Galleria mellonella as a Model System for Studying Listeria Pathogenesis

Abstract

Essential aspects of the innate immune response to microbial infection are conserved between insects and mammals. This has generated interest in using insects as model organisms to study host-microbe interactions. We used the greater wax moth Galleria mellonella, which can be reared at 37°C, as a model host for examining the virulence potential of Listeria spp. Here we report that Galleria is an excellent surrogate model of listerial septic infection, capable of clearly distinguishing between pathogenic and nonpathogenic Listeria strains and even between virulent and attenuated Listeria monocytogenes strains. Virulence required listerial genes hitherto implicated in the mouse infection model and was linked to strong antimicrobial activities in both hemolymph and hemocytes of infected larvae. Following Listeria infection, the expression of immune defense genes such as those for lysozyme, galiomycin, gallerimycin, and insect metalloproteinase inhibitor (IMPI) was sequentially induced. Preinduction of antimicrobial activity by treatment of larvae with lipopolysaccharide (LPS) significantly improved survival against subsequent L. monocytogenes challenge and strong antilisterial activity was detected in the hemolymph of LPS pretreated larvae. We conclude that the severity of septic infection with L. monocytogenes is modulated primarily by innate immune responses, and we suggest the use of Galleria as a relatively simple, nonmammalian model system that can be used to assess the virulence of strains of Listeria spp. isolated from a wide variety of settings from both the clinic and the environment.

Listeriae are rod-shaped, motile, facultative, anaerobic Gram-positive bacteria that are ubiquitously distributed in the environment (28). Of the six species that comprise the genus Listeria, only L. monocytogenes and L. ivanovii are pathogenic and cause disease, while strains of the species L. innocua, L. welshimeri, L. seeligeri, and L. grayi are generally considered to be nonpathogenic (26). L. monocytogenes is a major food-borne pathogen, and listeriosis is an invasive disease that in its severest form can lead to meningitis, meningoencephalitis, septicemia, and abortions (38). Listeriosis occurs primarily in pregnant women, newborn infants, and the elderly as well as in immunocompromised patients, with a mortality rate of about 30% (22, 36). The virulence of L. monocytogenes has been linked to a 9.6-kb pathogenicity island designated vgc (virulence gene cluster) that comprises six genes encoding its major virulence determinants. These are (i) prfA, a master regulator of many known listerial virulence genes; (ii) hly, encoding listeriolysin, a hemolysin required for bacterial escape from the host primary vacuole to the host cytoplasm; (iii) two phospholipase genes denoted plcA and plcB, for facilitating lysis of host cell membranes; (iv) actA, encoding a surface bound protein that directs polymerization of host cell actin and is required for intracellular motility; and (v) mpl, encoding a metalloproteinase which is thought to work together with the plcB product to facilitate cell-to-cell spread (28). Presently, identification and characterization of novel virulence factors rely on assessing mutant bacteria for growth in the organs of infected mice. Nevertheless, the dependence on mouse infection models limits large-scale screening for additional mutants defective in their ability to grow in the host intracellularly or for those required to overcome host innate defenses (33).

The possibility of addressing many aspects of mammalian innate immunity in invertebrates has opened a new arena for developing invertebrate models to study human infections. Recently the use of invertebrate models, in particular the fruit fly Drosophila melanogaster, has been introduced for the study of septic listerial infections (37). Listeria mutants attenuated for virulence in a mouse model exhibited lowered virulence in this model. The Drosophila model system has powerful genetic tools available and has thus provided deeper insights into molecular mechanisms of the interactions between Listeria and the insect innate immune system (1, 8-10, 18, 24). However, a recent study has shown that even nonpathogenic L. innocua strains cause lethal infections of Drosophila, limiting it use as a discerning model for the study of virulence potential among pathogenic L. monocytogenes isolates (32).

We have a longstanding interest in host-pathogen interactions of the greater wax moth, Galleria mellonella, in particular with entomopathogenic microbes (55). Recently, Galleria has also emerged as a reliable model host to study the pathogenesis of many human pathogens (7, 11, 12, 17, 21, 30, 31, 39-42, 44, 46, 48-51). Among the advantages provided by the Galleria model host (e.g., low rearing costs, convenient injection feasibility, and status as an ethically acceptable animal model), it is of particular importance that Galleria has a growth optimum at 37°C, to which human pathogens are adapted and which is essential for synthesis of many virulence/pathogenicity factors. Significantly, a correlation between the virulence of a pathogen in G. mellonella and that in mammalian models has been established (16, 25).

The innate immunity of Galleria is a complex, multicomponent response involving hemolymph coagulation, cellular phagocytosis, and phenol oxidase-based melanization. Importantly, killing of pathogens is achieved similarly to that in mammals, i.e., by enzymes (e.g., lysozymes), reactive oxygen species, and antimicrobial peptides (e.g., defensins). Galleria employs recognition of nonself microbe-associated molecular patterns by germ line-encoded receptors (e.g., Toll and peptidoglycan recognition proteins) (52). Recently, we have found that Galleria also senses pathogens by danger signaling, by detecting either nucleic acids released from damaged cells or peptides resulting from proteolytic cleavage of self proteins by matrix metalloproteinases (3-6).

In this work we examined the Galleria model of septic infection for its ability to differentially distinguish between infections caused by strains with different virulence potentials in the mouse infection model, as well as in avirulent strains of Listeria. We found that the Galleria model is highly discriminatory in assessing the pathogenic potential of Listeria spp., and we observed a strong correlation with the virulence previously determined in the mouse model of infection. Here, we present data indicating that the Galleria model also replicates many aspects of innate immune function, such as the constitutive expressions of potential antimicrobial factors following infection. Also, prior induction of immunity in Galleria can protect larvae from septic infection with highly pathogenic L. monocytogenes.

MATERIALS AND METHODS

Insects, bacteria, and media.

G. mellonella larvae were reared on an artificial diet (22% maize meal, 22% wheat germ, 11% dry yeast, 17.5% bees wax, 11% honey, and 11% glycerin) at 32°C in darkness prior to use. Last-instar larvae, each weighing between 250 and 350 mg, were used in all experiments. The different Listeria species, serotypes, and mutants used in this experiment are listed in Table 1. The wild-type strain Listeria monocytogenes EGD-e used in this study belongs to serotype 1/2a (23). The bacterial cultures were grown aerobically in brain heart infusion medium (BHI) (Difco, Franklin Lakes, NY) at 37°C and on BHI agar plates. For long-term storage, Listeria strains were frozen in BHI with 30% glycerol at −80°C. For injection experiments, Listeria cultures with a density of 10⁹ CFU/ml in 10 ml of BHI broth growing in logarithmic phase were used. Bacterial inoculums were washed and serially diluted using 0.9% NaCl to appropriate concentrations. Fifty microliters of each dilution was plated out on BHI agar plates and incubated at 37°C for 24 h, and the bacterial CFU were used to calculate the inoculum injected. Cultures of L. innocua harboring the pUvBBAC vector containing the vgc1 locus from L. monocytogenes EGD-e were grown in the presence of 5 μg/ml erythromycin and 5 μg/ml kanamycin (27). The Escherichia coli host for plasmid constructions was INVαF′. Plasmid DNA was transferred to INVαF′ using the method of Hanahan (29). The electroporation protocol of Park and Stewart (43) was utilized for transformation of L. monocytogenes strains.

TABLE 1.

Bacterial strains used in this study

Species and strain	Serotype	Reference
L. monocytogenes
EGD-e	1/2a	23
L99	4a
L312	4b
SLCC2376/ATCC 19116	4c
ATCC 19117	4d
EGD-e Δvgc		Present study
EGD-e ΔuhpT		Present study
EGD-e ΔprfA		14
EGD-e Δhly		25
EGD-e ΔactA		13
EGD-e ΔplcA		45
EGD-e ΔplcB		25
EGD-e Δmpl		25
EGD-e ΔinlAB		35
L. innocua
CLIP 11262	6a
CLIP 11262 vgc1		27
L. welshimeri SLCC 5334/ATCC 35897	6b
L. grayi CLIP 12515
L. ivanovii PAM55	5
L. seeligeri SLCC 3954/ATCC 35967	1/2b

Deletion of the virulence gene cluster (vgc) comprising the genes prfA, plcA, hly, mpl, actA, and plcB in L. monocytogenes EGD-e.

A PCR product of approximately 2,500 bp was generated with the forward primer 5′-TCTAATCGTGAACTAGCTG-3′ and the reverse primer 5′-CGTAAGTGTTCGTGATGCAGCTTATG-3′ using chromosomal DNA of the nonpathogenic strain L. innocua NCTC 11289. The PCR product was cloned into plasmid pAUL-A and transformed into L. monocytogenes EGD-e, and the isogenic vgc mutant strain was generated as described previously (47). A 12-kb fragment comprising the genes prfA, plcA, hly, mpl, actA, and plcB was replaced by the 2.5-kb genomic fragment present between prs and ldh of L. innocua NCTC 11289. The loss of vgc was confirmed by sequencing and by immunoblotting with monoclonal antibodies directed against proteins PlcA, Hly, Mpl, ActA, and PlcB.

Construction of a chromosomal ΔuhpT deletion mutant of L. monocytogenes EGD-e.

A ΔuhpT (or Δhpt) mutant harboring only the first 22 amino acid residues of UhpT was obtained as follows. Appropriate regions flanking the uhpT gene were PCR amplified with oligonucleotide primers uhpT-for1 (5′-AGAAACGGAGCTCGTGATTC-3′) and uhp-rev2 (5′-AAAGTGTTGGATCCATTGTTG-3′) or uhpT-for3 (5′-TAAGTTGGATCCAATGAGTG-3′) and uhpT-rev4 (5′-GCTAAGTCGACTCAATCCG-3′), respectively. Both PCR products were digested with BamHI and ligated to each other. The ligation product containing the deletion was selectively amplified with oligonucleotide primers uhpT-for1 and uhpT-rev4. The corresponding DNA fragment flanked by SacI and SalI restriction sites was inserted into the temperature-sensitive shuttle vector pAUL-A (36). The L. monocytogenes wild-type strain EGD-e was transformed with this construct, and chromosomal integration of the plasmid and plasmid excision and curing were carried out as previously described (35). Replacement of the wild-type allele by its truncated ΔuhpT derivative was confirmed by sequencing of the PCR product obtained with oligonucleotide primers uhpT-for1 and uhpT-rev4.

G. mellonella injection and CFU count of L. monocytogenes.

Bacterial inoculums were injected dorsolaterally into the hemocoel of last-instar larvae using 1-ml disposable syringes and 0.4- by 20-mm needles mounted on a microapplicator. After injection, larvae were incubated at 37°C. Caterpillars were considered dead when they showed no movement in response to touch. No mortality of Galleria larvae was recorded when they were injected with 0.9% NaCl. For CFU counting, Galleria larvae were infected with L. monocytogenes (10⁶ CFU/larva) and were homogenized in BHI medium with 1% Triton X-100. Homogenates were plated onto Palcam Listeria selective agar plates (Heipha Diagnostika), and colonies were counted after incubation at 37°C for 48 h. For each time point, homogenates of 10 larvae were plated individually for CFU count.

Preimmune activation of G. mellonella larvae and antibacterial activity assays.

Last-instar larvae were injected independently with 10 mg/ml lipopolysaccharide (LPS) (purified Escherichia coli endotoxin 0111:B4) (catalog no. L2630; Sigma, Taufkirchen, Germany) or heat-killed L. monocytogenes to trigger strong immune responses. The heat-killed preparation of bacteria was obtained as follows. An exponential-phase bacterial culture was harvested, centrifuged, and washed three times in 0.9% NaCl. The recovered bacteria were resuspended in NaCl and incubated at 85°C for 1 h. After two additional washes in NaCl, the wet weight of the bacterial pellet was adjusted to 10 mg/ml in 0.9% NaCl. Each 10 μl was administered directly into the hemolymph of the larvae to induce an immune response. At 24 h after administration of LPS or heat-killed Listeria, 10⁶ CFU of L. monocytogenes strain EGD-e was injected into each larva for survival counts. To investigate the presence of antimicrobial activities in Galleria killing viable Listeria, we used the inhibition zone assay. In brief, petri dishes (100 mm) were filled with 7 ml BHI medium containing 0.7% high-purity agar-agar (Roth, Karlsruhe, Germany), and subsequently, 10⁴ CFU of viable bacteria in logarithmic growth phase was plated. Hemolymph samples from larvae were extracted at 24 h following immune induction and inlaid into 4-mm-diameter wells previously punched into the agar. The diameters of clear zones were measured after 24 h of incubation at 37°C.

Ex vivo infection of Galleria hemocytes.

The hemocytes from Galleria were maintained at 37°C in Schneider medium (BioWhittaker) supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Bio West). Intracellular growth of L. monocytogenes in primary Galleria hemocytes was monitored by using cell monolayers on sterile coverslips for immune fluorescence microscopy observation and on microtiter plates for estimation of bacterial CFU. Briefly, bacterial cultures logarithmically grown in BHI medium at 37°C were washed with 0.9% NaCl before infection. After 60 min of infection at 37°C, the hemocytes were carefully washed three times with cell culture medium, followed by the addition of 1 ml of Schneider medium containing 50 μg/ml of gentamicin. To quantify bacterial intracellular growth, cell monolayers were lysed by sterile water containing 0.2% Triton X-100 for 4 h after L. monocytogenes infection, and CFU were determined by plating dilutions of cell lysates on BHI plates followed by overnight incubation at 37°C. For microscopic analysis, cells were fixed by placing a drop of 3.7% paraformaldehyde on the coverslips and incubating at room temperature for 10 min. Coverslips were washed by dipping them into sterile phosphate-buffered saline (PBS), and then the hemocytes were permeabilized with 1 ml 0.2% Triton X-100 in PBS for 1 min and again washed in PBS. The coverslips were incubated with ActA N4 and ActA N81 (1:1) monoclonal antibodies (Helmholtz Zentrum for Infection Braunschweig; prepared by Jürgen Wehland) for 30 min at 33°C. After being washed three times with PBS, coverslips were incubated with Cy3-labeled secondary anti-mouse antibody (1:100) (Dianova, Hamburg, Germany) and Alexa Fluor 488 conjugated to phalloidin (1:100) (Molecular Probes, Invitrogen, Carlsbad, CA) in PBS containing 1% bovine serum albumin for 30 min at 33°C. Subsequently, coverslips were washed three times with PBS and mounted using Prolong Gold antifade reagent (Invitrogen).

Quantitative real time RT-PCR.

Three larvae per treatment for each time point were homogenized in 1 ml of Trizol reagent (Sigma), and whole animal RNA was extracted according to the manufacturer's recommendations. RNA integrity was confirmed by ethidium bromide gel staining, and quantities were determined spectrophotometrically. Quantitative real time reverse transcription-PCR (RT-PCR) was performed with the real-time PCR system Mx3000P (Stratagene) using the FullVelocity SYBR green quantitative RT-PCR master mix (Stratagene) according to the protocols of the manufacturer. We used appropriate primers along with the 10 ng RNA per reaction to amplify the genes for 18S RNA (2), actin (3), IMPI (3), galiomycin (5), gallerimycin (3), and lysozyme (3).

Data analysis.

All experiments were performed a minimum of three times. Significant differences between two values were compared with a paired Student's t test. Values were considered significantly different when the P value was less than 0.05.

RESULTS

Mortality in Listeria-infected Galleria larvae depends on the pathogen load.

We examined the susceptibility of G. mellonella to a known pathogenic strain and a nonpathogenic strain of Listeria. Larvae were injected with 10⁷, 10⁶, 10⁵, and 10⁴ bacteria of either L. monocytogenes strain EGD-e or L. innocua, and mortality was recorded up to 7 days postinjection. At 10⁷ CFU we observed killing of Galleria irrespective of whether the pathogenic or nonpathogenic Listeria strain was used. At doses below 10⁶ CFU, clear differences in lethality between L. monocytogenes and L. innocua were observed (Fig. 1A and B). Differences in mortality were less apparent at lower doses (10⁵ and below), and no deaths were recorded when larvae were injected with 0.9% saline alone. Thus, for subsequent experimental assays we used 10⁶ CFU/larva as the inoculating dose to study septic infection by Listeria spp.

FIG. 1.

Dose-dependent survival of Galleria caterpillars after inoculation with L. monocytogenes and L. innocua. Bacteria were grown to log phase in BHI medium at 37°C. The time course of survival of the larvae when inoculated with pathogenic L. monocytogenes strain EGD-e (A) and/or apathogenic L. innocua (B) depended on the amount of CFU injected. Injection of 10⁷, 10⁶, 10⁵, or 10⁴ CFU/larvae resulted in higher mortality with EGD-e than with L. innocua. Results represent means of at least three independent determinations ± standard deviations for 10 animals per treatment.

Listeria infection of G. mellonella resembles that seen with vertebrates.

To examine whether cellular aspects of infection are similar to those observed with vertebrate cells, we isolated hemocytes from naive larvae and subjected them to infection with L. monocytogenes. Bacteria were incubated with hemocytes for 1 h to allow for invasion. Subsequently the supernatant of cultures were replaced with fresh medium supplemented with 50 μg/ml of gentamicin to kill extracellular Listeria cells. Bacteria growing intracellularly were monitored by immunofluorescence microscopy using an ActA-specific monoclonal antibody. Actin-based motility of bacteria was detected by colocalization of intracellular bacteria with fluorescence derived by actin-specific Alexa Fluor 488-conjugated phalloidin. In infected cells, we detected intracellular bacteria either covered by actin “clouds” or undergoing rapid movement as judged by the lengths of their respective actin “comet tails,” thus resembling the infection process seen previously in vertebrate cells (Fig. 2A).

FIG. 2.

Multiplication of L. monocytogenes in Galleria. (A) L. monocytogenes cells were stained using ActA antibodies (resulting in red fluorescence), and host actin of hemocytes was stained using Alexa-phalloidin (resulting in green fluorescence). Note that Listeria organisms are spreading throughout the cytosol of the hemocyte, and actin tails at the poles of some of the bacteria are visible. (B) To determine rate of multiplication of L. monocytogenes in Galleria larvae, we determined the listerial load from infected larvae at several time points postinfection. For each time point, homogenates of 10 larvae were plated individually for CFU count on Listeria selective Palcam agar plates. These results are shown as one dot, and resulting mean values of are shown in red. Surviving animals contained reduced listerial load, whereas dying larvae contained about 2 × 10⁵ CFU, as indicated by a circle. The experiment was repeated three times with similar results.

We also addressed the question of whether mortality of Galleria is associated with the growth of L. monocytogenes in infected larvae. Galleria larvae infected with EGD-e at 10⁶ CFU/larva were homogenized in BHI medium containing 1% Triton X-100 and then plated onto Listeria-selective Palcam plates. L. monocytogenes colonies were counted following incubation at 37°C for 48 h. We observed a rapid decrease in the CFU count at 1 h postinfection, indicating that the constitutive immune defenses of Galleria were highly effective in reducing the pathogen load (Fig. 2B). At 24 h and 48 h postinfection, successive increases in bacterial CFU were recorded. In Galleria larvae that had succumbed to infection at 96 h postinjection, high numbers of bacteria were detected.

Listeria shows species-specific pathogenesis for Galleria.

We investigated the Galleria model for its ability to distinguish between pathogenic and nonpathogenic strains comprising all known Listeria species. There was a clear difference between pathogenic L. monocytogenes and the nonpathogenic L. innocua, L. seeligeri, L. welshimeri, and L. grayi strains. L. ivanovii was clearly less pathogenic in the Galleria model than L. monocytogenes but nevertheless still demonstrated a small but significant difference in mortality compared to the nonpathogenic L. innocua (Fig. 3A).

FIG. 3.

Time-dependent survival of Galleria larvae after inoculation with different Listeria species and L. monocytogenes serotypes. The time course of survival of the larvae varies with the type of Listeria species employed for inoculation. (A) Inoculation with 10⁶ CFU/larva EGD-e resulted in a significantly higher rate of killing of larvae than inoculation with L. ivanovii, L. innocua, L. seeligeri, L. welshimeri, or L. grayi. Only L. ivanovii had a tendency for enhanced killing of Galleria with respect to L. innocua (P < 0.05). (B) Inoculation with different L. monocytogenes serotypes resulted in various rate of killing. Serotype 4b showed a significant high rate of killing of larvae, whereas the pathogeneses of 4a, 4c, and 4d were strongly attenuated, with respect to the pathogenic serotype 1/2a strain EGD-e. Results represent means of at least three independent determinations ± standard deviations for 10 animals per treatment.

Serotype-specific virulence in Galleria.

The ability to distinguish between L. monocytogenes strains previously characterized as being either highly virulent or attenuated in the mouse model of infection was examined next. Among the different serotypes tested, a serotype 4b strain was the most pathogenic, causing a significantly higher rate of killing than the serotype 1/2a strain (EGD-e) (Fig. 3B). However strains of other serotypes of L. monocytogenes, such as 4a, 4c, and 4d, exhibited significantly lower virulence, mirroring their reduced pathogenic potential previously seen in the mouse infection model (Fig. 3B).

Septic infection of Galleria is dependent on the vgc locus of L. monocytogenes.

The vgc locus encodes the major factors required for virulence of L. monocytogenes (54). We generated and used an EGD-e mutant with the vgc locus deleted, EGD-e Δvgc (see Materials and Methods), and examined its ability to kill Galleria. The EGD-e Δvgc strain was highly attenuated for killing ability in comparison to the wild-type EGD-e in the Galleria model (P < 0.0005) (Fig. 4A). Isogenic strains lacking prfA, hly, actA, plcB, and mpl were highly attenuated for the killing of infected larvae (Fig. 4B and C). Interestingly, deletion of plcA revealed no virulence attenuation. This was also the case with a mutant lacking internalins A and B.

FIG. 4.

Contributions of major virulence-related genes of L. monocytogenes in the mortality of Galleria. (A) The vgc locus is responsible for the pathogenicity of L. monocytogenes. L. monocytogenes with vgc deleted had a significant reduction of killing capacity in comparison to EGD-e (P < 0.005) but still showed greater killing ability than L. innocua (P < 0.05). (B and C) Deletion of single virulence genes prfA, hly, actA, plcB, mpl, and uhpT (hexose-phosphate transporter gene) resulted in significantly reduced mortality in the Galleria model system. However deletion of vgc-associated plcA or inlA and intB caused no significant reduction in mortality rates. The P value for the mortality rates between ΔuhpT and ΔinlAB was found to be 0.005, and that between ΔuhpT and ΔplcB was 0.05. Results represent means of at least three independent determinations ± standard deviations for 10 animals per treatment.

We also assessed the contribution of an additional PrfA-regulated factor, the hexose phosphate transporter UhpT, which is required for efficient growth and survival of the bacterium in infected vertebrate cells. The ΔuhpT mutant was also found to be attenuated for killing of infected larvae (Fig. 4C).

We used a recombinant L. innocua strain engineered to harbor the vgc locus and found that following infection it was significantly more pathogenic as observed by the higher rates of killing of the larvae than with L. innocua alone (Fig. 5). Nevertheless, it was significantly less pathogenic than the EGD-e strain from which the vgc locus was derived (P < 0.05).

FIG. 5.

Insertion of EGD-e-derived vgc in L. innocua results in induced virulence. Artificial introduction of the vgc1 locus into otherwise nonpathogenic L. innocua resulted in a significant increase of virulence with respect to that of wild-type L. innocua. Results represent means of at least three independent determinations ± standard deviations. Each repetition contained 30 larvae per treatment.

Expression patterns of genes encoding antimicrobial peptides.

Galleria is capable of synthesizing a broad spectrum of antimicrobial peptides in response to septic injury (43). Consequently, we were interested in examining whether antimicrobial peptides are induced in Galleria upon septic Listeria infections. Larvae were infected with L. monocytogenes, and RNA was extracted at 1, 6, and 24 h postinfection. Transcriptional activation is represented as the fold change of expression of immune-related genes in infected Galleria relative to the mock-injected control larvae and normalized using the housekeeping 18S RNA gene (Fig. 6). Increased levels of lysozyme expression were recorded throughout the whole period of L. monocytogenes infection. The amounts of immune-related gallerimycin and lysozyme mRNAs were found to be induced about 4.0-fold and 2.0-fold at 1 h postinfection. At 6 h postinfection we observed increased galiomycin (∼11-fold), gallerimycin (∼80-fold), and lysozyme (∼10-fold) mRNA levels. Induced expression of host actin (2.0-fold) was also found at 6 h following L. monocytogenes infection. Interestingly, IMPI mRNA levels were only induced at 24 h postinfection, whereas expression levels of galiomycin, gallerimycin, and lysozyme were reduced at 6 h postinfection.

FIG. 6.

Transcriptional activation of actin and immune-responsive genes following infection. The transcription levels of actin, galiomycin, gallerimycin, IMPI, and lysozyme were determined by quantitative real-time RT-PCR analysis and are shown relative to the expression levels in mock-injected animals. Results were normalized to expression of the housekeeping 18S RNA gene and represent means of three independent determinations ± standard deviations.

Activation of immunity in Galleria enhances the host defense against L. monocytogenes infection.

Previous studies provide evidence for the presence of inducible immune defense molecules in Galleria that provide relatively long-lasting antimicrobial responses to repeated infections (12, 45). To examine whether the prior induction of immune responses in Galleria would protect against subsequent infection by L. monocytogenes, we injected larvae with 100 μg LPS and then challenged them by injecting a dose of 10⁶ CFU of L. monocytogenes 24 h later. LPS-mediated induction of immune responses provided vigorous protection against subsequent infection by a lethal dose of L. monocytogenes (Fig. 7A). To examine the basis of reduced L. monocytogenes growth, we isolated hemolymphs from preimmunized and naive larvae and used them in an inhibition zone assay (see Materials and Methods) that indicates the presence of antimicrobial activity. On BHI agar plates plated with L. monocytogenes, we observed zones of inhibitory growth that were dependent on the concentration of LPS used for preimmune activation, indicating inducible antilisterial activity in hemolymph (Fig. 7B). Similar results were observed when using heat-killed Listeria cells instead of LPS (data not shown).

FIG. 7.

Effects of preimmune activation on subsequent challenge with L. monocytogenes. (A) Activation of the immune system by injecting 10 mg/ml of LPS 24 h prior Listeria infection resulted in a significant increase of survival of Galleria larvae (•) in comparison to untreated larvae (○). From totals of 10 mg/ml and 1 mg/ml of LPS stock solution 100 μg and 10 μg of LPS were injected into each larva for immune induction. (B) Hemolymph samples of the preimmune activated larvae produce antimicrobial effectors that inhibit the growth of L. monocytogenes. The size of the inhibition zone increased with the concentration of LPS used for preimmune activation. Similar results were obtained using heat-killed Listeria cells for immune activation prior to infection (data not shown). Results represent means of at least three independent determinations ± standard deviations. Each repetition contained 30 larvae per treatment. Statistically differences are indicated (*, P < 0.05; ***, P < 0.01; **, P < 0.005).

DISCUSSION

Invertebrate infection models have been recently employed to investigate the pathogenesis of L. monocytogenes (18), but a comparative analysis of the pathogenic potentials of various Listeria species and serotypes obtained from various human and environmental sources has not been previously addressed in these models. In this work we show that the Galleria model was able to clearly distinguish between pathogenic and nonpathogenic Listeria species and to discriminate between L. monocytogenes serotypes exhibiting attenuated virulence properties. In addition, we report that mutants of L. monocytogenes lacking either single or multiple virulence factors are attenuated for pathogenicity in Galleria. Conversely, an avirulent strain of L. innocua engineered to express the vgc locus of L. monocytogenes exhibited enhanced virulence in the Galleria model. Thus, the invertebrate host Galleria emulates many aspects of Listeria infection seen in vertebrates.

Previous studies with Drosophila have revealed the relative contribution of virulence factors of L. monocytogenes to septic infection and demonstrated that the Drosophila S2 cell line can be used to examine intracellular growth of Listeria (16). However, a recent study described some limitations of D. melanogaster as a heterologous host for the study of pathogenesis of several Gram-positive bacteria (26). Also, because Drosophila cannot be maintained at 37°C, it does allow experimental analysis at a temperature to which mammalian pathogens are adapted. In another invertebrate infection model employing Caenorhabditis elegans, a Listeria mutant lacking actA was found to be lethal, thus also limiting the utility of C. elegans to study Listeria pathogenicity (53).

Listerial virulence in Galleria is influenced by the concentration of the inoculum injected. At high concentrations, i.e., 10⁷ CFU/larva, even nonpathogenic Listeria species such as L. innocua induce septic death in Galleria. This is probably due to a threshold over which processes leading to larval death are induced via the overwhelming activation of the innate immune system. At 10⁶ CFU/larva, nonpathogenic Listeria species lacking virulence factors are probably engaged by cellular receptors recognizing bacterial pathogen-associated molecular patterns (PAMPs), such as peptidoglycan, leading to the activation of the innate immune system and bacterial clearance. On the other hand, L. monocytogenes showed significant pathogenesis in Galleria. We show here that this can be attributed to the expression of specific virulence factors responsible for the survival within the invertebrate host, e.g., through cytotoxicity by listeriolysin or engaging cellular pathways that utilize components of the host cytoskeleton by ActA, to affect the course of infection. Hence, bacteria lacking the hexose phosphate transporter (uhpT) were significantly reduced for virulence against infected Galleria larvae, implying that energy-rich phosphorylated derivatives of glucose are also important substrates for bacterial growth in invertebrate cells.

However, some virulence factors, such as PlcA, appear to be dispensable for pathogenesis in Galleria. The lack of pathogenic potential has been previously observed for a plcA mutant in human umbilical vein endothelial cell (HUVEC) monolayers (20), suggesting that the listerial phosphatidylinositol phospholipase may have host- and cell-type-specific properties. As has also previously been observed, intravenous (i.v.) infection of mice with either the inlA or inlB mutant did not reveal any pathogenic potential. This is not unexpected, as these are cell-tropic factors that are required for overcoming epithelial and endothelial barriers following oral infection (19).

The role of specific virulence factors of L. monocytogenes in Galleria infection was also illustrated by the increase in mortality caused by nonpathogenic L. innocua harboring the vgc locus from L. monocytogenes. We note, however, that there is a significant difference in the mortality rates of the pathogenic EGD-e strain and the virulent L. innocua vgc recombinant strain, suggesting the presence of additional specific factors encoded by the EGD-e genome that contribute to listerial pathogenesis in Galleria.

In conclusion, we note that while some listerial virulence genes are generally needed for infection in mammals as well as in invertebrates, others have evolved for different hosts as well as tissue-specific infections. Recently, chitinases capable of hydrolyzing α-chitin from arthropods were found in some L. monocytogenes strains, which may be of importance for invertebrate infections (34).

We differentiated numerous species and serotypes of Listeria based on their ability to infect Galleria, showing that only the human pathogen species L. monocytogenes was lethal for Galleria. Among the L. monocytogenes serotypes, 4b is reported to be the most invasive and pathogenic to mammals (15). Indeed, the heightened virulence of this serotype in Galleria underlines the discerning properties of this model system.

The ability of L. monocytogenes to overcome host immune responses and multiply within the host system was confirmed by monitoring bacterial CFU following infection. We observed a strong reduction of L. monocytogenes in larvae at very early times, i.e., at 1 h postinfection, suggesting the presence of effective constitutively expressed components of innate immune responses in Galleria. Apart from the efficient constitutive immune system that Galleria employs to limit microbial growth, we show here that an induced response comprising sequential and overlapping expression of antimicrobial peptides, lysozyme, and inhibitors of host and bacterial metalloproteinases is required for complete elimination of bacteria causing septic infections. This inducible immunity in Galleria against Listeria infection seems to be nonspecific and can be induced by products that are not part of the infecting pathogen. Thus, as we show here, preactivation with LPS (which is not present in Gram-positive bacteria) or heat-killed preparations of L. monocytogenes can induce immune responses similar to those observed upon lethal challenge with pathogenic Listeria.

Despite the clear utility of Galleria as a surrogate model to assess infections with L. monocytogenes, several limitations remain. The relatively long time required to monitor killing of larvae and the inability to assess oral infections are impediments that need to be overcome. A further impediment is the lack of a genome sequence for Galleria and of a well-established method to generate mutants. In this study we have used death as an end point to monitor progress of infection. However additional phenotype and cellular assays, such as signs of melanization, nodulation, inducibility of pupa formation, and clotting phenotypes, need to be incorporated to improve the discerning power of the model system. The processes that are reproduced in mice and Galleria may represent ancient mechanisms of cell-cell interactions. However, the enormous evolutionary distance between these models also makes it clear that many host-specific phenomena are likely to exist.

In conclusion, here we demonstrate that G. mellonella is a simple yet powerful model system for assessing virulence of L. monocytogenes. Our data indicate that following infection, pathogenic listeriae are able to overcome both constitutive and inducible components of invertebrate innate immunity. By generating additional mutants, we can now further explore this model system to identify further bacterial factors that modulate innate immunity to promote bacterial growth during infection.