Activity of Posaconazole Combined with Amphotericin B against Aspergillus flavus Infection in Mice: Comparative Studies in Two Laboratories

Posaconazole and/or amphotericin B was given to mice pretreatedwith a steroid and then infected by inhalation of Aspergillusflavus conidia. Two laboratories conducted studies using almostidentical protocols to evaluate both survival and lung tissueburdens 8 days after infection. The results of the in vivo studiesperformed at both laboratories were consistent. We found that(i) up to 5 mg of amphotericin B per kg of body weight was poorlyeffective in treating invasive aspergillosis; (ii) posaconazoleat 2 or 10 mg/kg/dose prolonged survival and reduced lung tissueCFU; and (iii) there was generally no antagonistic interactionof the drugs in combination, even when the experiments weredesigned to maximize the likelihood of antagonism. These studiesdo not confirm the antagonistic interaction of triazoles andpolyenes reported by others.

INTRODUCTION

Top

Introduction

Acute invasive aspergillosis (AIA) is one of the most fearedinfections of immunosuppressed patients. Both corticosteroidsand neutropenia predispose to this infection, which almost invariablyoccurs after inhalation of infectious conidia (1). The speedof dissemination or pulmonary spread of the infection, and ultimatelysurvival, depend in large part on the nature and severity ofthe predisposing host immune defect(s). In patients with themost fulminating forms of AIA, it has been difficult to demonstrateefficacy of antifungal therapy (7, 11). To date, there havebeen three approaches in the management of aspergillosis. Thefirst is reversal of the predisposing conditions if possible.The second is antifungal therapy, and the third is resectionwhen possible (18). Animal studies have been increasingly usedto help determine whether a particular antifungal drug is effectiveagainst invasive aspergillosis. Some advantages of animal studiesinclude (i) information in ascertaining the relative efficacyand dose-dependent toxicity of antifungal agents; (ii) assessingthe contributions of immune deficiency to the outcome of aspergillosis;(iii) evaluating combinations of antifungals; and (iv) controllingfor a variety of conditions that are thought to contribute tohuman invasive aspergillosis and its outcome but that cannotbe controlled in clinical studies.

The experiment quoted most often, an experiment using amphotericinB and ketoconazole, was conducted by Schaffner and Frick (15)some years ago. When mice with AIA were treated with ketoconazolebefore amphotericin B, Schaffner and Frick noted a marked decreasein the efficacy of amphotericin B (15). They theorized thatthe azole blocked the synthesis of the ergosterol target necessaryfor the binding of amphotericin B. Polak et al. also found somewhatsimilar results (13, 14). These studies led to the concern thatazole antifungal agents would antagonize the effects of amphotericinB in humans and that therefore they should not be used in combinationwith amphotericin B clinically.

In a large review of 595 cases of aspergillosis, Patterson etal. found that patients treated with amphotericin B and thenwith itraconazole fared no worse than those given amphotericinB alone (11). However, most of these patients had begun amphotericinB therapy before itraconazole was administered. The sequenceused by Schaffner and Frick has seen limited clinical use. Thisis likely to change. The use of itraconazole antifungal prophylaxisand the recent phase III study of voriconazole, indicating thesuperiority of voriconazole over amphotericin B in AIA, allsuggest that we shall be seeing increasing numbers of patientswho receive a triazole before amphotericin B (R. Herbrecht,D. W. Denning, T. F. Patterson, W. V. Kern, K. A. Marr, D. Caillot,E. Theil, and P. Ribaud, Abstr. 41st Intersci. Conf. Antimicrob.Agents Chemother. 41:378, 2001).

The present studies were undertaken (i) to check for interactionsof amphotericin B and posaconazole and (ii) to explore the potentialreproducibility of triazole-polyene combinations in the sameanimal model of AIA studied in two different laboratories. Inorder to control the variables and to limit the number of animalsused, we chose to use infection by inhalation of conidia, becausethis most closely approximates clinical infection. We chosecorticosteroids for immune suppression because they predisposeto a lethal infection in mice and are a major predisposing factorfor AIA in humans. We anticipated that the results of thesestudies would give us some guidance on whether new triazoleswith activity against Aspergillus can be used safely togetherwith amphotericin B or should not be used in specific circumstances.Such information may ultimately help guide the use of theseagents in clinical therapy.

MATERIALS AND METHODS

Top

Materials and Methods

Sites Studies were conducted at two different laboratories, ScheringPlough Research Institute (SPRI) and The University of TexasHealth Science Center at San Antonio (UTHSCSA). Both institutionsutilized as close to identical methods as was possible.

Description of model The pathogen used was Aspergillus flavus strain ND83, obtainedfrom SPRI. This is available from the American Type CultureCollection as ATCC MYA-1004. Prior to infection, 11- to 13-gCF-1 (SPRI) or outbred ICR (UTHSCSA) male mice were treatedwith cortisone acetate (100 mg/kg of body weight) given subcutaneouslyon day -1, day 0 (day the mice were infected), and day +1. One-literErlenmeyer flasks with eight circumventing tubes were used forthe inhalation infection. Each tube was approximately 4 in.long and had a 1-in. internal diameter, and the tubes were placedequidistant from each other. A small, 3- to 4-mm-long, orificewas made in the wall of the flask at the center of each tubeto allow communication between the tube and flask. The outerends of the tubes around each flask were plugged with cotton,and the flasks were sterilized. This inhalation chamber is aSPRI modification of the chamber first described by Piggot andEmmons (12) and routinely used at SPRI (8).

Two weeks before infection, 150 ml of liquid malt extract (SPRI)or potato flake (UTHSCSA) agar was added to the sterile flask.The agar was allowed to cool 3 h or longer and then heavilyinoculated with A. flavus conidia. A rubber stopper with a glasstube, curved 45 degrees at the lower end (2 to 4 cm above theagar) and a second central glass tube was placed on top of theagar, and the flask was incubated for 13 days at room temperatureto allow a dense mycelial mat to form.

At the time of infection, eight unanesthetized mice were insertedinto the tubes. One mouse was inserted into each tube. Eachmouse was inserted far enough that their nostrils protrudedinto the central chamber. The tubes were plugged with cottonto prevent the mice and inoculum from leaving the tube. A 60-mlsyringe was attached to the central glass tube, and one vigorousblast of air was delivered into the chamber by plunging thesyringe. This disrupted the mycelia, forming a cloud of conidia.Mice were allowed to breathe the conidial suspension for 1 minand were then rapidly removed from the chamber. The tubes werethen refilled with a second cohort of mice. The glass tube wasrotated so that the end was pointed at undisrupted mycelia,and the infection process was repeated. A third cohort of micewas similarly exposed. No more than three cohorts of mice wereused for any one flask.

From each chamber, one mouse was sacrificed at 1 h after pulmonaryinfection for quantitative counts of Aspergillus conidia, doneby serial colony count dilutions of homogenates of the totallungs. These were considered the baseline counts. In order toexamine the natural course of this infection, in two studies,two mice from each flask were also sacrificed at days 3 and5 after infection. Lungs were removed and homogenized in 2 mlof saline, and serial dilution colony counts were made.

To minimize variability from different flasks, mice from eachflask were randomly assigned to different cages, and the cageswere randomly assigned to control or treatment groups. Therewere 8 to 12 mice/group. Treatment began 24 h after infection.A clinical oral suspension of posaconazole was obtained fromSPRI. Further dilutions were made in sterile water. Both laboratoriesused amphotericin B (Fungizone [Bristol-Myers Squibb Co., Princeton,N.J.]) prepared according to the manufacturer's directions.UTHSCSA further diluted amphotericin B with 5% dextrose, whileSPRI further diluted amphotericin B in sterile water for injection.In some studies, amphotericin B, posaconazole, or both weregiven on day 1 after infection (concurrent treatment). In otherstudies, posaconazole therapy was begun on day 1, and amphotericinB therapy was not begun until day 2 (sequential treatment).Drug doses used included 2 or 10 mg of posaconazole per kg administeredonce a day orally or 1 or 5 mg of amphotericin B per kg administeredintraperitoneally once a daily. Treatment continued throughday 7. For tissue burden studies, quantitative cultures of thelungs of mice were performed for all mice when they died orwere sacrificed on day 8. For survival studies, treatment continuedthrough day 7, and mice were terminated on day 8.

In vitro MICs were determined by using the National Committeefor Clinical Laboratory Standards (NCCLS) method modified formycelial fungi, with incubation at 24 and 48 h for end points(10). At 48 h, the posaconazole MIC was 0.06 µg/ml, andthe amphotericin B MIC was 2 µg/ml.

Statistics The Mann-Whitney U test was used to compare tissue counts. Thelog rank test was used to compare survival of the groups. Becauseeach group was compared both with the untreated control andwith groups given the individual drugs for efficacy, a P of<0.03 was used to determine significant differences.

RESULTS

Top

Results

It was our initial intention to study AIA with Aspergillus fumigatus,as this is the most significant pathogen of the genus Aspergillus.In preliminary studies, eight isolates of A. fumigatus wereevaluated using the inhalation model. A. fumigatus caused afulminating infection which was lethal too quickly to allowstudies of combined antifungal therapy. In other models, posaconazoleis effective against A. fumigatus (3).

A less virulent species, A. flavus, was then chosen. Using isolateND83 (ATCC MYA-1004), we found that untreated mice usually succumbed6 to 10 days after infection. To determine the actual inoculumsize, the quantity of viable conidia inhaled was measured. Therefore,in several studies, one group of mice was sacrificed 1 h afterinfection, and the lungs were quantitatively cultured. As shownin Fig. 1A (data from UTHSCSA), the fungal burden fell about1 log unit between 1 h and 1 day after infection and was thenstable for 5 days after infection. However, as the conidia germinateinto mycelia within 24 h, measurements after 1 day were in essencefragmented mycelia from lung abscesses. Figure 1B shows theresults of a typical experiment at SPRI and UTHSCSA, using micetreated with water (controls). Note that the median tissue countfor the SPRI mice was 1 log higher than for the UTHSCSA mice.We could not determine whether this was due to more vigorousdispersion of the conidia, more fragile mycelial cultures, orother aspects of culture. This systematic difference was maintainedfrom 1 h to 8 days after infection.

View larger version (14K):
[in this window]
[in a new window]
FIG. 1. Total lung tissue burdens of mice infected with A. flavus and sacrificed after infection. (A) Mice were infected with 5.57 log₁₀ CFU/lung (geometric mean) and sacrificed at 1 h, 1 day (D1), 3 days, and 5 days after infection. (B) Mice were infected with 6.34 log₁₀ CFU/lung at SPRI and 4.99 log₁₀ CFU/lung at UTHSCSA (both values are geometric means). Lung tissue burdens were measured 1 h after infection. The lung tissue burdens in the controls were measured at the time of death or 8 days after infection. Data from two studies each were combined. Control mice were treated orally with water (0.2 ml/day) from days 1 to 7.

Survival Survival studies were conducted with posaconazole given at 2or 10 mg/kg/day and with amphotericin B given at 1 or 5 mg/kg/day.Figure 2 presents the results of posaconazole (10 mg/kg/day)and amphotericin B (5 mg/kg/day) therapy, with both drugs begunconcurrently 1 day after infection. Control mice at SPRI, inoculatedmore heavily, began to succumb about a day earlier than thoseat UTHSCSA. Mice at SPRI and UTHSCSA were not protected by amphotericinB. This result is similar to the results of previously publishedstudies where we found inconsistent protection by amphotericinB in steroid immunosuppressed animals challenged intranasallywith A. fumigatus (3). Posaconazole gave almost 100% survivalboth at SPRI and UTHSCSA when it was given alone at a dose of10 mg/kg or combined with amphotericin B (5 mg/kg). Becauseof the high efficacy of posaconazole, this study would havebeen able to detect only antagonism, not any additive effectsof the combination therapy. There was no antagonism seen withthe combination.

View larger version (14K):
[in this window]
[in a new window]
FIG. 2. Concurrent therapy survival study. The results of survival studies with posaconazole (10 mg/kg) ( ${blacksquare}$ ), amphotericin B (5 mg/kg) ( ${blacktriangleup}$ ), and the combination of posaconazole and amphotericin B (•) are shown. Both drugs were begun concurrently 1 day after infection and continued daily through day 7 postinfection. Posaconazole (10 mg/kg) and combined posaconazole-amphotericin B treatment prolonged survival over that of controls ( ${blacklozenge}$ ) at both laboratories (P $<=$ 0.03). Amphotericin B was ineffective in prolonging survival. The combination of posaconazole and amphotericin B gave results similar to those of POS given alone.

Lower drug doses were also studied (results not shown). An amphotericinB dose of 1 mg/kg conferred no protection to mice when the survivalof treated and control mice was compared (P > 0.3 for bothlaboratories). Mice receiving posaconazole (2 mg/kg) were significantlyprotected compared to the controls, with survival percentagesof 60 and 67 at the study termination (P < 0.008 for bothlaboratories). At SPRI and UTHSCSA, mice concurrently treatedwith both posaconazole and amphotericin B had survival ratessimilar to those given posaconazole alone (P > 0.4). Therefore,at these lower drug doses, there were neither additive nor antagonisticeffects.

Tissue burden Figure 3 shows the lung tissue burden results of mice treatedwith the same doses used in Fig. 2. Two studies done at UTHSCSAwith similar results were combined, while only one study wasdone at SPRI. The mean control counts were 5.95 log₁₀ CFU/lungat SPRI and 4.66 log₁₀ CFU/lung at UTHSCSA. Posaconazole treatmentat 10 mg/kg significantly reduced tissue burden levels belowthose of the control mice in both laboratories. At SPRI, amphotericinB treatment at doses of 5 mg/kg (5.88 log₁₀ CFU/lung) did notsignificantly reduce the tissue counts relative to those ofcontrols, but at UTHSCSA, the same dose of amphotericin B significantlyreduced the mean CFU to 4.31 log₁₀ CFU/lung. Mice treated withthe combination of posaconazole at 10 mg/kg and amphotericinB at 5 mg/kg had counts significantly below those of controlmice and mice given amphotericin B alone. Both at SPRI and UTHSCSA,the combination did not give results not significantly differentfrom the results when posaconazole alone was given. These studieswere also done with posaconazole at 2 mg/kg and amphotericinB at 1 mg/kg. The results (not shown) were qualitatively similarto those at the higher doses.

View larger version (14K):
[in this window]
[in a new window]
FIG. 3. Concurrent therapy tissue burden survival study. The lung tissue burden results of mice treated with posaconazole at 10 mg/kg (POS 10), amphotericin B at 5 mg/kg (AMB 5), or a combination of POS and AMB (COMBO). Both drugs were begun concurrently 1 day after infection and continued daily through day 7 postinfection. POS and COMBO groups reduced lung fungal burden from those of the control groups at both laboratories. AMB reduced counts only at UTHSCSA and not at SPRI. The COMBO results were similar to those of POS alone. Values that are significantly different (P $<=$ 0.03) from the control values are indicated by an asterisk.

Figure 4 shows the results of survival studies in mice treatedsequentially. In these studies, posaconazole treatment (10 mg/kg/day)was initiated on the first day after infection, and amphotericinB treatment (5 mg/kg/day) was begun on the second day afterinfection. This was similar to the initiation of ketoconazoletreatment before amphotericin B treatment in the studies ofSchaffner and Frick (15). Sequential therapy survival resultsat SPRI were similar to those given concurrent therapy (Fig.2). Control mice survived a median of 5 days in both laboratories.Posaconazole treatment at both laboratories prolonged survivalto 8 days (P < 0.0001). Amphotericin B treatment was ineffectiveat SPRI (median survival time, 5 days; P = 0.5010) but prolongedsurvival significantly over that of controls at UTHSCSA (mediansurvival time, 8 days; P = 0.0095). Combination therapy recipientshad significantly prolonged survival compared to the survivalof controls at SPRI and UTHSCSA, but not compared to mice givenposaconazole alone.

View larger version (14K):
[in this window]
[in a new window]
FIG. 4. Sequential therapy survival study. The results of survival studies in mice treated sequentially with posaconazole at 10 mg/kg initiated on day 1 after infection ( ${blacksquare}$ ), amphotericin B at 5 mg/kg begun on day 2 after infection ( ${blacktriangleup}$ ), or the combination of posaconazole and amphotericin B (•). Therapy continued daily for 7 days. The groups of mice treated with posaconazole and the posaconazole-amphotericin B combination showed prolonged survival compared to control groups ( ${blacklozenge}$ ) (P $<=$ 0.03) at both laboratories. However, amphotericin B at 5 mg/kg prolonged survival (P $<=$ 0.03) only at UTHSCSA. The survival of mice treated with the posaconazole-amphotericin B combination was similar to that of mice treated with posaconazole alone.

Both laboratories found prolonged survival for mice given thelower dose of posaconazole (2 mg/kg) compared to that of thecontrols (median survival time, 8 days at SPRI and 7 days atUTHSCSA; P < 0.0001) (data not shown). Amphotericin B at1 mg/kg was not protective. In both laboratories, the survivaltimes of mice given the combination of posaconazole (2 mg/kg)and amphotericin B (1 mg/kg) were similar to that of mice givenposaconazole alone.

Figure 5 shows lung tissue counts in mice treated sequentially.Treating mice with 5 mg of amphotericin B per kg did not affectthe lung tissue counts of mice compared with those of controlmice at both laboratories. At SPRI and UTHSCSA, treating micewith 10 mg of posaconazole per kg significantly lowered lungtissue burden from that of the controls. The lung tissue countsof mice given combination therapy were significantly lower thanthe counts of the controls but similar to the counts for micegiven posaconazole alone.

View larger version (16K):
[in this window]
[in a new window]
FIG. 5. Sequential therapy tissue burden study. Lung tissue counts in mice treated sequentially with posaconazole at 10 mg/kg (POS 10) initiated on day 1 after infection, amphotericin B at 5 mg/kg (AMB 5) begun on day 2 after infection, or the POS-AMB combination (COMBO). Therapy continued daily for 7 days. Both POS- and COMBO-treated groups showed prolonged survival (P $<=$ 0.03) over that of controls at both laboratories. AMB was not effective. Results with COMBO treatment were similar to those of POS alone. Values that were significantly different from the control values (P $<=$ 0.03) are indicated by an asterisk.

In summary, we found generally similar results for studies performedin the two laboratories. Posaconazole treatment (2 or 10 mg/kg)significantly prolonged survival and reduced lung tissue burdenof A. flavus compared with those of controls. Amphotericin Btreatment (1 or 5 mg/kg) was irregularly effective and lessprotective than posaconazole treatment alone. Combined therapy,either concurrent or sequential, gave similar results to posaconazoletherapy alone. Therefore, in these experiments there were noantagonistic or additive effects of amphotericin B (in dosesup to 5 mg/kg) on posaconazole.

DISCUSSION

Top

Discussion

The present studies illustrate both the benefits and limitationsof animal studies in reproducing clinical disease. One of thebenefits of the present studies was that the natural inhalationroute of infection was used in mice pretreated with corticosteroids(a predisposing factor). This is similar to the clinical environment.The target organ is the lungs, just as in clinical disease.However, A. fumigatus is a more significant cause of clinicaldisease than A. flavus. In the present inhalation model, wewere unable to use A. fumigatus due to its consistently highvirulence and had to rely on A. flavus, a less common pathogen.

Another benefit of animal studies is the standardization ofinoculum size and use of reference isolates of fungal pathogens.Standardizing these variables is necessary for the reproducibilityof experimental aspergillosis, but the clinical outcome of aspergillosisdepends in part on the inoculum size and virulence of the infectingfungal isolate. Thus, at the time treatment is initiated, patientsmay have much heavier fungal burdens and later stages of fungaldisease than animal models can reproduce.

One advantage of animal models is that the influence of specificpredisposing immune defects can be studied. However, patientsoften have combined immune deficiencies (i.e., neutropenia andsteroid administration). Most animal models are limited by useof a single immune deficiency (neutropenia or steroid administration).

Additional considerations include variable drug absorption,distribution, and metabolism among animal species. As one example,voriconazole (and to a lesser degree itraconazole) is clearedso rapidly in mice that it is very difficult to show efficacy,whereas in guinea pigs, clearance is slow and voriconazole isefficacious in aspergillosis (4, 16, 17). In the present studies,we used posaconazole, which is cleared slowly in mice and ismore effective than either voriconazole or itraconazole.

Finally, in animal models, both the doses and sequence of drugsin combinations can be tightly controlled. However, patientsoften present with a treatment history including multiple prioror current antifungals.

All of these factors, while allowing us to reproduce a modelof a serious fungal infection in vivo and manipulate it in multipleways, significantly limit the degree to which one can extrapolateresults of animal studies to the clinical situation.

Finally, there is the as yet unanswered question of reproducibilityof in vivo studies. The NCCLS has spent considerable effortin standardizing in vitro antifungal testing, but little hasbeen done in vivo (9, 10). We considered that a similar effortat standardization might be undertaken using an in vivo murinemodel of aspergillosis, with single and combination drug regimens.Even with multiple controlled parameters are used in animalstudies, the same drug may yield highly disparate results indifferent laboratories. This may be due to different Aspergillusisolates, inoculation routes, treatment regimens, methods ofassessing outcome, or strains or species of animals used (2).These problems are further confounded when one attempts to evaluatecombinations of antifungals. To search for additive effectsin combination studies, one must use a suboptimal dose of eachdrug and then show that the combination is superior to eachcomponent. This is often very difficult to achieve, especiallywhen drug efficacy depends highly on inoculum size and timingof initiation of therapy in a very short window. Conversely,to search for antagonism, one should use highly effective regimensand then seek to determine whether efficacy is compromised bythe combination. Therefore, the sequence of drugs further influencesresults (5, 6).

In recent years, there has also arisen an ethical emphasis onlimiting the numbers of animals in each treatment group andin limiting the repetition of studies for confirmation and statisticalsignificance. Therefore, it is not surprising that differentlaboratories may report additive, neutral, or antagonistic effectswith the same drugs used against the same fungal species.

In the present studies, we found that similar protocols generallygave similar results in the two different laboratories. Threesystematic variables, different mouse strains, different mediaused for culture of the A. flavus isolate, and a higher inoculumat SPRI (approximately 10 times higher), did not affect theoverall similarity of responses to antifungal therapy. In bothlaboratories, treatment results were strikingly similar. Posaconazoletreatment at doses of 2 and 10 mg/kg was effective, prolongingsurvival and reducing lung tissue count. Treatment with amphotericinB alone at both 1 and 5 mg/kg was mostly ineffective. In theinhalation model of steroid-immunosuppressed mice, amphotericinB is often poorly effective (3). Both drugs thus performed similarlyat both laboratories.

These studies are also the first to evaluate the combinationof posaconazole and amphotericin B, using both concurrentlyinitiated and sequential treatment regimens. The sequentialstudies were performed in regimens set to maximize the potentialfor antagonism, per the studies of Schaffner and Frick (15),Kontoyiannis et al. (5), and Lewis et al. (6). In the presentstudies, there was no evidence that amphotericin B treatment,at doses as high at 5 mg/kg, had an antagonistic effect to posaconazole.

It is not clear why the results of our studies conflict withthose of Kontoyiannis et al. (5). We are assured by our results,conducted both with concurrently initiated and with sequentialtherapy, using both survival and reduction of lung tissue burden,that there is no antagonism of the triazole antifungal posaconazoleand the polyene amphotericin B. Therefore, we conclude thatin vivo models can be standardized to give predictable resultsin murine aspergillosis. Just as in vitro studies, a numberof parameters need to be standardized as far as possible.

How do the results of these studies translate to the clinician?First, they should encourage the clinical investigator thatposaconazole has potential as an effective antifungal againstAIA. AIA carries such a high mortality that many cliniciansare now utilizing combination therapy as treatment. This isin part the result of a number of in vitro studies supportingadditive or synergistic effects and in spite of conflictinganimal studies. Whether combinations are superior to individualantifungal drugs is a matter that may require clinical studiesfor resolution. Our studies revealed no combination effect ofposaconazole and amphotericin B in this model. There was littleevidence of antagonism of posaconazole and amphotericin B. Therefore,these studies provide no reason to avoid clinical studies ofposaconazole and amphotericin B combination therapy.

ACKNOWLEDGMENTS

These studies were supported in part by Schering Plough ResearchInstitute.

FOOTNOTES

* Corresponding author. Mailing address: Department of Medicine, Division of Infectious Disease (7881), The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78229-3900. Phone: (210) 567-0990. Fax: (210) 567-0962. E-mail: Najvar{at}uthscsa.edu.

REFERENCES

Top