INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

The use of amikacin for first-line therapy is limited, despite its excellent bactericidal activity against gram-negative organisms. Poor oral absorption, a short half-life, and serious renal, auditory, and vestibular toxicities necessitate frequent intravenous administration, with concomitant monitoring of drug levels to ensure that the levels in plasma remain within a narrow therapeutic window (26).

Encapsulation of a drug in liposomes can increase a drug's therapeutic index by reducing the level of drug delivered to sites where it is toxic relative to the level at sites of efficacy within the body (18). Aminoglycosides have been encapsulated in liposomes with the primary objective of targeting intracellular infections (12, 25). While these attempts demonstrated in vivo activity against intracellular infections, the efficacies of liposomal aminoglycosides against extracellular organisms largely remain unexplored. In a clinical trial, gentamicin in large plurilamellar liposomes had a half-life of 9 h (29). Trough (24-h) levels of total gentamicin were below 1 µg/ml, and the levels of free gentamicin exceeded 30% of the total levels in plasma, indicating that most of the liposomal drug had leaked from the formulation. Another formulation of gentamicin, one in sterically stabilized liposomes, caused deaths in rats receiving lipid doses above 350 µmol/kg (3). The characteristics of these formulations suggest that neither formulation would safely provide the high sustained levels of antibiotic desirable for the treatment of extracellular infections.

Long-circulating (low-clearance) liposomes could improve antibacterial therapy by prolonging the time that the antibiotic remains present at therapeutic levels. This would increase the amount of drug delivered to sites of infection and decrease the level of renal exposure. To test this hypothesis, amikacin was formulated in small unilamellar liposomes consisting of phospholipids that have high phase-transition temperatures and that remain in the gel state under physiologic conditions. These liposomes had both curative and prophylactic activities in animal models of Klebsiella pneumonia and Pseudomonas endocarditis (14, 34). In a single-dose study, this formulation increased the amikacin concentrations in plasma and tissues and prolonged the time that amikacin concentrations remained elevated compared to those for the conventional formulation of amikacin (16), suggesting that infrequent dosing might be feasible. To establish an optimal dosing regimen, the clearance, half-life, and fluctuations in the levels of liposomal amikacin in plasma should also be determined after repeated dosing. In addition, the fraction of liposomal amikacin cleared by the kidneys and the level of exposure of plasma to free amikacin may be important factors in establishing the safety profile of low-clearance liposomal amikacin. To accomplish these objectives, we studied the pharmacokinetics of low-clearance liposomal amikacin in rhesus monkeys after the intravenous administration of single and repeated doses in a clinically relevant dosing regimen.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Drug formulation. Low-clearance liposomal amikacin (MiKasome; NeXstar Pharmaceuticals, San Dimas, Calif.) consisted of amikacin hydrochloride (12.6 mg/ml) encapsulated in small unilamellar liposomes composed of hydrogenated soybean phosphatidylcholine, cholesterol, and distearoylphosphatidylglycerol in a mole ratio of 2:1:0.1, respectively. The ratio of drug to total lipid was approximately 1:5 (by weight). The lipids were spray dried, and the liposomes were formed by rehydration in amikacin-containing formulation buffer (sterile 9% sucrose, succinate buffer [pH 6.5]). The liposomes were homogenized to produce small, unilamellar liposomes (median diameter, <100 nm by laser light scattering) and were then dialyzed against the formulation buffer to remove unencapsulated drug. Of the amikacin contained in the final formulation, more than 85% was encapsulated within the liposomes, while the remainder of the amikacin was externally associated with the liposomes.

Animals. Young adult rhesus monkeys (Macaca mulatta; body weight, 4.5 to 5 kg) were supplied by Sierra Biomedical, Inc. (Sparks, Nev.) and were acclimated for at least 14 days prior to being placed in the study. The animals were housed individually in stainless steel cages, fed Teklad Primate Diet (Harlan, Indianapolis, Ind.) supplemented with fresh fruit and vegetables, and had free access to water. The room where the animals were housed was ventilated (>10 changes/h) with fresh air and was maintained at between 65 and 84°F, with a 12-h light and 12-h dark cycle. A veterinary examination, which included respiratory, cardiovascular, serum chemistry, hematology, and fecal parasite evaluations and tuberculosis tests, was performed for each animal prior to the study. Four healthy animals (two males and two females) were placed in the study. All animals were treated in accordance with the U.S. Animal Welfare Act (9 CFR 1-3).

Treatment. Each animal received a single intravenous dose of liposomal amikacin (20 mg of amikacin/kg of body weight) that was diluted 1:1 (by volume) in dextrose, 5% injection, and that was infused at a constant rate over a period of 60 min through an indwelling catheter placed in a saphenous or cephalic vein. The animals were fasted for approximately 8 h prior to dosing and were not sedated during the infusion period. Catheters were flushed with dextrose, 5% injection, at the conclusion of the infusion period. Seven months after the single-dose treatment, the same animals each received a series of liposomal amikacin infusions (20 mg/kg over 60 min) administered every 48 h for a total of eight doses. Blood samples (1.5 to 2 ml) were obtained by venipuncture (femoral vein) at various times after the administration of each dose (see Results) and were placed into EDTA-containing tubes. The plasma was separated by centrifugation and was then frozen at 70°C prior to analysis for amikacin levels. At some time points, 0.5 ml of freshly obtained plasma was placed into an ultrafiltration device (Microcon-100; Amicon, Beverly, Mass.) which was centrifuged (3,000 × g) for 15 to 25 min at 15°C to obtain plasma ultrafiltrates (100 to 150 µl), which were frozen at 70°C prior to analysis for amikacin levels. After administration of the doses, the animals were placed into stainless steel cages designed for urine collection, and the total urine output was collected over various time intervals in polyethylene containers placed in an ice bath. Urine volumes were recorded, and an aliquot of each urine collection was frozen at 70°C prior to analysis for amikacin levels. Serum and urine samples for determination of creatinine levels were obtained prior to and at various times after dosing.

Assay. The total amikacin concentration was measured in all plasma, plasma ultrafiltrate, and urine samples by a commercial fluorescence polarization immunoassay (TDx/FLx; Abbott Diagnostics, Abbott Park, Ill.). The assay was performed with reagents supplied by the manufacturer, except that Triton X-100 (final concentration, 0.05%) was added to the assay dilution buffer to release liposome-associated amikacin. The validated lower limits of quantitation for the modified TDx assay were 0.3 µg/ml for plasma, 0.3 µg/ml for plasma ultrafiltrates, and 0.6 µg/ml for urine samples. Samples containing unknown concentrations were serially diluted (1:10) in sodium phosphate buffer (50 mM; pH 7.0) so that the concentrations would fall within the linear range of the assay (0.3 to 7.0 µg/ml). Relative standard deviation (RSD) and percent bias determined from liposomal amikacin-spiked plasma samples were 6.7 and 5.5%, respectively, at 5.5 µg/ml, 9.1 and 1.8%, respectively, at 2 µg/ml, and 2.9 and +6.4%, respectively, at 0.925 µg/ml. Liposomal amikacin quality control standards (2,000, 555, and 0.925 µg/ml) were prepared in blank plasma and were run with each assay to ensure the recovery of encapsulated amikacin. Runs were accepted if the values for the quality control samples were within ±20% of the expected values.

Pharmacokinetic analysis. Each animal's plasma concentration-versus-time data obtained after the single-dose regimen, during the 48-h interval after the administration of the first dose of the repeated-dose regimen, and after the administration of the last dose of the repeated-dose regimen were fit to multiexponential equations by a nonlinear, least-squares procedure (RSTRIP; MicroMath, Salt Lake City, Utah). Weighting was adjusted (unweighted, 1/y, or 1/y²) to obtain the best curve fits as judged by a normalized Akaike information criterion. Intercepts obtained from the fitting of postinfusion data were corrected by the method of Loo and Riegelman (28). The values of standard pharmacokinetic parameters were determined from the resulting slopes and intercepts (24). In addition, portions of the elimination phase (before and after 120 h for the single dose and before and after 144 h for repeated dosing) were fit separately to determine early- and late-phase elimination half-lives. After repeated dosing, the steady-state clearance (CL_SS) was determined as CL_SS = dose/AUC_SS, where AUC_SS is the area under the plasma concentration-versus-time curve during the last (48-h) dosing interval. The predicted C_SS was determined as C_SS = AUC/dosing interval, where AUC was calculated to infinity after administration of the single dose and over one dosing interval after the repeated-dose regimen. The fluctuation in the level in plasma was quantitated as the ratio of the maximum to minimum concentrations observed during the final 48-h dosing interval. The renal clearance (CL_R) was calculated for each urine collection interval as CL_R = A_ur/AUC_t1-t2, where A_ur was the amount of amikacin recovered in the urine, and AUC_t1-t2 was the area under the plasma total amikacin concentration-time curve during the collection interval. Creatinine clearance (CL_CR) was calculated as CL_CR = A_ur/(C_CR · 16 h), where A_ur was the amount of creatinine recovered in the urine between 8 and 24 h after the dose, and C_CR was the serum creatinine concentration measured 24 h after administration of the dose. The statistical significance of differences in pharmacokinetic parameters between the single-dose and repeated-dose regimens was determined by the paired, two-sample t test (two-tailed), because all parameters were found to be normally distributed. Although both male and female animals were studied, the data were summarized for all the animals, because no apparent differences by gender were observed for any of the measured parameters. Changes in the serum creatinine level and CL_CR were evaluated by a one-way repeated-measures analysis of variance, followed by Dunnett's test for the single-dose study (values obtained on day 1, day 2, and day 6 were compared to prestudy values) and by the paired t test for the multiple-dose study (values obtained after administration of the first dose and the eighth dose were compared). Both parameters were normally distributed.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

Pharmacokinetics in plasma after administration of a single dose. Plasma amikacin concentrations declined biexponentially following the administration of a single dose of liposomal amikacin (Fig. 1). The levels in plasma initially decreased with a half-life of 0.75 h, but this phase accounted for only 1.3% of the total AUC, with the remainder of the elimination occurring much more slowly. Maximum concentrations in plasma of 378 ± 24 µg/ml were achieved at the end of the 1-h infusion, and total plasma amikacin concentrations remained above 100 µg/ml for more than 48 h after dosing. During the elimination phase, the half-life of liposomal amikacin appeared to decrease over time. The elimination half-life observed prior to 120 h appeared to be longer than the half-life observed after 120 h (Table 1). A similar time-dependent elimination of liposomal amikacin was observed in rats after the administration of a single intravenous dose of 50 mg/kg (15).

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Total amikacin concentrations (mean ± standard deviation; n = 4) in the plasma of rhesus monkeys after the administration of low-clearance liposomal amikacin. (A) After the administration of a single intravenous dose (20 mg/kg). (B) During and after the administration of eight consecutive intravenous doses (20 mg/kg) administered at 48-h intervals.

1

1

1

1

Pharmacokinetics in plasma after repeated dosing. Profiles of the concentration in plasma following the administration of the first dose of the repeated-dose regimen were similar, in all animals, to those observed after the single-dose regimen (Fig. 1). After a brief initial phase, the levels in plasma fell with a half-life of 50 ± 13 h during the 47-h interval prior to administration of the second dose. Peak (0.5-h) levels of total amikacin in plasma increased over the course of the repeated-dose regimen from 298 ± 36 µg/ml (first dose) to 728 ± 72 µg/ml (last dose). Trough (47-h) levels in plasma also rose over this period, from 110 ± 21 to 418 ± 60 µg/ml (Fig. 1).

Amikacin concentrations in plasma ultrafiltrates. To estimate the level of exposure to free (unencapsulated) amikacin, the concentrations of amikacin were measured in ultrafiltrates of plasma obtained during the study. It was previously shown that >90% of the free amikacin added to blank plasma or liposomal amikacin-containing plasma was recovered in the ultrafiltrate (data not shown). During the conduct of this study, it was observed that the centrifugal ultrafiltration devices did not always produce a clear, protein-free ultrafiltrate of plasma. It was later determined that the design of the devices used in this study could allow bulk flow around the ultrafiltration membrane when concentrated solutions such as liposome-containing plasma were filtered. Subsequent design changes by the manufacturer have reportedly rectified this problem. As a result, we determined the volume, protein content (Chemstrip 10; Boehringer Mannheim, Indianapolis, Ind.), and visual appearance of each ultrafiltrate. Ultrafiltrates that exhibited evidence of filtration device failure (samples that were cloudy or colored, that had a volume of >200 µl, or that had a protein concentration of >30 mg/dl) were excluded from analysis.

1

View larger version (9K): [in this window] [in a new window]   FIG. 2.   Amikacin concentrations in plasma ultrafiltrates. (A) After administration of a single intravenous dose (20 mg/kg) of low-clearance liposomal amikacin (mean ± standard deviation; n = 3 to 4). (B) Following administration of the last of eight consecutive doses (20 mg/kg) administered at 48-h intervals. The values presented in panel B are medians; some individual values were below the assay's limit of quantitation. The lines indicate the exponential fits used to determine half-lives.

1

Urinary recovery and clearance. Following the administration of a single dose of liposomal amikacin, most of the amikacin that was recovered unchanged in the urine appeared within 24 h (19.4% ± 10.1% of the dose). Amikacin continued to be slowly excreted (<1% of the dose/day), so that by the end of the 2-week study the cumulative urinary recovery had risen to 29.8% ± 10.1% of the dose (Fig. 3A). The mean urinary clearances of amikacin over the course of the study ranged from 0.006 to 0.013 ml min¹ kg¹. Mean CL_CRs were 2.7 to 3.7 ml min¹ kg¹. The urinary clearance of liposomal amikacin was less than 0.4% of the CL_CR and accounted for only 27 to 56% of the total clearance of liposomal amikacin. The urinary recovery of amikacin for one female monkey exceeded those for the other three animals as a result of the recovery of greater amounts (30.5%) during the 0- to 8-h collection period. However, this apparent difference in urinary recovery was not reflected in the values of the animal's plasma pharmacokinetic parameters.

View larger version (13K): [in this window] [in a new window]   FIG. 3.   Cumulative urinary recovery of amikacin (mean ± standard deviation; n = 4). (A) After administration of a single intravenous dose (20 mg/kg) of low-clearance liposomal amikacin. (B) Following administration of the last of eight consecutive doses (20 mg/kg) administered at 48-h intervals.

1

1

1

1

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

The pharmacokinetic profile of a low-clearance liposomal amikacin formulation in nonhuman primates was markedly different from that of conventional amikacin in humans and other animals. While the clearance of conventional amikacin approximates the CL_CR in all species studied, the clearance of liposomal amikacin was over 100-fold lower than the CL_CR in this study with primates. As a result, liposomal amikacin had plasma half-lives (22.9 to 81.7 h) much longer than the half-life of conventional amikacin (2.3 h in humans [4]; 0.2 h in rats [16]). Conventional amikacin has a V close to that of the extracellular space, while the V of liposomal amikacin was lower, approximating the volume of plasma. While these data cannot precisely define the physiologic space to which the liposomes distribute, they suggest that much of the amikacin remains sequestered for an extended period of time in the circulating liposomes, which are slowly cleared from the plasma.

The pharmacokinetic profile of these small, low-clearance liposomes appeared to differ from those of other antibiotic-containing liposomes. A large plurilamellar liposome formulation of gentamicin had an apparent half-life of less than 10 h, with 24-h trough levels generally of <1 µg/ml after the administration of a 5.1-mg/kg dose to humans (29). The free plasma gentamicin level (measured by ultrafiltration) was 33% of the total plasma gentamicin level, suggesting that after administration a much larger fraction of drug leaked from these liposomes than from the low-clearance liposomes that we studied. Formulations of ciprofloxacin in large unilamellar liposomes also had short half-lives (approximately 3 h in mice), with evidence of significant drug leakage (33). Even the clearance of small, sterically stabilized liposomes (0.03 ml min¹ kg¹ [19]) was not lower than that of the low-clearance liposomal amikacin formulation. The clearances that we observed imply that extraction ratios for low-clearance liposomal amikacin could not exceed 1% for any organ (9). This is in contrast to other liposomes, which had hepatic extraction ratios of up to 60% in rodents (20, 27).

The combination of small size and low rate of clearance may represent a significant clinical advantage for the treatment of extracellular infections with liposomal amikacin. Reduced clearance by the mononuclear phagocyte system (MPS) results in a longer residence time in the body, and this may facilitate the accumulation of increased levels of liposomal drug in tissues outside the MPS (16, 33), including sites of inflammation. Reductions in liposomal clearance have been shown to correlate with increased uptake by tumors and other tissues (23), and the permeability of inflamed tissues may even be higher than that of tumors (10). Large liposomes and liposomes that are rapidly cleared by the MPS cannot readily penetrate other sites in the body. Although this study did not measure drug concentrations in tissue, low-clearance liposomal amikacin was shown to increase and prolong the uptake of amikacin by the tissue of rats (16). These observations also suggest the need for extrarenal safety monitoring, since the altered disposition and prolonged residence time of low-clearance liposomal amikacin could result in new target tissues to which the formulation is toxic.

Although peak and trough concentrations increased during repeated dosing with liposomal amikacin, the levels in plasma appeared to approach steady state by the end of the 2-week dosing period. From the clearances that we observed after the repeated-dose regimen, a steady-state amikacin concentration in plasma of 511 ± 56 µg/ml was predicted for this dosing regimen. Low-clearance liposomal amikacin produced very high total exposures in plasma: the levels in plasma remained above 100 µg/ml during the entire dosing regimen and for at least 6 days thereafter. Thus, weekly dosing would be expected to maintain mean levels in plasma of greater than 100 µg/ml. The levels in plasma during the last dosing interval fluctuated by a factor of only 1.4, indicating that a dosing interval of 48 h or longer would not result in large fluctuations in the levels in plasma.

Conventional amikacin is almost entirely excreted unchanged in the urine within hours after administration in all species studied. The urinary excretion of unchanged amikacin was reduced for low-clearance liposomal amikacin (29.8% after the administration of a single dose; 38.5 to 43.6% during repeated dosing). In addition, the time course of urinary excretion was delayed, reflecting the prolonged residence of liposomal amikacin in the body, so that small amounts of amikacin were still detected in the urine 2 weeks after dosing ceased. By interpolating urinary recoveries between the single- and repeated-dose regimens, we estimated that the entire urinary recovery of amikacin during and for 2 weeks after the repeated-dose regimen was 40.3% of the total dose administered. This was in close agreement with the urinary recoveries observed during the last two dosing intervals, when the plasma amikacin levels were approaching steady state. The reduced urinary clearance and recovery that we observed for liposomal amikacin imply that renal tubular exposure resulting from glomerular filtration of free amikacin in plasma was correspondingly reduced for this formulation compared to that for conventional amikacin. Since the proximal tubular epithelium is the principal target of aminoglycoside-induced nephrotoxicity, low-clearance liposomal amikacin may offer an improved safety profile versus that of conventional amikacin. This would especially be true if the liposomal drug is administered less frequently, because in this study exposure to free drug occurred only during the first few hours after the administration of each dose.

The possibility that liposomal amikacin could produce prolonged or toxic concentrations of amikacin in the kidneys by other mechanisms remains to be explored, but increased renal toxicity was not observed after a 1-month exposure to low-clearance liposomal amikacin in dogs (17). In the present study, serum creatinine concentrations and CL_CRs were measured prior to exposure and over the course of each regimen. No clinically or statistically significant (P > 0.05) changes in either parameter were observed as a result of exposure to liposomal amikacin.

The failure to recover more than half of the administered amikacin suggests that the remainder of the dose had been metabolized or distributed to deep compartments within the body. Conventional aminoglycosides, which distribute mainly to the extracellular space, are not metabolized, but there is evidence that aminoglycoside metabolism may occur within some cells (8). The small fraction of conventionally delivered aminoglycosides that penetrates the deep intracellular compartment is released very slowly (31). Both observations imply that if a portion of the liposomal amikacin entered the intracellular compartment, it may have been retained and/or metabolized. Since even low-clearance liposomes are subject to phagocytosis, it is possible that low-clearance liposomal amikacin entered cells by this mechanism.

The convexity observed in the plasma concentration-versus-time curve suggests that a time- or dose-dependent process is involved in the disposition of low-clearance liposomal amikacin. This could represent the recovery of a saturable clearance process or the stimulation of phagocytic activity (2). While the clearance of conventional liposomes is markedly dose dependent (22) and is associated with a blockade of the MPS (13), the clearance of low-clearance liposomal amikacin remains low across a wide range of doses (17). The depletion of opsonin in plasma can also limit liposome clearance (21), and the concentration of surface-adsorbed proteins may be directly related to liposome clearance (7). The much larger surface areas presented by small, unilamellar liposomes compared to those presented by multilamellar liposomes at a given lipid dose suggests that this may play a role in their low rates of clearance, since the available opsonins in plasma would be distributed over a larger surface area. Another mechanism for the low rate of clearance of these liposomes could involve stabilization of the bilayer by encapsulated amikacin, because amikacin has been shown to decrease the fluidity of phosphatidylglycerol-containing bilayers through ionic interaction (6).

A rapid, early phase during which the levels in plasma fell to about two-thirds of their initial values was observed. A similar biphasic disposition has been observed with other liposomes and for the purposes of parameter calculation has sometimes been treated as a distributional phase (30). This yields central-compartment V's close to the plasma volume but steady-state volumes that are up to severalfold higher. Another common approach has been to report plasma liposome concentrations in units of the percentage of the dose in plasma (1). This normalizes the measured concentrations to a constant, assumed V (usually the plasma volume), while it makes no attempt to calculate a V. Unfortunately, neither approach fully addresses the question of the actual liposomal V. Although the proteins in plasma slowly distribute into the extravascular fluid space (11), there is little evidence that particles as large as liposomes could achieve rapid diffusional equilibrium with a fluid space outside the plasma compartment. Thus, it may be reasonable to assume that circulating liposomes are confined to a volume equal to that of plasma (approximately 40 to 50 ml/kg [32]), in which case alternative explanations of the rapid disappearance phase should be considered. These might include a nonuniform distribution within the vascular space (i.e., binding to endothelial surfaces or fixed tissue macrophages), reversible uptake into the lymphatic system, rapid initial uptake or release of liposome contents mediated by complement or other proteins or lipoproteins in plasma, and rapid clearance of a specific fraction of the injected liposomes due to size or other nonhomogeneity within the liposome population. For these reasons, only the V of the central compartment is reported here.

If circulating liposomal amikacin is confined to a V similar to that of plasma, it may seem paradoxical that increased levels of exposure to tissue are observed with this formulation (16). For small drug molecules that diffuse readily across endothelial barriers, more extensive distribution into tissues is reflected in larger observed V's. However, if liposomes enter tissues via phagocytosis or another irreversible mechanism, then such a relationship between the level of exposure to tissue and V would not exist. Studies to define the extent and time course of exposure to tissue after repeated dosing will be required to interpret the altered safety and efficacy profiles of this formulation.

The unique pharmacokinetic profile of low-clearance liposomal amikacin, with a prolonged residence time and altered disposition compared to those of conventionally administered amikacin, has already been shown to alter the safety and efficacy profiles of this antibiotic (14, 17, 34). It has also been observed that the in vitro MICs of low-clearance liposomal amikacin underestimate in vivo susceptibility, as reported for other liposomal antibiotics (3). These facts imply that new pharmacokinetic and pharmacodynamic correlates by which the dose, the frequency and duration of exposure, and the resulting exposures in plasma, tissues, and sites of infection can be related to the intensity, onset, duration, and antimicrobial spectrum of the clinical activity of low-clearance liposomal amikacin may need to be devised.

In conclusion, a small unilamellar liposomal formulation of amikacin had plasma and urinary clearances markedly lower than those of other dosage forms when the formulation was studied in clinically relevant single- and repeated-dose regimens in rhesus monkeys. By sequestering its antibiotic payload in liposomes with long circulation times, this formulation maintained antibiotic levels in the body for over 1 week after treatment. The long plasma half-life and prolonged urinary excretion that we observed suggest that low-clearance liposomal amikacin could be administered as infrequently as once weekly and that it significantly alters the disposition of amikacin within the body while it decreases the potentially toxic level of renal tubular exposure resulting from the glomerular filtration of free amikacin. The level of exposure to free amikacin in plasma was low (<20 µg/ml) and occurred only briefly after the administration of each dose of liposomal amikacin. Decreased dosing frequency and increased safety suggest that low-clearance liposomal amikacin could be clinically useful for outpatient therapy or prophylaxis for serious bacterial infections.