OUP user menu

The pursuit of cryptococcal pathogenesis: heterologous hosts and the study of cryptococcal host–pathogen interactions

Roanna London, Benjamin Samuel Orozco, Eleftherios Mylonakis
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00056.x 567-573 First published online: 1 June 2006


Analysis of the molecular mechanisms by which a pathogen interacts with the human host is most commonly performed using a mammalian model of infection. However, several virulence-related genes previously shown to be involved in mammalian infection with Cryptococcus neoformans have also been shown to play a role in the interaction of these pathogens with invertebrates, such as Acanthamoeba castellanii, Caenorhabditis elegans, Dictyostelium discoideum, Drosophila melanogaster and Galleria mellonella. The study of host–pathogen interactions using these model hosts has allowed rapid screening of mutant libraries and can be used for the study of evolutionarily preserved aspects of microbial virulence and host response.

  • Caenorhabditis elegans
  • Cryptococcus neoformans
  • Drosophila melanogaster
  • fungal infection
  • Galleria mellonella
  • Toll-pathway


Cryptococcus neoformans is a significant cause of morbidity and mortality, with the majority of infections occurring in those with either immune suppression or dysfunction, including organ transplant recipients and human immunodeficiency virus (HIV)-infected individuals. Recent studies suggest that the study of microbial pathogenesis, particularly the identification and expression of virulence genes and the host responses, can be accomplished through a multimodal approach in increasingly relevant nonhuman models that identifies key genes in a facile manner for further investigation (Thomas, 1998; Tan, 1999a, b; Mahajan-Miklos, 2000; Rahme, 2000; Mylonakis & Aballay, 2005).

The aim of this review is to present three models that have been used in the study of C. neoformans pathogenesis. Caenorhabditis elegans, the well-studied microscopic nematode, Drosophila melanogaster, a fruit fly, and the larva of the greater wax moth Galleria mellonella are such organisms and their role in the study of C. neoformans will be discussed.

Caenorhabditis elegans

Background response of Caenorhabditis elegans to fungi

The free-living nematode Cn. elegans has proven to be a valuable model host for studying the pathogenesis of mammalian infections for a number of reasons. The short reproductive rate, large brood size and ease with which it is propagated in vitro contribute to the usefulness of Cn. elegans in the laboratory. However, the system has limitations for modeling C. neoformans infection. The killing assay relies on ingestion of the pathogen and not on systemic proliferation of the pathogen within the pseudocoele of the nematode. The immune response of the nematode is not as well studied, and there is no mechanism of host response described as there is in mammals or even in some other model organisms, such as D. melanogaster.

The discovery of the Toll/interleukin-1 pathway in Drosophila led to the discovery of Toll-like receptors in mammalian immunity. This discovery paved the way for studies of the host response to pathogens in a variety of nonmammalian organisms by demonstrating the conservation of host responses across very diverse organisms. Although in Cn. elegans there are no macrophages for clearance of microorganisms, genetic screens in bacteria have demonstrated that Cn. elegans response to Pseudomonas aeruginosa is mediated by a highly conserved p38 mitogen-activated protein kinase (MAPK) pathway (Kim, 2002). A microarray approach was also employed to identify effectors of the innate immune system important in the defense response against the fungal nematode pathogen Drechmeria coniospora. Various fungal-inducible genes were identified, and the expression of two antimicrobial peptides, NLP-29 and NLP-31, is regulated by Dr. coniospora (or bacterial) infection (Couillault, 2004; Liberati, 2004). NLP-31 demonstrated strong antifungal activity in vitro, comparable with that of drosomycin of Drosophila. Expression of NLP-29 and NLP-31 is controlled in part by tir-1, a gene encoding a highly conserved Toll–interleukin-1 receptor (TIR) domain protein (Couillault, 2004; Liberati, 2004). This suggests a TIR-1/p38 MAPK pathway that may represent a highly conserved defense response by Cn. elegans against the fungal pathogen (Couillault, 2004; Liberati, 2004). However, preliminary results suggest that C. neoformans does not up-regulate NLP-31 (Couillault, 2004).


Caenorhabditis elegans, commonly the wild-type strain Bristol N2, are easily propagated in the laboratory on nematode growth medium (NGM; consisting of NaCl, agar, peptone, cholesterol, CaCl2, MgSO4 and potassium phosphate) and feed on a lawn of Escherichia coli, typically the auxotrophic strain OP50. Pathogen killing assays can be performed by transferring Cn. elegans from a lawn of E. coli (OP50) to a plate containing the pathogenic bacteria or yeast of interest. It is of note that, should a particular microorganism under study allow for the worms to produce progeny, nematodes must be transferred to a fresh plate every 48 h. This prevents the original worm from being lost among the growing number of progeny worms.

For C. neoformans, strains are inoculated into yeast peptone dextrose (YPD) liquid media and grown with aeration at 37°C for 24 h. Then, using the smooth surface of a bent sterile pipette tip, 10 μL of C. neoformans culture is spread onto 35 mm tissue culture plates containing brain heart infusion (BHI) agar with antibiotics to suppress growth of microorganisms other than yeast, particularly of E. coli OP50 that may be present with the worms upon transfer. The plates are then incubated overnight. Transfer of Cn. elegans nematodes onto a lawn of C. neoformans is achieved using a platinum wire flattened on one end to make a ‘pick’ (Mylonakis, 2002). For most killing assays, 120–150 nematodes (divided across three plates) are used for each strain. Periodically the worms are monitored to check for viability on the cryptococcal plates and are considered to be dead when they no longer respond to the touch of the pick. This method proves particularly useful for screening libraries of mutant pathogens.

Cryptococcus neoformans–Caenorhabditis elegans system

In the laboratory setting it has been demonstrated that Cn. elegans can live and propagate on lawns of nonpathogenic yeasts, such as Cryptococcus laurentii and Cryptococcus kuetzingii, and have a lifespan similar to that of nematodes grown under normal conditions (NGM and E. coli OP50). Interestingly, on C. kuetzingii lawns, the nematodes are longer-lived than on E. coli OP50. However, all tested serotypes of C. neoformans accumulate within the intestine of the nematode (Fig. 1), and nematodes exposed to C. neoformans produce no progeny and die. Acapsular mutants of C. neoformans that do not accumulate in the intestine can also kill the nematodes. This is possibly through a separate mechanism involving a reaction to fragments of the yeast cell (such as cell wall components) or through the production of toxins. This nematode model has been shown to be useful in the study of the pathogenesis of C. neoformans owing to the potential of screening large numbers of mutants simultaneously to identify genes important in fungal virulence (Mylonakis, 2002). The polysaccharide capsule and several genes known to be involved in mammalian pathogenesis of C. neoformans, such as the G-α protein-cAMP–protein kinase A (PKA) signaling cascade that regulates melanin and capsule production, have also been shown to play an important role in its virulence in Cn. elegans (Fig. 2; Mylonakis, 2002).

Figure 1

Wild-type Cryptococcus neoformans accumulates in the gastrointestinal tract. Intact yeast cells are present in the distended gastrointestinal tract of Caenorhabditis elegans after feeding for 48 h on C. neoformans strain KN99α.

Figure 2

Cryptococcus neoformans virulence factors for mammalian infection also enhance killing of Caenorhabditis elegans. Survival of Cn. elegans N2 animals feeding on C. neoformans mutants with disruptions in the genes encoding the G protein–cAMP–protein kinase A and the RAS1-controlled signal transduction cascades demonstrated hypovirulence (gpa1, ras1 and pka1) or hypervirulence (pkr1), similar to results in mammalian models. P<0.001 for each of the mutants compared with the parental strain H99 (from Mylonakis, 2002).

A screen of a library of random C. neoformans mutants has been reported (Mylonakis, 2004). Of 350 mutants tested, seven were identified with attenuated virulence that persisted after evaluating linkage of the phenotype to the insertion by screening progeny after crossing the mutation back into a wild-type strain. Genetic analysis of one strain revealed an insertion in a gene homologous to Saccharomyces cerevisiae KIN1, which encodes a serine/threonine protein kinase. Cryptococcus neoformans kin1 mutants exhibited significant defects in virulence in murine models and resulted in increased binding to alveolar and peritoneal macrophages.

Recent observations suggest that some C. neoformans mutants may enable progeny production (Tang, 2005). This simplifies the screening process as a plate is scored for the presence of progeny without having to assess the presence of live and dead worms in high enough numbers to see statistically significant differences between the wild-type and mutant yeast under investigation. Evaluating brood size of Cn. elegans grown on mutant C. neoformans revealed that a ras1 mutant and acapsular strain, but not a laccase-negative strain, produced significantly more progeny than wild-type strains (Tang, 2005). A screen of ∼1500 random C. neoformans mutants generated in an F99 (MATαura5) background using biolistic transformation to generate mutants identified three mutants with increased progeny production, one of which was found to be disrupted in a gene homologous to S. cerevisiae Rom2p, required for cell integrity under both heat and osmolar stress. Importantly, the C. neoformans rom2 mutant was also significantly attenuated in the mouse nasal inhalation infection model (Tang, 2005).

Drosophila melanogaster

Background response of Drosophila to fungi

The role of the host–pathogen interaction with Drosophila has been well studied, especially since 1996 when it was demonstrated that Toll receptor activation was crucial in antifungal defenses of the fly and culminated in the synthesis of, among other things, antifungal peptides (Lemaitre, 1996). Work since then has shown that the innate immune system in Drosophila is mediated by Toll and Imd signaling pathways involved in mycoses and bacterial infections, respectively (although there appears to be significant cross-talk). A fungal infection triggers Toll receptors through a cytokine-like protein Spatzle. Spatzle is cleaved from the fat-body cells and the interaction of Spatzle and Toll causes Pelle, a threonine-serine kinase, to phosphorylate Cactus. Cactus thus targeted for degradation allows for the translocation from the cytoplasm to the nucleus of the transcription factors Dorsal and Dif. This leads to the transcription of antimicrobial peptides. Wild-type Drosophila is able to survive injection of Aspergillus fumigatus, C. neoformans or Candida spp. (Lemaitre, 1996; Alarco, 2004; Apidianakis, 2005). However, Toll mutant flies do not produce antimicrobial peptides and are sensitive to Aspergillus (Bhabhra, 2004) and Candida (Alarco, 2004). In addition, it should be noted that the site of inoculation may play a role in the host response via Toll signaling in Drosophila (Apidianakis, 2005).

Cellular responses in Drosophila are also involved in the fly response to fungi. Drosophila hemocytes include plasmatocytes (professional phagocytes), crystal cells (which contain the enzymes necessary for humoral melanization that accompanies a number of immune reactions) and lamellocytes (cells that differentiate after parasitism and form a capsule around the invader) (Irving, 2005). However, the activation of the cellular response in Drosophila is complex and affected by multiple signaling pathways (Zettervall, 2004).


This system allows both systemic and local inoculation of the microorganism under investigation. Furthermore, there are multiple means of exposing Drosophila to these microorganisms to study their pathogenesis. Direct injection is commonly used to inoculate the fly with a microorganism that cannot penetrate the exoskeleton and cause infection. This can be accomplished by pricking the dorsal thorax or abdomen with a needle dipped in a liquid culture of the pathogen. Injection into the abdomen may allow for larger inocula, but also risks more trauma than injection to the thorax. An alternative method uses microinjection to inoculate exact inocula of microbes into the body cavity. Some pathogens are able to penetrate the first line of defenses in Drosophila. Alternatively, such a pathogen may be administered directly to the fly in its food or sprayed onto the outer surfaces of the fly; the fly can also be rolled for 1–2 min on a lawn of the pathogen grown on an agar plate.

Cryptococcus neoformans–Drosophila melanogaster system

Similar to the infection with A. fumigatus, wild-type Drosophila is resistant to systemic infection with C. neoformans (Apidianakis, 2004). For example, wild-type flies are able to eliminate 400 CFU of C. neoformans strain H99 3–4 days following injection into the hemolymph. However, C. neoformans (and not C. laurentii or C. kuetzingii) are pathogenic to wild-type flies when ingested by the fly. Cryptococcus-mediated killing of Drosophila following ingestion involves a variety of factors relevant to mammalian pathogenesis, including genes associated with the PKA and RAS signal transduction pathways. The Toll pathway seems to play no role in resistance to infection by ingested C. neoformans, but similar to other fungal infections (Lemaitre, 1996; Alarco, 2004) remains crucial in host defenses to systemic inoculation by C. neoformans. Similar to data from studies in Cn. elegans, acapsular mutants are hypovirulent but can still kill Drosophila. However, mutations in the PKA or Ras1 pathways have a more severe effect on virulence (Apidianakis, 2004).

Galleria mellonella

Background response of Galleria mellonella to fungi

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. The insect immune response has a number of structural and functional similarities to the innate immune response of mammals. Six types of hemocytes have been identified in G. mellonella, and the insect response includes phagocytosis, ‘nodulation’ (encapsulation of large invading pathogens by layers of hemocytes) and melanization (reviewed in Kavanagh & Reeves, 2004). Hemocytes of G. mellonella are capable of phagocytosing Candida cells, and the kinetics of phagocytosis and microbial killing are similar in the insect hemocytes and human neutrophils (Slepneva, 2003; Bergin, 2005). Immunoblotting of G. mellonella hemocytes with antibodies raised against human neutrophil phox proteins revealed the presence of proteins homologous to gp91phox, p67phox, p47phox and the GTP-binding protein rac 2, proteins which function in the generation of reactive oxygen species in phagocytosis (Slepneva, 2003; Bergin, 2005).

As noted above, humoral responses are also involved in the G. mellonella immune response. For example, a defensin-like peptide was recently identified in G. mellonella. The amino acid sequence of the defensin-like peptide exhibits similarities to drosomycin and the recombinant peptide is active against the entomopathogenic fungus Metarhizium anisopliae (but not against yeast, or gram-negative or gram-positive bacteria) (Schuhmann, 2003).

In Candida albicans there is significant correlation between the virulence in G. mellonella and mice (Brennan, 2002; Bergin, 2003). Virulent Ca. albicans strains grow rapidly in larval hemolymph, limit nodulation, and produce toxins that damage hemocytes and the fat body (Dunphy, 2003). Hyphal transition is associated with Candida-mediated killing of G. mellonella and avirulent myosin-I-defective yeast cells are rapidly removed from the hemolymph in vivo (Dunphy, 2003).

Galleria mellonella has also been used to study the pathogenicity of Aspergillus spp. Conidia of Aspergillus flavus are not virulent when applied to the surface of healthy caterpillars, but kill the caterpillars when injected, within 72 h of host death covering the dead caterpillars in a thick coat of conidia (St Leger, 2000). Aspergillus flavus as well as spores of the four isolates of A. fumigatus studied by St Leger et al. are rapidly phagocytosed. However, unlike A. flavus, strains of A. fumigatus fail to germinate within hemocytes and produce little or no symptoms, and no hyphal bodies are observed free in the hemolymph (St Leger, 2000). Recently, Reeves (2004) reported that G. mellonella is susceptible to A. fumigatus strain ATCC #26 933 and suggested a role in virulence for gliotoxin in promoting tissue penetration.


In order to conduct a virulence assay with G. mellonella a Hamilton 10 μL syringe is used to inoculate a G. mellonella final instar larva with the culture of the pathogen of interest or a control solution of phosphate-buffered saline (PBS). The injection site is generally the last left proleg. The area is usually swabbed with alcohol and a broad-spectrum antibacterial can be administered to prevent infection by bacteria that colonize the larvae. The larvae may be immobilized by placing them at 4°C for 30 min prior to injection. After inoculation larvae are stored in containers and the number of dead is recorded daily. Furthermore, for the study of antifungal agents, one may inject (usually into a different proleg) antifungal agents (Mylonakis, 2005).

Cryptococcus–Galleria mellonella system

Cryptococcus neoformans proliferates within the hemocoel of G. mellonella and, in spite of phagocytosis by hemocytes, strains from all varieties of C. neoformans kill G. mellonella. The rate and extent of killing depends on the cryptococcal strain and the number of colony-forming units injected into the larva. Similar to the behavior in the Cn. elegans and the D. melanogaster systems, C. neoformans maintains yeast-like morphology and although the development of capsule and melanization has not been documented within these invertebrate hosts, the genes that enable C. neoformans to make capsule and become melanized are involved in killing of the host. More specifically, the killing process is also dependent on many of the same factors proven to play a role in mammalian virulence, such as CAP59, GPA1, RAS1 and PKA1. Interesting similarities also exist in the role of the mating locus in pathogenesis in G. mellonella and in mammals. The MFa1 gene is induced during the proliferative stage of the infection models. Killing of G. mellonella caterpillars that received the MATα strain JEC21 was significantly faster than in the group that received the otherwise isogenic MATa strain JEC20. Similarly, we found that there was no pMFα1::GFP expression in 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 MFα1 gene is induced during the proliferative stage of the infection, similar to the findings in mammals (del Poeta, 1999). Finally, killing of G. mellonella by C. neoformans var. grubii strains H99, KN99a, KN99α and KN99-5 was identical, similar to findings in mammals (del Poeta, 1999; Nielsen, 2003).

Interestingly, the efficacy of antifungals in the G. mellonellaC. neoformans system also correlates with what is observed in mammalian systems or human treatment. Amphotericin B, flucytosine or fluconazole given to a larva post injection of C. neoformans demonstrated antimicrobial properties as survival was improved. Monotherapy with amphotericin B or flucytosine increased the mean survival time of G. mellonella to 9 days compared with 6 days in the control group. Furthermore, the combination of amphotericin B with flucytosine was more effective than the single drug regimen of amphotericin B or flucytosine alone (Mylonakis, 2005).


Three organisms, Caenorhabditis elegans, Drosophila melanogaster and Galleria mellonella, have been presented as models for the study of cryptococcal virulence and host–pathogen interactions (Table 1). These organisms have proven to be particularly promising for the study of fungal pathogens, and can provide useful insights for the study of Cryptococcus neoformans pathogenesis. Of note, in addition to these model hosts, Acanthamoeba spp. (Levitz, 2001; Steenbergen, 2001), a free-living amoeba, and the slime mold Dictyostelium discoideum (Steenbergen, 2003) have also been used to model cryptococcal infection.

View this table:
Table 1

Comparison of Caenorhabditis elegans, Drosophila melanogaster and Galleria mellonella systems for study of Cryptococcus neoformans pathogenesis

Cn. elegansD. melanogasterG. mellonella
Reproductive capacity (rate/number)HighModerateNot applicable
Similarity between model and mammalian system in host–pathogen interactionModerately studiedWell studiedFurther study needed
Available mutants and protocolsManyManyLimited
Genome sequenced and annotatedYesYesNot available
Compatibility with genetic tools (RNAi library, microarrays, etc.)HighHighLow
Ease of genetic manipulation (forward, reverse)HighModerateLow

Moreover, the interactions of these model organisms with C. neoformans may suggest that the virulence factors important in human disease pathogenesis probably evolved through interactions with more primitive organisms (Levitz, 2001; Steenbergen & Casadevall, 2003; Casadevall, 2005).

First, Cn. elegans is a genetically tractable model with properties that make it attractive to the laboratory setting. It has a short reproductive cycle, produces a large number of offspring, and is inexpensively procured and maintained in a laboratory. The Cn. elegans model knowledge base is expanding as it has been in use for several decades. There are a number of genetic tools available, including a sequenced and annotated genome, an RNAi feeding library and microarrays. In addition, mutants and protocols are also readily available for order. Finally, it is transparent, and easily visualized and manipulated in a dissecting microscope.

Drosophila melanogaster is also extensively studied and shares many of the same technical advantages as Cn. elegans. For instance, this model also provides a sequenced and annotated genome, RNAi library and microarrays. It is inexpensive, has a relatively short reproductive life cycle, and many protocols and mutants are available. Moreover, it allows both systemic and local inoculation.

Galleria mellonella is inexpensive and readily available. In addition, it has the advantage of being injected easily with precise amounts of antibiotics and numbers of pathogens, the ability to model higher temperature systems, and ease of handling.

However, each model has its drawbacks. Galleria mellonella does not have a sequenced genome and wild-type Drosophila are resistant to C. neoformans infection. The host response of Cn. elegans is not as easily paralleled to a mammalian system, while Drosophila is not as easily manipulated in forward and reverse genetics. However, each model has the ability to bring new insights to the study of fungal pathogenesis and may in fact have the very type of interactions that initially sparked and maintained evolutionary pressure for the virulence factors seen in mammals.


Financial support was provided by a K08 award AI63084-01 from NIH and a New Scholar Award in Global Infectious Diseases of the Ellison Medical Foundation to E.M.


  • Editor: Stuart Levitz


View Abstract