Galleria mellonella was recently established as a reproducible, cheap, and ethically acceptable infection model for the Mycobacterium tuberculosis complex. Here we describe and demonstrate the steps taken to establish successful infection of G. mellonella with bioluminescent Mycobacterium bovis BCG lux.
Tuberculosis is the leading global cause of infectious disease mortality and roughly a quarter of the world’s population is believed to be infected with Mycobacterium tuberculosis. Despite decades of research, many of the mechanisms behind the success of M. tuberculosis as a pathogenic organism remain to be investigated, and the development of safer, more effective antimycobacterial drugs are urgently needed to tackle the rise and spread of drug resistant tuberculosis. However, the progression of tuberculosis research is bottlenecked by traditional mammalian infection models that are expensive, time consuming, and ethically challenging. Previously we established the larvae of the insect Galleria mellonella (greater wax moth) as a novel, reproducible, low cost, high-throughput and ethically acceptable infection model for members of the M. tuberculosis complex. Here we describe the maintenance, preparation, and infection of G. mellonella with bioluminescent Mycobacterium bovis BCG lux. Using this infection model, mycobacterial dose dependent virulence can be observed, and a rapid readout of in vivo mycobacterial burden using bioluminescence measurements is easily achievable and reproducible. Although limitations exist, such as the lack of a fully annotated genome for transcriptomic analysis, ontological analysis against genetically similar insects can be carried out. As a low cost, rapid, and ethically acceptable model for tuberculosis, G. mellonella can be used as a pre-screen to determine drug efficacy and toxicity, and to determine comparative mycobacterial virulence prior to the use of conventional mammalian models. The use of the G. mellonella-mycobacteria model will lead to a reduction in the substantial number of animals currently used in tuberculosis research.
Tuberculosis (TB) is a major threat to global public health with 9 million new cases per year and 1.5 million deaths1. In addition, it is estimated that one quarter of the world’s population is infected with the causative agent of the disease, Mycobacterium tuberculosis (Mtb). Amongst the infected population, 5−10% will develop active TB disease over their lifetime. Furthermore, the emergence and spread of multi-drug resistant (MDR) and extensively-drug (XDR) resistant Mtb poses a serious threat to disease control, with 123 countries reporting at least one XDR case1. Treatment of TB requires a cocktail of at least four anti-mycobacterial drugs, of which isoniazid and rifampicin are prescribed for a minimum duration of six months; treatment is often associated with complex side effects and toxicities. Protection from the only licensed vaccine against TB, Mycobacterium bovis Bacillus Calmette-Guérin (BCG), is variable2. An incomplete understanding of the pathogenesis of TB significantly hampers the development of new therapeutic and vaccination strategies.
For decades animal infection models have been vital for TB research to understand the basic pathogenesis and host response to infection, and to evaluate novel anti-mycobacterial agents, immuno-therapeutics and new vaccine candidates3,4. However, research using animal infection models of TB is notoriously difficult as the pathogenesis and progression of TB infection are complex, and there is no single animal model that mimics the full spectrum and important features of the disease5,6. Furthermore, animal experiments are expensive, time consuming to undertake and require full ethical justification. Nevertheless, animal infection models of TB have been described in non-human primates (e.g., macaques), guinea pigs, rabbits, cattle, pigs, mice and zebrafish, with each having their limitations3,4. The murine model is the most commonly used model due to cost, availability of inbred lines, reproducibility of infection and abundance of immunological reagents. However, they do not typically form granulomas associated with areas of hypoxia that are characteristic of latent tuberculosis infection (LTBI)6. Guinea pigs are highly susceptible to Mtb infection, with pathology and early granuloma formation similar to those in humans, and are widely used in vaccine testing; yet the lack of immunological reagents hampers their use as an infection model7. Zebrafish are suitable for large-scale screening in early-stage preclinical studies due to their small size, rapid reproduction and advanced genetic tools, but are anatomically and physiologically different to humans and are only susceptible to Mycobacterium marinum infection3. The animal models most closely resembling human Mtb infection are non-human primates (e.g., the macaque), but they are expensive and have significant ethical and practical considerations which considerably limits their use8.
The insect larva of the greater wax moth or honeycomb moth, Galleria mellonella, have become increasingly popular as an infection model for a variety of bacterial and fungal pathogens9, and as a screen for novel antimicrobial drug candidates10. G. mellonella is a successful invertebrate model due to its sophisticated innate immune system (comprised of cellular and humoral defenses) that shares a high degree of structural and functional similarity to that of vertebrates11. This includes immune mechanisms such as the phagocytosis of pathogens by hemocytes (functionally similar to mammalian macrophage and neutrophils)12,13, the production and circulation of anti-microbial peptides (AMPs) and complement-like proteins within the hemolymph (analogous to mammalian blood) of G. mellonella11. Other advantages9,14,15 of G. mellonella larvae as a model include 1) their large size (20−30 mm) which allows for easy manipulation and infection, as well as the collection of tissue and hemolymph for analyses, 2) easy maintenance at 37 °C, compatible for studying human pathogens, 3) precise infection by injection without the need for anesthesia, 4) efficacy of antimicrobial agents can be assessed utilizing less drug for evaluation, 5) lack of ethical constraints compared to the use of mammals, 6) large group sizes can be used compared to animal models allowing greater reproducibility, and 7) shorter times for infection experiments are required.
In a recent study, we demonstrated that G. mellonella can be used as a novel infection model for studying the pathogenesis of infection by bioluminescent M. bovis BCG lux, a genetically modified version of the vaccine strain and member of the Mtb complex (MTBC)16. While G. mellonella has previously been used as an infection model for non-tuberculous mycobacteria (NTM), mainly M. marinum and Mycobacterium abscessus17,18, studies using MTBC are limited to that of Li et al.16. Bioluminescent non-pathogenic mycobacterial strains, which can be used at containment level (CL) 2 as a surrogate for Mtb, offer the advantages of safety and practicality over pathogenic mycobacteria. Following infection with BCG lux, larvae begin to develop early granuloma-like structures, which could provide valuable insight into the role of innate immunity in the establishment of TB infection16. In addition, this simple invertebrate infection model has the potential to provide a rapid, low-cost, and reliable evaluation of TB pathogenesis incorporating controlled challenge and multiple replicates for reproducibility. Furthermore, the model has the potential to be used to screen novel anti-TB drug and vaccine candidates in early development, reducing the overall number of animals in experimentation. The ability to measure changes in host and pathogen structure, transcriptome and proteome to determine drug targets and assess mechanisms of action of novel drugs and therapeutic vaccines, are also advantageous.
Here we describe the experimental protocols for the preparation of a bioluminescent M. bovis BCG lux inoculum and G. mellonella larvae for mycobacterial infection, as well as the determination of both larval and mycobacterial survival in response to infection.
NOTE: All work described below are to be carried out in a CL2 laboratory within a class 2 microbiological safety cabinet (MSC) following local health and safety guidelines.
1. Preparation of M. bovis BCG lux for Infection
2. Preparation of G. mellonella Larvae
3. Infecting G. mellonella with BCG lux
4. Monitoring the Survival of G. mellonella Following Infection
5. Measuring the In Vivo Burden of BCG lux in G. mellonella
6. Statistical Analysis
Here we present representative data that can be obtained using the G. mellonella — BCG lux infection model and highlight the benefits of G. mellonella as an infection model for members of the MTBC (Figure 1). Experimental procedures with key technical points are outlined in Figure 2.
Figure 1: The benefits of G. mellonella as an infection model. This figure has been adapted from Kavanagh and Sheehan22. Please click here to view a larger version of this figure.
Figure 2: Outline of experimental procedures. (A) Maintenance and preparation of G. mellonella for infection with M. bovis BCG lux. (B) Preparation of BCG lux culture and inoculum for infection. (C) Infection of G. mellonella with BCG lux. (D) Measurement of virulence and in vivo burden of BCG lux in G. mellonella larvae. Please click here to view a larger version of this figure.
BCG lux dose dependent virulence was observed in G. mellonella larvae over a 96 h incubation period (Figure 3), and the lethal dose required for 50% larval mortality (LD50) was determined to be 1 x 107 CFU. The survival distribution reflecting the virulence of the BCG lux doses tested was significantly different (p < 0.0001). Control groups injected with a 10 µL dose of PBS-T or those simply pricked simulating needle injuries, did not affect larval health or lead to an increase in mortality as determined by observational checks on motility and melanization at different time points.
Figure 3: Kaplan-Meir survival curve of G. mellonella in response to varying inocula of M. bovis BCG lux. Healthy larvae (n ≥ 10 per group), were infected with varying doses of BCG lux. Larvae were incubated at 37 °C and monitored for survival every 24 h for up to 96 h. The uninfected group was injected with PBS, and a pricked group (insertion of needle only) demonstrated the effect of needle injury on larval health. The means of two independent experiments are shown, with 95% confidence interval, represented as dotted lines in corresponding color to the inoculum. This figure has been adapted from Li et al.16. Please click here to view a larger version of this figure.
All larvae infected with BCG lux displayed physiological changes over time and, for larvae infected with a 2 x 107 CFU dose of BCG lux, melanization of the larval dorsal line was observed from 48 h post infection (pi), and systematic melanization was observed from 96 h pi (Figure 4). Furthermore, the motility of the larvae reduced with the severity of melanization and the ability of larvae to pupate was lost upon infection in comparison with uninfected controls.
Figure 4: Melanization of G. mellonella in response to infection with M. bovis BCG lux. Healthy larvae at 0 h were infected with a 2 x 107 CFU dose of BCG lux. At 48 h and 96 h pi, melanization along the larval dorsal line and systematic melanization, respectively, was observed.
Survival of BCG lux within the G. mellonella larvae was determined over 2 weeks through bioluminescence measurement of larval homogenates. Infection with a 1 x 107 CFU dose of BCG lux resulted in an initial decline of BCG lux bioluminescence from 0−72 h pi. However, from 72−144 h pi, the bioluminescence of BCG lux plateaued, indicating the establishment of persistent infection (Figure 5).
Figure 5: In vivo burden of M. bovis BCG lux in G. mellonella quantified using bioluminescence (relative light unit, RLU/mL) over a two week time course. Healthy larvae (n = 30) were infected with 1 x 107 CFU dose of BCG lux. In vivo burden was quantified by homogenizing five larvae at each time point (0, 24, 48, 72, 96, 168, 336 h), and measuring the bioluminescence of the homogenate. The means of three independent experiments are shown, and the error bars represent the standard deviation of the mean. This figure has been reprinted from Li et al.16. Please click here to view a larger version of this figure.
As the rapid and reproducible quantification of in vivo mycobacterial growth in these studies was determined by the measurement of bioluminescence, the ratio of RLU and CFU should also be determined in vivo. In our particular infection system, the in vivo ratio of RLU and CFU ranged from 2:1-5:1, with an average of 4:1 over the 168-h time course (Figure 6).
Figure 6: Determining the in vivo RLU/CFU ratio of M. bovis BCG lux in G. mellonella. Healthy larvae (n = 30) were infected with 1 x 107 CFU dose of BCG lux. At each time point (0, 24, 96 and 168 h), four infected/control (PBS-T) larvae were homogenized, and the homogenates were measured for bioluminescence and plated out onto 7H11 agar to enumerate CFU counts. The means of two independent experiments are shown and the error bars represent the standard deviation of the mean. Please click here to view a larger version of this figure.
The use of G. mellonella as an infection model has been established for a number of bacterial and fungal pathogens for the study of virulence, host-pathogen interaction, and as a screen for novel therapeutics10,22. The following discussion is based on the experimental procedure for the use of G. mellonella as an infection model for the MTBC.
The health of the naïve larvae prior to experimentation can have a considerable impact on the outcome of the experiment. Therefore, it is vital that any discolored and/or injured larvae are removed upon arrival and are not used for any experimentation. If a large number of larvae are found dead within the same container upon arrival, it is advisable to discard the batch as pre-existing infections may be the cause of death. When possible perform experiments using the larvae as close to the day of purchase/arrival. Before use ensure to store the larvae at 18 °C to prevent pupation and to maximize the number of larvae available for experimentation. Healthy larvae can be used for up to 7 days following delivery/purchase. Following infection, ensure to remove any dead larvae from the Petri dish as, for reasons yet unknown, the presence of dead larvae appears to increase the rate of mortality in the sample population. For users of self-reared larvae, it is important to be aware of biological variability in comparison to purchased larvae, as variance in diet and growth conditions can have an impact on larval immunity21,23. Inter-experimental variabilities can be limited by keeping the source, if purchased, or the feed and growth conditions of the reared larvae consistent between experimentation. In all experiments, the inclusion of ‘blank’ and ‘pricked’ negative controls are essential; the blank control is an indicator of contamination or toxicity of the suspension matrix (PBS-T or media broth), and the pricked control mimics the effect of the needle injury on larval health. Furthermore, these controls normalize any biological variation between batches of larvae, ensuring reproducibility and accuracy between experiments.
For injection of G. mellonella larvae, the use of a 25 G needle is recommended as larger gauge needles can cause excessive bleeding and sharper smaller gauge needles can easily puncture the gut of the larvae, leading to larval mortality and false positive results. Conventional methods of needle injection typically immobilize the larvae by hand, which increases the risk of needle injury. By using tweezers to immobilize the larvae, the risk of needle stick injury is significantly reduced as the hand is not in close proximity to the needle at any point during infection. Alternatively, larvae could be immobilized by cooling. However, cold shock at 12 °C for 15 min prior to infection has been documented to enhance the innate immune response to infection. Therefore, the analysis of the results obtained via cooling should carefully consider its impact24. Using our technique, medium throughput of injection (2−3 larvae per minute) can be achieved with user practice; this is comparable to the speed of conventional injection from our experience (3−4 larvae per min). Furthermore, our method can be adapted for higher throughput injection by utilizing a pedal operated injection platform comprised of an infusion pump with a disposable syringe connected to a 25 G butterfly cannula.
The preparation of BCG lux inoculum is based on the rapid estimation of CFU using RLU of the mycobacterial culture20. In our experience with broth culture, the ratio of RLU to CFU is 3:120, contrasting with BCG lux growing in G. mellonella where the RLU to CFU ratio is 5:120. Mycobacterial cell aggregation or ‘clumping’ commonly seen in dense cultures can have an impact on the RLU measurements, as clumping can result in unreliable bioluminescence measurements. As such, the use of clumped mycobacterial culture for inoculum preparation is not recommended. However, the addition of polysorbate 80 to the growth media minimizes cell clumping without altering the growth of BCG lux.16 Any minor cell clumping can be resolved by washing the culture with PBS-T and is vital for preparing an accurate inoculum using RLU as the readout. Furthermore, the PBS-T wash is vital for removing any extracellular virulence factor secreted during growth. Mycobacterial strain and passage number should also be taken into consideration, as this can impact the severity of mycobacterial aggregation25. In all cases, the inoculum should be enumerated by CFU on 7H11 agar plates to ensure that the correct CFU was prepared by bioluminescence measurement. Additionally, the RLU/CFU ratio in vivo should be determined with new bioluminescent reporter or stocks, as the ratio will likely vary.
G. mellonella holds several advantages over conventional models of TB infection, including cheaper acquisition and maintenance costs, ease of infection and research throughput, especially in combination with bioluminescent strains16. The lack of ethical constraints allows for greater sample size (minimum of 20−30 larvae per group) in comparison to mammalian models, giving greater confidence and reliability in the results obtained26,27. However, there are a number of limitations in the use of G. mellonella as an infection model. As an invertebrate, they naturally lack adaptive immunity making them unsuitable for antigenicity or immunological studies10. Cellular innate immune responses of G. mellonella are comprised of a number of hemocyte types, and plasmatocytes and granulocytes have been reported to function similarly to mammalian phagocytes (neutrophils and macrophages)12. However, the role and mechanisms of these cell types remain under characterized, and direct comparative studies between mammalian and insect phagocytes are yet to be carried out. Furthermore, the lack of an annotated G. mellonella genome hinders the analysis of the host response to infection, and this is currently reliant on gene ontology analysis against transcriptomic libraries of other invertebrates, such as Drosophila melanogaster and Bombyx mori28.
As a TB infection model, G. mellonella holds a promising future with its ability to develop granuloma-like structures in response to BCG lux infection16, which are vital pathophysiological hallmarks of TB infection and are a key feature in the development of LTBI. Future work will aim to characterize this model with a particular interest in the formation of granuloma-like structures using reference, clinical and mutant Mtb isolates under CL3 conditions. Additionally, we anticipate that it may also be useful for novel anti-mycobacterial agent screening, as a similar G. mellonella model was used for NTMs18, but this remains to be determined. The adoption of this model has the ability to significantly reduce the number of animals used within the TB research community, while simultaneously accelerating in vivo TB research output under ethically more acceptable conditions.
The authors have nothing to disclose.
This project was supported by grants from the Biotechnology and Biological Science Research Council (BBSRC), awarded to PRL and YL (BB/P001262/1), and the National Center for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs) awarded to PRL, SMN, BDR, and YL (NC/R001596/1).
1.5ml reaction tube (Eppendorf) | Eppendorf | 22431021 | |
20, 200 and 1000 µl pipette and filtered tips | Any supplier | n/a | |
24 well culture plate | Greiner | 662160 | |
25 ml pipettes and pipette boy | Any supplier | n/a | |
3 compartment Petri dish (94/15mm) | Greiner | 637102 | |
Centrifuge | Any supplier | n/a | |
Class II saftey cabinet | Any supplier | n/a | |
Erlenmeyer flask with vented cap (250 ml) | Corning | CLS40183 | |
Ethanol (>99.7%) | VWR | 208221.321 | |
Galleria mellonella (250 per pk) | Livefood Direct UK | W250 | |
Glycerol | Sigma-Aldrich | G5150 | |
Homogeniser (FastPrep-24 5G ) | MP Biomedicals | 116005500 | |
Hygromycin B | Corning | 30-240CR | |
Luminometer (Autolumat LB 953) | Berthold | 34622 | |
Luminometer tubes | Corning | 352054 | |
Lysing matrix (S, 2.0ml) | MP Biomedicals | 116925500 | |
Micro syringe (25 µl, 25 ga) | SGE | 3000 | |
Microcentrifuge | Any supplier | n/a | |
Middlebrook 7H11 agar | BD Bioscience | 283810 | |
Middlebrook 7H9 broth | BD Bioscience | 271310 | |
Middlebrook ADC enrichment | BD Bioscience | 212352 | |
Middlebrook OADC enrichment | BD Bioscience | 212240 | |
Mycobacterium bovis BCG lux | Various | n/a | |
n-decyl aldehyde | Sigma-Aldrich | D7384-100G | |
Orbital shaking incubator | Any supplier | n/a | |
Phosphate buffered saline | Sigma-Aldrich | P4417-100TAB | |
Polysorbate 80 (Tween-80) | Sigma-Aldrich | P8074-500ml | |
Small box | Any supplier | n/a | dark vented or non-sealed box recommended |
Tweezer | Any supplier | n/a | Short and narrow tipped/Blunt long tweezers |
Winterm (V1.08) | Berthold | n/a | Program LB953.TTB |
Petri dish (94/15mm) | Greiner | 633181 | |
Filter paper (94mm) | Any supplier | n/a | Cut to fit |