Direct neuraxial administration of narcotics was described in rats in 1976  and in humans 3 yr later [2,3]. This method has since been used to provide reliable pain relief for countless patients. Because the duration of analgesia produced by a single dose of narcotic is limited, repeated injections or catheterization techniques are required to obtain longlasting analgesia. Because analgesic duration is related directly to narcotic concentration in the dorsal horn, strategies to prolong the duration of analgesia have focused on extending drug residence time in the neuraxis.
Opiates have been coupled with various delivery systems to achieve slow release of the drugs. Intrathecal injection of some slow-release formulations have increased the duration of drug action. Iophendylate has been used as a vehicle to prolong meperidine action in rabbits , and beta-cyclodextrins prolong the duration of action of opiates in rats [5,6]. More recently, two studies have used liposomes as a vehicle for alfentanil administered intrathecally in rats [7,8]. In one study, liposomal encapsulation prolonged alfentanil analgesia , although the duration of analgesia was less than that produced by plain morphine in the same model by another investigator . Analgesic effect was not significantly prolonged in the second study .
We have shown that a single dose of liposomal morphine significantly prolonged duration of analgesia and reduced toxicity compared to plain morphine after intraperitoneal administration in mice . This study was designed to evaluate the effect of various doses of liposomal morphine administered intrathecally on duration of analgesia, drug toxicity, and drug distribution within the neuraxis.
The protocol was approved by New York University's Animal Care and Use Committee. Male Swiss-Webster mice weighing 27 +/- 2 g were used. The mice were housed in a temperature- and humidity-controlled animal facility with a 12-h light-dark cycle, and were given free access to food and water.
Dimyristoyl-sn-glycero-3-phosphocholine (Avanti Polar Lipids, Alabaster, AL) and cholesterol (Sigma, St. Louis, MO) were dissolved in chloroform in a 2:1 molar ratio. The chloroform was evaporated under vacuum to yield a thin lipid film. Multilammellar liposomes were prepared by hydrating the film with different concentrations of isosmotic morphine sulfate or normal saline. The resulting liposomes were subjected to five freeze-and-thaw cycles. Nonencapsulated morphine was removed by successive washings with normal saline in the centrifuge at 2500 g. High-performance liquid chromatography, with a Waters Nova-Pak CN Registered Trademark column (Millipore, Philadelphia, PA) at 210 nm, was used to quantify morphine concentration in the liposomes and supernatant. A mobile phase of acetonitrile:potassium phosphate 25 mM in a 70:30 ratio was used. For analgesia studies, the final concentration of liposomal morphine was adjusted to 5.0, 2.5, or 1.25 mg/mL by dilution with appropriate amounts of normal saline. For toxicity studies, liposomal morphine formulations from 5.9 to 37.1 mg/mL were prepared, and for distribution studies liposomal morphine was prepared to have a final concentration of 2 mg/mL.
Ten microliters of test solution was injected intrathecally using the method of Hylden and Wilcox . After mice were lightly anesthetized with halothane, the skin overlying the dorsal lumbar spine was opened using an 8- to 10-mm transverse incision. The spinal space was identified by inserting a 30-gauge needle cephalad at a 20 degrees angle between adjacent lumbar spinal processes (L5-6). Resistance encountered at a depth of approximately 5 mm indicated proper position of the needle tip within the spinal canal. Ten microliters were injected using a 100-micro Liter automated syringe (Hamilton Co., Reno, NV). The L5-6 interspace approximates the termination of the spinal cord and the origins of the cauda equina . Injection at this site maximizes intervertebral accessibility and minimizes the likelihood of spinal cord damage. Mice were given injections in the prone position and were then transferred to their cages in this position, where they were allowed to ambulate freely for the duration of the study.
The lethality of several concentrations of plain or liposomal morphine was assessed using the up-and-down technique . Briefly, a dose approximating the LD50 was selected based on preliminary results for plain morphine (100 micro gram), and administered intrathecally to one mouse. The animal was observed for 24 h, and the next mouse was given an injection with a dose either 1.3 times greater or lesser, depending on survival or death, respectively, of the initial animal. For the liposomal formulations, the starting dose was also 100 micro gram, since preliminary studies did not reveal toxicity with any liposomal dose.
The duration of analgesia was quantified by serial tail-flick testing. Mice were placed in individual metal cylinders with an opening to allow the tail to protrude. The central portion of the tail was heated from below with a 150-W projector lamp. A single-control switch simultaneously activated the lamp and a timer. The time interval between switching on the light to the flick of the tail was recorded as tail-flick latency (TFL). The light intensity was adjusted to obtain a baseline TFL of 2.0 +/- 0.2 s. Failure to flick the tail by 5 s was taken to indicate analgesia. The 5-s cutoff time was used to avoid thermal injury. Testing was halted after TFL was < 5 s for two successive intervals. The duration of analgesia was defined as the time from injection to the last testing interval where TFL was > 5 s.
After dorsal skin incision (see above), mice were allowed 1 h to recover from the halothane anesthetic. Baseline TFL was then determined. Thirty minutes after baseline TFL determination, eight mice per group received intrathecal injections with either liposomal morphine (50, 25, or 12.5 micro gram), plain morphine (50, 25, or 12.5 micro gram), empty (normal saline) liposomes, or normal saline. Analgesia was assessed by measuring TFL at hourly intervals for 11 h, then every 3 h thereafter until TFL was < 5 s for two successive intervals. The tail-flick observer was blinded to the nature of all injections.
Morphine distribution and residence time within the neuraxis was determined by assaying for morphine in spinal cord and brain segments after injection of plain or liposomal morphine, 20 micro gram in 10 micro Liter. Four groups of six mice received intrathecal injections with 10 micro Liter of either formulation, and were killed by 100% CO2 inhalation at 1 min and at 1, 4, or 8 h after completion of injection. An additional group of six mice received injections with the liposomal formulation and were killed at 24 h.
The vertebral column was removed from each mouse and divided into three segments of equal length. Each cord segment (low, middle, or high spinal cord) was extruded from the vertebral canal using normal saline. The brain was removed from the cranium and divided into hindbrain and forebrain. After removal, the tissue samples were not washed prior to morphine assay. The wet weight of all specimens was determined before tissue samples were homogenized with acidic isopropanol. Debris was removed from the homogenates by centrifuging at 4000 times g at 4 degrees C. Morphine concentration was determined in the supernatants by radioimmunoassay (Coat-a-Count Registered Trademark; Diagnostics Products Corp., Los Angeles, CA). Samples were diluted with normal saline as necessary to bring the final concentration within the range of the assay (0.1-250 ng/mL). Results were expressed as nanograms of morphine per milligram of wet tissue.
The LD50 determined by the up-and-down method was analyzed by the method of maximum likelihood using nonlinear regression . Dose-response curves for the duration of analgesia were compared between plain and liposomal morphine using a regression model with formulations as a two-category grouping variable and dose as a continuous linear covariate. To further elucidate this interaction, formulations were compared separately for each dose. Repeated-measures analysis of variance was used to analyze segmental morphine concentration as a function of formulation, time, and segment. Formulation and time were modeled as "between subject" factors and segment as the "within subject" factor. Two-way analysis of variance was used to compare total drug recovered at each interval between formulations.
For plain morphine, the estimated LD50 was 200 micro gram (confidence interval 151-257 micro gram). For liposomal morphine, death did not result with the maximum dose administered, 371 micro gram. The next dose level (482 micro gram) could not be assessed, since it was not technically possible to deliver this dose in a 10-micro Liter volume.
The mean duration of analgesia after intrathecal injection of plain morphine was 2.6 +/- 0.4 (SE) for 12.5-micro gram, 4.1 +/- 0.5 for 25-micro gram, and 4.6 +/- 1.0 h for the 50-micro gram dose. For liposomal morphine formulations, the mean duration of analgesia was 4.3 +/- 1.0 for 12.5-micro gram, 13.4 +/- 1.6 for 25-micro gram, and 16.8 +/- 4.0 h for the 50-micro gram dose Figure 1. Regression analysis of dose-response curves for plain and liposomal morphine demonstrated significant interaction between dose and formulation (P < 0.004); duration of analgesia increased at a greater rate with increasing dose for liposomal morphine than for plain morphine. When each dose of the two formulations was compared separately, mean duration of analgesia was significantly prolonged for 25- and 50-micro gram doses of liposomal morphine (P < 0.001) but not for the 12.5-micro gram dose. No analgesia was detected after injection of saline or "empty" (normal saline) liposomes.
The total morphine recovered from the neuraxis Figure 2 1 min after injection was similar for plain and liposomal morphine. However, by 1 h, significantly more morphine was recovered from mice given the liposomal formulation (P < 0.005). This difference persisted at 4 and 8 h (P < 0.001). In mice given the liposomal morphine, neuraxial morphine levels were still detectable 24 h after injection.
(Figure 3) depicts segmental morphine distribution within the neuraxial segments. At 1 min, mice given the liposomal formulation had a greater concentration of morphine in the low spinal cord and a lower concentration in the high spinal cord than mice given plain morphine. Significant interactions between segment times time and segment times formulation were observed (each P < 0.0001). The presence of a segment times formulation interaction signifies that the pattern of observed segment concentration depended on the formulation used. For example, with liposomal morphine, the highest concentration at each time interval was found in the low spinal cord segment, whereas concentration in the low spinal cord segment was not significantly increased compared to other segments using plain morphine.
An ideal narcotic for neuraxial administration would remain localized at the injection site for a sustained time to facilitate interaction with segmental spinal opiate receptors, and minimize the likelihood of supraspinal side effects. This may be achieved by altering the pharmacokinetics of a known narcotic by incorporating it into a drug delivery system. Various delivery systems have been evaluated including iophendylate , beta cyclodextrins [5,6], and most recently, liposomes [7,8]. Alfentanil has been encapsulated in liposomes in an attempt to modify its action after intrathecal administration [7,8]. Although liposomal alfentanil resulted in lower plasma drug levels compared to the plain drug, the duration of analgesia was prolonged only modestly in one study . It is not clear whether this failure to prolong analgesia was due to injection of an insufficient drug depot and/or to rapid release of alfentanil from the liposomes.
We have previously reported that morphine is slowly released from liposomes in vitro, and that this is associated with an increased duration of analgesia after systemic administration compared to plain drug . In the current study, similar effects were observed after intrathecal injection of sufficiently large doses of liposomal morphine. As in our previous study, the extended duration of analgesia is explained by the slow passage of the poorly lipid soluble morphine across the liposomal membrane. Sequestration of drug within liposomes also explains the observed differences in lethality between plain and liposomal morphine, which presumably was due to respiratory depression. Because the maximum dose of liposomal morphine which could be administered in a 10-micro Liter volume did not result in death, it was not possible to determine the LD50 for that formulation.
Residence time within the neuraxis was significantly different for the two formulations, with plain morphine being rapidly cleared, but with liposomal morphine persisting in the neuraxis Figure 2. Prolonged neuraxial residence time has also been observed after intrathecal injection of other liposomal drugs [15,16], and has been attributed to the inability of large intact liposomes to pass through the subarachnoid villi . Thus, because only free morphine is able to pass into the systemic circulation, the rate-limiting factor in morphine removal from the neuraxis was likely the rate of narcotic release from the liposomal depot.
Although the duration of analgesia was not quantified in the mice used for the segmental distribution studies, an estimate of the expected duration of analgesia produced by the 20-micro gram dose can be derived by interpolation of the results presented in Figure 1. The mean analgesic duration produced by 20 micro gram of plain or liposomal morphine would have been approximately 4 and 14 h, respectively. The decline in measured segmental morphine concentration paralleled these estimated analgesic durations. For plain morphine, only 5% of the dose of plain morphine recovered at 1 min remained at 4 h, whereas 44% of the 1-min liposomal dose was present at 4 h, and 29% remained at 8 h. The failure of increasing doses of plain morphine to significantly prolong analgesic duration indicates that the concentration of morphine in the neuraxis declined below a critical level, as free drug was removed from the site. The prolonged analgesia produced by increasing doses of liposomal morphine is reflected by the persistent neuraxial morphine levels.
The segmental distribution of morphine was quite different for plain and liposomal formulations. In mice given plain morphine, drug concentration was not significantly larger in any particular segment. After liposomal morphine, however, segmental drug levels were significantly larger in the low spinal cord segment at all intervals tested. However, not all liposomal morphine remained confined to the low spinal cord region. Brain morphine levels at 1, 4, and 8 h were greatest in mice given the liposomal formulations, confirming cephalad spread of drug. Because the methodology used to assay segmental morphine concentration was designed to disrupt liposomes, the data do not differentiate between free or liposomal-bound morphine. Thus, we cannot state with certainty whether the brain morphine levels represented cephalad spread of intact liposomes or free drug. However, the rapid disappearance of morphine from the neuraxis after injection of plain drug suggests that free morphine is rapidly cleared after its release from liposomes. Therefore, in mice given the liposomal formulation, the morphine recovered likely represented mostly liposome-encapsulated drug, and the higher morphine levels detected in brain probably also represented liposomal drug. This is supported by lack of toxicity seen in the liposomal group. Some liposomes, clearly, were able to spread cephalad to the brain.
In a recent study of intrathecal liposomal alfentanil in rats, allodynia was present after empty liposomes made of dipalmitoyl phosphocholine were administered . A follow-up study demonstrated that allodynia was caused by the levo, but not the dextro, isomer of the phospholipid . Previous work has demonstrated demyelination after intrathecally administered lysophosphatidyl choline , but minimal toxicity after intrathecal administration of neutral liposomes . We did not observe allodynia in the current study. However, we did not specifically assess the tissue toxicity of any of our formulations. Clearly, thorough evaluation of the potential for neurotoxic effects of liposomal formulations must be done prior to human use.
In summary, these data suggest that the prolonged analgesia produced by liposomal morphine results from slow release of drug from a lipid narcotic depot. Furthermore, the encapsulation of morphine within liposomes markedly decreased its potential for lethality. These formulations appear to represent a simple, safe, and reliable method to achieve prolonged analgesia with a single neuraxial injection.
1. Yaksh TL, Rudy TA. Analgesia mediated by a direct spinal action of narcotics. Science 1976;192:1357-8.
2. Behar M, Olshwang D, Magora F, et al. Epidural morphine in treatment of pain. Lancet 1979;1:527-8.
3. Wang JK, Nauss LA, Thomas JE. Pain relief by intrathecally applied morphine in man. Anesthesiology 1979;50:149-51.
4. Langerman L, Golomb E, Benita S. Prolongation of the pharmacologic effect of intrathecal meperidine by the use of a lipid solution of it. Anesth Analg 1991;72:635-8.
5. Yaksh TL, Jang J, Nishiuchi Y, et al. The utility of 2-hydroxypropyl-beta-cyclodextrin as a vehicle for the intracerebral and intrathecal administration of drugs. Life Sci 1991;48:623-33.
6. Jang J, Yaksh TL, Hill HF. Use of 2-hydroxypropyl-beta-cyclodextrin as an intrathecal drug vehicle with opioids. J Pharmacol Exp Ther 1992;261:592-600.
7. Bernards CM, Luger TJ, Malmberg AB, et al. Liposome encapsulation prolongs alfentanil spinal analgesia and alters systemic redistribution in the rat. Anesthesiology 1992;77:529-35.
8. Wallace MS, Yanez AM, Ho RJY, et al. Antinociception and side effects of liposome-encapsulated alfentanil after spinal delivery in rats. Anesth Analg 1994;79:778-86.
9. Rattan AK, McDonald JS, Tejwani GA. Differential effects of intrathecal midazolam on morphine-induced antinociception in the rat: role of spinal opioid receptors. Anesth Analg 1991;73:124-31.
10. Grant GJ, Vermeulen K, Zakowski MI, et al. Prolonged analgesia and decreased toxicity with liposomal morphine in a mouse model. Anesth Analg 1994;79:706-9.
11. Hylden JLK, Wilcox GL. Intrathecal morphine in mice: a new technique. Eur J Pharmacol 1980;67:313-6.
12. Sidman RL, Angevine JB Jr, Pierce ET. Atlas of the mouse brain and spinal cord. Cambridge: Harvard University Press, 1971.
13. Dixon WJ. The up-and-down method for small samples. J Am Statis Assoc 1965;60:967-78.
14. Yaksh TL, Noueihed R. The physiology and pharmacology of spinal opiates. Annu Rev Pharmacol Toxicol 1993;25:433-62.
15. Kim S, Scheerer S, Geyer MA, et al. Direct cerebrospinal fluid delivery of an antiretroviral agent using multivesicular liposomes. J Infect Dis 1990;162:750-2.
16. Kim S, Kim DJ, Geyer MA, et al. Multivesicular liposomes containing 1-beta-D-arabinofuranosylcytosine for slow-release intrathecal therapy. Cancer Res 1987;47:3935-7.
17. Isackson J, Wallace M, Ho RJY, et al. Antinociception and side effects of L and D isomer liposome-encapsulated alfentanil after spinal delivery in rats [Abstract]. Anesthesiology 1994;81:A914.
18. Hall SM. The effect of injections of lysophosphatidyl choline into white matter of the adult mouse spinal cord. J Cell Sci 1972;10:535-46.
19. Adams DH, Joyce G, Richardson VJ, et al. Liposome toxicity in the mouse central nervous system. J Neurol Sci 1977;31:173-9.