INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Candida species are important causes of invasive fungal infections in neutropenic patients (1, 15). Standard treatment with amphotericin B deoxycholate (D-AmB) is associated with a crude mortality rate of 20 to 50% (3, 9, 21, 22, 25, 35), and the crude mortality rate reaches almost 100% in the presence of deep tissue involvement and persistent neutropenia (8, 9, 20). Although the current antifungal azoles have expanded our therapeutic options, they are not considered first-line agents in the neutropenic patient due to deficiencies in either their pharmacokinetics or their antifungal spectra (7, 16). Thus, there is a continuing need for novel antifungal compounds that have potent, broad-spectrum antifungal activity, favorable pharmacokinetic properties, and tolerable toxicity.

Nystatin was discovered in the early 1950s as the first polyene antifungal antibiotic (18). It binds to ergosterol, the main sterol in the cell membrane of fungi and Leishmania spp., leading to channel formation, efflux of protons and cations, and concentration-dependent cell death (11, 17, 19). Although nystatin has potent, broad-spectrum fungicidal activity in vitro, problems with solubilization and systemic toxicity precluded its development as a parenteral therapeutic agent (17). In the late 1980s, however, laboratory investigators at the University of Texas incorporated nystatin into a multilamellar liposome preparation consisting of dimyristoyl phosphatidylcholine (DMPC) and dimyristoyl phosphatidylglycerol (DMPG) in a 7:3 molar ratio (23). While it preserved antifungal activity in vitro, the liposomal formulation of nystatin (L-Nys) was considerably less toxic to mammalian cells than the free drug (23) and prolonged survival in a murine model of disseminated candidiasis (24). More recently, a multilamellar liposome formulation of nystatin with the same constituents but with a defined particle size of 0.1 to 3 µm (34) has been launched for clinical development. This liposomal formulation has been shown to improve survival in a murine model of disseminated aspergillosis (28) and to provide effective microbiological clearance in a persistently neutropenic rabbit model of invasive pulmonary aspergillosis (12). It was well tolerated in patients without dose-limiting toxicity at dosages of up to 8 mg/kg of body weight/day (4) and is undergoing phase II and III clinical trials in nonneutropenic patients with candidemia, in neutropenic cancer patients who require empirical antifungal therapy, and as salvage therapy for patients with refractory invasive candidiasis and other invasive mycoses.

Little is known, however, about the in vivo activity of L-Nys against deeply invasive candidiasis in the setting of profound and persistent neutropenia. We therefore investigated the efficacy and safety of this novel polyene formulation in a persistently neutropenic rabbit model of subacute disseminated candidiasis and compared it to therapy with amphotericin B deoxycholate and fluconazole.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. New Zealand White rabbits (weight, 2.5 to 3.5 kg; Hazleton, Denver, Pa.) were used for all experiments. They were individually housed and were given water and standard rabbit feed ad libitum according to National Institutes of Health guidelines (5). Nontraumatic vascular access was established in each rabbit by the surgical placement of a subcutaneous silastic central venous catheter (31).

Immunosuppressive regimen and supportive care. Cytosine arabinoside (ara-C; Upjohn Pharmaceuticals, Kalamazoo, Mich.) was administered intravenously at 440 mg/m² on days 1 through 5 and on days 8 to 9 and 13 to 14 to produce profound and persistent neutropenia, respectively. The mean granulocyte counts on the day of inoculation ranged from 154 to 237/µl in all five cohorts (differences were not significant by analysis of variance [ANOVA]) and were below 25/µl on days 13 and 17. Concomitant platelet counts ranged from 13,000 to 26,000/µl between cohorts (the difference was not significant). Starting on day 4, ceftazidime (Glaxo Pharmaceuticals, Research Triangle Park, N.C.) was administered at 75 mg/kg intravenously twice daily, gentamicin (Baxter Health Care Corp., Deerfield, Ill.) was administered at 5 mg/kg intravenously every other day, and vancomycin (Eli Lilly & Co., Indianapolis, Ind.) was administered at 15 mg/kg intravenously daily to prevent the occurrence of invasive bacterial infections during neutropenia.

Test strain. Candida albicans NIH-8621 was from a granulocytopenic patient with autopsy-confirmed disseminated candidiasis and was used for all experiments. The isolate was subcultured from a frozen stock culture (stored at 70°C in a 10% skim milk, 15% glycerol solution) on Sabouraud dextrose agar plates, incubated for 24 h at 35°C, and maintained during the course of the experiments at 4°C.

Preparation of inoculum. The inoculum was grown in Emmon's modified Sabouraud dextrose broth (pH 7.0; Media Department, Clinical Center, National Institutes of Health) as described in detail elsewhere (33), incubated in a gyratory water bath at 37°C for 18 h, and then centrifuged and washed. The concentration of the suspension was adjusted by use of a hemacytometer and was retrospectively confirmed with quantitative cultures of 10-fold serial dilutions. Inoculation was performed on day 6 of the experiment. The inoculum was 10³ CFU and was administered in a volume of 5 ml of sterile normal saline over 1 min via the indwelling silastic central venous catheter.

Antifungal therapy. L-Nys (Nyotran; 50 mg of nystatin USP in a mixture of 350 mg of DMPC and 150 mg of DMPG; Aronex Pharmaceuticals, The Woodlands, Tex.) was provided as a lyophilized powder that was maintained at 4°C. Prior to use the drug was freshly reconstituted with 50 ml of sterile normal saline to a 1-mg/ml solution, as recommended by the manufacturer, and was administered intravenously once daily over 10 min at either 2 mg/kg/day (L-Nys2) or 4 mg/kg/day (L-Nys4). These dosages have been selected for clinical investigation of the efficacy of L-Nys against candidemia and for empirical antifungal therapy in cancer patients with neutropenia and persistent fever. Amphotericin B (50-mg vials; Fungizone, Bristol Myers-Squibb, Princeton, N.J.) was prepared as required and was administered intravenously once daily over 10 min. A dosage of 1 mg/kg/day was used. Flu (2 mg/ml; Diflucan solution for infusion; Pfizer, Groton, Conn.) was administered once daily as a steady intravenous bolus over 1 min at a dosage of 10 mg/kg/day.

Assessment of antifungal efficacy. The main endpoint for the assessment of antifungal efficacy of L-Nys was determination of the quantitative clearance of C. albicans from tissue. Further endpoints included the time to the clearance of the organism as determined with cultures of blood, the definite clearance of the organism at the end of the experiment as determined with cultures of blood, and survival (in days postinoculation). The pattern of infection of subacute disseminated candidiasis, however, permitted survival of nearly all rabbits until the termination of the experiment.

2

4

Monitoring of toxicity. Serum samples were collected from each rabbit before administration of the first dose of ara-C and on days 7, 9, 11, and 17 of the experiment. Samples were stored at 80°C until they were assayed. Clinical chemistry values were analyzed with the last sample drawn from each rabbit and were compared among groups and to the range of values observed in normal, healthy rabbits. A complete blood count was determined before administration of the first dose of ara-C and on day 17 prior to euthanasia.

Statistical considerations. Differences between the means of continuous data among groups were evaluated by nonparametric ANOVA with Bonferroni's correction for multiple comparisons, as appropriate. Differences between proportions were analyzed by the Fisher exact test. All values were tabulated as means ± standard errors of the means (SEMs). A two-tailed P value of 0.05 was considered statistically significant.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Antifungal efficacy. L-Nys demonstrated dose-dependent antifungal efficacy in the treatment of subacute disseminated candidiasis. While all eight control animals were heavily infected at the postmortem examination, for four of eight rabbits treated with L-Nys2 (P < 0.05 versus controls) and none of 8 rabbits treated with L-Nys4 (P < 0.001 versus controls), C. albicans was recovered from at least one tissue site (P < 0.05 for comparison of both dosage groups). In the cohorts of D-AmB-treated and Flu-treated rabbits, one rabbit each was found to be infected (P < 0.001 versus untreated controls; the difference was not significant versus the result for animals treated with L-Nys).

Toxicity. Treatment with L-Nys was associated with significantly less nephrotoxicity than treatment with D-AmB, as assessed by the mean serum creatinine and blood urea nitrogen levels at the end of treatment (P < 0.01 to P < 0.001 for all comparisons) (Table 2). No significant differences between treatment groups were noted with regard to serum potassium levels. There were, however, increases in serum magnesium levels in D-AmB- and L-Nys4-treated rabbits, which reached statistical significance in the latter group, and a trend toward low serum bicarbonate levels in D-AmB- and L-Nys4-treated rabbits. These observations may reflect impaired renal clearance of magnesium and early renal tubular acidosis, respectively.

Survival. All rabbits treated with L-Nys2 and L-Nys4 survived until the termination of the experiment. There was one premature death, on day 10 postinoculation, in each of the D-AmB- and Flu-treated cohorts. Both animals had completely cleared the organism from their bloodstreams and tissues, and there was no evidence of bacterial superinfection. The serum creatinine level was elevated in the D-AmB-treated rabbit but not in the Flu-treated animal. As to be expected by the study design, the majority of untreated control rabbits (n = 5 of 8 [62.5%]) survived until the end of the experiment.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The results of this study demonstrate that L-Nys2 or L-Nys4 is as active as standard dosages of AmB-D or Flu for the early treatment of subacute disseminated candidiasis in persistently neutropenic rabbits. Responses to treatment with L-Nys were dosage dependent, with complete microbiological clearance of all investigated tissue sites by treatment with L-Nys4. Overall, L-Nys was less nephrotoxic than conventional amphotericin B, and rabbits treated with L-Nys displayed no apparent hepatic, hematological, or metabolic toxicities.

These findings are important, as experimental data on the in vivo antifungal activity of the currently developed liposomal formulation of nystatin against Candida species are scant, are limited to survival studies, and have not been published. Our study demonstrates for the first time the principal ability of L-Nys to clear Candida spp. from blood and tissues in a manner equivalent to that of standard therapies. This lends important scientific support to the evaluation of the compound in clinical trials with patients with proven or documented invasive Candida infections. Indeed, in an ongoing phase II and III multicenter study with nonneutropenic patients with candidemia, L-Nys2 showed promising antifungal activity, and the rate of successful treatment with L-Nys4 was comparable to the rate of successful treatment with conventional amphotericin B in a historical control group of patients (10, 27).

The pharmacokinetics of L-Nys in plasma after intravenous administration to rabbits (13) are overall comparable to those in the blood of human subjects reported previously (6). As shown previously (13), on the basis of the area under the plasma concentration-versus-time curve from time zero to 24 h and the observation that concentrations in plasma are approximately twice as high as the corresponding levels in blood (14), a 4-mg/kg dose in a healthy rabbit approximately corresponds to a 2-mg/kg dose in humans and a 6-mg/kg dose in the rabbit approximately corresponds to a 3-mg/kg dose in humans. The up to twofold lower projected mean peak levels in the plasma of humans after the administration of equivalent doses may be accounted for mainly by the relatively rapid infusion rate of 10 min in the rabbit but also by the high interindividual variability in human patients (13). These pharmacokinetic considerations may serve to validate the efficacy data from our experiments with regard to their implications for clinical studies. In a broader sense, they also underscore the essential need for integration of pharmacokinetic data into the preclinical evaluation of investigational drugs.

With the introduction and maintenance of profound neutropenia, broad-spectrum antibiotic therapy, a relatively low inoculum, and the start of treatment within 24 h after inoculation, our infection model most closely resembles the clinical situation in a neutropenic patient who presents with persistent fever and early, occult Candida infection or early phases of overt candidemia, which is probably the most frequently encountered clinical situation. The results of this study cannot be directly applied to preventive modalities and more advanced stages of acute disseminated candidiasis and chronic disseminated candidiasis, for which separate experimental investigations would be required (30). It bears notice, however, that cultures of blood from only three of the five surviving untreated animals were positive for C. albicans at the termination of the experiment, despite uncontrolled and extensive tissue infection. This observation may serve to emphasize the fact that overt candidemia represents only part of the clinical spectrum of invasive candidiasis (2, 9) and that new, non-culture-based diagnostic approaches continue to be urgently needed.

In healthy rabbits, a dosage of 6 mg of L-Nys per kg per day administered intravenously for 14 days produced only mild increases in serum creatinine levels, which did not exceed 1.5 times the baseline levels, and no increases in the blood urea nitrogen values (13). The 4-mg/kg/day dosage was tolerated without signs of nephrotoxicity and was therefore used as the maximum dosage in this study. Although nephrotoxicity occurred in some animals that received 4 mg/kg/day in the setting of disseminated candidiasis, dehydration, catabolism, immunosuppression, and vancomycin and gentamicin therapy, L-Nys was overall significantly less nephrotoxic than D-AmB. Our toxicity data are overall consistent with results from a clinical phase I study with 32 neutropenic patients with hematological malignancies and fever refractory to broad-spectrum antibiotics. The patients received multiple doses of L-Nys at escalating dosages of up to 8 mg/kg/day without dose-limiting nephrotoxicity (4). Similarly, in the ongoing phase II and III study with nonneutropenic patients with candidemia who are receiving L-Nys2, no patient has experienced therapy-limiting nephrotoxicity (27). A doubling of the serum creatinine levels from those at the baseline was observed for 14% of patients; this rate is less than that reported for patients treated with D-AmB (10) and is approximately equivalent to that for neutropenic patients treated with liposomal amphotericin B at a dosage of 3 mg/kg/day and investigated in the setting of empirical antifungal treatment (32).

In conclusion, L-Nys demonstrated potent, dose-dependent antifungal activity in the early treatment of subacute disseminated candidiasis in persistently neutropenic rabbits. It was as effective as Flu and amphotericin B but was less nephrotoxic than the latter. Its demonstrable efficacy and safety as well as its broad spectrum of fungicidal activity support the further clinical investigation of L-Nys for the treatment of candidemia and for empirical antifungal therapy in persistently febrile neutropenic patients.