Previous Article | Next Article ![]()
Infection and Immunity, July 2005, p. 3842-3850, Vol. 73, No. 7
0019-9567/05/$08.00+0 doi:10.1128/IAI.73.7.3842-3850.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Division of Infectious Diseases,1 Department of Molecular Biology,2 Center for Immunology and Inflammatory Diseases,3 Division of Rheumatology, Allergy, and Immunology, Massachusetts General Hospital, Boston, Massachusetts 02114,4 Department of Molecular Genetics and Microbiology,5 Division of Infectious Diseases,6 Department of Medicine,7 Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina 27710,8 Department of Microbiology and Molecular Genetics,9 Department of Genetics, Harvard Medical School, Boston, Massachusetts 02115,10
Received 27 November 2004/ Returned for modification 7 January 2005/ Accepted 26 January 2005
|
|
|---|
|
|
|---|
To address these limitations, we developed a new system using caterpillars of the greater wax moth (Galleria mellonella) as an alternative model for the study of C. neoformans. We report here that various serotypes of C. neoformans proliferate inside the hemocoel and kill the caterpillar, even though the fungi undergo phagocytosis by G. mellonella hemocytes. Caterpillar killing correlates with the number of CFU of C. neoformans inoculated, and virulence factors involved in mammalian cryptococcal infection play a significant role in G. mellonella killing. The efficacy of antifungal agents in this system is similar to results obtained in studies with humans, suggesting that the model may be developed for study of the in vivo efficacy of antifungal therapy.
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Fungal strains used and their interaction with G. mellonellaa
|
A 10-µl Hamilton syringe was used to inject 10-µl aliquots of the inoculum into the hemocoel of each caterpillar via the last left proleg (9). Before injection, the area was cleaned using an alcohol swab. A total of 20 mg of ampicillin/kg of body weight was coadministered to prevent infection by bacteria naturally present on the surface of the caterpillar. After injection, caterpillars were incubated in plastic containers, and the number of dead caterpillars was scored daily. Caterpillars were considered dead when they displayed no movement in response to touch.
Antifungal drugs were injected using the same technique. For experiments that required multiple injections, a different proleg was used for each injection, starting from the left last proleg and rotating left to right and moving proximally (i.e., injecting through the left last proleg, right last proleg, penultimate left proleg, and penultimate right proleg, as needed). Antifungal agents used were amphotericin B (Sigma), fluconazole (Toronto Research Chemicals), and flucytosine (InvivoGen). All antifungal drugs were diluted in water, and each antifungal agent was administered once by a separate injection.
To heat kill C. neoformans, yeast cells were exposed to 60°C for 45 min. Heat killed C. neoformans cells were washed three times, resuspended in sterile phosphate-buffered saline (PBS), diluted to the appropriate density, and inoculated into the caterpillars. For preparation of the inocula, cultures were diluted as needed, using a hemocytometer to determine the exact dose administered. The concentration of cryptococcal cells in the inoculum was confirmed by plating serial dilutions on plates with YPD medium containing ampicillin (100 µg/ml) and streptomycin (100 µg/ml). After inoculation, caterpillars were incubated at 37°C, unless otherwise specified. A mock inoculation with PBS was performed in each experiment to monitor killing due to physical injury or infection by pathogenic contaminants. In most experiments, no caterpillars in this control group died. Rare experiments with more than two dead caterpillars in the control group were discarded and repeated. For simplicity, the control group is not included in some figures.
Killing curves were plotted and estimation of differences in survival (log rank and Wilcoxon tests) analyzed by the Kaplan-Meier method using STATA 6 statistical software (Stata). The same software program was used for the statistical analysis of the CFU of G. mellonella in the hemocoel (Mann-Whitney and Kruskal-Wallis tests). A P value of <0.05 was considered significant. Each experiment was repeated at least three times, and each independent experiment gave similar results. Data presented here are from a representative experiment.
Tissue burden culture studies. For the evaluation of the tissue burden of C. neoformans in caterpillars over time, five caterpillars per group were weighed and homogenized in 1 ml sterile PBS with a Tissue Tearor (model 398; Biospec Products), and serial dilutions of the homogenates were plated on YPD agar plates containing ampicillin (100 µg/ml), streptomycin (100 µg/ml), and nourseothricin (100 µg/ml). Plates were incubated at 30°C for 72 h before colonies were counted.
Collection of hemocytes. At selected time points after inoculation, 6 to 10 caterpillars per group were bled by insertion of a lancet into the hemocoel. Hemolymph was collected into ice-cold anticoagulant saline (54) with or without EDTA (Fluka). Hemolymph was centrifuged, the supernatant was discarded, and cells were diluted in Grace's insect medium (Gibco). Specimens were fixed in 4% paraformaldehyde (Electron Microscopy Sciences). For clarity in experiments, we used cryptococcal cells that were stained before inoculation with fluorescein isothiocyanate (FITC; Molecular Probes) or, after specimen collection, with calcofluor white (Polysciences).
Some experiments involving cryptococcal strains containing the pYGFP3 plasmid (expressing the synthetic green fluorescent protein [GFP]) (11) used undiluted hemolymph. For these latter experiments, caterpillars were bled into an ice-cold test tube containing 1.5 mg phenylthiourea (Sigma) to prevent melanization and then used for microscopy without any further processing.
To determine if the cryptococcal cells interacting with the hemocytes were intracellular or extracellular, we incubated the mixture of cells containing the hemocytes and FITC-labeled C. neoformans with the dye trypan blue (Sigma). Trypan blue quenches FITC fluorescence but does not get inside the hemocytes, so only intracellular FITC-labeled C. neoformans were visualized in this manner, as previously described (48).
Light and fluorescence microscopy were performed using an Olympus BX51 microscope, and images were captured and processed using the digital camera Qimaging Retiga (Burnaby) and Olympus Microsuite Software. Vectashield mounting medium was used for fluorescence microscopy (Vector Laboratories).
|
|
|---|
![]() View larger version (15K): [in a new window] |
FIG. 1. Killing of G. mellonella caterpillars by C. neoformans depends on the number of cryptococcal cells inoculated. Kaplan-Meier plots of G. mellonella survival after injection of different inocula of C. neoformans strain H99 (ATCC 208821). There was no killing of caterpillars that received PBS or heat-killed cryptococcal cells of the same C. neoformans strain (105 cells/larva).
|
Killing of G. mellonella is not a unique feature of C. neoformans strain H99. We determined whether different varieties of C. neoformans could kill G. mellonella in addition to H99. As shown in Table 1, all C. neoformans strains tested were able to kill G. mellonella caterpillars, although strain H99, the most widely used C. neoformans strain for the study of pathogenesis, was the most virulent among the strains tested. Strain H99 killed caterpillars faster than other C. neoformans strains tested and at inocula as low as 20 CFU/larva. Substantially higher numbers of inocula were needed for killing of the caterpillars by other strains. For example, strain NIH 312 (ATCC 34880; CDC B3182) was avirulent at 5 x 103 CFU/larva or less. Killing reached 60% by day 14 at inocula of 1.5 x 104 CFU/larva and reached 75% at 1.75 x 104 CFU/larva (with mean time to mortality of 11 days at this inoculum). Similar results were obtained with several other C. neoformans strains (Table 1), including the clinical isolates of C. neoformans var. grubii strain ATCC 62068 (serotype A), Cryptococcus neoformans var. gattii strains ATCC 34877 and ATCC 32609 (NIH 444) (serotype B) (22), and Cryptococcus neoformans var. neoformans strain ATCC 96909 (serotype D).
Tissue burden of C. neoformans in G. mellonella and the role of hemocytes.
Because G. mellonella larvae might be colonized by other fungi, we used a strain resistant to the antimicrobial nourseothricin (KN99
[kin1 + KIN1], designated KN99
kin1 + KIN1) to evaluate the fungal burden in G. mellonella. This enabled us to plate homogenates of caterpillars on medium containing nourseothricin, along with broad-spectrum antibacterials, to inhibit growth of other yeasts and bacteria. Strain KN99
kin1 + KIN1 is congenic with H99 and has a level of virulence similar to that of H99 with mice and C. elegans (32), as well as with G. mellonella (Table 1). After injection of 2.2 x 103 CFU/larva, the proliferation in G. mellonella was slow at the beginning, but by day 3, the organism burden reached levels (mean ± standard deviation) of 4.2 x 104± 4.1 x 103 CFU/mg. The cryptococcal burden increased only slightly by the following day, when the caterpillars began to die. A control group of caterpillars was injected with PBS, and homogenates from these caterpillars were plated using a similar protocol. No organisms grew from this control group.
We also monitored the interaction between C. neoformans strain H99 and hemocytes. The host response of G. mellonella to infection consists of structural and passive barriers, as well as cellular and humoral responses that are performed by hemocytes within the hemolymph. Six types of hemocytes in G. mellonella have been identified, and the insect response includes phagocytosis by plasmatocytes and granulocytes and nodulation by layers of hemocytes, to encapsulate large invading pathogens (reviewed in reference 18). Over the first hours after injection, cryptococcal cells were bound to hemocytes (Fig. 2A), followed by a progressively increasing number of C. neoformans cells within hemocytes, indicating successful phagocytosis (Fig. 2B and C). By 24 h, we found a significant number of cryptococcal cells surrounded by layers of hemocytes (Fig. 2D) in a process known as nodulation, which has been described previously with G. mellonella in association with pathogens (21) and that may have some similarities with the granulomas associated with cryptococcal infection in mammals (16-18, 32).
![]() View larger version (138K): [in a new window] |
FIG. 2. Binding, phagocytosis, and nodulation of C. neoformans cells by G. mellonella hemocytes. C. neoformans fungal cells binding to (A) or phagocytosed by G. mellonella hemocytes (B and C) are shown. (D) A C. neoformans fungal cell is surrounded by layers of hemocytes in a process known as nodulation that has been described previously for G. mellonella in association with pathogens. Fungal cells (B) were stained with calcofluor white, and a phagocytosed C. neoformans is identified by a white arrowhead. (C and D) C. neoformans cells were stained with FITC before inoculation as described in Materials and Methods. Original magnification in all panels, x100; bar, 50 µM.
|
![]() View larger version (10K): [in a new window] |
FIG. 3. C. neoformans virulence factors important for mammalian infection also enhance killing of C. elegans. Survival of G. mellonella after injection of 1.5 x 104 CFU/larva of wild-type C. neoformans strain H99 or mutants with disruptions in the genes encoding CAP59 (essential for capsule formation) or the G protein-cAMP-PKA or the RAS1-controlled signal transduction cascades demonstrated hypovirulence (cap59, gpa1, ras1, and pka1). There was a significant decrease in virulence between the wild type and the mutants at both 37°C (A) and 30°C (B). P values were <0.01 for each of the mutants compared to the parental strain H99 (Table 1).
|
![]() View larger version (15K): [in a new window] |
FIG. 4. Melanin biosynthesis is involved in G. mellonella killing by C. neoformans. A mutation involving the gene that encodes laccase (an enzyme essential for melanin biosynthesis) renders C. neoformans less virulent in the G. mellonella system at 37°C than the wild type (P = 0.001) (in this experiment, caterpillars received 2.5 x 104 CFU/larva).
|
and MATa) undergo conjugation. Studies with mammalian models (23) and C. elegans (31) demonstrate that strain JEC21, the model
mating strain of C. neoformans var. neoformans, is more virulent than the otherwise isogenic a mating parent JEC20 (23). However, in C. neoformans var. grubii, congenic a and
mating type strains demonstrate equivalent virulence in both mammalian models (35) and C. elegans (E. Mylonakis, K. Nielsen, and J. Heitman, unpublished data); it is not known why this difference between C. neoformans var. neoformans and C. neoformans var. grubii strains exists. Even in the case of C. neoformans var. grubii, cells containing the pMF
1::GFP reporter gene demonstrate specific GFP expression in the central nervous system (CNS) of infected rabbits (pMF
1 regulates production of the
mating type pheromone), suggesting that the MF
1 gene is induced during the proliferative stage of a CNS infection (11).
Our findings in G. mellonella were similar to those in mammalian models in all these aspects. Specifically, the rate of killing of G. mellonella caterpillars that received MAT
strain JEC21 was significantly faster than in the group that received the otherwise-isogenic MATa strain JEC20 (Table 1). Similarly, we found that there was no pMF
1::GFP expression of C. neoformans cells in G. mellonella hemolymph during the first 48 h after injection but that there was significant GFP expression by day 3 after injection, suggesting that in G. mellonella, the MF
1 gene is induced during the proliferative stage of the infection, similar to the findings with rabbit CNS (11) and C. elegans (31) models (Fig. 5). Finally, killing of G. mellonella by C. neoformans var. grubii strains H99, KN99a, KN99
, and KN99-5 was identical, similar to findings with mammals (Table 1) (35).
![]() View larger version (156K): [in a new window] |
FIG. 5. The C. neoformans MF 1 promoter is expressed in the G. mellonella hemocoel during the proliferative stage. Fluorescent microscopy of G. mellonella hemolymph on day 3 (proliferative stage) after injection with 1.25 x 104 CFU/larva of C. neoformans strain H99 expressing GFP fused to the MF 1 promoter. (A and B) Hemolymph was undiluted to provide a better appreciation of the fungal burden within the insect hemolymph. (C and D) Cells were fixed as detailed in Materials and Methods. Images in panels A and C are fluorescence images, and images in panels B and D are the corresponding confocal images. There was no fluorescence observed on day 1 or 2 of this experiment.
|
![]() View larger version (16K): [in a new window] |
FIG. 6. Antifungal drugs prolong the survival of G. mellonella caterpillars after challenge with C. neoformans. We examined the role of the most commonly used agents for C. neoformans infection by administering a single dose of amphotericin B (AMB; 1.5 mg/kg), fluconazole (FLU; 14 mg/kg), or flucytosine (5-FC; 20 mg/kg) alone or in combination 48 h after the inoculation of caterpillars with 1.2 x 103 CFU of C. neoformans strain H99 per larva. A control group received the C. neoformans inoculum and PBS instead of antifungal drugs. Monotherapy with amphotericin B prolonged the survival of G. mellonella caterpillars (P = 0.001 compared to control). Also, there was a trend suggesting that fluconazole was effective (P = 0.072). The combination of amphotericin B plus flucytosine was significantly more effective than amphotericin B alone (P = 0.0002).
|
In addition to monitoring the survival of G. mellonella caterpillars, we also evaluated the impact of antifungal agents on the tissue burden of C. neoformans within the hemocoel. As noted above, because G. mellonella larvae are often colonized by other fungi, we used a nourseothricin-resistant strain (KN99
kin1 + KIN1) to evaluate the fungal burden in G. mellonella. For this, we injected caterpillars with C. neoformans (2.4 x 103 CFU of C. neoformans strain KN99
kin1 + KIN1 per caterpillar). After 48 h, we injected caterpillars (five caterpillars per group) with PBS (control group) or antifungal drugs (amphotericin B, fluconazole, and flucytosine either as monotherapy or in combination). All antifungal agents significantly decreased the number of fungi in the caterpillars, but fluconazole was significantly less effective than the other antifungal therapies we studied. The fungal load (mean ± standard deviation) in the control group that received no antifungal drugs was 5.5 x 104± 1.4 x 103 CFU/mg but was only somewhat lower in the fluconazole-treated group (2.1 x 104± 9.9 x 103 CFU/mg; P = 0.016 compared to control). In contrast, both flucytosine (356 ± 225 CFU/mg; P = 0.009 compared to control) and amphotericin B (384 ± 234 CFU/mg; P = 0.009 compared to control) dramatically lowered the fungal load. The combination of amphotericin B with flucytosine was more effective at decreasing the tissue burden of C. neoformans than either agent alone (the tissue burden in the amphotericin B-plus-flucytosine group was 54 ± 33 CFU/mg; P = 0.008 compared to the groups that received amphotericin B or flucytosine monotherapy). Interestingly, triple therapy was the most effective therapy in decreasing the tissue burden of C. neoformans in the caterpillars (9 ± 6 CFU/mg; P = 0.035 compared to the amphotericin B with flucytosine group), even though as noted above, this therapy did not prolong the survival of G. mellonella compared to dual therapy with amphotericin B plus flucytosine. None of the therapeutic regimens studied sterilized the caterpillars following infection; instead, the prolongation of survival meant that infected caterpillars treated with antifungals had delayed mortality.
|
|
|---|
Studies of microbial pathogenesis in nonvertebrate hosts during the past decade have resulted in important insights into the molecular mechanisms of microbial pathogenesis and host defense. It is now apparent that many of the same microbial virulence factors are involved in pathogenesis in evolutionarily disparate hosts (3, 31, 32, 44, 46). G. mellonella caterpillars have previously been used to study infection by various pathogens, including Pseudomonas aeruginosa, Proteus mirabilis, Escherichia coli, Bacillus cereus, and the insect pathogenic fungus Metarhizium anisopliae (13, 20, 27-29, 50, 51). Among human fungal pathogens, Candida spp. were also shown to kill G. mellonella when injected into the hemocoel of the insect caterpillars (6, 10, 14, 42, 43). Brennan et al. reported that in the case of Candida albicans, there is a correlation between virulence in G. mellonella and virulence measured by systemic infection of mice (6). Of note is that for most strains of Candida spp., at least 2 x 105 cells/larva are usually needed for killing (10), whereas C. neoformans is considerably more virulent.
G. mellonella has also been used to study Aspergillus flavus pathogenicity. Conidia of A. flavus were not virulent when applied to the surface of healthy caterpillars but killed the caterpillars (100% mortality within 48 h) when injected (47). Recently, Reeves et al. reported that G. mellonella is susceptible to Aspergillus fumigatus strain ATCC 26933 and suggested a role in virulence for gliotoxin in promoting tissue penetration (38).
Our findings suggest that, similar to the findings reported by St. Leger and coworkers with A. flavus and A. fumigatus (47), G. mellonella hemocytes are able to phagocytose C. neoformans. However, the fact that inocula as low as 20 fungal cells of C. neoformans strain H99 were able to kill caterpillars suggests that successful phagocytosis does not necessary translate into fungal cell clearance. Possibly, phagocytosis is more effective with other C. neoformans strains than H99; this may explain why significantly higher numbers of inocula were needed for killing of caterpillars by a variety of C. neoformans strains.
The work reported here provides several examples of correlations between C. neoformans factors involved in mammalian pathogenesis and those necessary for G. mellonella killing. The C. neoformans polysaccharide capsule, as well as several C. neoformans genes previously shown to be involved in mammalian virulence (such as GPA1, PKA1, and RAS1), were shown to play a role in G. mellonella killing. Of note is that not only was strain cap59 attenuated compared to parent strain H99 (P < 0.001 at 37°C and P = 0.003 at 30°C), but killing of caterpillars by this mutant in capsule production was similar to that in the control groups that received PBS or heat-killed H99. This finding resembles findings with mammalian models (5, 8) and the amoeba system (25, 44), where acapsular strains of C. neoformans are completely avirulent. In the C. elegans (31) and D. melanogaster (3) systems, in contrast, although cap59 is attenuated, it results in killing of more than half of the animals. Also, heat-killed C. neoformans cells kill C. elegans but not wax moth caterpillars. Similarly, a ras1 mutant was avirulent in G. mellonella (Fig. 3 and Table 1), while pka1 and gpa1 mutants that were avirulent in mammalian models were still able to kill more than half of the caterpillars. This suggests that the RAS1 pathway may be more important in the caterpillar model than the G
protein-cAMP-PKA signaling pathway (2, 12). Previous work with D. melanogaster demonstrated that these virulence traits are involved in the killing of nonmammalian hosts after feeding (3). Coupled with the current studies, this suggests that C. neoformans factors involved in mammalian pathogenesis are involved in the survival of C. neoformans within nonmammalian hosts.
The choice of treatment for human C. neoformans infection depends on both the anatomic site of involvement and the immune status of the host. For management of severe cryptococcal infection involving the CNS, guidelines (39) based on a number of clinical trials suggest the initial use of amphotericin B plus flucytosine, followed by fluconazole consolidation therapy (4, 7, 24, 49). In the G. mellonella system reported here, amphotericin B, flucytosine, and fluconazole all prolonged survival of caterpillars, but combination therapy with amphotericin B plus flucytosine was the most effective therapy. Further studies evaluating the tropism of C. neoformans within G. mellonella and the concentration of antifungal drugs within tissues are needed to obtain a more thorough understanding of the efficacy of antifungal agents in this model.
Larvae of G. mellonella are inexpensive and relatively easy to manipulate, and their use may reduce the need to employ mammals for some in vivo studies for the efficacy of antifungal agents. This model may be a particularly useful addition to laboratories that study fungal pathogenesis but do not want to allocate the resources and commitment necessary to study C. elegans or D. melanogaster. However, important limitations of the G. mellonella model need to be considered. First, compared to other invertebrate models (such as C. elegans and D. melanogaster), the G. mellonella system has limited genetic tractability; the genome of this organism has not been sequenced. G. mellonella (similar to flies and nematodes) does not have an adaptive immune system. It is unlikely that all virulence traits that are important for mammalian infection will be significant in G. mellonella killing. For example, a clinical isolate of C. neoformans var. gattii that was associated with an outbreak of severe cryptococcal infection in Vancouver Island (15; http://ftp.cdc.gov/pub/infectious_diseases/iceid/2002/pdf/starr.pdf) was not considerably more virulent in the caterpillar model than other strains we tested (Table 1). Moreover, the impact of increased temperature on the host response of G. mellonella has not been studied in detail, and this may contribute to the increased killing of larvae at mammalian temperature that we noted in our experiments. Finally, a practical consideration for the use of G. mellonella is that caterpillars are unable to undergo an indefinite number of multiple injections, as this increases the mortality from trauma or infection from pathogenic contaminants on the insect cuticle.
In conclusion, a positive correlation exists between the pathogenicity of C. neoformans evaluated in the insect G. mellonella and in other model systems, including mammals. The G. mellonella model allows the administration of precise fungal inocula and the study of cryptococcal virulence at mammalian temperatures. The correlation between virulence in G. mellonella and mammalian models suggests that the G. mellonella-C. neoformans system can be used for the identification of new genes in C. neoformans involved in virulence, as well as for the in vivo evaluation of new antifungal agents.
Financial support was provided by the New Scholar Award in Global Infectious Diseases of the Ellison Medical Foundation and the Pfizer Fellowship in Medical Mycology from the Infectious Diseases Society of America to E.M. and by a grant from Aventis, SA, to F.M.A. and S.B.C.
|
|
|---|
isolates. Infect. Immun. 71:4831-4841.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»