点击显示 收起
Departments of Infectious Diseases, Infection Control, and Employee Health, Laboratory Medicine
Biochemistry and Molecular Biology, The University of Texas M. D. Anderson Cancer Center
College of Pharmacy, University of Houston, Houston
Invasive aspergillosis (IA) is the most important opportunistic mycosis in immunosuppressed patients. The lack of a sufficient number of effective antifungals and our incomplete understanding of the pathogenesis of IA contribute to its overall unfavorable prognosis. Studies of drug efficacy against IA and Aspergillus virulence rely on conventional animal models that are laborious and use limited numbers of animals; alternative, less cumbersome in vivo models are desirable. Using different inoculation models of IA, we found that Toll-deficient Drosophila flies exposed to voriconazole (VRC), the preferred drug for the treatment of IA in humans, had significantly better survival rates and lower tissue fungal burdens than did those not exposed to VRC. Furthermore, Toll-deficient Drosophila flies infected with an alb1-deleted hypovirulent Aspergillus mutant had significantly better survival rates than did those infected with a wild-type Aspergillus strain. Therefore, the Drosophila fly is a fast, high-throughput in vivo model for the study of drug efficacy against IA and Aspergillus virulence.
Since the 1990s, invasive aspergillosis (IA) has emerged as a leading cause of infection-related death in patients with leukemia and recipients of bone-marrow and solid-organ transplants [1]. The prognosis for IA is unfavorable, because modern antifungals have mediocre anti-Aspergillus activity in vivo [1]. This suboptimal efficacy has necessitated new drug development and novel therapeutic strategies, such as combination antifungal therapy [1, 2].
In recent years, several antifungals with improved in vitro anti-Aspergillus activity have been introduced [1]. However, in vitro susceptibility testing of antifungals alone or in combination has been plagued by methodological problems and has demonstrated limited correlation with clinical efficacy [2, 3]. Therefore, assessment of the activity of antifungals relies on studies with animal models of IA that have been established in mice, rabbits, guinea pigs, and rats [4, 5]. These models are expensive, laborious, and (because of logistical constraints) usually rely on testing drug activity against only 1 Aspergillus isolate in a small number of animals.
Furthermore, little is known about Aspergillus virulence factors. As the A. fumigatus genome-sequencing project is being completed [6] and molecular tools for the study of the biological processes of A. fumigatus are rapidly being developed [7], it is anticipated that several candidate Aspergillus mutants will be available in the near future with which to study the pathogenesis of IA. However, the study of the pathogenetic potential of such mutants in conventional host systems faces the same logistical challenges as does the testing of antifungals.
Because the limitations of the larger-animal models described above are not restricted to the study of Aspergillus infection, there has been a surge in studies of microbial pathogenesis in simpler nonvertebrate hosts that have well-characterized genetic systems. Thus, Drosophila melanogaster, Caenorhabditis elegans, and Acanthamoeba castellanii have been successfully used to study the pathogenesis of gram-positive bacterial and yeast infections [812]. Furthermore, extensive research with D. melanogaster has resulted in important insights into the molecular mechanisms of innate immunity. D. melanogaster flies have 2 conserved signaling pathways that are activated during immune responses: imd and Toll [1315]. The Toll pathway is critical for the protection of Drosophila flies against fungi, and Drosophila flies with various Toll mutations rapidly succumb to Aspergillus species and other fungi [8, 9, 14, 15].
Here, we used D. melanogaster as a host model to test drug activity against IA and Aspergillus virulence. Because (1) the Toll pathway is significantly conserved between Drosophila flies and humans and (2) these flies offer the advantage that they can be grown, manipulated, and analyzed in large numbers with significantly less labor and expense than with conventional host systems, Drosophila flies could provide a useful model system for high-throughput screening of compounds for activity against IA and for testing Aspergillus virulence.
MATERIALS AND METHODS
Drosophila Stocks
We used Oregon R flies (gift from T. Y. Ip, University of Massachusetts Medical Center, Worchester) as wild-type (wt) flies. Tlr632/TlI-RXA Drosophila mutant flies (hereafter, "Tl mutant flies") were generated by crossing flies carrying a thermosensitive allele of Toll (Tlr632) with flies carrying a null allele of Toll (TlI-RXA) (gift from T. Y. Ip) [9, 15]. We used standard procedures for the manipulation, housing, and feeding of Drosophila flies.
Fungal Isolates and Antifungal Agents
We used the A. fumigatus clinical isolate AF293 (gift from D. W. Denning, University of Manchester, Manchester, United Kingdom). For the Aspergillus virulence experiments, we used the alb1-deleted A. fumigatus strain B-5233/RGD12-8 (hereafter, "alb1 A. fumigatus mutant") and its isogenic wt Aspergillus strain, B-5233 (gift from K. A. Marr, Fred Hutchinson Cancer Research Center, University of Washington, Seattle) [16]. We prepared voriconazole (VRC; Pfizer) and terbinafine (TRB; Novartis) dilutions in distilled water and dimethyl sulfoxide, respectively, and stored them at -80°C until use.
Drosophila Infection Models of IA
Injection assay.
We punctured the dorsal side of the thorax of 4060 CO2-anesthetized Drosophila flies with a thin sterile needle that had been dipped in concentrated solutions of conidia (range, 1 × 107 to 1 × 1010 conidia/mL), using a procedure described elsewhere [14]. We considered flies that died within 3 h of the injection (<5%) to have died as a result of the procedure and did not include them in the survival analysis. We used adult flies 24 days old and housed them at 29°C, at which Tl susceptibility to microbial challenge is maximal [14]. We transferred the flies into fresh vials every 2 days and assessed survival daily by visual inspection until day 8 after infection. To provide a control, we inserted a sterile needle into flies, to determine the effect of the procedure on survival [14]. We performed each experiment at least in triplicate. We quantified the conidial inocula introduced into flies by injection by transferring conidia from the tip of a needle previously dipped in either a 1 × 107 or a 1 × 1010 conidia/mL solution to 1 mL of sterile saline. We then performed serial dilutions (in triplicate), plated 100 L of the solution on yeast agar glucose (YAG) plates at 37°C, and counted the colony-forming units after 48 h.
Rolling assay.
We rolled 4060 CO2-anesthetized Drosophila flies for 2 min on YAG plates that contained a carpet of conidia, which we had produced by growing 2 × 107 conidia for 3 days. We then transferred the flies into fresh vials every 2 days and left them for 10 days at 29°C. We considered flies that died within 3 h of rolling (<1%) to have died as a result of the procedure and did not include them in the survival analysis. To provide a control, we rolled flies on empty petri dishes, to assess the effect of rolling-associated injury on survival. We performed each experiment at least in triplicate.
Ingestion assay.
We prepared special fly-food vials that contained YAG medium. We grew 2 × 107 conidia for 3 days, to produce a conidial carpet on the surface of these vials. Next, we placed 4060 flies into the vials and left them for 68 h to feed on the conidia. We then transferred the flies into vials without conidia every 2 days and left them for 10 days at 29°C. To provide a control, we placed flies into vials that contained YAG medium without conidia. We performed each experiment at least in triplicate.
Drug protection experiments.
For the VRC protection experiments, we used female Tl mutant flies 24 days old. We put flies in empty vials (without food) for 68 h to starve them and then transferred them to vials with VRC-containing food (concentration, 1 mg/mL). After 24 h of VRC exposure, we infected flies with conidia using 1 of the 3 assay methods described above and then transferred them daily into new vials with VRC-containing food and left them for 10 days at 29°C. We starved another group of flies for 68 h and infected them with conidia but did not expose them to VRC either before or after the infection; these were the control flies. We assessed the efficacy of VRC against Aspergillus infection daily until day 10 after infection. We performed each experiment at least in triplicate. For the combination drug experiments, we prepared vials containing either VRC alone, TRB alone, or VRC plus TRB at 64 g/mL each.
Tissue fungal burden analysis by real-time quantitative polymerase chain reaction (qPCR).
We stored groups of 20 flies at -80°C and washed them twice with 0.85% NaCl to remove conidia from their exterior before homogenization and DNA extraction by use of the DNeasy Tissue Kit (Qiagen). We tested flies from representative experiments using the rolling assay at different time points (i.e., days 1, 4, and 8 after infection). We performed all of the qPCR experiments in triplicate and analyzed each DNA sample in duplicate using the ABI PRISM 7000 sequence detection system (Applied Biosystems), according to a procedure described elsewhere [17]. We reported the results as conidial equivalents of A. fumigatus DNA [17].
Bioassay.
We homogenized groups of 20 VRC-exposed flies on day 8 after infection by rolling in 0.85% NaCl using a bead-beater homogenizer (Mini Bead) and instilled 200 L of the homogenate into an opening on the surface of RPMI 1640 agar plates (with 0.165 mol/L morpholinepropanesulfonic acid [pH 7.0]) that had been inoculated with 1 × 104 cfu of Candida kefyr (ATCC 66028). We then incubated the plates for 24 h at 37°C and compared zones of growth inhibition on the basis of a standard curve that had been generated by use of the same method with known VRC concentrations (range, 0.258.00 g/mL). We performed each experiment in triplicate.
Histopathological and scanning electron microscopy (SEM) analysis.
On day 8 after infection by rolling, we fixed flies using 10% formaldehyde, processed them, and embedded them in paraffin wax. We stained tissue sections with Grocott-Gomori methenaminesilver nitrate (GMS) and examined them for visible hyphal burden. We performed SEM analysis of Tl mutant flies on day 8 after infection by rolling, as described elsewhere [18].
Statistical analyses.
We compared treatment groups using the Kruskal-Wallis test (nonparametric analysis of variance) and the Mann-Whitney U test, as appropriate. Also, we plotted survival using Kaplan-Meier analysis. We analyzed differences in survival rates between treatment groups using the log-rank test (GraphPad Prism software; version 3; GraphPad Software). We considered P .05 to be statistically significant.
RESULTS
Establishment of reproducible models of experimental IA in Drosophila flies.
To reliably study drug activity against IA and Aspergillus virulence in Drosophila flies, we first sought to establish reproducible models of experimental IA. Hence, we tested different sites of inoculation of Aspergillus conidia in flies: via injection into the hemolymph [14], via penetration though the exoskeleton [15], and via ingestion through the gastrointestinal tract.
Aspergillus infection of Tl mutant flies resulted in significantly higher mortality rates (P < .001), compared with those in wt flies, in all assays (figure 1A1C). Indeed, Tl mutant flies had a higher tissue fungal burden as determined by qPCR in representative experiments (P < .001, for the rolling assay) (figure 1D). We verified the presence of conidia in the fly gut by examining histopathological sections (stained with GMS) of flies killed 12 h after infection by ingestion (figure 1E). Because tissue fungal burden has an effect on IA outcome in humans [1], we wanted to evaluate whether survival of Tl mutant flies after Aspergillus infection was similarly dependent on the conidial inoculum size. Indeed, survival rates were significantly lower (P < .01) in flies injected via a needle dipped in a 1 × 1010 conidia/mL solution and inoculated with 20,000 conidia (21% and 0% on days 3 and 6 after infection), compared with those in flies injected via a needle dipped in a 1 × 107 conidia/mL solution and inoculated with 700800 conidia (54% and 21% on days 3 and 6 after infection). After the effects of inoculum sizes and routes of inoculation on survival were characterized, the optimized models were used to examine the effects of antifungal therapy and Aspergillus virulence in Drosophila flies.
Protection of Tl mutant flies against IA provided by VRC alone and in combination with TRB.
To establish Drosophila flies as a fast and economical model for the mass screening of compounds for anti-Aspergillus activity, we sought to determine whether VRC protects Tl mutant flies against IA. VRC is a triazole with proven anti-Aspergillus activity in vitro [19], in conventional animal models [19], and in patients with IA [20].
We next used the Drosophila model to test the efficacy of the combination of VRC and TRB against IA. These 2 drugs block sequential steps in the ergosterol synthesis pathway and have synergistic effects against Aspergillus species and other fungi in vitro [21, 22]. In pilot experiments, concentrations of TRB up to 2 mg/mL were not toxic in adult flies (data not shown). We found that Tl mutant flies exposed to VRC plus TRB had a significantly better survival rate than did those exposed to either antifungal alone (P < .001) (figure 2G).
DISCUSSION
Because conventional animal models are expensive, labor intensive, and create a bottleneck for the mass screening of new compounds with which IA could potentially be treated, we evaluated whether Drosophila flies could be advantageous for high-throughput testing of candidate antifungals. Using 3 infection assays, we found that VRC significantly protects Tl mutant flies from Aspergillus infection. In addition to experiencing a substantial reduction in mortality, VRC-fed flies also had significantly decreased tissue fungal burdens than did flies not exposed to VRC.
To further examine whether Drosophila flies are a good model for testing drug activity against IA, we evaluated the interaction of VRC and TRB in Tl mutant flies and found that flies exposed to both drugs had better survival rates than did flies exposed to either drug alone. The study of antifungal drugs administered in combination or sequentially in Drosophila flies is especially appealing, given that in vivo testing of various antifungal combinations (sometimes referred to as "the in vivo checkerboard method") is particularly laborious and expensive in conventional animal models because many combinations of different concentrations should be tested [2].
The study of Aspergillus virulence in large-animal models is hampered by the same logistical difficulties encountered when drugs are tested. As we have demonstrated, Drosophila flies can be a valuable tool for the evaluation of the virulence of hundreds of putative Aspergillus mutants. As proof of principle, we used the alb1 A. fumigatus mutant, which lacks melanin [16], an Aspergillus virulence factor [23]. It has been shown that this strain has attenuated virulence in a murine model of IAit caused a mortality rate of 10%, compared with a mortality rate of 80% caused by the isogenic wt Aspergillus strain [16]. Similar to what was observed in mice, a much more complex host, we found that, in Drosophila flies infected by ingestion or rolling, the alb1 A. fumigatus mutant has substantially reduced virulence, compared with that of the isogenic wt Aspergillus strain.
However, the alb1 A. fumigatus mutant caused a mortality rate comparable to that caused by the isogenic wt Aspergillus strain when Tl mutant flies were infected by injection. Although little is known about the differences that may underlie Drosophila immune responses after introduction of conidia via different mucosal surfaces, it is possible that the IA that develops after direct inoculation of conidia in the fly hemolymph results in an overwhelming, rapidly progressing infection, which does not allow for an analysis of the difference in virulence between the alb1 A. fumigatus mutant and the isogenic wt Aspergillus strain. Alternatively, the induction of epithelial immune responses against conidia could be different in various epithelia and may explain the attenuated virulence of the alb1 A. fumigatus mutant in the rolling and ingestion assays. The induction of antimicrobial peptides on mucosal surfaces is controlled by the imd pathway in Drosophila flies [24], and this pathway was intact in the Tl mutant flies used in the present study. Thus, these mutants could more effectively defend against Aspergillus infection introduced by ingestion or rolling. Nevertheless, this differing behavior of the alb1 A. fumigatus mutant as a function of the mode of introduction of infection emphasizes the effect that the site of inoculation of conidia has on the acuity of Aspergillus infection in Drosophila flies and other hosts. This difference should be taken into consideration in future studies of Aspergillus virulence.
Given the complexity of the immune responses in humans with IA and the evolution of adaptive immunityrelated functions of the Toll pathway in humans, we argue that focusing on simple hosts such as Drosophila flies that have well-defined effector immune responses will be instructive, as was shown recently with Candida albicans [9]. The logistical benefits of working with Drosophila flies make them an appropriate alternative model for the study of Aspergillus virulence, drug activity against IA, and the interaction between drugs and the biological fitness of Aspergillus mutants.
Despite having the advantage of making it possible to test a large number of animals in a fast and inexpensive manner, the Drosophila model has its limitations. For example, it is difficult to precisely quantify the number of conidia that infect each fly as well as the amount of antifungal drug ingested by individual flies. In addition, we tested the efficacy of orally absorbed antifungals (VRC and TRB) only in the present study; we are currently testing the efficacy of the administration of nonorally absorbable agents by injection using micropipettes into Drosophila flies, to explore the potential this model has in screening for nonorally absorbable agents that may have promising anti-Aspergillus activity.
In summary, despite its limitations, the Drosophila model is fast, inexpensive, and easy to use. It has the technical benefit of making possible the use of hundreds of animals for both the mass screening of compounds for anti-Aspergillus activity and the evaluation of the virulence of large numbers of Aspergillus mutants.
Acknowledgments
We thank G. Chamilos and J. Childress, for assistance.
References
1. Kontoyiannis DP, Bodey GP. Invasive aspergillosis in 2002: an update. Eur J Clin Microbiol Infect Dis 2002; 21:16172. First citation in article
2. Kontoyiannis DP, Lewis RE. Combination chemotherapy for invasive fungal infections: what laboratory and clinical studies tell us so far. Drug Resist Updat 2003; 6:25769. First citation in article
3. Pfaller MA, Rinaldi MG. Antifungal susceptibility testing: current state of technology, limitations, and standardization. Infect Dis Clin North Am 1993; 7:43544. First citation in article
4. Lewis RE, Prince RA, Chi J, Kontoyiannis DP. Itraconazole pre-exposure attenuates the efficacy of subsequent amphotericin B therapy in a murine model of acute invasive pulmonary aspergillosis. Antimicrob Agents Chemother 2002; 46:320814. First citation in article
5. Kirkpatrick WR, Perea S, Coco BJ, Patterson TF. Efficacy of caspofungin alone and in combination with voriconazole in a guinea pig model of invasive aspergillosis. Antimicrob Agents Chemother 2002; 46:25648. First citation in article
6. Denning DW, Anderson MJ, Turner G, Latge JP, Bennett JW. Sequencing the Aspergillus fumigatus genome. Lancet Infect Dis 2002; 2:2513. First citation in article
7. Brakhage AA, Langfelder K. Menacing mold: the molecular biology of Aspergillus fumigatus. Annu Rev Microbiol 2002; 56:43355. First citation in article
8. Apidianakis Y, Rahme LG, Heitman J, Ausubel FM, Calderwood SB, Mylonakis E. Challenge of Drosophila melanogaster with Cryptococcus neoformans and role of the innate immune response. Eukaryot Cell 2004; 3:4139. First citation in article
9. Alarco A, Marcil A, Chen J, Suter B, Thomas D, Whiteway M. Immune-deficient Drosophila melanogaster: a model for the innate immune response to human fungal pathogens. J Immunol 2004; 172:56228. First citation in article
10. Steenbergen JN, Shuman HA, Casadevall A. Cryptococcus neoformans interactions with amoebae suggest an explanation for its virulence and intracellular pathogenic strategy in macrophages. Proc Natl Acad Sci USA 2001; 98:1524550. First citation in article
11. Garsin DA, Sifri CD, Mylonakis E, et al. A simple model host for identifying Gram-positive virulence factors. Proc Natl Acad Sci USA 2001; 98:108927. First citation in article
12. Mylonakis E, Ausubel FM, Perfect JR, Heitman J, Calderwood SB. Killing of Caenorhabditis elegans by Cryptococcus neoformans as a model of yeast pathogenesis. Proc Natl Acad Sci USA 2002; 99:1567580. First citation in article
13. Hoffmann JA, Reichhart JM. Drosophila innate immunity: an evolutionary perspective. Nat Immunol 2002; 3:1216. First citation in article
14. Lemaitre B, Nicolas E, Michaut L, Reichhart JM, Hoffmann JA. The dorsoventral regulatory gene cassette spatzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 1996; 86:97383. First citation in article
15. Lemaitre B, Reichhart JM, Hoffmann JA. Drosophila host defense: differential induction of antimicrobial peptide genes after infection by various classes of microorganisms. Proc Natl Acad Sci USA 1997; 94:146149. First citation in article
16. Tsai HF, Chang YC, Washburn RG, Wheeler MH, Kwon-Chung KJ. The developmentally regulated alb1 gene of Aspergillus fumigatus: its role in modulation of conidial morphology and virulence. J Bacteriol 1998; 180:30318. First citation in article
17. Bowman JC, Abruzzo GK, Anderson JW, et al. Quantitative PCR assay to measure Aspergillus fumigatus burden in a murine model of disseminated aspergillosis: demonstration of efficacy of caspofungin acetate. Antimicrob Agents Chemother 2001; 45:347481. First citation in article
18. Kango-Singh M, Nolo R, Tao C, et al. Shar-pei mediates cell proliferation arrest during imaginal disc growth in Drosophila. Development 2002; 129:571930. First citation in article
19. Murphy M, Bernard EM, Ishimaru T, Armstrong D. Activity of voriconazole (UK-109,496) against clinical isolates of Aspergillus species and its effectiveness in an experimental model of invasive pulmonary aspergillosis. Antimicrob Agents Chemother 1997; 41:6968. First citation in article
20. Herbrecht R, Denning DW, Patterson TF, et al. Voriconazole versus amphotericin B for primary therapy of invasive aspergillosis. N Engl J Med 2002; 347:40815. First citation in article
21. Kontoyiannis DP, Lewis RE. Antifungal drug resistance of pathogenic fungi. Lancet 2002; 359:113544. First citation in article
22. Ryder NS, Leitner I. Synergistic interaction of terbinafine with triazoles or amphotericin B against Aspergillus species. Med Mycol 2001; 39:915. First citation in article
23. Langfelder K, Streibel M, Jahn B, Haase G, Brakhage AA. Biosynthesis of fungal melanins and their importance for human pathogenic fungi. Fungal Genet Biol 2003; 38:14358. First citation in article
24. Tzou P, Ohresser S, Ferrandon D, et al. Tissue-specific inducible expression of antimicrobial peptide genes in Drosophila surface epithelia. Immunity 2000; 13:73748. First citation in article