Distribution of free and liposome-entrapped [3H]methotrexate in the central nervous system after intracerebroventricular injection in a primate.

[ -H] Methotrexate ([ -H]MTX) was entrapped inside sonically disrupted, positively charged, [14C]cholesterol-labeled liposomes. Its distribution and metabolic integrity was then studied after a single injection into the cérébrospinal fluid (CSF) in the lateral cerebroventricles of cynomolgus monkeys (Macaca fascicularis). Between 24 and 96 hr after injection, a maximal 4-fold increase in average tissue levels, compared with the levels found after injection of free [3H]MTX, was found for total central nervous system (brain plus spinal cord). Greater in creased levels of [3H]MTX after injection of liposomeentrapped [3H]MTX compared with free [3H]MTX were found in CSF, but considerable variation was noted ac cording to the site of sampling. Metabolic degradation of free MIX, as measured by the difference between MTX determined by 'H-specific activity and MTX determined by the dihydrofolate reductase assay, was detected at 24 hr, and 96 hr after injection only 11% of the free MTX in CSF represented intact MTX. In contrast, after injection of liposome-entrapped MTX 72% was found to be intact at the same time. Consistent with this protection against metabolic degradation, a 4C:'H ratio of 0.9 to 1.1 in CSF for up to 24 hr after injection was found, suggesting that most of the liposome-entrapped [3H]MTX injected had remained in intact liposomes during this time period. Distribution within brain tissue was also studied at 24 hr postinjection, and it was found that [3H]MTX levels after injection of liposome-entrapped [3H]MTXwere higher than after injection of the free form near tissue-CSF interfaces, but less than the free form in deep areas of the brain. Considerable leakage from the central nervous system to plasma was found to occur soon after injection for both the free and the liposome-entrapped [3H]MTX, but the liposome-entrapped [3H]MTX showed 10to 20-fold higher steady-state plasma levels. This was attributed to a much lower rate of removal from plasma due to the much lower renal excretion of liposome-entrapped MTX. These find ings are compared with the behavior found after systemic injection of liposome-entrapped MTX and other com pounds. The implications of these results for a possible chemotherapeutic use of liposome-entrapped drugs di rectly injected into the central nervous system are briefly discussed. INTRODUCTION Following initial work (10) an increasing number of more recent studies (for recent reviews see Refs. 9 and 24) has shown that entrapment of substances inside phospholipid liposomes can markedly modify the distribution of the entrapped species after injection in vivo. The magnitude of the effects observed depends both on the nature of the entrapped species, as well as the composition of the lipo somes (see Ref. 2 for methods of preparation and properties of liposomes). In general, however, injection of liposomeentrapped substances i.V., or via other systemic routes, retards the clearance of the entrapped species from the body and increases their uptake into tissues, especially tissues of the reticuloendothelial system (6, 8, 10, 12, 14). The potential advantages for chemotherapy of increased specific uptake by tumor tissue and decreased toxicity to normal tissue (9, 24) would also apply to injection of liposome-entrapped material into the CMS,2 and further more preferential clearance by the reticuloendothelial sys tem would be expected to be diminished. Also, since we have found no significant uptake of intact liposomes into the CMS after i.v. injection (14), direct administration into the CNS would seem to be required. However, although a report studying the toxicity of liposomes injected intracerebrally into mice has appeared (1), there have been no published studies on the distribution of liposomes and liposome-entrapped substances after direct injection into the CNS. We have recently found that i.v. injection of liposome-entrapped MTX into cynomolgus monkeys (Ma caca fascicularis) resulted in greatly increased plasma lev els and tissue uptake, and reduced renal excretion com pared with free [3H]MTX. Entrapment inside liposomes was also found to reduce the otherwise considerable metabolic degradation of [3H]MTX to non-MTX products found in these animals after i.v. injection (13), resulting in up to 103fold increases in plasma levels of the intact drug (12, 14). Administration of material into the CNS is preferably done by injection into the CSF. This is much easier to do in larger animals, and we therefore studied the distribution of free and liposome-entrapped [3H]MTX after injection into the lateral cerebral ventricles of cynomolgus monkeys. A pre liminary report on this work has appeared (15). MATERIALS AND METHODS Preparation of Liposome-encapsulated MTX. This was 1This investigation was supported by Grant CA 17516 awarded by the National Cancer Institute, Department of Health, Education and Welfare. Received May 31, 1977; accepted December 13. 1977. 2 The abbreviations used are: CNS, central nervous system; MTX, methotrexate; [3H]MTX, [3',5',9(n)-3H]methotrexate; CSF, cerebrospinal fluid; PC, phosphatidylcholine. 706 CANCER RESEARCH VOL. 38 on May 3, 2017. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from Free and Liposome-entrapped [3H]MTX in Primate CNS performed as previously described (12, 14) except that, in order to keep the volumes small for intracerebroventricular injection, free MIX was removed by 24-hr bulk dialysis. Liposomes consisted of 43 /¿molesof purified egg PC, 22 AmÃ3les of recrystallized cholesterol, and 10 ¿¿moles of stearylamine (1:0.52:0.22), with 5 /¿Ciof [14C]cholesterol sometimes added. Taking an approximate molecular weight for the egg PC of 800, the weight of lipids present was 34 mg of egg PC, 8.5 mg of cholesterol, and 2.7 mg of stearylamine for a total lipid weight of 45.2 mg. They were made up in a total volume of 1 ml and sonically disrupted for 1 to 1.5 hr to near clarity under N2, which has been shown to reduce their size to a range of 250to 500-Õ diameter (17). Furthermore, they did not centrifuge down at 100,000 x g for 1 hr, and fractionation on a Sepharose 4B column showed a clear separation between the sonically disrupted and nonsonically disrupted liposomes, as shown in Chart 1. This behavior further suggests that the sonically disrupted liposomes constitute a separate population of unito bilamellar vesicles with a size range of 250 to 500 A (11,18). As mentioned above the initial small volume of the lipo some suspensions were maintained by removing nonentrapped [3H]MTX by 24-hr bulk dialysis, and the average capture per 43 timÃ3les of added PC after dialysis for the sonically disrupted liposomes was 5.6% of the added MTX (100 /¿Ciof [3H]MTX and 15 to 18 mg of MTX). This is similar to the previous value of 0.10% captured per /¿mole of PC for the same liposomes found after more rapid removal of nonincorporated [3H]MTX by gel filtration (14). Long-term bulk dialysis is possible because these lipo somes show very low permeability to [3H]MTX (12). In this study it was found that after the initial 24-hr bulk dialysis, an average 5.4% leaked out during the first subsequent 1.5-hr dialysis period, dropping to 1.7% during the fourth 1.5-hr dialysis period. Preparation of Animals and Injection Procedures. Young adult male or female cynomolgus monkeys (M. fascicularis, 2.9 to 3.7 kg) were obtained commercially and used after a minimum of 1 week of conditioning. They were anesthetized by i.m. administration of phencyclidine hydrochloride (2.0 mg/kg; Sernylan; Bioceutic Laboratories, Inc., St. Joseph, Mo.), with appropriate supplemental doses given as needed. Anesthetic use of Sernylan was accompa nied by 0.1 mg of atropine sulfate. Femoral arterial and venous catheters and urinary catheters were inserted, blood pressure was recorded continuously from the femoral ar tery, and temperature recorded rectally and maintained at 37 to 38°as previously described (13,14). For intracerebroventricular injections, the animal was placed in a stereotaxic frame in the prone, head-up position (5, 13). Small, bilateral frontal craniectomies permitting placement of a needle in each lateral ventricle were per formed. For short-term experiments (<4 hr), a 20-gauge needle was positioned above both the right and left lateral ventricles in holders on calibrated bars attached to the stereotaxic unit according to the dimensions given by Fenstermacher (7). Access to the ventricular fluid was confirmed by appearance of CSF in the needle hub and by pulsation of a column of artificial CSF (13) in tubing at tached to the needle. The area around the needle was Jj s .UNSONICATED