The magnesium ion is a noncompetitive N-methyl-D-aspartate (NMDA) receptor antagonist . The NMDA receptor channels are blocked by magnesium in a voltage-dependent manner at a site deep within the ion channel, most likely distinct from the MK801 binding site [2,3]. Magnesium sulfate (MAG) administered intrathecally to neuropathic rats suppresses neuropathic pain responses at the spinal level , but antinociception is only attained with large doses, and efficacy is limited by side effects. Patients receiving magnesium systemically for the treatment of peripheral neuropathic pain did not have significantly reduced pain or allodynia . Although magnesium alone is unlikely to provide adequate analgesia, the combination of systemic MAG and morphine (MOR) has been demonstrated to decrease post-operative MOR requirements, which suggests a potentiating effect of magnesium .
The systemic co-administration of the noncompetitive NMDA receptor antagonist MK801 attenuated the development of MOR tolerance in rats [7,8]. Continuous intrathecal co-infusion of MK801 and MOR has also been demonstrated to delay the development of MOR tolerance in rats . Despite the efficacy of both systemic and intrathecally administered MK801 in blocking the development of MOR tolerance, neurotoxicity limits the clinical application of MK801 .
The specific aims of this study were 1) to compare the development of tolerance (a decreased drug effect with repeated administration) to large-dose intrathecal MOR (20 nmol/h) in normal rats with co-administration of intrathecal MAG or MK801, 2) to evaluate the effect of MAG co-infusion on MOR dependence (an altered state that necessitates continued exposure to the drug to prevent the stereotypic responses characteristic of withdrawal of the drug) by observing withdrawal signs after systemic naloxone administration, 3) to characterize the potentiating and the tolerance attenuating effect of the combination of MAG and moderate-dose (1 nmol/h) MOR, and 4) to measure cerebrospinal fluid (CSF) and blood magnesium levels after prolonged intrathecal MAG infusion.
The institutional animal care and use committee approved the study. Experiments were performed on 205 male Sprague-Dawley rats (Harlan, Indianapolis, IN) weighing 200-225 g. Animals were anesthetized with halothane and intrathecal polyethylene catheters (0.28 mm inner diameter, 0.61 mm outer diameter) implanted via the cisterna magna and threaded 8 cm caudally so that the tip was at the rostral portion of the lumbar enlargement . The other end of the catheter, which had been heat-sealed to a larger polyethylene catheter (0.76 mm inner diameter, 1.22 mm outer diameter), was sealed with a stylet and buried subcutaneously . All skin margins were closed, and the animals were allowed to recover for 4 days. Animals exhibiting neurologic impairment before intrathecal drug administration were excluded from the study. The animals were then briefly anesthetized with halothane and an osmotic minipump (Model 2ML2; Alza, Palo Alto, CA; 5 [micro sign]L/h flow rate) was attached to the catheter (15 [micro sign]L dead space) and placed in a subcutaneous pocket on the dorsum of the neck.
The development of tolerance to large-dose intrathecal MOR (20 nmol/h) with co-administration of intrathecal MAG or MK801 was examined by implanting pumps containing solutions designed to deliver MK801 10 nmol/h, MOR 20 nmol/h, MK801 10 nmol/h + MOR 20 nmol/h, and MAG 50 [micro sign]g/h or MAG 50 [micro sign]g/h + MOR 20 nmol/h. To evaluate the dose-response of MAG on the development of large-dose MOR tolerance, pumps contained solutions to deliver MOR 20 nmol/h alone or combined with MAG 20, 40, 60, and 90 [micro sign]g/h. To examine the potentiating effect and the attenuation of opioid tolerance of a moderate dose of MOR and MAG, pumps were filled to deliver MOR 1 nmol/h, MAG 60 [micro sign]g/h + MOR 1 nmol/h, MAG 60 [micro sign]g/mL, or saline. The MOR infusion rate of 1 nmol/h was chosen because it produced an approximate percent maximal possible effect (%MPE) of 50 in pilot experiments. The solutions contained in all pumps were randomly assigned. The specific agents used in this study were morphine sulfate hypodermic tablets for injection (Eli Lilly and Co., Indianapolis, IN), magnesium sulfate injection USP (SoloPak Labs, Elk Grove Village, IL), and MK801 and naloxone hydrochloride (Research Biochemicals International, Natick, MA). All drug dilutions were made with 0.9% sodium chloride injection.
Tail-flick testing was performed the morning of the implantation of the osmotic pump and was repeated at 11 AM daily for 7 consecutive days. Before each tail-flick test, the animals were evaluated for motor function using a 0-4 grading scale of gait . The grading of neurological status was scored as follows: 0 = total paralysis of hindlimbs, 1 = walking-like movements of hindlimbs but no weight-bearing, 2 = walking with marked impairment or ataxia, 3 = walking well with slight deficit, 4 = normal. Thermal latencies were measured as the time required for the rat to move its tail from a heat source, with the latency to withdrawal determined automatically using a photoelectric sensor (Omnitech, Columbus, OH). Starting from an ambient temperature of 24[degree sign]C, the heat source reaches a temperature of 88.9 +/- 3.0[degree sign]C 6 s after its activation, the typical response latency in a normal rat. Data for latencies before pump implantation were compared using a one-way analysis of variance (ANOVA) with post hoc testing performed when appropriate using the Tukey-a method (alpha = 0.05). The personnel performing all antinociceptive and withdrawal testing were blinded to the solutions contained in the pumps. Data were converted to %MPE using the formula : %MPE = 100 x (latency - baseline)/(cutoff - baseline). A 15-s cutoff was used to prevent thermal injury to the tail. The time course of antinociception was analyzed using ANOVA for repeated measures with post hoc testing performed when appropriate using the Tukey-a method (alpha =0.05).
Animals receiving MOR 20 nmol/h, MAG 40 [micro sign]g/h + MOR 20 nmol/h, and MAG 60 [micro sign]g/h + MOR 20 nmol/h were evaluated for naloxone-induced withdrawal signs after 7 days of drug infusion. After antinociceptive testing on Day 7, the osmotic pump was removed under halothane anesthesia, and 24 h later, an intraperitoneal injection of naloxone 0.3 mg/kg was administered. The presence or absence of withdrawal signs --e.g., vocalization in response to light touch with a piece of polyethylene tubing; spontaneous vocalization; abnormal posture; ejaculation; "wet dog head shakes"; escape attempts-were assessed for 1 h. The frequency of each of the withdrawal signs was compared using a chi squared statistic. To examine the persistence of the antinociceptive effect, one additional group receiving MAG 60 [micro sign]g/h + MOR 20 nmol/h had the osmotic pump removed on Day 7 as described above, and tail-flick latencies were monitored daily until latencies had returned to baseline.
The animals receiving MOR 1 nmol/h, MAG 60 [micro sign]g/h + MOR 1 nmol/h, MAG 60 [micro sign]g/h, or saline were evaluated for antinociception using the tail-flick method for the first 7 days. These %MPE data were also analyzed using ANOVA for repeated measures with post hoc testing performed when appropriate using the Tukey-a method (alpha = 0.05). After testing on the 7th day, the pump and the wider polyethylene tubing were removed under halothane anesthesia so that probe MOR bolus doses could be administered to evaluate the degree of tolerance. The narrow tubing was then externalized through the dorsum of the neck, flushed with saline (10 [micro sign]L), and closed with a stylet. Antinociception was evaluated 24 h later, then a single intrathecal bolus injection of MOR 3, 10, 30, or 100 nmol was administered. Tail-flick testing was repeated 30 min after the bolus. MOR antinociceptive 50% effective dose (ED50) values for each of the drug groups were derived using linear regression of %MPE with the log of the probe dose. Differences in the ED50 estimations were determined using the confidence interval method at P = 0.05 .
In a separate group of rats (n = 10), CSF and serum magnesium concentrations were determined after a 2-day intrathecal MAG infusion at 60 [micro sign]g/h. Measurement of magnesium ion concentration was by a formazan dye binding method with reflectance spectrophotometry using an automated chemistry analyzer (Ektacem 700 XR; Kodak, Rochester, NY). The intrathecal catheter system was modified by attachment of a microdialysis loop . Quantification of the in vivo dialysis membrane recovery was performed by retrodialysis . The catheter dead space was filled with saline, and the pump flow rate (0.5 [micro sign]L/h flow rate) was chosen to permit determination of a baseline CSF magnesium level at 24 h after pump attachment, before exposure to the MAG solution. Microdialysis samples were collected for 60 min at 5 [micro sign]L/min perfusion rate. Determination of CSF magnesium level with MAG exposure was obtained 72 h after pump attachment. A serum magnesium concentration was determined by direct cardiac puncture (1 mL; approximately 400 [micro sign]L of serum) under halothane anesthesia after the final microdialysis. In a separate control group of rats, baseline serum magnesium levels were also determined by direct cardiac puncture under halothane anesthesia. CSF magnesium levels before and after the intrathecal MAG infusions were compared using Wilcoxon's signed rank test. Serum magnesium levels between groups of animals were compared using the Mann-Whitney U-test. P < 0.05 was considered statistically significant.
Tail-flick responses to MOR 20 nmol/h on Days 1 and 2 (assessed by %MPE) reflected near maximal antinociception, but by Days 4-7, antinociception was significantly reduced compared with Day 2 (Figure 1). Neither MK801 nor MAG alone increased the %MPE on any day compared with Day 0. Relative to Day 2, the combination of MK801 and MOR maintained antinociception through Day 7, whereas the combination of MAG with MOR maintained antinociception through Day 6. Compared with MOR, the combination of MK801 and MOR resulted in a significant increase in %MPE on Days 4-7, whereas the combination of MAG with MOR resulted in a significant increase in the %MPE on Days 3-7.
The effect of different doses of MAG in combination with MOR on the decrease in response latencies over time is shown in Figure 2. The effect of a co-infusion of MAG 20 [micro sign]g/h did not differ from that of MOR alone (data not shown). MAG 40 [micro sign]g/h resulted in a significant increase in %MPE on Days 3-5 compared with MOR alone, whereas MAG 60 [micro sign]g/h produced significantly higher %MPE on Days 3-7 compared with MOR alone. When MOR was combined with MAG 90 [micro sign]g/h, five of eight animals became ataxic (neurologic status grade range 2-3), and antinociception could not be reliably assessed. Neurological impairment was not seen during the 7-day infusion at any smaller magnesium doses. Withdrawal signs after systemic naloxone administration were not significantly changed by the co-infusion of MAG 40 [micro sign]g/h or 60 [micro sign]g/h (Figure 3). After the intrathecal infusion of MAG 60 [micro sign]g/h + MOR 20 nmol/h, antinociceptive responses were maintained above than baseline for 3 days. On the 4th day after drug infusion, the tailflick latencies had returned to baseline (%MPE = -1.3 +/- 4.1 on Day 11) (Figure 4).
Compared with Day 1, there was a decrease in the antinociceptive response to a small dose of MOR (1 nmol/h) by Day 4 of the infusion. The antinociceptive response to MAG 60 [micro sign]g/h + MOR 1 nmol/h noted on Day 1 did not decrease until Day 7. MAG alone produced no increase in %MPE at any time. The combination of MAG 60 [micro sign]g/h + MOR 1 nmol/h produced a significantly greater overall antinociceptive effect on Days 1-7 compared with MOR alone (Figure 5).
Withdrawal latencies (in seconds) 24 h after pump disconnection were not different from preinfusion baseline values in the saline (5.89 +/- 0.08), MOR 1 nmol/h (5.47 +/- 0.15), and MAG 60 [micro sign]g/h (6.29 +/- 0.18) groups, but they were prolonged in the MAG 60 [micro sign]g/h + MOR 1 nmol/h group (from 6.00 +/- 0.12 to 6.89 +/- 0.35 [alpha = 0.05]), similar to those in the MAG 60 [micro sign]g/h + MOR 20 nmol/h group (see Figure 4 and Figure 5). MOR bolus probe doses administered on Day 8 demonstrated a significant shift to the right in the dose-response curve for animals in the MOR group compared with those receiving saline, MAG, or the combination of MAG + MOR (Figure 6). The antinociceptive ED50 (95% confidence intervals) values were 10.9 (3.6-33.1) nmol for saline, 10.9 (4.8-31.3) nmol for MAG, 11.2 (2.6-48.5) nmol for MAG + MOR, and 109.7 (50.1-216.7) nmol for MOR.
Magnesium concentrations in CSF obtained during saline infusion were 17.0 +/- 1.0 [micro sign]g/mL, and they increased to 41.4 +/- 23.6 [micro sign]g/mL after 48 h of an intrathecal MAG 60 [micro sign]g/h infusion (P = 0.04). Serum magnesium concentrations obtained from control rats were 26.0 +/- 2.2 [micro sign]g/mL, and after 48 h of an intrathecal MAG infusion (60 [micro sign]g/h), they were 26.0 +/- 3.2 [micro sign]g/mL (P = 0.73).
Magnesium can block calcium influx and, in particular, noncompetitively antagonize NMDA receptor channels. These effects have prompted the investigation of magnesium as an analgesic agent. In rats with a peripheral mononeuropathic lesion of the sciatic nerve, a reduction in heat hyperalgesia occurred with intrathecal MAG (185 and 375 [micro sign]g), but large doses (>or=to500 [micro sign]g) produced lethargy and ataxia . On the contralateral sham-operated (control) hindlimb, antinociception was not observed even at doses of 750 [micro sign]g. In normal rats, the intrathecal administration of MAG (1260 [micro sign]g) initially produced a sensory blockade of the lower torso and hindlimbs but resulted in unconsciousness by 12 minutes after injection . Therefore, magnesium may not be useful as a sole analgesic agent, because analgesia is evident only at large doses that are limited by serious side effects.
Studies suggesting an interaction of NMDA receptor antagonists with the response to opioids led to our investigation of MAG and its interaction with MOR. The systemic administration of the NMDA noncompetitive antagonist MK801 to rats has been shown to enhance MOR antinociception in some studies [18,19], but not in others . Intrathecally administered MK801 did not increase the antinociceptive response to MOR when co-administered in rats [9,20], although intrathecal ketamine reduced intrathecal dose requirements of MOR in patients with terminal cancer pain . The co-administration of IV magnesium decreased systemic MOR requirements for the first 48 hours postoperatively . In the present study, we demonstrate that the intrathecal co-infusion of MAG and MOR produces a potentiation of the antinociceptive effect, which is most clearly exhibited at doses of MOR that alone produced approximately 50% MPE. Specifically, on Day 1 of drug infusion, MOR 1 nmol/h alone produced a %MPE of 63.2 +/- 7.6, and MAG 60 [micro sign]g/h alone produced a %MPE of 7.5 +/- 7.2, whereas the combination resulted in a %MPE of 99.8 +/- 0.2, which was statistically greater than MOR.
NMDA antagonists have also been shown to delay the development of MOR tolerance. Systemic MK801 attenuated the development of tolerance to systemic MOR in the rat [7,8], and continuous intrathecal MK801 blocked the development of tolerance to continuous intrathecal MOR . In the current study, MK801 10 nmol/h in combination with MOR 20 nmol/h maintained the antinociceptive response, compared with MOR 20 nmol/h alone for seven days (see Figure 1), similar to the previous study. We have demonstrated that MAG 50 [micro sign]g/h or 60 [micro sign]g/h, in combination with MOR 20 nmol/h, also attenuates the decrease in %MPE response latencies observed with prolonged infusion of MOR 20 nmol/h alone. The increase in tail-flick latencies for the combination of MAG 60 [micro sign]g/h + MOR 20 nmol/h persisted for three days after the infusion but returned to baseline by Day 4. This demonstrates that the combination of MAG + MOR did not permanently alter the tail-flick with-drawal response in the rat at the doses studied.
Probe MOR bolus doses demonstrated a reduced antinociceptive effect in animals that received MOR 1 nmol/h, confirming that tolerance had developed to the moderate dose of MOR. The addition of MAG 60 [micro sign]g/h to MOR 1 nmol/h preserved MOR probe dose responsiveness so that it did not differ from that in animals that received saline or MAG 60 [micro sign]g/h (Figure 6). This effect was different than that previously reported with MK801, which either alone or in combination with MOR resulted in a leftward shift of the dose-response curve compared with saline . In addition, unlike the findings with MK801, we did not observe significant alteration in withdrawal responses after naloxone administration to animals that had received the combination of MAG and MOR (Figure 3). Similarly, dextromethorphan, another NMDA antagonist, was less effective than MK801 in reducing the severity of naloxone-induced withdrawal symptoms after systemic co-infusion with MOR . Although the mechanisms for these differences between MK801 and magnesium are not apparent, they may relate to a difference in binding sites between magnesium and MK801 at the NMDA receptor complex .
There are no studies in rats examining the relationship between blood and CSF concentrations of magnesium after MAG infusions. The effects of MAG on MOR antinociceptive potentiation and tolerance observed in this study occurred in association with CSF magnesium concentrations that would be difficult to achieve with systemic administration in humans. In the treatment of preeclampsia, a MAG loading dose of 6 g and infusion of 2 g/h only increased CSF magnesium ion concentrations by 19%, but serum concentrations increased almost 300% . The administration of larger amounts of MAG (3-6 g/h) increased serum concentrations markedly (800%) and was associated with increases in PR intervals, gastric hypomotility, and peripheral muscle flaccidity . In our study, the intrathecal infusion of MAG 60 [micro sign]g/h produced a 144% increase in CSF magnesium ion concentrations without altering serum concentrations. It seems that there is a significant pharmacokinetic advantage to administering MAG intrathecally to achieve effective CSF levels that could not be safely achieved via systemic administration. A potential limitation to the use of intrathecal MAG may be the narrow range at which effective antinociceptive potentiation and tolerance reduction may occur without side effects. At a dose of 20 [micro sign]g/h, no difference in %MPE was observed when MAG was added to MOR. The continuous intrathecal infusion of MAG at a dose of 90 [micro sign]g/h affected motor function in the rat, and although not evident with doses of 40-60 [micro sign]g/h, it is possible that MAG at these doses may be causing subtle effects on locomotion not detected in this study.
Although the exact mechanism of the interaction between the NMDA receptor complex and opioid antinociception has not been fully elucidated, the similarity of our findings to observations with MK801 suggests that the noncompetitive antagonism of the NMDA receptor complex by magnesium may be involved. Observed differences in antinociceptive responses, tolerance, and withdrawal may be a result of differences between magnesium and MK801 in the binding sites within the NMDA complex. They may also represent magnesium activity at other presynaptic or postsynaptic cation channels, although the NMDA site seems to be the locus in the spinal cord most sensitive to magnesium . In summary, the combination of intrathecal MAG and MOR resulted in a potentiation of antinociception and a delay in the development of MOR tolerance in rats. These results suggest that the intrathecal administration of MAG may be a useful adjunct to spinal morphine analgesia, although careful, long-term neurotoxicity and side effect studies are necessary before clinical application could be considered.
The authors thank Lisa Albu, Laurie Sears, Aleksandra Stupairic-Stancic, and Bohung Chen, for their technical assistance.
1. Cotman CW, Iversen LL. Excitatory amino acids in the brain: focus on NMDA receptors. Trends Neurosci 1987;10:263-5.
2. Ascher P, Nowak L. Electrophysiological studies of NMDA receptors. Trends Neurosci 1987;10:284-8.
3. Wong EHF, Kemp JA. Sites for antagonism on the N-methyl-D-aspartate receptor channel complex. Annu Rev Pharmacol Toxicol 1991;31:401-25.
4. Xiao WH, Bennett GJ. Magnesium suppresses neuropathic pain responses in rats via a spinal site of action. Brain Res 1994;666:168-72.
5. Felsby S, Nielsen J, Arendt-Nielsen L, Jensen TS. NMDA receptor blockade in chronic neuropathic pain: a combination of ketamine and magnesium chloride. Pain 1995;64:283-91.
6. Tramer MR, Schneider J, Rene-Andreas M, Rifat K. Role of magnesium sulfate in postoperative analgesia. Anesthesiology 1996;84:340-7.
7. Trujillo KA, Akil H. Inhibition of morphine tolerance and dependence by the NMDA receptor antagonist MK-801. Science 1991;251:85-7.
8. Marek P, Ben Eliyahu S, Gold M, Liebeskind JS. Excitatory amino acid antagonists (kynurenic acid and MK-801) attenuate the development of morphine tolerance in the rat. Brain Res 1991;547:77-81.
9. Dunbar S, Yaksh TL. Concurrent spinal infusion of MK801 blocks spinal tolerance and dependence induced by chronic intrathecal morphine in the rat. Anesthesiology 1996;84:1177-88.
10. Olney JW, Labruyere J, Price MT. Pathological changes induced in cerebrocortical neurons by phencyclidine and related drugs. Science 1989;244:1360-2.
11. Stevens CW, Yaksh TL. Dynorphin A and related peptides administered intrathecally in the rat: a search for putative kappa opiate receptor activity. J Pharmacol Exp Ther 1986;238:833-8.
12. Kroin JS, Bianchi GD, Penn RD. Intrathecal baclofen down-regulates GABAB
receptors in the rat substantia gelatinosa. J Neurosurg 1993;79:544-9.
13. Yaksh TL, Kohl RL, Rudy TA. Induction of tolerance and withdrawal in rats receiving morphine in the spinal subarachnoid space. Eur J Pharmacol 1977;42:275-84.
14. Tallarida RJ, Murray RB. Manual of pharmacologic calculations with computer programs. 2nd ed. New York: Springer-Verlag, 1987:26-31.
15. Marsala M, Malmberg AB, Yaksh TL. The spinal loop dialysis catheter: characterization of use in the unanesthetized rat. J Neurosci Methods 1995;62:43-53.
16. Wang Y, Wong SL, Sawchuk RJ. Microdialysis calibration using retrodialysis and zero-net flux: application to a study of the distribution of zidovudine to rabbit cerebrospinal fluid and thalamus. Pharmacol Res 1993;10:1411-9.
17. Bahar M, Chanimov M, Grinspun E, et al. Spinal anesthesia induced by intrathecal magnesium sulfate. Anaesthesia 1996;51:627-33.
18. Ben Eliyahu S, Marek P, Vaccarino AL, et al. The NMDA receptor antagonist MK-801 prevents long-lasting non-associative morphine tolerance in the rat. Brain Res 1992;575:304-8.
19. Kest B, Marek P, Liebeskind JC. The specific N-methyl-D-aspartate (NMDA) receptor antagonist MK-801 blocks U-50,488 but not morphine antinociception. Brain Res 1992;589:139-42.
20. Yamamoto T, Yaksh TL. Studies on the spinal interaction of morphine and the NMDA antagonist MK801 on the hyperesthesia observed in the rat model of sciatic mononeuropathy. Neurosci Let 1992;135:67-70.
21. Yang CY, Wong CS, Chang JY, Ho ST. Intrathecal ketamine reduces morphine requirements in patients with terminal cancer pain. Can J Anaesth 1996;43:379-83.
22. Manning BH, Mao J, Frenk H, et al. Continuous coadministration of dextromethorphan or MK-801 with morphine: attenuation of morphine dependence and naloxone-reversible attenuation of morphine tolerance. Pain 1996;67:79-88.
23. Thurnau GR, Kemp DB, Jarvis A. Cerebrospinal fluid levels of magnesium in patients with preeclampsia after treatment with intravenous magnesium sulfate: a preliminary report. Am J Obstet Gynecol 1987;157:1435-8.
24. Fisher RS, Kaplan PW, Krumholz A, et al. Failure of high-dose intravenous magnesium sulfate to control myoclonic status epilepticus. Clin Neuropharmacol 1988;11:537-44.
25. Ault B, Evans RH, Francis AA, et al. Selective depression of excitatory amino acid induced depolarizations by magnesium ions in isolated spinal cord preparations. J Physiol 1980;307:413-28.