Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, September 2005, p. 3697-3701, Vol. 49, No. 9
0066-4804/05/$08.00+0 doi:10.1128/AAC.49.9.3697-3701.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Aberdeen Fungal Group, School of Medical Sciences, Institute of Medical Sciences, University of Aberdeen, Aberdeen AB25 2ZD, United Kingdom
Received 6 May 2005/ Returned for modification 4 June 2005/ Accepted 7 June 2005
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
Caspofungin presents particular problems for tests with Aspergillus spp. in vitro. Determination of the minimum effective concentration, i.e., the lowest concentration that least to microscopically evident stunted hyphal growth, is more reproducible than a traditional MIC test (1, 6, 19). However, the recommended method for susceptibility testing of molds with voriconazole and other triazoles is to define MICs on the basis of a complete growth inhibition end point (17). The outcome of tests with combinations of these drugs in vitro is therefore likely to depend on the choice of end point.
Evidence for favorable interactions between caspofungin and voriconazole in a guinea pig model of disseminated aspergillosis was obtained by Kirkpatrick et al. (12). They found that combinations of the two agents but not monotherapy with either agent alone led to a significantly reduced proportion of tissues positive for A. fumigatus in culture. For all other parameters monitored (survival time and mean tissue burdens), no superior efficacy could be demonstrated for the combination over monotherapy.
We undertook to reexamine caspofungin-voriconazole interactions in experimental aspergillosis in an effort to provide a more robust conclusion than was obtained in the previous study (12). The design of interaction experiments in vivo is complicated by several restraints. However, the cited published data gave us confidence that antagonistic effects between caspofungin and voriconazole had never been noted, either in terms of antifungal effects in vitro and in vivo or in terms of toxicity in vivo. We therefore sought a design that involved minimizing the possible therapeutic effects of each drug alone, in an effort to demonstrate enhanced effects of the combination. The guinea pig is the preferred animal for experiments with voriconazole because the drug is rapidly eliminated in mice (24). We used real-time PCR, since it is superior to culture for accurate measurement of tissue burdens of A. fumigatus (3). We used a higher level of intravenous challenge with A. fumigatus and a longer-sustained immunosuppression than in the previous study in an effort to reduce further the potential efficacy of the monotherapies, and we included measurement of levels of the two drugs in plasma to ensure that combination treatment did not adversely affect the pharmacokinetics of either component. Treatment was continued for only 7 days, and consequences of infection monitored until 21 days after challenge. We wanted to use an infecting A. fumigatus isolate that was resistant to one or other of the two drugs, but we were unable to obtain any such isolate. The multiresistant isolate of A. fumigatus described by Warris et al. (25) has the characteristics of the newly described species Aspergillus lentulus (2), and other resistant isolates have not been described.
Our results confirm a benefit of caspofungin-voriconazole combinations over monotherapies, but the benefit is small and suggestive of additive benefit, rather than synergistic interactions between the two agents.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Antifungal agents. Caspofungin and voriconazole pure powders for tests in vitro were the gifts of Merck & Co, Inc. (Rahway, NJ) and Pfizer Central Research (Sandwich, Kent, United Kingdom), respectively. For animal experiments the commercial intravenous formulations of caspofungin (Cancidas) and voriconazole (Vfend) were purchased from the local hospital pharmacy.
Animals. Male guinea pigs (Harlan) weighing 450 to 550 g were housed in groups of four per cage and were supplied with food and water ad libitum. Care, maintenance, and handling of the animals were in accordance with United Kingdom government license conditions for animal experimentation. The guinea pigs were rendered temporarily neutropenic by intraperitoneal injection of cyclophosphamide, 150 mg/kg, 3 days and 1 h before challenge and 3 days after challenge. Neutropenia sustained for 9 days was confirmed by microscopic examination of blood smears. During the period of neutropenia, doxycycline (1 g/liter) was added to the animals' drinking water to minimize the possibility of bacterial infections.
Tests for antifungal interactions in vitro. To obtain data for interactive effects of caspofungin and voriconazole against A. fumigatus, we attempted to determine time-kill curves for cultures inoculated with conidia or hyphae and with growth assayed by ATP measurement, without success. We successfully attempted to measure growth inhibition in microdilution plate assays with the aid of the redox indicator 3-(4,5-dimethyl-2-thiazyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT)(16), but the results so closely paralleled those obtained by spectrophotometry of culture turbidity that we abandoned the extra effort involved in the indicator step. Interactions were finally determined by checkerboard layouts of doubling dilutions of caspofungin and voriconazole at final concentrations of from 1.0 to 0.0078 µg/ml in microdilution plates. Cultures were inoculated to 104 conidia/ml in antibiotic medium 3 (Becton Dickinson, Cowley, Oxford, United Kingdom) and incubated at 35°C for 48 h. We favored this medium over RPMI 1640 because it supports higher growth yields of A. fumigatus, giving a clearer spectrophotometric distinction between control and inhibited growth. Turbidity in the wells was measured by spectrophotometry at 405 nm, and growth was expressed as a percentage of control growth in drug-free wells, corrected for background absorbance. For each test isolate the fractional inhibitory concentration (FIC) index was calculated (4) for inhibition end points of 50% and 20% of control growth.
Determination of fungal burdens in kidneys. The quantitative PCR (qPCR) method was based on that described by Bowman et al. (3). Guinea pig kidneys were weighed, 3.6 volumes of sterile saline were added per gram of kidney, and the mixture was homogenized with an Ultra-Turrax apparatus. A 900-µl sample of homogenate was transferred into a 1.8-ml screw cap microcentrifuge tube, and 0.3 g of 0.5-mm glass beads was added. Samples were mechanically disrupted by vigorous agitation in a bead beater (three 1-min bursts, with incubation on ice in between to ensure that samples did not overheat and destroy DNA). A secondary homogenate was collected by centrifugation at 800 x g at 4°C for 5 min. DNA was extracted from the secondary homogenates with the QIAGEN (West Sussex, United Kingdom) DNeasy tissue kit, according to the manufacturer's instructions for animal tissue extraction, and the DNA was eluted from the DNeasy column in 200 µl of elution buffer. DNA samples were stored at 20°C.
Oligonucleotide primer sequences and a dual-labeled fluorogenic oligonucleotide probe sequence were as previously described (3). All qPCR reagents and consumables were purchased from Applied Biosystems (Warrington, Cheshire, United Kingdom). DNA samples were analyzed in triplicate with the ABI Prism 7700 sequence detection system. Each 25-µl qPCR mixture contained 5 µl DNA sample, 1x TaqMan Universal PCR Mastermix, and primers and probe at final concentrations of 900 nM and 200 nM, respectively. Reactions were performed in MicroAmp optical 96-well reaction plates sealed with MicroAmp optical caps. qPCRs were run according to the manufacturer's recommendations. For each sample, the cycle number where the fluorescence exceeded the threshold level assigned by the analysis software was determined. This value was used to calculate conidial equivalents from a standard curve generated from known numbers of conidia spiked into uninfected kidney homogenate. Samples for standard curve preparation were prepared in parallel with experimental samples and run in triplicate on each 96-well qPCR plate. All qPCR results for samples are expressed as conidial equivalents per gram of tissue.
HPLC for determination of caspofungin and voriconazole in guinea pig plasma. Resources for this project did not allow for a full-scale development of a high-pressure liquid chromatography (HPLC) method able to determine both agents in a plasma sample. The method used was based on a simple procedure described for detection of voriconazole alone (7). However, guinea pig plasma samples extracted with acetonitrile proved to contain considerable UV-absorbing material not found in human plasma. Our final conditions satisfactorily detected caspofungin, but the sensitivity of the assay for voriconazole was considerably reduced from that of the original published method. The column was a 10-cm-by-4.6-mm (inner diameter) column packed with 3-µm BDS-C18 Hypersil (Alltech Associates, Carnforth, United Kingdom) at 25°C in an Agilent series 1100 apparatus. The running buffer was 45% acetonitrile and 55% 10 mM sodium phosphate, pH 3.0, run at 8 ml/min. Caspofungin concentrations were determined from peak areas detected by fluorescence with an excitation wavelength of 224 nm and a detection wavelength of 302 nm. The retention time was 2.9 min, and the maximum sensitivity was 1 µg/ml. Voriconazole concentrations were determined from peak areas detected by UV absorption at 255 nm with a retention time of 3.6 min; the maximum sensitivity was 5 µg/ml. Samples of guinea pig plasma were extracted with 2 volumes of acetonitrile and centrifuged at full speed in a microcentrifuge for 10 min. The supernatants were applied to the column in 15-µl sample volumes.
Animal model of aspergillosis. The temporarily neutropenic animals were injected intravenously with conidial suspensions of A. fumigatus J980659/3 at challenge doses of 104 CFU/g and 103 CFU/g (viable counts were verified by culture of samples from the inoculum suspension). Animals were monitored twice daily, and any that showed signs of severe illness or distress were euthanatized and recorded as having died on the following day. Tissue burdens were determined by viable counting at the time of demise of an animal. Pilot experiments showed that the challenge doses led to mean survival times of approximately 3 and 4 days, respectively. For 7 days, starting at 1 h postchallenge, animals were treated twice daily (b.i.d.) with voriconazole by gavage and once daily with caspofungin by intraperitoneal (i.p.) injection, or with either drug alone along with intraperitoneal injection of sterile saline as a caspofungin placebo for voriconazole-treated animals or gavage with 17% hydroxypropyl-ß-cyclodextrin as a voriconazole placebo for caspofungin-treated animals. One group of animals received both placebo treatments and served as controls.
Pilot experiments with groups of four animals indicated that voriconazole treatment at 5 mg/kg for 7 days led to survival to day 21 of all animals given the high challenge dose, while at 1 mg/kg voriconazole showed only marginal prolongation of survival and no reduction in organ burdens. Caspofungin at 1 mg/kg did not significantly prolong survival or reduce kidney burdens. The experiments were therefore done with both agents at doses of 1 mg/kg.
Results were aggregated from two separate experiments for each challenge dose. Cages were randomly assigned to receive treatments on all occasions. Initial total group sizes were 12 for each treatment, but some animals were terminated prior to or within 1 day of challenge with signs of illness resulting from the immunosuppressive treatment. These animals were excluded from follow-up and analysis.
Blood samples were obtained by ear pricks at various times relative to treatments to allow estimation of levels of voriconazole and caspofungin in plasma.
Analysis of data. Survival curves were compared by the log rank statistic with a Kaplan-Meier analysis conducted with the SPSS statistical package. Organ burden data were compared in two-tailed t tests done with Microsoft Excel software.
| RESULTS |
|---|
|
|
|---|
Effects of caspofungin and voriconazole alone and in combination in experimental aspergillosis. Survival curves for the two challenge doses and the four therapeutic regimens are shown in Fig. 1. For animals challenged at the higher dose of 104 conidia/g body weight (Fig. 1a), Kaplan-Meier/log rank analysis showed significant prolongation of survival over placebo-treated controls for the group treated with caspofungin at 1 mg/kg (P = 0.002) and for the group treated with the combination therapy (P = 0.004). No other comparisons with these survival data showed statistically significant differences. For the animals challenged with 103 conidia/g (Fig. 1b), survival was significantly prolonged over that of control animals for the groups treated with caspofungin (P = 0.001), voriconazole (P = 0.014), and the combination (P < 0.001). The combination treatment also extended survival beyond that of the caspofungin group at the P = 0.048 level, while the survival difference between the voriconazole-treated and combination therapy groups was not significant.
|
|
Levels of caspofungin and voriconazole in plasma. Our method for HPLC detection of voriconazole was unable to detect concentrations of this agent lower than 5 µg/ml, and no sample contained detectable voriconazole under these test conditions. Neither voriconazole nor caspofungin was detectable in any sample from animals given placebo treatment. Levels of caspofungin were less than 1 µg/ml in three plasma samples collected just before treatment with caspofungin alone and in combination with voriconazole. In all samples taken 2 h after treatment with caspofungin alone or the caspofungin-voriconazole combination on days 2, 3, and 7, caspofungin was measured at detectable levels. Caspofungin levels (mean ± standard deviation) for animals receiving caspofungin alone were 6.2 ± 2.0 µg/ml on day 2 (n = 3), 3.4 ± 2.2 on day 3 (n = 3), and 5.8 µg/ml day 7 (n = 1). For animals treated with the caspofungin-voriconazole combination, the equivalent results were 5.7 ± 0.4 µg/ml on day 2 (n = 5), 3.3 ± 2.0 on day 3 (n = 3), and 8.4 ± 3.3 on day 7 (n = 4).
| DISCUSSION |
|---|
|
|
|---|
The survival curves in Fig. 1, the kidney burden data for the higher challenge dose in Table 1, and the level of negative conventional kidney cultures for the animals challenged at the higher dose all self-evidently reveal a trend towards higher efficacy for combination therapy than for either monotherapy. However, it is equally self-evident that the extent of the advantage for the combination therapy is small in real terms. We interpret the data as indicating a small extra benefit for the combination therapy rather than a truly synergistic interaction, at least with the challenge strain and experimental conditions we used. Since we also found no synergy between caspofungin and voriconazole against A. fumigatus in our tests in vitro, we consider that our conclusion of minor additional benefit in vivo is supported by the available evidence. The previous report of synergistic interactions in vitro (20) showed FIC indices of less than 0.5 in 14/24 A. fumigatus isolates tested. The lowest FIC index for the species was 0.26. These data do not indicate particularly potent synergy. Considered in conjunction with the data from the present study and two others that found no synergistic interactions (5, 15), they suggest that synergy between caspofungin and voriconazole at a level that would suggest a strong potential clinical benefit for the combination seldom occurs. The emergence of isolates of A. fumigatus that are resistant to either agent might alter this conclusion.
Checkerboard methodology for determining antifungal interactions is relatively unsophisticated and prone to methodological and interpretive difficulties (4, 18). Killing curves and surface boundary analyses are likely to provide better information on antifungal interactions, but we were unable within the course of this study to devise reliable systems for such determinations with A. fumigatus.
The tissue burdens of A. fumigatus measured by qPCR were uniformly higher than were seen by culture, in which some organs were negative, particularly those from animals that received the combination treatment. The culture negativity parallels data reported by Kirkpatrick et al. for a similar animal model (12). The higher sensitivity of detection of A. fumigatus DNA offered by PCR reduces the apparent quantitative therapeutic effect of the antifungals tested, although significant reductions of burden in treated animals were consistently detected.
The high mean plasma caspofungin levels that we measured 2 h after i.p. treatment with monotherapy and in combination with voriconazole showed no obvious pharmacokinetic disadvantage with respect to the combination treatment. Peak levels in plasma well in excess of 10 µg/ml have been determined radiologically after intravenous administration of caspofungin in other animal species (22), and our data are compatible with levels in plasma determined by HPLC for other animal species dosed at 1 mg/kg (8, 26). Published data for voriconazole in guinea pigs indicated a peak level in plasma of 4 µg/ml after a 10-mg/kg oral dose (21). It is therefore unsurprising that we were unable to detect voriconazole with our low-sensitivity assay in the plasma of guinea pigs given 1/10 this dose.
Our data confirm the known efficacy of caspofungin and voriconazole in the treatment of experimental aspergillosis and indicate that the small efficacy benefits seen from combination therapy are not of a genuinely synergistic nature.
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Clin. Vaccine Immunol. | Clin. Microbiol. Rev. |
|---|---|
| J. Clin. Microbiol. | ALL ASM JOURNALS |