Exploring the impact of probiotic route of administration on its protective effects against pathogenic infection in Galleria mellonella

ABSTRACT

Identifying promising probiotic candidates for further development remains challenging. Traditional mammalian models are invaluable for assessing efficacy, but the associated limitations slow the preliminary screening process. The Galleria mellonella larvae have emerged as a powerful in vivo screening model. Most probiotic studies on G. mellonella use intra-hemocelic injection, bypassing the natural gut entry point, making it analogous to introducing probiotics into the mammalian blood. Therefore, despite their advantages, discrepancies exist between G. mellonella and mammalian models, particularly regarding the route of probiotic administration. This study bridges this gap by investigating the differential effects of the commonly studied intra-hemocelic injection and the less common oral administration of a probiotic, Escherichia coli Nissle 1917 (EcN1917), on its protective efficacy against a gastrointestinal pathogen, Salmonella Typhimurium ATCC 14028 (ST14028). This study demonstrated that oral EcN1917 pre-treatment significantly increased survivability against ST14028 infection compared to the control group and alleviated the pathogen gut burden. Notably, injection pre-treatment decreased survivability. This discrepancy is attributed to the dual nature of probiotics, exhibiting beneficial effects in the gut but acting as pathogens in non-native locations like the hemolymph, concluding that the route of probiotic administration in G. mellonella significantly impacts the protective effects of probiotics.

KEYWORDS:

• Greater wax moth
• prophylactic
• E. coli Nissle 1917
• S. Typhimurium
• inoculation route

Introduction

Traditionally, murine models are often used for the in vivo study of probiotic intervention in managing microbial infections. Galleria mellonella larvae have recently served as ideal surrogate models for high-throughput studies (Köhler 2015) to overcome some ethical, economic and logistical restrictions associated with traditional mammalian models, often slowing research progress. Although not as genetically well established as Drosophila melanogaster or Caenorhabditis elegans, the ease of manipulation, survivability at incubation temperatures of 25–37°C and similarity of larval innate immune systems to mammalian models increase the attractiveness of this model system for studying protective effects of probiotics and the virulence of pathogens in a human physiological environment (Nathan 2014). With the rapid rise in antimicrobial resistance in many pathogens (Ten et al. 2021; Ting et al. 2021), probiotics have been proposed as potential alternative antibiotics.

Probiotics are live, non-pathogenic microorganisms that benefit the host, often by residing in the gut after consumption of appropriate doses (Rossoni et al. 2017). However, most probiotics screening studies using G. mellonella have been conducted via intra-hemocelic injection, whereby the probiotics are introduced into the larval hemolymph, which is analogous to introducing probiotics into the mammalian blood (Rossoni et al. 2017; Scalfaro et al. 2017; Jorjão et al. 2018; Maslova et al. 2023). This approach is misleading and might give an inaccurate interpretation of the protective effects of the tested probiotics. Therefore, our study aims to investigate and compare the differences in the degree of protection conferred by a probiotic against a gastrointestinal pathogen via different routes of administration (force-feeding vs injection).

In this study, Escherichia coli Nissle 1917 (EcN1917), a commercialised and well-characterised therapeutic with over a century-long history of human use (Pradhan and Weiss 2020), was used as the model probiotic. Salmonella Typhimurium, a widespread gastrointestinal pathogen with increasing outbreaks of drug-resistant infections (Park et al. 2019) was used to establish a gut infection of the larval model. This research showed that the protective effects of probiotics in G. mellonella are influenced significantly by the method of administration, emphasising the importance of considering administration routes during probiotic screening.

When comparing routes of inoculation, G. mellonella were expected to be more susceptible to health decline, with decreased survivability when ST14028 were injected directly into the hemolymph, rather than into the gut via force-feeding. The probiotic EcN1917 was expected to also cause a decline in larval survivability through injection while showing no impacts via force-feeding. Consequently, the prophylactic protective effects of the probiotic were expected to be more pronounced through oral force-feeding. With these outcomes, it was expected that the probiotic oral pre-treatment would be capable of suppressing the larval ST14028 burden. The major implication of this study is that the route of probiotic administration significantly impacts its protective effects and should be considered when screening for probiotics in G. mellonella.

Materials and methods

Bacterial strains, culture preparations, and growth curves

Probiotic EcN1917 and pathogenic ST14028 were cultured and routinely maintained on nutrient agar. Growth kinetics for each bacterial strain were determined to prepare active bacterial suspensions with the highest concentration of viable cells for larval inoculation. Concisely, the bacterial overnight cultures were sub-cultured in Luria Bertani (LB) broth and incubated at 37˚C, shaking at 200 rpm. The optical density (OD 600 nm) was determined every 30 minutes for 6 hours until consistent consecutive measurements were obtained. The spectrophotometric measurements from three replicates were used to plot growth curves and determine the respective lag, exponential and stationary phases.

G. mellonella challenge assays

Batches of G. mellonella larvae were sourced from Carolina Biological, USA and maintained as per established protocol (Ten et al. 2023). Larvae without melanisation were incubated overnight and acclimatised to 37°C before use in the challenge assays. Bacteria were grown to the mid-log phase, harvested by centrifugation (6700 g, 7 minutes) and washed to remove traces of media or compounds produced in vitro. They were serially diluted 10-fold using 1X PBS (Ten et al. 2023). Larvae were sterilised using 70% ethanol and injected or force-fed with 10 µL of bacterial inocula using a Hamilton syringe. For force-feeding, blunt-end needles were used to prevent internal punctures. Two additional control groups were included: a PBS-only group accounting for physical trauma caused by the inoculation process and a no-inoculation group accounting for natural deaths. The larvae were incubated at 37˚ C; melanisation and survival were monitored and recorded at 0-, 3-, 6-, 24-, 48-, 72-, 96-, 120 – and 144 hours post-inoculation. Larvae that appeared dark-brown or black with no response upon stimulation were considered dead. The amount of bacteria (colony-forming units: CFU/mL) administered was determined by plating the prepared inocula on LB agar. The assays were performed in duplicates for each bacterial strain, with 5 larvae per treatment group, via force-feeding and injection.

G. mellonella prophylactic protection assays

Bacterial inocula (probiotic EcN1917 and pathogen ST14028) were prepared as decribed above. Larvae were inoculated according to treatment groups (as specified in Table 1) and incubated at 37°C for 2 hours before the second inoculation. Larval survival, activity, and melanisation were observed and scored at various time points post-second inoculation based on the scoring table provided in Figure 2(a). Protection assays were performed in triplicates for both injection and force-feeding techniques.

Table 1. Experimental groups were used for the protection assays. The different experimental groups have been presented, along with the site of inoculation on the larvae for the groups in the bacteria were administered intra-hemocelically via injection. Group 1* represents the treatment group (pre-treated with probiotics before infection with the pathogen). Groups 2–4 were the infection-only, probiotic-only, and trauma-control groups, respectively.

Group	Number of Larvae	1^st Inoculation	Inoculation Site (for Injection)	2nd Inoculation (2 hours after 1st inoculation)	Inoculation Site (for Injection)
1 (*)	10	10uL EcN1917	Last left proleg	10uL ST14028	Last right proleg
2	10	10uL PBS		10uL ST14028
3	10	10uL EcN1917		10uL PBS
4	10	10uL PBS		10uL PBS

Bacterial transformation and G. mellonella bacteria burden assays

EcN1917 and ST14028 were transformed using a heat-shock approach with selection plasmids. Chemically competent ST14028 and EcN1917 cells were transformed with pULTRA_RFP_KAN and pULTRA_CmR_KAN, respectively. Briefly, the steps involved were: preparation of chemically competent cells, extraction of desired plasmid, confirmation of plasmid size and transformation of competent cells with selected plasmid.

For determining the bacterial burden, the larvae were treated and then infected with the transformed EcN1917 and ST14028 following the protection assay protocol via oral force-feeding. At 0-, 4-, and 24 hours post-inoculation, the hemolymph and gut were extracted and dissected from 2 randomly selected larvae (24 hours was chosen as the final time point to ensure possible extraction). To release intracellular bacteria, 1 µL of 5 mg/mL Digitonin was added to the extractions, followed by centrifugation (2300 g, 4°C, 7 minutes). The bacterial supernatants were serially diluted and plated on different selective agars. 50 µg/mL kanamycin (KAN) and RFP expression were used for ST14028 selection. Cell extracts were plated on agar containing 50 µg/mL KAN and 15 µg/mL chloramphenicol (CHL) to select for EcN1917. Plates were incubated at 37˚ C, and the number of CFU was determined and normalised to the weight of the larval extractions. Assays were carried out in triplicates.

Statistical analysis

Results from replicates of experiments were pooled and statistically analysed using GraphPad Prism 8 (GraphPad Software, La Jolla, USA). Kaplan-Meier survival curves were plotted for the challenge and protection assays, and the significance differences were predicted using the Mantel–Cox log-rank tests (significance at p ≤ 0.01). Multiple t-tests (significance at p ≤ 0.05) were conducted to compare the differences between health index scores and bacterial burden presented in clustered bar graphs. Statistical significance between two groups is visualised as asterisks in figures (∗p ≤ 0.05; ∗∗p ≤ 0.01; ∗∗∗p ≤ 0.001; ∗∗∗∗p ≤ 0.0001).

Results and discussion

Route of probiotic and pathogen administration alter G. mellonella survivability

Both oral force-feeding (Figure 1a) and intra-hemocelic injection (Figure 1b) were tested in the challenge assays to identify the difference in potencies of the bacterial strains. Dose–response curves were plotted for each bacterial strain and inoculation route using data from the 72-hour time point by which maximal deaths were observed. While injections caused a concentration-dependent decline in survivability at lethal dose LD50 of 10⁶ CFU/mL (Figure 1c), larval survivability remained at 100% after force-feeding with all concentrations (10⁵–10⁸ CFU/mL) of EcN1917 (Figure 1d). While probiotics are safe in the gut, positively modulating host microbiota and immune responses following oral administration (Jorjão et al. 2018), they can cause disease in systemic circulation. Probiotic bacteraemia, though rare, has been documented in immunocompromised individuals (Kulkarni 2019; Han et al. 2021). Similarly, EcN1917, despite being considered safe, was found to harbour the pks pathogenicity island, encoding non-ribosomal and polyketide synthases involved in colibactin production (Nougayrède et al. 2021). This genotoxic factor triggers DNA damage during systemic infection. Hence, introducing EcN1917 into the larval hemolymph via injection may cause both bacteraemia and septicaemia, contributing to larval mortality at higher concentrations (Nougayrède et al. 2021; Lynch et al. 2022). This observation underscores the dynamic nature of responses triggered following EcN1917 administration through the various routes of inoculation.

Figure 1. Key results from the G. mellonella challenge assays. (a) Larval inoculation via intra-hemocelic injection. (b) Larval inoculation via oral force-feeding. Kaplan Meier Survival curves showing the effects of inoculating various concentrations of the EcN1917 via (c) Injection (d) Force-feeding (survival remained at 100% in treatment groups) and ST14028 via (e) Injection (f) Force-feeding. For each bacterial strain and route of inoculation, all controls (No treatment and PBS) and concentrations (10⁸, 10⁷, 10⁶ and 10⁵ CFU/mL) were tested. All differences in survivability between treatment groups were significant (p-value < 0.01), (Mantel-Cox log-rank test). (n = 5 per treatment group, 2 Biological replicates). Dose-response curves, with the lethal dose to 50% of the larval population, have been presented.

In contrast, inoculation with ST14028, an enteric pathogen, decreased larvae survivability through both routes of inoculation (Figure 1e and f). Notably, the larvae were more susceptible to ST14028 when injected intra-hemocelically (LD50 of 10⁵ CFU/mL) than force-fed into the gut (LD50 of 10⁷ CFU/mL). The LD50 values indicate that a lower dosage of the pathogen ST14028 was required to cause larval death following injection. The presence of pathogens in the larval hemolymph triggers the immune cells to produce reactive oxygen species (ROS) to eradicate them. This immune response mechanism mirrored that in mice models with systemic S. Typhimurium infection (Rhen 2019). When infected with high doses of pathogens, elevated hemolymph ROS levels are detrimental to the larvae (Kazek et al. 2020), which aligns with our observation that higher ST14028 inocula concentrations increase larvae deaths. The eicosanoids in G. mellonella possibly lose functionality at high ROS levels, disrupting anti-oxidative enzymes, leading to ROS accumulation and increased oxidative stress, inducing lipid peroxidation and immune cell cytoskeleton disruption, damaging essential proteins that further deteriorate larval health (Büyükgüzel et al. 2007). Impacts of S. Typhimurium in the larvae parallel observations in humans, where infections affecting the systemic circulation elevate mortality risk and severity of symptoms compared to those limited to the gastrointestinal tract (Pulford et al. 2021). These findings underscore the importance of selecting an appropriate larval inoculation route to closely replicate mammalian models’ internal physiological conditions and increase the relevance of generated results to practical applications.

Oral administration of EcN1917 conferred protection against ST14028 infection

The concentrations of EcN1917 and ST14028 inocula used in the protection assays via injection were 10⁵ CFU/mL, while for force-feeding they were 10⁸ CFU/mL. These doses of probiotic did not harm the larvae yet maintained the pathogen's infectivity to cause larval death. The protection assays were carried out via both intra-hemocelic injection, the common route of inoculation used in larval probiotic studies, and oral force-feeding to highlight the differences in outcomes. In these, assays, the larvae were pre-treated with the probiotic and then infected with the pathogen 2 hours later to study the prophylactic effects of EcN1917 against ST14028. Incubation for 2 hours with the probiotic before infection allowed for the establishment of the probiotic bacteria in the larvae, priming them for the subsequent pathogen challenge. Their survivability and health were monitored using a previously established G. mellonella Health Index Scoring (Tsai et al. 2016) (Figure 2a). The maximum score of 7 indicated healthy larvae with no physical signs of melanisation or decline in activity.

Figure 2. Key results from the G. mellonella prophylactic protection assays. (a) Health Index Scoring system for monitoring larval health. (b) Health Index Scores for first 3 days via force-feeding (c) Kaplan Meier Survival curve for protection assay via force-feeding (d) Health Index Scores for first 3 days via injection (c) Kaplan Meier Survival curve for protection assay via injection. All differences in survivability between treatment groups were significant (p < 0.01, Mantel-Cox log-rank test) (n = 10 for each treatment group, 3 Biological replicates). Health scores of the treatment group (EcN + ST), which were significantly different (p < 0.05) from the infection-only group (PBS + ST), have been indicated (*).

The differences in larval health index scores between the force-fed EcN + ST and PBS + ST were significant at 3, 24, 48 and 72 hours (p < 0.05), with higher mean health scores for the pre-treated group (Figure 2b). Concordantly, the larvae survivability dropped to 50% by day 7 in the ST14028 infection-only group, while the larvae group that received prophylactic pre-treatment with EcN1917 showed elevated survivability to 85% (p-value 0.0027) (Figure 2c). The protective effects observed following oral probiotic pre-treatment could be due to several mechanisms including limitation of micronutrient availability, production of antibacterial compounds by the probiotic or the host and colonisation resistance. Some of the relevant mechanisms have been discussed further in the later section.

Conversely, probiotic pre-treatment administered through injection lowered larval survivability to 15%, compared to injection with pathogen alone, which saw 45% survival (p-value 0.0024). The EcN + ST group had lower total health scores than the PBS + ST group at 3, 24, 48 and 72 hours (Figures 2d and e). The health scores of the treatment group were also significantly lower than the controls (EcN + PBS and PBS + PBS). These findings corroborated with the challenge assays, where the probiotic could induce larval death when injected alone (Figure 1c), and the larvae faced an elevated risk of mortality when the pathogen was injected rather than force-fed (Figure 1e and f). The synergy in larval mortality with probiotic pre-treatment and pathogen infection via injection suggests probiotics cannot exert protective effects if administered into hemolymph instead of the gut of G. mellonella.

Rossoni et al. (2017) revealed that injection of some probiotic Lactobacillus paracasei strains, although capable of prolonging the life of larvae challenged with Candida albicans, was incapable of increasing the percentage survivability by the end of the experimental period. In other words, the rate at which the fungal infection caused death was slowed down. Similarly, Jorjão et al. (2018) demonstrated that pre-treatment of larvae with Lactobacillus rhamnosus via injection decreased the survivability of larvae infected with Staphylococcus aureus or E. coli compared to those that were not pre-treated. The findings from these two authors are consistent with this study, suggesting that inoculation of probiotics into the larval hemolymph should be avoided as it may not reflect the real protective effects of probiotics, particularly when investigating gut probiotics against food-borne pathogens..

Prophylactic pre-treatment of larvae with EcN1917 decreases ST14028 burden

Bacterial burden assays were conducted using transformed EcN1917 (EcN1917* carrying a chloramphenicol selectable marker in a plasmid) and ST14028 (ST14028* carrying a Kanamycin selectable marker and red fluorescence protein in a plasmid) to ascertain the correlation between the observed protective effects of EcN1917 with the quantities of ST14028 and EcN1917 within the larvae following pre-treatment and infection via oral-force-feeding (Figure 3a). The virulence of EcN1917* and ST14028* was similar to the corresponding untransformed strains (Figure 3b), indicating that the introduction of plasmids carrying the antibiotics’ selectable marker had no impact on bacterial virulence and can be used for subsequent bacterial burden experiments.

Figure 3. Key results from the G. mellonella bacterial burden assays. (a) Simplified schematic of plasmids with selection markers used for enumeration of EcN1917 and ST14028 (b) Kaplan Meier Survival curve comparing native and transformed(*) EcN1917 and ST14028 (c) ST14028 burden in the larval guts (d) EcN1917 burden in the larval guts (e) ST14028 burden in larval hemolymph (f) EcN1917 burden in the larval hemolymph.

The bacterial burden of ST14028 in the gut (Figure 3c) of the infection-only group (PBS + ST) was higher than the probiotic pre-treated group (EcN + ST) at all time points, with a notable difference at 4 hours (p-value 0.012). The pathogen numbers also increased drastically between 0 and 24 hours without probiotic pre-treatment (p-value 0.034). Conversely, the pathogen burden in the probiotic-treated group remained almost consistent throughout the 24-hour experimental period. Moreover, no significant differences in probiotic abundance within the gut were observed between the treatment groups (EcN + ST and EcN + PBS) (Figure 3d). However, the levels of EcN1917 in the gut of the larvae pre-treated with EcN1917 before the ST14028 challenge were significantly increased (p-value 0.030).

The reduced pathogen abundance and the concurrent increase in probiotic levels observed with probiotic pre-treatment can be attributed to EcN1917's capacity to limit the availability of micronutrients, such as iron, which are crucial for Salmonella proliferation and pathogenicity (Banville et al. 2012; Deriu et al. 2013). The availability of free iron in healthy larvae is naturally low due to iron homeostasis by ferritin and transferrin (Andrews and Schmidt 2007). During infections, hemocytes produce lipocalin, a protein that reduces free iron levels, while sequestering iron from the pathogen’s siderophores. The introduction of probiotic EcN1917 further aggravates the stringency of iron availability in vivo by intercepting iron-bound Salmonella siderophores and concurrently producing an array of its siderophores competing with the pathogen for free iron uptake (Deriu et al. 2013; Sargun et al. 2021).

Besides, an in vitro study reported that EcN1917 could produce microcin MccH47 exhibiting antibacterial effects against S. Typhimurium, which is enhanced in iron-limited niches (Ma et al. 2023). Azpiroz and Lavina (2004) stated that in the presence of pathogens, EcN1917 produces pre-microcin MccH47, which promotes an increase in enterobactin siderophore expression. This event triggers the post-translational siderophore moiety attachment to the pre-microcin. MccH47 then mimics iron-siderophore complexes and is taken up by the pathogen, leading to unregulated proton entry and causing cell lysis (Massip and Oswald 2020).

The probiotic EcN was not detectable in the hemolymph of the larvae of all treatment groups (Figure 3f), indicating that the colonisation was restricted to the gut. Studies have shown that the gut probiotic EcN only undergoes translocation from the gut into the bloodstream in rare cases in individuals with compromised immune function or defective microbiota (Gronbach et al. 2010).

On the other hand, the hemolymph pathogen burden in untreated larvae was higher than treated at all time points. More than 10⁴ CFU/mg of ST14028 was present in the larval hemolymph for the group orally infected with pathogen ST14028 only (PBS + ST) at 4 hours post-infection (Figure 3e). This group's survivability and health scores dropped to 60% and Health Index of 4 by 24 hours (Figure 2c and d). In contrast, at 0 hours, there was no recovery of ST in the probiotic pre-treated group (EcN + ST), indicating the potential ability of the probiotic to restrict translocation of the pathogen from the gut into the hemolymph, particularly during the early stages of infection. Approximately 10² CFU/mg, were detected in the hemolymph of treated larvae at 4 hours post-infection (Figure 3e), but this did not lead to larval melanisation or significant deaths by 24 hours. 10² CFU/mL is well below the threshold of 10⁴ CFU/mL of S. Typhimurium required for lethality as suggested by Bismuth et al. (2021). This could potentially also explain the absence of a significant difference in the health score for the probiotic pre-treated (EcN + ST) group compared to the controls (EcN + PBS and PBS + PBS) 24 hours post-2nd inoculation (Figure 2d).

The probiotic disrupts SiiE adhesin-mediated attachment of the pathogen to intestinal porcine enterocytes (Schierack et al. 2011). At the same time, EcN1917 F1C fimbrae promote adhesion to host cells, limiting availability of binding sites for S. Typhimurium (Hu et al. 2020). Similar responses could occur within the larvae as the midgut architecture of G. mellonella is similar to that of mammals (Emery et al. 2019). It has also been shown that EcN1917 is capable of enhancing the intestinal gut barrier integrity. EcN1917 positively influences mammalian tight junctions, decreasing the likelihood of intestinal barrier dysfunction to restrict movement of S. Typhimurium from the gut into the circulation; the larvae have structures known as septate junctions, which are analogous to mammalian tight junctions (Trebichavsky et al. 2010).

Extending from the findings of this study, further work investigating the larval health in terms of changes in expression of immune markers or antimicrobial compounds could be carried out to gain a more comprehensive understanding of the protective mechanisms in the larvae and potentially identify the cause of differences in responses observed via oral force-feeding and intra-hemocelic injection. This will also overcome the limitation of the current study, whereby larval health was monitored subjectively, based on visual observation of physical characteristics. A future study investigating the prophylactic potential of EcN1917 against multiple clinical strains of S. Typhimurium, with increased representation of the species, could provide a better understanding and help generalise the protective effects observed in this study. Overall, this study confirms the suitability of the G. mellonella infection model in the study of probiotics during preliminary research to decrease dependency on traditional models. Importantly, it emphasises the need for careful consideration of the route of administration to increase the relevance of generated data to other in vivo models.

Conclusion

This study revealed dose – and route-dependent effects of probiotic EcN1917 and pathogenic ST14028 in G. mellonella larvae. Injection with probiotics led to significant larval mortality and health decline at concentrations above 10⁵ CFU/mL, while force-feeding showed no adverse effects. The protection assays demonstrated enhanced survival with oral pre-treatment, contrasting the decline in survivability through injection. This work is the first probiotic protection study in G. mellonella larvae that has reported the direct comparison between outcomes from oral and intra-hemocelic inoculation. Future research could explore the possible mechanisms by which the probiotic enhances the intestinal gut barrier and resists pathogen colonisation, emphasising the importance of route and dose of probiotic G. mellonella larval administration.

Data accessibility

All supporting data can be accessed from the respective figures and tables. Other relevant digital research materials have been listed in the references list.

All statistical analyses were carried out using the GraphPad Prism 8 by GraphPad Software, La Jolla, USA.

Ethical statement

The use of the G. mellonella invertebrate animal model does not require any ethical approval.

Acknowledgments

This study was supported by the School of Science, Monash University Malaysia.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by School of Science, Monash University Malaysia.