- Email: [email protected]
Liposome-Encapsulated Ampicillin: Physicochemical and Antibacterial Properties
Liposome-Encapsulated Ampicillin: Physicochemical and Antibacterial Properties ILANA SCHUMACHER
AND
RIMONA MARGALITX
Received August 31, 1995, from the Department of Biochemistry, the George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel. Final revised manuscript received December 9, 1996. Accepted for publication December 17, 1996X. Abstract 0 The objectives of this study were to develop high-performance liquid chromatography (HPLC) and antibacterial assays for ampicillin encapsulated in multilamellar liposomes (MLV) and investigate the physicochemical and antibacterial properties of ampicillin-liposome systems. The major findings were fourfold. First, ammonium acetate (0.575%) in methanol:water (450:550; v/v), adjusted to pH 7.2, was suitable as the mobile phase for the HPLC determinations of ampicillin in both aqueous and liposomal systems. This mobile phase also provided (alone or with additional methanol) complete dissolution of liposomal assay samples in a precolumn treatment that made all of the encapsulated drug available for chromatography. Multiple samples were assayed without any technical limitations. Second, the growth-inhibition antibacterial assay developed, which used the USP test organism Micrococcus Luteus and paper disks, was quantitative for both free and liposome-encapsulated ampicillin. Third, the physicochemical properties of encapsulated ampicillin include encapsulation efficiencies of 10 to 50% for liposome concentrations in the range 10−200 mM (lipid), and a single rate constant to sufficiently and quantitatively describe the diffusion of encapsulated ampicillin, with half-lives in the range of 40 h. Fourth, the biological properties include the first direct evidence that encapsulated ampicillin retains full biological activity; that is, liposome-encapsulated ampicillin was active against extracellular bacterial colonies of Micrococcus luteus. Furthermore, encapsulation enhanced ampicillin stability. For example, free ampicillin in an aqueous solution that was stored for 5 weeks at 4 °C lost 50% of its initial activity, whereas liposome-encapsulated ampicillin (freed from unencapsulated drug) stored under the same conditions lost only 17% of its initial activity. The findings of this study, provide strong support for ampicillin-liposome formulations as valid dosage forms for this drug that are worthy of further experimental evaluations.
Introduction Ampicillin, a potent antibiotic with relatively short-termed stability in aqueous solutions,1-10 is used clinically for the treatment of a broad range of bacterial infections.1,2,11-17 Liposomes are under consideration as delivery systems for ampicillin,18-23 with the anticipation that the change in dosage form from free to liposome-encapsulated drug will result in significant improvements of clinical outcomes. For this change in clinical outcome to be justified, the liposomal formulations should meet the following criteria: (1) efficient drug encapsulation; (2) control over the rates of drug release to best fit the rate to the requirements of the therapy; (3) retention or improvement of drug stability; (4) retention, in the encapsulated state, of antibacterial activity; (5) retention of the broad spectrum activity against both intracellular and extracellular infections; and (6) targeting. Few of these criteria were addressed in previous studies.18-23 Little, for example, was provided in terms of physicochemical properties, and where such data were made available, it was not encouraging (e.g., encapsulation efficiencies as low as 4%). X
Abstract published in Advance ACS Abstracts, February 15, 1997.
© 1997, American Chemical Society and American Pharmaceutical Association
The ampicillin in the liposomal formulations showed antibacterial activity, but the reports did not specify the relative contributions of the encapsulated and unencapsulated drug to the measured activity. Testing was against bacteria residing in the reticuloendothelial system (RES) where passive targeting operates and drug release does not rely on diffusion alone. Therefore, we explored the potential use of ampicillinliposome formulations by assessing the ability of these formulations to meet the criteria just listed. The results of the first stage of such studies, conducted at the molecular and in vitro levels, are reported in this communication. The following issues are addressed: (1) the efficiency of encapsulation, the kinetics of drug release, and the effects of liposome concentration on both; (2) the activity of liposome-encapsulated ampicillin against extracellular-residing bacteria; and (3) the effects of liposome encapsulation on the stability and the retention of biological activity of the encapsulated ampicillin. These investigations required chemical, such as highperformance liquid chromatography (HPLC),18 and antibacterial2 assays that could be applied directly to the encapsulated ampicillin. The established HPLC protocols for ampicillin assay are for the free, not the encapsulated, drug. Furthermore, with respect to liposomal formulations, previous use of such an assay was restricted to the unencapsulated fraction alone, after its separation from the liposomes.18 An HPLC assay of the encapsulated fraction in any liposome-drug formulation would require sample pretreatment to release all the encapsulated drug to make it available for chromatographic analysis. Obviously, such pretreatment should not damage the drug nor should it clog the column or interfere with components of the device itself. The development of an HPLC assay for the direct determination of liposomeencapsulated ampicillin was an essential aspect of this investigation. Also reported here is a method for antibacterial assay of liposome-encapsulated ampicillin that is based on an established growth inhibition assay of free ampicillin, with the USP test organism Micrococcus luteus. This assay also served as a model system for testing whether the previously reported activity of liposomal ampicillin against bacteria residing in phagocytic cells,18-23 could be extended to extracellular bacterial colonies.
Materials and Methods MaterialssHigh-purity soybean phosphatidylcholine (PC) was from American Lecithin Company (Atlanta, GA) and high-purity cholesterol (CH) was from Sigma Chemical Company (St. Louis, MO). Ampicillin was a gift from Baxter Healthcare Company [3H]Cholesterol ([3H]CH) was from Amersham (Arlington Heights, IL). Blood agar base, trypton, agar-agar, and Muiller-Hunton broth were from Difco Laboratories (Detroit, MI). Spectra/Por4 dialysis tubings (molecular weight cutoff of 12 000 to 14 000) were from Spectrum Medical Industries (Los Angeles, CA). Lyophilized Micrococcus luteus (ATCC 9341) was a kind gift from Teva Pharmaceuticals (Israel). All other reagents were of analytical and, where relevant, HPLC grade. High-speed centrifugation was performed with a Sorvall RC-58 centrifuge. Liquid scintillation counting was performed with a
S0022-3549(95)00369-8 CCC: $14.00
Journal of Pharmaceutical Sciences / 635 Vol. 86, No. 5, May 1997
Kontron Analytical Betamatic I. The HPLC was conducted with a Hewlett-Packard Company (Palo Alto, CA) chromatograph was from (model HP-1090L), with a Waters µBondapack C-18, 3.9 × 300 mm column. Ampicillin was detected spectrophotometrically with a UVDiode array spectrophotometer (Hewlett-Packard, model 1090L). Liposome Preparation and Drug Encapsulationsmultilamellar liposomes (multilamellar vesicles, MLV) were prepared as previously described24-29 from lipid compositions of PC:CH at 3:1 mole ratios. Briefly, the lipids were dissolved in a methanol:chloroform (1:2, v/v) solution in a round-bottomed flask, and the organic solvent was evaporated to dryness under low pressure in a rotary evaporator. The swelling solution was composed of ampicillin dissolved in pH 7.2 phosphate-buffered saline (PBS) at concentrations ranging from 4 to 20 mg/mL. The swelling solution was added to the dry lipid film, and the preparation was agitated vigorously (with a vortex) for several minutes and then incubated in a shaker bath for 2 h at 37 °C. Separation of the unencapsulated drug from the liposome fraction (containing the encapsulated fraction) was by high-speed centrifugation for 45 min at 4 °C and 27 000g. “Empty”, drug-free liposomes were prepared in a similar manner, except their swelling solution was buffer alone. Drug ReleasesThe kinetics of drug release were studied according to previously reported procedures.24-29 Briefly, a suspension of liposomes (1.0 mL) was placed in a dialysis sac that was immersed in a constantly stirred receiver vessel containing 10 mL of drug-free buffer (PBS). At designated periods, the dialysis sac was moved from one receiver vessel to another containing fresh (i.e., drug-free) buffer. Drug concentration was determined by HPLC in each dialysate and in the sac (at the beginning and end of each experiment). These data were used to calculate the cumulative fraction of released drug (normalized to the total amount of ampicillin in the preparation), which is denoted “f” for each time point. Efficiency of Drug EncapsulationsThe efficiency of drug encapsulation, defined as the fraction of the total drug in the system encapsulated within the liposomes, was determined by centrifugation24-29 and by processing data of kinetic studies. Samples of the liposome preparation were subjected to high-speed centrifugation. The supernatant containing the unencapsulated drug was isolated, and the pellet containing the liposomes and their encapsulated drug was resuspended in drug-free buffer. The ampicillin content was assayed by HPLC and/or by the antibacterial activity assay in the complete preparation (precentrifugation), in the separated supernatant and in the resuspended pellets. These data were then used to calculate the percent encapsulation. The fraction of ampicillin encapsulated in the liposomes at timezero was obtained by analysis of the kinetics of drug release. For cases in which these experiments were performed on complete liposomedrug preparations, the magnitude of this encapsulated fraction corresponds to the efficiency of encapsulation. Ampicillin Stability as Function of TimesImmediately following preparation, liposome-ampicillin samples were separated by centrifugation as previously described. Aliquots from the swelling solution, the complete liposome preparation, the supernatant, and the resuspended pellets were analyzed for their antibacterial activity. Samples from the swelling solutions and the resuspended pellets were stored at 4 °C until they were reanalyzed for their antibacterial activity. These analyses were completed within 3 to 8 weeks. Quantitative DeterminationssLiposomessThe molecular weight of liposomes, especially MLV, cannot be determined accurately. However, for a given liposome species (type and lipid composition), the lipid concentration, which can be determined accurately, is proportional to the liposome concentration. Therefore, liposome concentrations are given in this communication in terms of mM lipid, with the understanding that the relevant parameter is the liposome concentration and the lipid concentration is the units used to define it. Quantitative determinations of lipid concentrations were performed by inclusion of a trace amount of [3H]CH in the formulation. HPLC Assay of AmpicillinsThe mobile phase was composed of 0.575% ammonium acetate in methanol:water (450:550, v/v), adjusted to pH 7.2. The ampicillin standard solutions covered the range 5-100 µg/mL. The calibration curves were linear, with r2 values in the range 0.93-1.0. For assays of liposome-encapsulated ampicillin, the liposome swelling solution served as the stock solution for the HPLC standards and was diluted (to provide the desired ampicillin concentrations) directly into the mobile phase.
636 / Journal of Pharmaceutical Sciences Vol. 86, No. 5, May 1997
Liposomal HPLC samples were pretreated in two ways. Samples from the complete liposome preparation (0.5 mL of liposome suspension) were transferred into a 20-mL volumetric flask, dissolved by the addition of 10 mL methanol, and made up to the mark by the addition of mobile phase. The solution was mixed for 5 min and then filtered through a 0.22-µm disposable filter. Samples after liposome centrifugation (2-mL) aliquots of the stock solution) were centrifuged and separated into supernatant and pellet. The pellet was resuspended in buffer and prepared as just described for complete liposome preparation samples. A 0.5-mL aliquot from the supernatant was transferred into a 20-mL volumetric flask and made up to the mark with mobile phase. The solution was mixed and filtered through a 0.22-µm disposable filter. The HPLC standard and sample injection volumes were 50-100 µL, the column was operated at a pressure of 100 bars, the flow rate of the mobile phase was 0.8-1.0 mL/min, and detection was at 230 nm. Triplicate samples were taken from each injected material and the area under the peak was determined for each run. The assays for each experimental system, including standards and controls, were all run on the same day. The typical total number of daily injections was 35-40, and the assays were run without encountering any technical problems. Antibacterial (Growth-Inhibition) Assay of AmpicillinsThe preparation of solutions and materials, as well as the assay itself, were performed under aseptic conditions. The assay was performed essentially according to the cylinder plate method.2 with paper disks instead of metal cylinders30 and the following modifications: (a) Petri dishes containing the solid agar were prepared in advance and stored at 4 °C until use. (b) Several days prior to performing the assay, a bacterial colony from the bacterial slant was seeded onto a previously prepared blood base agar (BBA) medium and incubated for 24 h at 37 °C. Several selected colonies were transferred into 20 mL of Mueller Hinton medium (MHM) and then further incubated in a shaker bath for 24 h at 37 °C. The suspension was then divided into several sterile tubes and centrifuged (with a bench centrifuge) for 10 min at 4 °C and 3000 rpm. The pellets were pooled and resuspended. The suspension was diluted to a T value (percent transmission) of 10% (at 580 nm), and used for the preparation of the bacteria-in-topagar component. Alternatively, the bacterial pellets were suspended in 2 mL of saline and stored at 4 °C (for no longer than a month). (c) The top agar was also prepared in advance and stored at room temperature until use. On the first day of the assay, the top agar was liquified (in a microwave oven or by autoclaving for 30 min at a pressure of 1.2 atm) and then incubated for 1 h in a water bath at 45 °C. For each petri dish, 2.5 mL of the liquified top agar were mixed with 0.7 mL of the bacterial suspension (see modification b, vortexed, and poured into the dishes discussed in modification a. The petri dishes were dried for 2 h prior to the placement of the disks and the beginning of the assay. (d) The paper disks were sterilized by autoclaving for 30 min at 1.2 atm. The sterilized disks were stored at room temperature until use. Four disks were placed into each petri dish and four dishes were used for each ampicillin dose. Twenty microliters of each standard or unknown were added in a dropwise manner to each disk, after which the plates were allowed to dry for 2.5 h and then incubated for 24 h at 37 °C. The diameters of the growth inhibition zones were measured for each disk with a caliber accurate to 0.01 mm. Controls included top agar without bacteria, disks to which no samples were added, PBS (added to petri dishes with and without disks) and “empty liposomes”. Controls were tested under the same conditions just described. Calibration curves were constructed for each experiment with the standards and controls. A fresh ampicillin solution (4 mg/mL) served as the stock solution for the calibration standards. The swelling solutions were used as the stock solutions to obtain the calibration curves for the liposomeencapsulated ampicillin assays. The ampicillin standards were prepared by dilution from the stock solutions, with concentrations ranging from 0.01 to 0.5 µg/mL. Liposomal samples pipetted onto the disks were diluted from the appropriate stocks, with dilution factors ranging from 1 × 104 to 2 × 103.
Results and Discussion HPLC Assay of Liposomal AmpicillinsThe ampicillin HPLC assay requires that the liposomes be dissolved in the
Table 1sLipid Recovery in Centrifuge-Separated Ampicillin-Liposome Preparations
Table 2sHPLC Determinations of Ampicillin in Liposomal Formulations
Total Lipid Concentration of the Preparation (pre-centrifugation) (mM)
Lipid Recovery in the Pelleta (% from total)
Batch Code
Direct Determination Performed on Liposomal Samplesa (% from initial ampicillin)c
Indirect Determination Performed on Free Ampicillinb (% from initial ampicillin)
10 20 30 45 60 200
93 ± 0.8 97 ± 1.2 98 ± 3.9 95 ± 0.4 99 ± 2.8 102 ± 3.5
IS-15 IS-23 IS-24
116.1 ± 4.7d 93.6 ± 9.1 80.0 ± 8.6
111.4 ± 8.7 85.2 ± 1.3 85.4 ± 0.5
a Each data is the average of three independent separations (± standard deviation).
assay sample to ensure all of the encapsulated drug was available for analysis. Detergents, such as Triton X-100, were unsatisfactory in dissolving the liposomes because they subsequently clogged the column. Attempts to dissolve the liposomes with the mobile phase formulation customary for free ampicillin2,18 resulted in incomplete dissolution of the liposomes. However, an alternative mobile phase formulation composed of 0.575% ammonium acetate in methanol:water (450:550, v/v), adjusted to pH 7.2, effectively dissolved the liposomes, freeing the ampicillin for analysis. HPLC chromatograms of free ampicillin gave a single sharp peak with a retention time of 4 min that was identified by UV spectra as ampicillin. Assay samples containing only the encapsulated drug were obtained by centrifuging aliquots of ampicillin-liposome preparations as previously described, separating the pellet from the supernatant, and resuspending the pellet in drug-free buffer. The separation was quantitative, as evidenced from the lipid contents of the pellet and the supernatant shown in Table 1. The following HPLC chromatograms were similar to each other and to those of free ampicillin: complete drug-liposome preparations (i.e., containing encapsulated and unencapsulated drug), liposomes containing encapsulated drug alone, and the (separated) unencapsulated fraction of drug-liposome preparations. There were no peaks in chromatograms of control systems; that is, water, PBS, or pretreated drug-free (i.e., empty) liposomes. These data indicate that the pretreatment with the modified mobile phase did not interfere with the ability to assay the ampicillin. By visual observation, the pretreatment seemed to provide complete liposome dissolution, but it was necessary to determine quantitatively that all the encapsulated drug was released and assayed. Therefore, the HPLC assay of a given liposome-ampicillin system was performed by two methods: (1) aliquots from the complete liposome preparation were assayed for their ampicillin concentration, and (2) aliquots from the liposome preparation were subjected to exhaustive dialysis over a period of at least 24 h with frequent replacements of medium, as described under methods for kinetics of drug release. The dialysis was terminated following complete depletion as verified experimentally. The dialysates containing only free ampicillin in aqueous solutions were individually subjected to HPLC analysis. Therefore, the cumulative determinations of ampicillin in the dialysates should contain all of the drug that was originally in the liposomal sample (encapsulated and unencapsulated). If the pretreatment devised for the liposomal suspensions effectively enables all of the drug to be available for chromatographic analysis, then there should be agreement between the results from the two independent determinations for the same liposomal preparation. This expectation was confirmed, as shown in Table 2, for three ampicillin-liposome batches. The percent ampicillin determined by each method was within experimental error for each batch.
a Results of HPLC determination of liposomal samples subjected to the pretreatment. b Results of HPLC determination of ampicillin dialysates obtained from complete depletion of liposome-ampicillin treatment in the course of exhaustive dialysis. c The initial quantity of ampicillin put into a given liposome preparation served as the 100% value. Values over and under 100% indicate water or material losses in the course of liposome preparation, respectively. d Each value in the table is the average and standard deviation of duplicate samples.
Antibacterial Assay of Liposomal AmpicillinsThe assay for liposome-encapsulated ampicillin had to meet three criteria: (1) the encapsulated drug has to retain its biological activity; (2) all of the drug in the liposomes had to become available for the assay, which required complete release of all the encapsulated drug within the time course of the assay; and (3) the assay has to be quantitative. Typical results of the antibacterial assays for the liposome-encapsulated ampicillin are shown in Figure 1. Dishes A and B contain several controls, including empty disks, disks plated with PBS, and disks plated with empty (i.e., drug-free) liposomes suspended in PBS. As expected, there are no growth inhibition zones for these controls. In contrast, growth inhibition zones were usually observed with ampicillin formulations (Figure 1, dishes C-F). Dish C is a typical response to an ampicillin standard solution. Dish D is characteristic of the responses for the complete liposome preparation, and dishes E and F are characteristic of the responses to the unencapsulated and encapsulated fractions of the separated liposome preparations, respectively. The ampicillin originating from liposomal systems, including the encapsulated drug, was biologically active, as evidenced by the presence of growth inhibition zones. Within the concentration range 0.15-0.5 µg/mL, the ampicillin calibration curves were linear, with r2 values in the range 0.96-0.98. To be considered a valid bioassay, all the liposomal ampicillin had to become available to the bacteria during the course of the assay without significant loss in sample activity. To determine the feasibility of a given liposome-ampicillin preparation to meet these requirements, the following components were assayed: (1) the swelling solution; (2) the complete liposome preparation; (3) the supernatant fraction of a separated drug-liposome preparation; and (4) the resuspended pellet fraction of a separated drug-liposome preparation. The data from these bioassays were used to calculate the total quantity of active ampicillin in each of the four systems. Results for duplicate liposome-ampicillin preparations are shown in Figure 2. Initially, 40 mg of active ampicillin were incorporated into each liposome preparation through the swelling solution. Bars A and B represent the amounts of active ampicillin in the swelling solution and in the complete liposome preparation respectively. The general agreement between the two independent determinations suggest there was no significant drug loss during the course of liposome preparation, and that ampicillin activity was not compromised. Bars C and D represent, respectively, the unencapsulated and encapsulated active ampicillin of the liposome preparation. The sums of the C+D bars represent another evaluation of the total amount of active ampicillin in the liposome preparation. The general agreement of the three independent determinations of total active ampicillin (bars A, B, and C+D) suggests that the encapsulated drug retains its biological activity. Furthermore, the assay demonstrates
Journal of Pharmaceutical Sciences / 637 Vol. 86, No. 5, May 1997
Figure 1sAntibacterial assay of free and of liposome-encapsulated ampicillin by the growth-inhibition assay (raw data): (A) no treatment (left-hand side) or PBS alone (right-hand side); (B) “empty” (i.e., drug-free) liposomes; (C) ampicillin standard solution (in PBS); (D) complete ampicillin-liposome systems (i.e., containing encapsulated and unencapsulated drug); (E) unencapsulated ampicillin only (i.e., the supernatant, postseparation of systems, such as in D, by centrifugation); and (F) liposomeencapsulated ampicillin only (i.e., the resuspended pellet, postseparation of systems, such as in D, by centrifugation).
Figure 2sEvaluation of the antibacterial assay for two similar ampicillin-liposome preparations coded L-31 and L-29: (A) swelling solution; (B) complete ampicillinliposome preparation; (C) the unencapsulated fraction of preparation after separation by centrifugation; (D) The encapsulated fraction of the preparation after separation by centrifugation; and (C+D) the sum of the (separate) fractions C and D. Each column is an average of 16 repeat samples, and the error bars represent the standard deviations.
that the encapsulation efficiency was ≈50%. Essentially all of the encapsulated ampicillin became available to the bacteria during the course of the assay, as evidenced by the similarities between bars A, B, and C+D. These similarities may be attributed to drug diffusion from intact liposomes under the extremely low liposome concentrations being assayed. The general phenomena of an inverse relationship between the kinetics of drug release and liposome concentration24,25,29 indicates that complete and fast depletion of the liposomes can take place under the conditions of the assay, where the liposome concentrations in the samples plated onto the disks are extremely low, with dilution factors ranging from 1 × 104 638 / Journal of Pharmaceutical Sciences Vol. 86, No. 5, May 1997
Figure 3sThe efficiency of ampicillin encapsulation in liposomes (MLV) as a function of liposome concentration (expressed in lipid molar units). The data points labeled HPLC and Bioassay were determined from centrifuge-separated ampicillinliposome preparations by HPLC or the antibacterial assay, respectively. The data points labeled “derived from kinetic data” were determined from kinetic experiments by HPLC as the method of quantitative determination. The points are the experimental data, and the linear line is drawn to emphasize the increasing trend. Each point on the figure represents an individual preparation, and the preparationto-preparation reproducibility can be appreciated from the encapsulation efficiencies obtained for the individual preparations at the same lipid concentration.
to 2 × 103. Finally, the encapsulated ampicillin was active against bacterial colonies residing outside cells. Efficiency of Ampicillin EncapsulationsTheoretical considerations24 supported by experimental data24-29 have shown that increased liposome concentrations result in increased encapsulation efficiencies. Typical ampicillin encapsulation efficiencies as a function of liposome concentration (Figure 3) support this trend. The encapsulation efficiencies increase from <10% to ≈50% as the liposome concentrations increase from 20 to 200 mM lipid. Encapsulation efficiencies
at high liposome concentrations are an improvement over the previously reported efficiencies.21 Kinetics of Drug ReleasesRegardless of the route of administration and of the anatomic destination of the drug, there should be minimal free drug surrounding the liposomes in vivo, despite continuous drug efflux from the liposomes, because of the cumulative effect of dilution of both liposomes and released drug, drug binding at the site of action, drug binding at sites other than its site of action, drug metabolism, and drug clearance. Therefore, there should be an electrochemical gradient driving the drug out from the liposomes. The experimental design for the kinetic studies was made to model these flux conditions.24 Previous studies have shown that the magnitude of the rate constant for encapsulated drug efflux decreases with increasing liposome concentrations.24,25,28 This relationship has been attributed to deviations from ideality.24 Therefore, the kinetic studies were undertaken with liposome preparations at low lipid concentrations to increase experimental sensitivity to diffusion parameters. The data were analyzed according to a previously derived and experimentally supported mechanism,24,25,27-29 that is based on Eyring's theory. Initially, at timezero, the system is assumed to contain encapsulated and unencapsulated ampicillin. Diffusion of encapsulated and unencapsulated ampicillin likely occurs as multistep processes, governed by a single rate-limiting step. Therefore, the overall mechanism can be described as a series of parallel first-order processes with a common product and different reactants.24,29 For unencapsulated and encapsulated ampicillin, the relationship between f (the cumulative amount of ampicillin released at time t normalized to the total amount in the preparation at time ) 0) and time t is given by eq 1: -kft
f ) ff(1 - e
-kft
) + fl(1 - e
)
(1)
The fractional distribution of ampicillin at timezero is represented by the fraction unencapsulated, ff, and the fraction encapsulated, fl, where ff + fl ) 1. The corresponding rate constants are kf and kl for the unencapsulated and encapsulated fractions, respectively. In those cases where there was no separation from unencapsulated drug prior to the experiment, fl also represents the fraction of encapsulated ampicillin. Results obtained for an ampicillin-liposome preparation of 30 mM lipid are shown in Figure 4. Complete liposome depletion occurred within 25 h. The magnitudes of fl, kf, and kl, obtained for liposome preparations of 30 and 45 mM lipid, are given in Table 3. As expected, the fraction of encapsulated drug is rather low, and the rate constant for the encapsulated fraction decreases with increasing liposome concentrations. The rate constant for the release of encapsulated ampicillin corresponds to a half-life in the range of 40 h, which implies that ampicillin-liposome systems have the potential for use as sustained-release drug depots. Stability of Liposome-Encapsulated AmpicillinsThe stability of the encapsulated ampicillin as a function of time and of the encapsulation efficiency was investigated by assessing the loss of ampicillin activity. Fresh preparations of ampicillin-encapsulating liposomes were separated into encapsulated and unencapsulated fractions by centrifugation. The liposomal pellets containing only encapsulated drug were resuspended in drug-free buffer and stored at 4 °C. Samples of the swelling solutions that contained free aqueous ampicillin and were used in these preparations were stored under similar conditions. At designated time periods, aliquots were withdrawn from the liposome suspension and from the corresponding swelling solution and analyzed for active ampicillin, by the antibacterial assay. The aliquots withdrawn from the stored liposomes may contain encapsulated and
Figure 4sThe kinetics of ampicillin release from MLV. The points are experimental data, which were averages of duplicate runs, with the error bars indicating the standard deviations. Quantitative determination was by HPLC of triplicate samples for each time point in each run. The curve was calculated from eq 1 with the parameters listed in Table 3 for the preparation L-24. Table 3sKinetic Parametersa of Ampicillin Release from Liposomes (MLV) BatchCode
[LIPID](mM)
fl(%)
kf(h-1)
kl(h-1)
L-23 L-24
30 45
6.0 ± 0.7 8.6 ± 0.4
0.51 ± 0.006 0.57 ± 0.005
0.015 ± 0.009 0.011 ± 0.004
a Triplicate quantitative determinations of the released drug were made for each time point, and duplicate runs were made from each preparation. The average of the duplicate runs and the triplicate determinations are shown in Figure 4, together with the standard deviations (SDs). The average values were obtained by nonlinear regression analysis according to eq 1 in the text. The results of these analyses are listed in this table, and the SDs reported here, provided by the analysis process, illustrate the good fit of the experimental data to the proposed mechanism.
unencapsulated ampicillin even though unencapsulated ampicillin was removed prior to storage. Note that the objective of this study was to determine ampicillin activity as related to time and encapsulation efficiency, not to evaluate aspects of shelf-life. Lyophilized powders are the rational dosage form for the long-term storage required for pharmaceutical products.27 The results shown in Figure 5 are for three different liposome preparations and four test periods. One preparation was tested at two different times. The percent of encapsulation in the original preparation is shown by the line graph and the losses of antibacterial activity with time are presented as bars. Suspension I-28 was prepared at the concentration of 100 mM lipid, which yielded an encapsulation efficiency of 18%. The stability of the stored samples from this batch and from its corresponding swelling solution was assessed 3 and 8 weeks post preparation. Suspensions L-29 and L-31 were duplicate batches prepared at 200 mM lipid concentrations, which yielded encapsulation efficiencies approaching 50%. The stability of those liposome suspensions and their corresponding swelling solutions was assessed 5 weeks postpreparation. The loss in ampicillin activity of the swelling solutions should be independent of the liposome suspensions prepared with these solutions. The losses in antibacterial activity (Figure 5) were significant for the swelling solution associated with suspension I-28 at 3 and 8 weeks, and the swelling solution associated with suspension of I-29 at 5 weeks. The losses corresponded to 40-60% of the activity measured
Journal of Pharmaceutical Sciences / 639 Vol. 86, No. 5, May 1997
Figure 5sThe effects of encapsulation on the stability of ampicillin with time, as determined by the antibacterial activity of ampicillin. The codes I-28, I-29 and I-31 designate three different batches at initial liposome concentrations of 100, 200, and 200 mM lipid, respectively. The line graph is the efficiencies of drug encapsulation in the original preparations. The points are the experimental data and are connected by a line to emphasize the trend. The bars represent the loss in antibacterial activity, at specified test periods, where 100% activity was assigned to the values measured in the fresh swelling solution or resuspended liposome pellet (i.e., immediately upon system preparation). The specific test periods (in weeks) are given in parenthesis next to the preparation codes. The lightly shaded bars are the data for the swelling solution and the darkly shaded bars are the data for the liposomes. Each bar is an average of 12−16 assay samples, and the magnitudes of the standard deviations (number of samples) are: for data set I-28 (3), 0.62 and 0.50 for the swelling solution and the pellet, respectively; for data set I-28 (8), 0.45 and 0.45 for the swelling solution and the pellet, respectively; for data set I-29 (5), 0.47 and 0.45 for the swelling solution and the pellet, respectively; and for data set I-31 (5), 0.39 and 0.29 for the swelling solution and the pellet, respectively.
immediately after preparation in the fresh systems. However, the corresponding loss in the ampicillin activity in the swelling solution associated with suspension I-31 was lower. Although we have no explanation for this observation, we stress that our major interest is drug stability in the liposomal formulations. As shown in Figure 5 for both test periods, the losses in antibacterial activity of ampicillin encapsulated and stored in the liposome suspension I-28 were similar to those of observed for (free) ampicillin in the corresponding swelling solution. In contrast, regardless of some preparation-topreparation variabilities, the loss in ampicillin activity was significantly lower for ampicillin encapsulated and stored in the liposomal formulations I-29 (17% loss) and I-31 (6% loss). As expected, ampicillin stability in aqueous media was time limited. Loss in activity, although significant, was not linear with time and plateaued over 8 weeks (compare the data for I-28 at 3 and 8 weeks of storage). Encapsulation within liposomes did not increase the loss in activity (liposomes of I-28 versus swelling solutions) but may have actually decreased the loss in ampicillin activity (liposomes of I-29 and I-31 versus swelling solutions). However, further studies are necessary to determine whether the improvement in stability (as suggested by preparations I-29 and I-31 but not by preparation I-28) may be associated with the higher encapsulation efficiencies (≈50% versus 20%), the higher liposome concentration (200 versus 100 mM lipid), or both.
Conclusions Development of HPLC and Antibacterial Assays for Encapsulated AmpicillinsThe simple pretreatment of the liposome suspensions with the modified mobile phase released all the encapsulated ampicillin making it available for quan640 / Journal of Pharmaceutical Sciences Vol. 86, No. 5, May 1997
titative assays. The analysis could be performed continuously for multiple samples. The pretreatment with the modified mobile phase did not result in significant ampicillin degradation and/or inactivation, in blockage of the HPLC column, or any interference with other components. These results suggest that this HPLC assay could be an effective method to quantify liposome-encapsulated ampicillin. The antibacterial assay demonstrated that the liposome-encapsulated ampicillin retained its biological activity, indicating that the encapsulation procedures for preparation of liposome suspensions did not result in any detectable drug destabilization. The results for both unencapsulated and liposome-encapsulated ampicillin show that the use of paper disks is an accurate and useful method for the zone-inhibition assay that could replace the use of metal cylinders. The HPLC method and the bioassay have the potential to be further developed as analytical assays for liposomeencapsulated ampicillin, as ampicillin-liposome formulations will mature into pharmaceutical products. Furthermore, both assays could be extended, with low to moderate investments in additional efforts, to assay other antibiotics encapsulated in liposomes. Liposome Encapsulated AmpicillinsAmpicillin was encapsulated into MLV suspensions as a function of lipid concentration, and then the release and activity of the ampicillin were determined. Encapsulation efficiencies increased as the lipid concentration of the suspension increased. Encapsulation efficiencies ranged from ≈50% with 200 mM lipid concentrations to <20% at lipid concentrations of <20 mM. The lower encapsulation efficiencies were still greater than those previously reported.21 The rate constant for released ampicillin varied with liposome concentration. A half-life in the range of 40 h was achieved with liposome concentrations on the order of 30-45 mM lipid. The ampicillin encapsulated within the liposomes was active against bacterial colonies without the involvement of other cells. Taken together with the previously reported activity of liposome-ampicillin dosage forms against intracellular organisms,18-23 this result suggests that the broad activity spectrum of ampicillin is not decreased when the drug is encapsulated within the liposomes. In addition, encapsulation within liposomes does not compromise ampicillin stability but may actually enhance it. Therefore, ampicillin-liposome suspensions merit further consideration for development into pharmaceutically relevant products.
Abbreviations Used BBA, blood base agar; CH, cholesterol; MHM, Mueller Hinton medium; MLV, multilamellar vesicles; PBS, phosphatebuffered saline; PC, phosphatidyl choline.
References and Notes 1. Remington's Pharmaceutical Sciences,17th ed., Osol, A., Ed.; Mack: Easton, PA, 1985; pp 1194. 2. USP XXIII-NF18 , U. S. Pharmacopeial Convention, Inc.: Mack: Easton, PA, 1995; pp 107, 1681-1696. 3. Nguyen, N. T.; Mortada, L. M.; Notary, R. E. Pharm. Res. 1988, 5, 288-296. 4. Bundgaard, H.; Larsen, C. J. Chromatogr. 1977, 132, 51-59. 5. Larsen, C.; Bundgaard, H. J. Chromatogr. 1978, 147, 143-150. 6. Bundgaard, H. Pharm. Pharmacol. 1974, 26, 385-392 . 7. Van Der Bijl, P.; Seifart, H. I.; Parkin, D. P.; Mattheyse, F. J. S. Afr. Med. J. 1988, 73, 453-455. 8. Aki, H.; Sawai, N.; Yamamoto, K.; Yamamoto, M. Pharm. Res. 1991, 8, 119-122. 9. Oliyai, R.; Lindenbaum, S. Int. J. Pharm. 1991, 73, 33-36. 10. Kong, K. M. Am. J. Hosp. Pharm. 1988, 45, 314.
11. Hartzen, S. H.; Frimodt-Moller, N.; Andreasen, J. APMIS 1988, 96, 584-588. 12. Dukin, K. T.; Jones, S.; Howard, A. J. J. Antimicrob. Chemother. 1988, 21, 405-411. 13. Jephcott, A. E.; Gough, K. J. Antimicrob. Chemother. 1988, 21 (suppl. B), 43-48. 14. Martindale, The Extra Pharmacopoeia, 30 ed.; Reynolds, J. E. F., Ed.; The Pharmaceutical Press: London, UK, 1993; pp 116118. 15. Principles and Practice of Infectious Diseases, 3rd ed.; Mandell, G. L.; Dougles, R. C.; Bennett, J. I., Eds.; Churchill Livingstone, Edinburgh, 1990; pp 240-242. 16. Nelson, J. D.; Haltalin, K. C. J. Infect. Dis. 1974, 129, 222227. 17. McCracken, G. H. In Clinical Reviews in Pediatric Infectious Disease; Nelson, J. D.; McCracken, G. H., Eds.; B.C. Decker: Toronto, 1984; pp 73-77. 18. Youssef, M.; Fattal, E.; Alonso, M. J.; Roblot-Treupel, L.; Sazieres, J.; Tancrede, C.; Omnes, A.; Couvreur, P.; Andremont, A. Antimicrob. Agents Chemother. 1988, 32, 1204-1207. 19. Couvreur, P.; Fattal, E.; Andermont, A. Pharm. Res. 1991, l8, 1079-1086. 20. Bakker-Wuodenberg, I.; Lokerse, A. F.; Vink van den Berg, J. C.; Roedink, F. H.; Michel, M. F. Antimicrob. Agents Chemother. 1986, 30, 295-300. 21. Fattal, E.; Rojas, J.; Yousseff, M.; Couvreur, P.; Andremont, A. Antimicrob. Agents Chemother. 1991, 35, 770-772. 22. Bakker-Wuodenberg, I.; Lokerse, A. F.; Roedink, F. H. J. Pharmacol. Exp. Ther. 1989, 251, 321-327. 23. Bakker-Wuodenberg, I.; Lokerse, A. F.; Roerdink, F. H.; Regts, D.; Michel, M. F. J. Infect. Dis. 1985, 151, 917-924.
24. Margalit, R.; Alon, R.; Linenberg, M.; Rubin, I.; Roseman, T. J.; Wood, R. W. J. Controlled Release 1991, 17, 285-296. 25. Yerushalmi, N.; Margalit, R. Biochim. Biophys. Acta 1994, 1189, 13-20. 26. Margalit, R. in Vesicles, Surfactant Science Series; Rossof, M., Ed.; Marcel Dekker: New York 1995; Chapter 13, pp 527-560. 27. Margalit, R.; Yerushalmi, N. In Microencapsulation Methods and Industrial Applications; Benita, S., Ed.; Marcel Dekker: New York, 1995; Chapter 10, pp 259-295. 28. Lichtenstein, A.; Margalit, R. J. Inorg. Biochem. 1995, 60, 187198. 29. Margalit, R. in Microparticulates - Preparation, Characterization and Application in Medicine; Cohen, S.; Berenstein, H., Eds.; Marcel Dekker: New York, 1996; Chapter 15, pp 425-461. 30. Akimoto, Y.; Mochizuki, Y.; Uda, A.; Omata, H.; Shibutani, J.; Nishimura, H.; Komiya, M.; Kaneko, S.; Kobayashi, S.; Kuboyama, N.; Yamane, J.; Furukawa, Y.; Fujii, A. J. Nihon Univ. Sch. Dent. 1990, 32, 4-18. 31. Margalit, R.; Okon, M.; Yerushalmi, N.; Avidor, E. J. Controlled Release 1992, 19, 275-288. 32. Yerushalmi, N.; Arad, A.; Margalit, R. Arch. Biochem. Biophys. 1994, 313, 267-273. 33. Margalit, R. Crit. Rev. Ther. Drug Carriers Systems 1995, 12, 233-261.
Acknowledgments This work was supported by a research grant to R. M. from Baxter Healthcare Company, Round Lake, IL.
JS9503690
Journal of Pharmaceutical Sciences / 641 Vol. 86, No. 5, May 1997
AND
RIMONA MARGALITX
Received August 31, 1995, from the Department of Biochemistry, the George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv 69978, Israel. Final revised manuscript received December 9, 1996. Accepted for publication December 17, 1996X. Abstract 0 The objectives of this study were to develop high-performance liquid chromatography (HPLC) and antibacterial assays for ampicillin encapsulated in multilamellar liposomes (MLV) and investigate the physicochemical and antibacterial properties of ampicillin-liposome systems. The major findings were fourfold. First, ammonium acetate (0.575%) in methanol:water (450:550; v/v), adjusted to pH 7.2, was suitable as the mobile phase for the HPLC determinations of ampicillin in both aqueous and liposomal systems. This mobile phase also provided (alone or with additional methanol) complete dissolution of liposomal assay samples in a precolumn treatment that made all of the encapsulated drug available for chromatography. Multiple samples were assayed without any technical limitations. Second, the growth-inhibition antibacterial assay developed, which used the USP test organism Micrococcus Luteus and paper disks, was quantitative for both free and liposome-encapsulated ampicillin. Third, the physicochemical properties of encapsulated ampicillin include encapsulation efficiencies of 10 to 50% for liposome concentrations in the range 10−200 mM (lipid), and a single rate constant to sufficiently and quantitatively describe the diffusion of encapsulated ampicillin, with half-lives in the range of 40 h. Fourth, the biological properties include the first direct evidence that encapsulated ampicillin retains full biological activity; that is, liposome-encapsulated ampicillin was active against extracellular bacterial colonies of Micrococcus luteus. Furthermore, encapsulation enhanced ampicillin stability. For example, free ampicillin in an aqueous solution that was stored for 5 weeks at 4 °C lost 50% of its initial activity, whereas liposome-encapsulated ampicillin (freed from unencapsulated drug) stored under the same conditions lost only 17% of its initial activity. The findings of this study, provide strong support for ampicillin-liposome formulations as valid dosage forms for this drug that are worthy of further experimental evaluations.
Introduction Ampicillin, a potent antibiotic with relatively short-termed stability in aqueous solutions,1-10 is used clinically for the treatment of a broad range of bacterial infections.1,2,11-17 Liposomes are under consideration as delivery systems for ampicillin,18-23 with the anticipation that the change in dosage form from free to liposome-encapsulated drug will result in significant improvements of clinical outcomes. For this change in clinical outcome to be justified, the liposomal formulations should meet the following criteria: (1) efficient drug encapsulation; (2) control over the rates of drug release to best fit the rate to the requirements of the therapy; (3) retention or improvement of drug stability; (4) retention, in the encapsulated state, of antibacterial activity; (5) retention of the broad spectrum activity against both intracellular and extracellular infections; and (6) targeting. Few of these criteria were addressed in previous studies.18-23 Little, for example, was provided in terms of physicochemical properties, and where such data were made available, it was not encouraging (e.g., encapsulation efficiencies as low as 4%). X
Abstract published in Advance ACS Abstracts, February 15, 1997.
© 1997, American Chemical Society and American Pharmaceutical Association
The ampicillin in the liposomal formulations showed antibacterial activity, but the reports did not specify the relative contributions of the encapsulated and unencapsulated drug to the measured activity. Testing was against bacteria residing in the reticuloendothelial system (RES) where passive targeting operates and drug release does not rely on diffusion alone. Therefore, we explored the potential use of ampicillinliposome formulations by assessing the ability of these formulations to meet the criteria just listed. The results of the first stage of such studies, conducted at the molecular and in vitro levels, are reported in this communication. The following issues are addressed: (1) the efficiency of encapsulation, the kinetics of drug release, and the effects of liposome concentration on both; (2) the activity of liposome-encapsulated ampicillin against extracellular-residing bacteria; and (3) the effects of liposome encapsulation on the stability and the retention of biological activity of the encapsulated ampicillin. These investigations required chemical, such as highperformance liquid chromatography (HPLC),18 and antibacterial2 assays that could be applied directly to the encapsulated ampicillin. The established HPLC protocols for ampicillin assay are for the free, not the encapsulated, drug. Furthermore, with respect to liposomal formulations, previous use of such an assay was restricted to the unencapsulated fraction alone, after its separation from the liposomes.18 An HPLC assay of the encapsulated fraction in any liposome-drug formulation would require sample pretreatment to release all the encapsulated drug to make it available for chromatographic analysis. Obviously, such pretreatment should not damage the drug nor should it clog the column or interfere with components of the device itself. The development of an HPLC assay for the direct determination of liposomeencapsulated ampicillin was an essential aspect of this investigation. Also reported here is a method for antibacterial assay of liposome-encapsulated ampicillin that is based on an established growth inhibition assay of free ampicillin, with the USP test organism Micrococcus luteus. This assay also served as a model system for testing whether the previously reported activity of liposomal ampicillin against bacteria residing in phagocytic cells,18-23 could be extended to extracellular bacterial colonies.
Materials and Methods MaterialssHigh-purity soybean phosphatidylcholine (PC) was from American Lecithin Company (Atlanta, GA) and high-purity cholesterol (CH) was from Sigma Chemical Company (St. Louis, MO). Ampicillin was a gift from Baxter Healthcare Company [3H]Cholesterol ([3H]CH) was from Amersham (Arlington Heights, IL). Blood agar base, trypton, agar-agar, and Muiller-Hunton broth were from Difco Laboratories (Detroit, MI). Spectra/Por4 dialysis tubings (molecular weight cutoff of 12 000 to 14 000) were from Spectrum Medical Industries (Los Angeles, CA). Lyophilized Micrococcus luteus (ATCC 9341) was a kind gift from Teva Pharmaceuticals (Israel). All other reagents were of analytical and, where relevant, HPLC grade. High-speed centrifugation was performed with a Sorvall RC-58 centrifuge. Liquid scintillation counting was performed with a
S0022-3549(95)00369-8 CCC: $14.00
Journal of Pharmaceutical Sciences / 635 Vol. 86, No. 5, May 1997
Kontron Analytical Betamatic I. The HPLC was conducted with a Hewlett-Packard Company (Palo Alto, CA) chromatograph was from (model HP-1090L), with a Waters µBondapack C-18, 3.9 × 300 mm column. Ampicillin was detected spectrophotometrically with a UVDiode array spectrophotometer (Hewlett-Packard, model 1090L). Liposome Preparation and Drug Encapsulationsmultilamellar liposomes (multilamellar vesicles, MLV) were prepared as previously described24-29 from lipid compositions of PC:CH at 3:1 mole ratios. Briefly, the lipids were dissolved in a methanol:chloroform (1:2, v/v) solution in a round-bottomed flask, and the organic solvent was evaporated to dryness under low pressure in a rotary evaporator. The swelling solution was composed of ampicillin dissolved in pH 7.2 phosphate-buffered saline (PBS) at concentrations ranging from 4 to 20 mg/mL. The swelling solution was added to the dry lipid film, and the preparation was agitated vigorously (with a vortex) for several minutes and then incubated in a shaker bath for 2 h at 37 °C. Separation of the unencapsulated drug from the liposome fraction (containing the encapsulated fraction) was by high-speed centrifugation for 45 min at 4 °C and 27 000g. “Empty”, drug-free liposomes were prepared in a similar manner, except their swelling solution was buffer alone. Drug ReleasesThe kinetics of drug release were studied according to previously reported procedures.24-29 Briefly, a suspension of liposomes (1.0 mL) was placed in a dialysis sac that was immersed in a constantly stirred receiver vessel containing 10 mL of drug-free buffer (PBS). At designated periods, the dialysis sac was moved from one receiver vessel to another containing fresh (i.e., drug-free) buffer. Drug concentration was determined by HPLC in each dialysate and in the sac (at the beginning and end of each experiment). These data were used to calculate the cumulative fraction of released drug (normalized to the total amount of ampicillin in the preparation), which is denoted “f” for each time point. Efficiency of Drug EncapsulationsThe efficiency of drug encapsulation, defined as the fraction of the total drug in the system encapsulated within the liposomes, was determined by centrifugation24-29 and by processing data of kinetic studies. Samples of the liposome preparation were subjected to high-speed centrifugation. The supernatant containing the unencapsulated drug was isolated, and the pellet containing the liposomes and their encapsulated drug was resuspended in drug-free buffer. The ampicillin content was assayed by HPLC and/or by the antibacterial activity assay in the complete preparation (precentrifugation), in the separated supernatant and in the resuspended pellets. These data were then used to calculate the percent encapsulation. The fraction of ampicillin encapsulated in the liposomes at timezero was obtained by analysis of the kinetics of drug release. For cases in which these experiments were performed on complete liposomedrug preparations, the magnitude of this encapsulated fraction corresponds to the efficiency of encapsulation. Ampicillin Stability as Function of TimesImmediately following preparation, liposome-ampicillin samples were separated by centrifugation as previously described. Aliquots from the swelling solution, the complete liposome preparation, the supernatant, and the resuspended pellets were analyzed for their antibacterial activity. Samples from the swelling solutions and the resuspended pellets were stored at 4 °C until they were reanalyzed for their antibacterial activity. These analyses were completed within 3 to 8 weeks. Quantitative DeterminationssLiposomessThe molecular weight of liposomes, especially MLV, cannot be determined accurately. However, for a given liposome species (type and lipid composition), the lipid concentration, which can be determined accurately, is proportional to the liposome concentration. Therefore, liposome concentrations are given in this communication in terms of mM lipid, with the understanding that the relevant parameter is the liposome concentration and the lipid concentration is the units used to define it. Quantitative determinations of lipid concentrations were performed by inclusion of a trace amount of [3H]CH in the formulation. HPLC Assay of AmpicillinsThe mobile phase was composed of 0.575% ammonium acetate in methanol:water (450:550, v/v), adjusted to pH 7.2. The ampicillin standard solutions covered the range 5-100 µg/mL. The calibration curves were linear, with r2 values in the range 0.93-1.0. For assays of liposome-encapsulated ampicillin, the liposome swelling solution served as the stock solution for the HPLC standards and was diluted (to provide the desired ampicillin concentrations) directly into the mobile phase.
636 / Journal of Pharmaceutical Sciences Vol. 86, No. 5, May 1997
Liposomal HPLC samples were pretreated in two ways. Samples from the complete liposome preparation (0.5 mL of liposome suspension) were transferred into a 20-mL volumetric flask, dissolved by the addition of 10 mL methanol, and made up to the mark by the addition of mobile phase. The solution was mixed for 5 min and then filtered through a 0.22-µm disposable filter. Samples after liposome centrifugation (2-mL) aliquots of the stock solution) were centrifuged and separated into supernatant and pellet. The pellet was resuspended in buffer and prepared as just described for complete liposome preparation samples. A 0.5-mL aliquot from the supernatant was transferred into a 20-mL volumetric flask and made up to the mark with mobile phase. The solution was mixed and filtered through a 0.22-µm disposable filter. The HPLC standard and sample injection volumes were 50-100 µL, the column was operated at a pressure of 100 bars, the flow rate of the mobile phase was 0.8-1.0 mL/min, and detection was at 230 nm. Triplicate samples were taken from each injected material and the area under the peak was determined for each run. The assays for each experimental system, including standards and controls, were all run on the same day. The typical total number of daily injections was 35-40, and the assays were run without encountering any technical problems. Antibacterial (Growth-Inhibition) Assay of AmpicillinsThe preparation of solutions and materials, as well as the assay itself, were performed under aseptic conditions. The assay was performed essentially according to the cylinder plate method.2 with paper disks instead of metal cylinders30 and the following modifications: (a) Petri dishes containing the solid agar were prepared in advance and stored at 4 °C until use. (b) Several days prior to performing the assay, a bacterial colony from the bacterial slant was seeded onto a previously prepared blood base agar (BBA) medium and incubated for 24 h at 37 °C. Several selected colonies were transferred into 20 mL of Mueller Hinton medium (MHM) and then further incubated in a shaker bath for 24 h at 37 °C. The suspension was then divided into several sterile tubes and centrifuged (with a bench centrifuge) for 10 min at 4 °C and 3000 rpm. The pellets were pooled and resuspended. The suspension was diluted to a T value (percent transmission) of 10% (at 580 nm), and used for the preparation of the bacteria-in-topagar component. Alternatively, the bacterial pellets were suspended in 2 mL of saline and stored at 4 °C (for no longer than a month). (c) The top agar was also prepared in advance and stored at room temperature until use. On the first day of the assay, the top agar was liquified (in a microwave oven or by autoclaving for 30 min at a pressure of 1.2 atm) and then incubated for 1 h in a water bath at 45 °C. For each petri dish, 2.5 mL of the liquified top agar were mixed with 0.7 mL of the bacterial suspension (see modification b, vortexed, and poured into the dishes discussed in modification a. The petri dishes were dried for 2 h prior to the placement of the disks and the beginning of the assay. (d) The paper disks were sterilized by autoclaving for 30 min at 1.2 atm. The sterilized disks were stored at room temperature until use. Four disks were placed into each petri dish and four dishes were used for each ampicillin dose. Twenty microliters of each standard or unknown were added in a dropwise manner to each disk, after which the plates were allowed to dry for 2.5 h and then incubated for 24 h at 37 °C. The diameters of the growth inhibition zones were measured for each disk with a caliber accurate to 0.01 mm. Controls included top agar without bacteria, disks to which no samples were added, PBS (added to petri dishes with and without disks) and “empty liposomes”. Controls were tested under the same conditions just described. Calibration curves were constructed for each experiment with the standards and controls. A fresh ampicillin solution (4 mg/mL) served as the stock solution for the calibration standards. The swelling solutions were used as the stock solutions to obtain the calibration curves for the liposomeencapsulated ampicillin assays. The ampicillin standards were prepared by dilution from the stock solutions, with concentrations ranging from 0.01 to 0.5 µg/mL. Liposomal samples pipetted onto the disks were diluted from the appropriate stocks, with dilution factors ranging from 1 × 104 to 2 × 103.
Results and Discussion HPLC Assay of Liposomal AmpicillinsThe ampicillin HPLC assay requires that the liposomes be dissolved in the
Table 1sLipid Recovery in Centrifuge-Separated Ampicillin-Liposome Preparations
Table 2sHPLC Determinations of Ampicillin in Liposomal Formulations
Total Lipid Concentration of the Preparation (pre-centrifugation) (mM)
Lipid Recovery in the Pelleta (% from total)
Batch Code
Direct Determination Performed on Liposomal Samplesa (% from initial ampicillin)c
Indirect Determination Performed on Free Ampicillinb (% from initial ampicillin)
10 20 30 45 60 200
93 ± 0.8 97 ± 1.2 98 ± 3.9 95 ± 0.4 99 ± 2.8 102 ± 3.5
IS-15 IS-23 IS-24
116.1 ± 4.7d 93.6 ± 9.1 80.0 ± 8.6
111.4 ± 8.7 85.2 ± 1.3 85.4 ± 0.5
a Each data is the average of three independent separations (± standard deviation).
assay sample to ensure all of the encapsulated drug was available for analysis. Detergents, such as Triton X-100, were unsatisfactory in dissolving the liposomes because they subsequently clogged the column. Attempts to dissolve the liposomes with the mobile phase formulation customary for free ampicillin2,18 resulted in incomplete dissolution of the liposomes. However, an alternative mobile phase formulation composed of 0.575% ammonium acetate in methanol:water (450:550, v/v), adjusted to pH 7.2, effectively dissolved the liposomes, freeing the ampicillin for analysis. HPLC chromatograms of free ampicillin gave a single sharp peak with a retention time of 4 min that was identified by UV spectra as ampicillin. Assay samples containing only the encapsulated drug were obtained by centrifuging aliquots of ampicillin-liposome preparations as previously described, separating the pellet from the supernatant, and resuspending the pellet in drug-free buffer. The separation was quantitative, as evidenced from the lipid contents of the pellet and the supernatant shown in Table 1. The following HPLC chromatograms were similar to each other and to those of free ampicillin: complete drug-liposome preparations (i.e., containing encapsulated and unencapsulated drug), liposomes containing encapsulated drug alone, and the (separated) unencapsulated fraction of drug-liposome preparations. There were no peaks in chromatograms of control systems; that is, water, PBS, or pretreated drug-free (i.e., empty) liposomes. These data indicate that the pretreatment with the modified mobile phase did not interfere with the ability to assay the ampicillin. By visual observation, the pretreatment seemed to provide complete liposome dissolution, but it was necessary to determine quantitatively that all the encapsulated drug was released and assayed. Therefore, the HPLC assay of a given liposome-ampicillin system was performed by two methods: (1) aliquots from the complete liposome preparation were assayed for their ampicillin concentration, and (2) aliquots from the liposome preparation were subjected to exhaustive dialysis over a period of at least 24 h with frequent replacements of medium, as described under methods for kinetics of drug release. The dialysis was terminated following complete depletion as verified experimentally. The dialysates containing only free ampicillin in aqueous solutions were individually subjected to HPLC analysis. Therefore, the cumulative determinations of ampicillin in the dialysates should contain all of the drug that was originally in the liposomal sample (encapsulated and unencapsulated). If the pretreatment devised for the liposomal suspensions effectively enables all of the drug to be available for chromatographic analysis, then there should be agreement between the results from the two independent determinations for the same liposomal preparation. This expectation was confirmed, as shown in Table 2, for three ampicillin-liposome batches. The percent ampicillin determined by each method was within experimental error for each batch.
a Results of HPLC determination of liposomal samples subjected to the pretreatment. b Results of HPLC determination of ampicillin dialysates obtained from complete depletion of liposome-ampicillin treatment in the course of exhaustive dialysis. c The initial quantity of ampicillin put into a given liposome preparation served as the 100% value. Values over and under 100% indicate water or material losses in the course of liposome preparation, respectively. d Each value in the table is the average and standard deviation of duplicate samples.
Antibacterial Assay of Liposomal AmpicillinsThe assay for liposome-encapsulated ampicillin had to meet three criteria: (1) the encapsulated drug has to retain its biological activity; (2) all of the drug in the liposomes had to become available for the assay, which required complete release of all the encapsulated drug within the time course of the assay; and (3) the assay has to be quantitative. Typical results of the antibacterial assays for the liposome-encapsulated ampicillin are shown in Figure 1. Dishes A and B contain several controls, including empty disks, disks plated with PBS, and disks plated with empty (i.e., drug-free) liposomes suspended in PBS. As expected, there are no growth inhibition zones for these controls. In contrast, growth inhibition zones were usually observed with ampicillin formulations (Figure 1, dishes C-F). Dish C is a typical response to an ampicillin standard solution. Dish D is characteristic of the responses for the complete liposome preparation, and dishes E and F are characteristic of the responses to the unencapsulated and encapsulated fractions of the separated liposome preparations, respectively. The ampicillin originating from liposomal systems, including the encapsulated drug, was biologically active, as evidenced by the presence of growth inhibition zones. Within the concentration range 0.15-0.5 µg/mL, the ampicillin calibration curves were linear, with r2 values in the range 0.96-0.98. To be considered a valid bioassay, all the liposomal ampicillin had to become available to the bacteria during the course of the assay without significant loss in sample activity. To determine the feasibility of a given liposome-ampicillin preparation to meet these requirements, the following components were assayed: (1) the swelling solution; (2) the complete liposome preparation; (3) the supernatant fraction of a separated drug-liposome preparation; and (4) the resuspended pellet fraction of a separated drug-liposome preparation. The data from these bioassays were used to calculate the total quantity of active ampicillin in each of the four systems. Results for duplicate liposome-ampicillin preparations are shown in Figure 2. Initially, 40 mg of active ampicillin were incorporated into each liposome preparation through the swelling solution. Bars A and B represent the amounts of active ampicillin in the swelling solution and in the complete liposome preparation respectively. The general agreement between the two independent determinations suggest there was no significant drug loss during the course of liposome preparation, and that ampicillin activity was not compromised. Bars C and D represent, respectively, the unencapsulated and encapsulated active ampicillin of the liposome preparation. The sums of the C+D bars represent another evaluation of the total amount of active ampicillin in the liposome preparation. The general agreement of the three independent determinations of total active ampicillin (bars A, B, and C+D) suggests that the encapsulated drug retains its biological activity. Furthermore, the assay demonstrates
Journal of Pharmaceutical Sciences / 637 Vol. 86, No. 5, May 1997
Figure 1sAntibacterial assay of free and of liposome-encapsulated ampicillin by the growth-inhibition assay (raw data): (A) no treatment (left-hand side) or PBS alone (right-hand side); (B) “empty” (i.e., drug-free) liposomes; (C) ampicillin standard solution (in PBS); (D) complete ampicillin-liposome systems (i.e., containing encapsulated and unencapsulated drug); (E) unencapsulated ampicillin only (i.e., the supernatant, postseparation of systems, such as in D, by centrifugation); and (F) liposomeencapsulated ampicillin only (i.e., the resuspended pellet, postseparation of systems, such as in D, by centrifugation).
Figure 2sEvaluation of the antibacterial assay for two similar ampicillin-liposome preparations coded L-31 and L-29: (A) swelling solution; (B) complete ampicillinliposome preparation; (C) the unencapsulated fraction of preparation after separation by centrifugation; (D) The encapsulated fraction of the preparation after separation by centrifugation; and (C+D) the sum of the (separate) fractions C and D. Each column is an average of 16 repeat samples, and the error bars represent the standard deviations.
that the encapsulation efficiency was ≈50%. Essentially all of the encapsulated ampicillin became available to the bacteria during the course of the assay, as evidenced by the similarities between bars A, B, and C+D. These similarities may be attributed to drug diffusion from intact liposomes under the extremely low liposome concentrations being assayed. The general phenomena of an inverse relationship between the kinetics of drug release and liposome concentration24,25,29 indicates that complete and fast depletion of the liposomes can take place under the conditions of the assay, where the liposome concentrations in the samples plated onto the disks are extremely low, with dilution factors ranging from 1 × 104 638 / Journal of Pharmaceutical Sciences Vol. 86, No. 5, May 1997
Figure 3sThe efficiency of ampicillin encapsulation in liposomes (MLV) as a function of liposome concentration (expressed in lipid molar units). The data points labeled HPLC and Bioassay were determined from centrifuge-separated ampicillinliposome preparations by HPLC or the antibacterial assay, respectively. The data points labeled “derived from kinetic data” were determined from kinetic experiments by HPLC as the method of quantitative determination. The points are the experimental data, and the linear line is drawn to emphasize the increasing trend. Each point on the figure represents an individual preparation, and the preparationto-preparation reproducibility can be appreciated from the encapsulation efficiencies obtained for the individual preparations at the same lipid concentration.
to 2 × 103. Finally, the encapsulated ampicillin was active against bacterial colonies residing outside cells. Efficiency of Ampicillin EncapsulationsTheoretical considerations24 supported by experimental data24-29 have shown that increased liposome concentrations result in increased encapsulation efficiencies. Typical ampicillin encapsulation efficiencies as a function of liposome concentration (Figure 3) support this trend. The encapsulation efficiencies increase from <10% to ≈50% as the liposome concentrations increase from 20 to 200 mM lipid. Encapsulation efficiencies
at high liposome concentrations are an improvement over the previously reported efficiencies.21 Kinetics of Drug ReleasesRegardless of the route of administration and of the anatomic destination of the drug, there should be minimal free drug surrounding the liposomes in vivo, despite continuous drug efflux from the liposomes, because of the cumulative effect of dilution of both liposomes and released drug, drug binding at the site of action, drug binding at sites other than its site of action, drug metabolism, and drug clearance. Therefore, there should be an electrochemical gradient driving the drug out from the liposomes. The experimental design for the kinetic studies was made to model these flux conditions.24 Previous studies have shown that the magnitude of the rate constant for encapsulated drug efflux decreases with increasing liposome concentrations.24,25,28 This relationship has been attributed to deviations from ideality.24 Therefore, the kinetic studies were undertaken with liposome preparations at low lipid concentrations to increase experimental sensitivity to diffusion parameters. The data were analyzed according to a previously derived and experimentally supported mechanism,24,25,27-29 that is based on Eyring's theory. Initially, at timezero, the system is assumed to contain encapsulated and unencapsulated ampicillin. Diffusion of encapsulated and unencapsulated ampicillin likely occurs as multistep processes, governed by a single rate-limiting step. Therefore, the overall mechanism can be described as a series of parallel first-order processes with a common product and different reactants.24,29 For unencapsulated and encapsulated ampicillin, the relationship between f (the cumulative amount of ampicillin released at time t normalized to the total amount in the preparation at time ) 0) and time t is given by eq 1: -kft
f ) ff(1 - e
-kft
) + fl(1 - e
)
(1)
The fractional distribution of ampicillin at timezero is represented by the fraction unencapsulated, ff, and the fraction encapsulated, fl, where ff + fl ) 1. The corresponding rate constants are kf and kl for the unencapsulated and encapsulated fractions, respectively. In those cases where there was no separation from unencapsulated drug prior to the experiment, fl also represents the fraction of encapsulated ampicillin. Results obtained for an ampicillin-liposome preparation of 30 mM lipid are shown in Figure 4. Complete liposome depletion occurred within 25 h. The magnitudes of fl, kf, and kl, obtained for liposome preparations of 30 and 45 mM lipid, are given in Table 3. As expected, the fraction of encapsulated drug is rather low, and the rate constant for the encapsulated fraction decreases with increasing liposome concentrations. The rate constant for the release of encapsulated ampicillin corresponds to a half-life in the range of 40 h, which implies that ampicillin-liposome systems have the potential for use as sustained-release drug depots. Stability of Liposome-Encapsulated AmpicillinsThe stability of the encapsulated ampicillin as a function of time and of the encapsulation efficiency was investigated by assessing the loss of ampicillin activity. Fresh preparations of ampicillin-encapsulating liposomes were separated into encapsulated and unencapsulated fractions by centrifugation. The liposomal pellets containing only encapsulated drug were resuspended in drug-free buffer and stored at 4 °C. Samples of the swelling solutions that contained free aqueous ampicillin and were used in these preparations were stored under similar conditions. At designated time periods, aliquots were withdrawn from the liposome suspension and from the corresponding swelling solution and analyzed for active ampicillin, by the antibacterial assay. The aliquots withdrawn from the stored liposomes may contain encapsulated and
Figure 4sThe kinetics of ampicillin release from MLV. The points are experimental data, which were averages of duplicate runs, with the error bars indicating the standard deviations. Quantitative determination was by HPLC of triplicate samples for each time point in each run. The curve was calculated from eq 1 with the parameters listed in Table 3 for the preparation L-24. Table 3sKinetic Parametersa of Ampicillin Release from Liposomes (MLV) BatchCode
[LIPID](mM)
fl(%)
kf(h-1)
kl(h-1)
L-23 L-24
30 45
6.0 ± 0.7 8.6 ± 0.4
0.51 ± 0.006 0.57 ± 0.005
0.015 ± 0.009 0.011 ± 0.004
a Triplicate quantitative determinations of the released drug were made for each time point, and duplicate runs were made from each preparation. The average of the duplicate runs and the triplicate determinations are shown in Figure 4, together with the standard deviations (SDs). The average values were obtained by nonlinear regression analysis according to eq 1 in the text. The results of these analyses are listed in this table, and the SDs reported here, provided by the analysis process, illustrate the good fit of the experimental data to the proposed mechanism.
unencapsulated ampicillin even though unencapsulated ampicillin was removed prior to storage. Note that the objective of this study was to determine ampicillin activity as related to time and encapsulation efficiency, not to evaluate aspects of shelf-life. Lyophilized powders are the rational dosage form for the long-term storage required for pharmaceutical products.27 The results shown in Figure 5 are for three different liposome preparations and four test periods. One preparation was tested at two different times. The percent of encapsulation in the original preparation is shown by the line graph and the losses of antibacterial activity with time are presented as bars. Suspension I-28 was prepared at the concentration of 100 mM lipid, which yielded an encapsulation efficiency of 18%. The stability of the stored samples from this batch and from its corresponding swelling solution was assessed 3 and 8 weeks post preparation. Suspensions L-29 and L-31 were duplicate batches prepared at 200 mM lipid concentrations, which yielded encapsulation efficiencies approaching 50%. The stability of those liposome suspensions and their corresponding swelling solutions was assessed 5 weeks postpreparation. The loss in ampicillin activity of the swelling solutions should be independent of the liposome suspensions prepared with these solutions. The losses in antibacterial activity (Figure 5) were significant for the swelling solution associated with suspension I-28 at 3 and 8 weeks, and the swelling solution associated with suspension of I-29 at 5 weeks. The losses corresponded to 40-60% of the activity measured
Journal of Pharmaceutical Sciences / 639 Vol. 86, No. 5, May 1997
Figure 5sThe effects of encapsulation on the stability of ampicillin with time, as determined by the antibacterial activity of ampicillin. The codes I-28, I-29 and I-31 designate three different batches at initial liposome concentrations of 100, 200, and 200 mM lipid, respectively. The line graph is the efficiencies of drug encapsulation in the original preparations. The points are the experimental data and are connected by a line to emphasize the trend. The bars represent the loss in antibacterial activity, at specified test periods, where 100% activity was assigned to the values measured in the fresh swelling solution or resuspended liposome pellet (i.e., immediately upon system preparation). The specific test periods (in weeks) are given in parenthesis next to the preparation codes. The lightly shaded bars are the data for the swelling solution and the darkly shaded bars are the data for the liposomes. Each bar is an average of 12−16 assay samples, and the magnitudes of the standard deviations (number of samples) are: for data set I-28 (3), 0.62 and 0.50 for the swelling solution and the pellet, respectively; for data set I-28 (8), 0.45 and 0.45 for the swelling solution and the pellet, respectively; for data set I-29 (5), 0.47 and 0.45 for the swelling solution and the pellet, respectively; and for data set I-31 (5), 0.39 and 0.29 for the swelling solution and the pellet, respectively.
immediately after preparation in the fresh systems. However, the corresponding loss in the ampicillin activity in the swelling solution associated with suspension I-31 was lower. Although we have no explanation for this observation, we stress that our major interest is drug stability in the liposomal formulations. As shown in Figure 5 for both test periods, the losses in antibacterial activity of ampicillin encapsulated and stored in the liposome suspension I-28 were similar to those of observed for (free) ampicillin in the corresponding swelling solution. In contrast, regardless of some preparation-topreparation variabilities, the loss in ampicillin activity was significantly lower for ampicillin encapsulated and stored in the liposomal formulations I-29 (17% loss) and I-31 (6% loss). As expected, ampicillin stability in aqueous media was time limited. Loss in activity, although significant, was not linear with time and plateaued over 8 weeks (compare the data for I-28 at 3 and 8 weeks of storage). Encapsulation within liposomes did not increase the loss in activity (liposomes of I-28 versus swelling solutions) but may have actually decreased the loss in ampicillin activity (liposomes of I-29 and I-31 versus swelling solutions). However, further studies are necessary to determine whether the improvement in stability (as suggested by preparations I-29 and I-31 but not by preparation I-28) may be associated with the higher encapsulation efficiencies (≈50% versus 20%), the higher liposome concentration (200 versus 100 mM lipid), or both.
Conclusions Development of HPLC and Antibacterial Assays for Encapsulated AmpicillinsThe simple pretreatment of the liposome suspensions with the modified mobile phase released all the encapsulated ampicillin making it available for quan640 / Journal of Pharmaceutical Sciences Vol. 86, No. 5, May 1997
titative assays. The analysis could be performed continuously for multiple samples. The pretreatment with the modified mobile phase did not result in significant ampicillin degradation and/or inactivation, in blockage of the HPLC column, or any interference with other components. These results suggest that this HPLC assay could be an effective method to quantify liposome-encapsulated ampicillin. The antibacterial assay demonstrated that the liposome-encapsulated ampicillin retained its biological activity, indicating that the encapsulation procedures for preparation of liposome suspensions did not result in any detectable drug destabilization. The results for both unencapsulated and liposome-encapsulated ampicillin show that the use of paper disks is an accurate and useful method for the zone-inhibition assay that could replace the use of metal cylinders. The HPLC method and the bioassay have the potential to be further developed as analytical assays for liposomeencapsulated ampicillin, as ampicillin-liposome formulations will mature into pharmaceutical products. Furthermore, both assays could be extended, with low to moderate investments in additional efforts, to assay other antibiotics encapsulated in liposomes. Liposome Encapsulated AmpicillinsAmpicillin was encapsulated into MLV suspensions as a function of lipid concentration, and then the release and activity of the ampicillin were determined. Encapsulation efficiencies increased as the lipid concentration of the suspension increased. Encapsulation efficiencies ranged from ≈50% with 200 mM lipid concentrations to <20% at lipid concentrations of <20 mM. The lower encapsulation efficiencies were still greater than those previously reported.21 The rate constant for released ampicillin varied with liposome concentration. A half-life in the range of 40 h was achieved with liposome concentrations on the order of 30-45 mM lipid. The ampicillin encapsulated within the liposomes was active against bacterial colonies without the involvement of other cells. Taken together with the previously reported activity of liposome-ampicillin dosage forms against intracellular organisms,18-23 this result suggests that the broad activity spectrum of ampicillin is not decreased when the drug is encapsulated within the liposomes. In addition, encapsulation within liposomes does not compromise ampicillin stability but may actually enhance it. Therefore, ampicillin-liposome suspensions merit further consideration for development into pharmaceutically relevant products.
Abbreviations Used BBA, blood base agar; CH, cholesterol; MHM, Mueller Hinton medium; MLV, multilamellar vesicles; PBS, phosphatebuffered saline; PC, phosphatidyl choline.
References and Notes 1. Remington's Pharmaceutical Sciences,17th ed., Osol, A., Ed.; Mack: Easton, PA, 1985; pp 1194. 2. USP XXIII-NF18 , U. S. Pharmacopeial Convention, Inc.: Mack: Easton, PA, 1995; pp 107, 1681-1696. 3. Nguyen, N. T.; Mortada, L. M.; Notary, R. E. Pharm. Res. 1988, 5, 288-296. 4. Bundgaard, H.; Larsen, C. J. Chromatogr. 1977, 132, 51-59. 5. Larsen, C.; Bundgaard, H. J. Chromatogr. 1978, 147, 143-150. 6. Bundgaard, H. Pharm. Pharmacol. 1974, 26, 385-392 . 7. Van Der Bijl, P.; Seifart, H. I.; Parkin, D. P.; Mattheyse, F. J. S. Afr. Med. J. 1988, 73, 453-455. 8. Aki, H.; Sawai, N.; Yamamoto, K.; Yamamoto, M. Pharm. Res. 1991, 8, 119-122. 9. Oliyai, R.; Lindenbaum, S. Int. J. Pharm. 1991, 73, 33-36. 10. Kong, K. M. Am. J. Hosp. Pharm. 1988, 45, 314.
11. Hartzen, S. H.; Frimodt-Moller, N.; Andreasen, J. APMIS 1988, 96, 584-588. 12. Dukin, K. T.; Jones, S.; Howard, A. J. J. Antimicrob. Chemother. 1988, 21, 405-411. 13. Jephcott, A. E.; Gough, K. J. Antimicrob. Chemother. 1988, 21 (suppl. B), 43-48. 14. Martindale, The Extra Pharmacopoeia, 30 ed.; Reynolds, J. E. F., Ed.; The Pharmaceutical Press: London, UK, 1993; pp 116118. 15. Principles and Practice of Infectious Diseases, 3rd ed.; Mandell, G. L.; Dougles, R. C.; Bennett, J. I., Eds.; Churchill Livingstone, Edinburgh, 1990; pp 240-242. 16. Nelson, J. D.; Haltalin, K. C. J. Infect. Dis. 1974, 129, 222227. 17. McCracken, G. H. In Clinical Reviews in Pediatric Infectious Disease; Nelson, J. D.; McCracken, G. H., Eds.; B.C. Decker: Toronto, 1984; pp 73-77. 18. Youssef, M.; Fattal, E.; Alonso, M. J.; Roblot-Treupel, L.; Sazieres, J.; Tancrede, C.; Omnes, A.; Couvreur, P.; Andremont, A. Antimicrob. Agents Chemother. 1988, 32, 1204-1207. 19. Couvreur, P.; Fattal, E.; Andermont, A. Pharm. Res. 1991, l8, 1079-1086. 20. Bakker-Wuodenberg, I.; Lokerse, A. F.; Vink van den Berg, J. C.; Roedink, F. H.; Michel, M. F. Antimicrob. Agents Chemother. 1986, 30, 295-300. 21. Fattal, E.; Rojas, J.; Yousseff, M.; Couvreur, P.; Andremont, A. Antimicrob. Agents Chemother. 1991, 35, 770-772. 22. Bakker-Wuodenberg, I.; Lokerse, A. F.; Roedink, F. H. J. Pharmacol. Exp. Ther. 1989, 251, 321-327. 23. Bakker-Wuodenberg, I.; Lokerse, A. F.; Roerdink, F. H.; Regts, D.; Michel, M. F. J. Infect. Dis. 1985, 151, 917-924.
24. Margalit, R.; Alon, R.; Linenberg, M.; Rubin, I.; Roseman, T. J.; Wood, R. W. J. Controlled Release 1991, 17, 285-296. 25. Yerushalmi, N.; Margalit, R. Biochim. Biophys. Acta 1994, 1189, 13-20. 26. Margalit, R. in Vesicles, Surfactant Science Series; Rossof, M., Ed.; Marcel Dekker: New York 1995; Chapter 13, pp 527-560. 27. Margalit, R.; Yerushalmi, N. In Microencapsulation Methods and Industrial Applications; Benita, S., Ed.; Marcel Dekker: New York, 1995; Chapter 10, pp 259-295. 28. Lichtenstein, A.; Margalit, R. J. Inorg. Biochem. 1995, 60, 187198. 29. Margalit, R. in Microparticulates - Preparation, Characterization and Application in Medicine; Cohen, S.; Berenstein, H., Eds.; Marcel Dekker: New York, 1996; Chapter 15, pp 425-461. 30. Akimoto, Y.; Mochizuki, Y.; Uda, A.; Omata, H.; Shibutani, J.; Nishimura, H.; Komiya, M.; Kaneko, S.; Kobayashi, S.; Kuboyama, N.; Yamane, J.; Furukawa, Y.; Fujii, A. J. Nihon Univ. Sch. Dent. 1990, 32, 4-18. 31. Margalit, R.; Okon, M.; Yerushalmi, N.; Avidor, E. J. Controlled Release 1992, 19, 275-288. 32. Yerushalmi, N.; Arad, A.; Margalit, R. Arch. Biochem. Biophys. 1994, 313, 267-273. 33. Margalit, R. Crit. Rev. Ther. Drug Carriers Systems 1995, 12, 233-261.
Acknowledgments This work was supported by a research grant to R. M. from Baxter Healthcare Company, Round Lake, IL.
JS9503690
Journal of Pharmaceutical Sciences / 641 Vol. 86, No. 5, May 1997