The intraspinal morphine ED⁵⁰value increased two times from 1.1 micro gram (0.8 to 1.5 micro gram; 95% CI) on day 1 to 2.1 micro gram (1.6 to 2.7 micro gram; 95% CI) on day 5 (Table 3). The intraspinal morphine ED sub 50 for the KET + MOR group on day 5 was significantly different from that of the SAL + MOR group on day 5, indicating an attenuation in the development of tolerance to intraspinal morphine (from a 46-fold increase in the ED⁵⁰value to a twofold increase in the ED⁵⁰value of intraspinal morphine). Control groups (KET + SAL; SAL + SAL) showed no difference in the ED⁵⁰value of intraspinal morphine on day 1 or day 5 compared with that of day 1 in the SAL + MOR group (Table 3). No differences were observed in baseline tail-flick latencies assessed before the morphine cumulative dose-response assessment between days 1 and 5 in any of the treatment groups.

No changes in tail-flick latencies were seen 10 min after the intrathecal administration of 12 micro gram ketamine and 10 min after the subsequent administration of saline (Table 4), confirming that intraspinal ketamine at this dose did not affect baseline tail-flick latencies. No behavioral effects were observed on days 1 through 5 after intraspinal ketamine (KET + SAL) was given.

Study 1 showed that coadministration of systemic ketamine with morphine dose-dependently attenuates the development of morphine tolerance (Table 1). Study 2 found that the systemic administration of 10 mg/kg ketamine could reverse established morphine tolerance (Table 2). This is perhaps a more clinically relevant issue, because many patients have some degree of opioid tolerance. These data correlate with the earlier finding that various NMDA receptor antagonists can attenuate and reverse morphine tolerance. [2,4-6]Trujillo and Akil [13]previously studied the effect of systemic ketamine on the attenuation and reversal of morphine tolerance in rats, and they attenuated but did not reverse morphine tolerance. These authors described "reversal" as the inhibition of the expression of morphine tolerance, and they tested this by administering a single bolus dose of ketamine (10 mg/kg sup -1 given intraperitoneally) after morphine tolerance had been established to determine the effect on morphine tolerance 60 min after ketamine administration. In our study, we administered ketamine subcutaneously at 10 mg/kg for 4 days (once on day 4 and 3 times per day on days 5 through 7) after established morphine tolerance was demonstrated by determining morphine ED⁵⁰values (Table 2). We assessed the reversal of morphine tolerance by repeating the ED⁵⁰determination on day 8 without ketamine pretreatment. The reversal of established morphine tolerance that we observed was not a result of the inhibition of expression of tolerance but rather the actual decrease in the magnitude of the acquired analgesic tolerance as expressed by a direct comparison of quantitative morphine ED⁵⁰shifts. Studies of morphine tolerance suggest that the adaptive changes that occur require days to develop. [13,16]The development of complete tolerance to 10 mg/kg morphine given subcutaneously twice daily, evaluated by the hot-plate test, required more than 8 days. [13]Continuous infusion of intraspinal morphine at rates of 2, 6, and 20 nmol/h elevated the hot-plate latencies in a concentration-dependent manner on day 1, and these latencies gradually returned to saline-infused values by 3 to 5 days for each dose of morphine. [16]Furthermore, the subcutaneous infusion of LY274614 to morphine-tolerant animals that continue to receive morphine showed a gradual return of morphine sensitivity over 2 to 3 days. [4]Thus reversal of established morphine tolerance probably requires repeated administration of ketamine over time. For this reason, we repeatedly administered ketamine for 4 days after morphine tolerance had been established.

The results of our study (see Figure 2and Figure 3and Table 1and Table 3) support the concept that functional NMDA receptors are necessary for the development of morphine tolerance. However, the exact role that they play has not been clearly defined. In vitro, micro receptor activation increases NMDA receptor-gated calcium currents and is mediated by protein kinase C. [17]This potentiation of the NMDA receptor-mediated response by protein kinase C is a result of an increase in the probability of opening and a reduction of the Magnesium sup 2+ block of the NMDA receptor channels. [18]This may be one mechanism by which morphine interacts with the NMDA receptor in the development of tolerance. Activation of the NMDA receptors increases Calcium²+ influx, thereby increasing intracellular Calcium²+ concentration. An increase in basal free intracellular calcium is seen in brain synaptosomes from morphine-tolerant mice. [19]The increase in intracellular Calcium²+ may then initiate a cascade of intracellular events leading to the development of morphine tolerance, such as the production of nitric oxide. N^G-nitro-L-arginine, a nitric oxide synthase inhibitor, attenuates morphine tolerance in mice. [2,20,21]In addition, rats tolerant to intraspinal morphine show an increase in membrane-bound protein kinase C (translocated protein kinase C) in dorsal horn neurons that is not seen after a single dose of morphine. GM1 ganglioside, a substance that blocks the translocation of protein kinase C, prevents morphine tolerance and the increase in membrane-bound protein kinase C. [22]These findings suggest that nitric oxide production and translocation of protein kinase C are some of the important intracellular events involved in the development of tolerance. Furthermore, changes in central nervous system levels of NMDAR1 mRNA are seen in morphine-tolerant rats, [7]and NMDA receptors are down-regulated in specific regions of the central nervous system in rats subjected to long-term treatment with morphine, [23]suggesting that changes in gene expression are involved in the development of morphine tolerance. Ketamine has been shown to block NMDA-evoked currents in voltage- and use-dependent manners in cultured mouse hippocampal neurons [11]and to attenuate an NMDA-evoked increase in intracellular Calcium²+ levels in rat cortical slices. [12]Ketamine probably modulates morphine tolerance by blocking the NMDA receptor-dependent ion channel and thereby prevents the subsequent intracellular events that lead to morphine tolerance.

The spinal cord is probably a major site of action of NMDA receptor antagonists in attenuating morphine tolerance. [24]Intraspinal administration of the NMDA receptor antagonist MK-801 attenuates morphine tolerance development produced by the intraspinal administration of morphine. [25,26]We found in study 3 that intraspinal administration of increasing doses of morphine produced tolerance, as shown by a 46-fold increase in the intraspinal morphine ED sub 50 value. This morphine tolerance was almost completely attenuated by the coadministration of intraspinal ketamine (Figure 3and Table 3). Ketamine, a compound with high lipid solubility, is rapidly accumulated by nervous system tissue. [27]Ketamine's site of action when given intraspinally probably is local, within the spinal cord. Thus, if ketamine acted spinally in attenuating tolerance to intraspinal morphine, spinal cord NMDA receptors are probably important in the development of tolerance to intraspinal morphine. The dose of intraspinal ketamine that we used, 12 micro gram or 50 nmol, although it had no effect on baseline tail-flick latencies (Table 4), has been associated with antinociceptive effects in rat models of central sensitization. [28-30] 

Subcutaneous ketamine given at 0.3, 3, or 10 mg/kg to morphine-naive mice did not affect the analgesic sensitivity to morphine, as demonstrated by no difference in the morphine ED⁵⁰value after ketamine was given subcutaneously compared with saline (Table 1). Trujillo and Akil [13]also showed that in morphine-tolerant rats, 10 mg/kg ketamine given intraperitoneally did not change morphine sensitivity. We did not evaluate the acute effects of intraspinal ketamine on analgesic sensitivity to intraspinal morphine. However, the long-term intraspinal administration of ketamine did not change analgesic sensitivity to intraspinal morphine (Table 3). Furthermore, the ED⁵⁰value of intraspinal morphine was evaluated more than 24 h after the last dose of ketamine was given, suggesting that the attenuation of tolerance to intraspinal morphine that we observed was not a result of the change in intraspinal morphine sensitivity caused by intraspinal ketamine.

In these studies in mice and rats, the investigators were not blinded to the treatments, and thus bias could have affected the results. However, these results correlate with previous blinded studies using other NMDA receptor antagonists. [1-6] 

Ketamine is a clinically available noncompetitive NMDA antagonist that has been used in the operating room as an intravenous anesthetic for almost 30 y. Ketamine, like other PCP-like drugs, blocks the NMDA-dependent ion channel as a consequence of binding to the PCP receptor. The PCP binding site may be located within the lumen of the ion channel, and PCP-like drugs can exert their effects on the channel most readily when the channel is open as a result of activation by agonists (i.e., they are open channel blockers). [11,31]The clinical significance of open channel blockers of the NMDA receptor channel may be that their effect is maximal with pathologic activation of NMDA receptors, as in disease states, and they may have a lesser effect in steady-state physiologic conditions. [32]Ketamine, which is effective as an NMDA receptor antagonist at doses lower than currently used as an anesthetic, [9]deserves consideration for clinical use in blocking pathologic NMDA receptor-mediated events, such as morphine tolerance. Ketamine at subanesthetic doses has been reported to have analgesic effects in humans [33-35]and may be of clinical use in managing neuropathic pain [36-38]and cancer pain. [39,40] 

The most probable side effects that may be encountered with the clinical use of ketamine at subanesthetic doses are dose-dependent psychotomimetic effects. [41]The coadministration of a benzodiazepine attenuates ketamine's postanesthetic emergence reactions [42]and may be useful in controlling this side effect. The use of S(+)-ketamine, which produces fewer psychotomimetic effects in humans compared with R(-)-ketamine at equipotent doses, [43]may also be an approach to minimize these psychotomimetic side effects in patients.

Neuropathologic evaluation after large doses of both competitive and noncompetitive NMDA receptor antagonists in the adult rat reveal neuronal vacuolation in the cingulate cortex and retrosplenial cortex and other neuronal regions suggestive of a neurotoxic effect of excessive NMDA receptor blockade. [44]These findings have not been replicated in the dose range that modulates nociceptive processing and the development of tolerance in rodents, nor have they been reproduced in primate models. We are unaware of any reported cases of neurotoxicity or permanent psychosis from the established clinical experience with ketamine. The safety of repeated administration of systemic ketamine has been demonstrated in patients with burns, in whom it is used for analgesia for daily debridement and dressing changes. [45-47] 

After repeated administration of intraspinal ketamine to rabbits, postmortem examination showed no evidence of histologic damage to the spinal cord. [48]However, in primates, only single-dose administration of intraspinal ketamine has been reported. [49]Ketamine has been given to patients in single doses for surgical anesthesia without any obvious sequela. [50]However, the safety of repeated administration of intrathecal ketamine in humans has not been determined.

We presented evidence of ketamine's efficacy in attenuating and reversing morphine tolerance and of the presence of a spinal site of action for ketamine. The clinical utility and safety of ketamine for modulating morphine tolerance in patients with chronic pain syndromes should be evaluated.