Galleria mellonella as a novel invertebrate model for studying Ehrlichia ruminantium pathogenesis and host–pathogen interactions

Abstract

Ehrlichia ruminantium, the causative agent of heartwater disease, is an obligate intracellular bacterium that poses significant economic threats to livestock production in endemic regions. Current research models present substantial ethical, logistical, and economic constraints, particularly for studying host–pathogen interactions within arthropod vectors. Here, we establish Galleria mellonella larvae as a tractable invertebrate infection model for Ehrlichia ruminantium, enabling experimental investigation of pathogen persistence and host–pathogen interactions in an arthropod system. Following infection, G. mellonella proved susceptible to E. ruminantium with moderate mortality and remarkable bacterial persistence. Using rhodamine-labeled bacteria and fluorescence microscopy, we tracked bacterial dissemination from injection sites to systemic distribution in characteristic segmental patterns throughout the larval body. Rhodamine-labeled bacteria were consistently associated with hemocytes, the primary immune cells of G. mellonella, although widefield imaging cannot establish intracellular localization. Quantitative PCR revealed stable bacterial loads over the study period, and bacteria recovered from infected larvae remained viable and infectious for endothelial cells. These findings show that E. ruminantium persists in G. mellonella for at least two weeks without inducing a detectable melanization response. The fluorescence adopted a reproducible segmental distribution whose biological basis remains to be determined. This model offers significant advantages: ethical acceptability, cost-effectiveness, experimental tractability, and compatibility with high-throughput screening approaches. The G. mellonella system represents a valuable complement to existing mammalian models and provides a unique platform for investigating arthropod-specific aspects of E. ruminantium biology, screening antimicrobial compounds, and testing hypotheses of immune evasion that may inform strategies for heartwater disease control.

We establish Galleria mellonella as a tractable invertebrate infection model demonstrating Ehrlichia ruminantium intracellular persistence within hemocytes and characteristic segmental dissemination patterns that recapitulate key aspects of heartwater pathogenesis.

Introduction

Ehrlichia ruminantium is a Gram-negative, obligate intracellular bacterium belonging to the family Anaplasmataceae (order Rickettsiales) that causes heartwater disease, one of the most economically important tick-borne diseases affecting ruminants worldwide (Thomas 2016, Rodrigues 2024). This severe and often fatal disease affects domestic and wild ruminants including cattle, sheep, goats, deer, giraffes, and camelids across sub-Saharan Africa, the Indian Ocean region, and the Caribbean (Guadeloupe and Antigua), with ongoing threats of expansion to the American mainland (Allsopp 2015, Marcelino et al. 2016, Dufleit et al. 2026).

The pathogen exhibits a complex biphasic developmental cycle, alternating between infectious extracellular elementary bodies and intracellular reticulate bodies (Meyer et al. 2023). Transmission occurs primarily through Amblyomma ticks during blood feeding, after which E. ruminantium replicates in regional lymph nodes before disseminating via the bloodstream to vascular endothelial cells (Rodrigues 2024). Clinical manifestations include neurological disorders, respiratory distress, hyperesthesia, and the pathognomonic hydropericardium that gives the disease its name (Allsopp 2015, World Organisation for Animal Health 2018).

Despite significant economic impact on livestock production, fundamental aspects of E. ruminantium pathogenesis, immune evasion mechanisms, and host–pathogen interactions remain poorly understood, particularly within arthropod vectors (Moumène and Meyer 2016). Traditional research approaches have relied predominantly on mammalian models, including laboratory mice and natural hosts (Brayton et al. 2003, Simbi et al. 2006, Seakamela et al. 2024). While these models have provided valuable insights, they present substantial limitations including ethical considerations, high costs, logistical complexity, and restricted applicability to arthropod-specific research questions.

The need for alternative experimental models is particularly acute given the limited understanding of E. ruminantium biology within tick vectors, where the pathogen must survive and potentially replicate before transmission to vertebrate hosts. Recent advances in alternative model systems have highlighted the utility of invertebrate models for studying intracellular bacterial pathogens (Asai et al. 2023, Villani et al. 2025).

Galleria mellonella, the greater wax moth larva, has emerged as a powerful, versatile, and cost-effective model organism for studying host–pathogen interactions (Ménard et al. 2021, Villani et al. 2025). This invertebrate system offers several compelling advantages: (i) a functional innate immune system with remarkable similarities to vertebrate immunity, including cellular and humoral responses; (ii) tolerance to mammalian physiological temperatures (37°C and higher), enabling optimal expression of bacterial virulence factors; (iii) ease of handling and manipulation; (iv) ethical acceptability and cost-effectiveness; and (v) compatibility with high-throughput experimental approaches (Kavanagh and Reeves 2004, Ménard et al. 2021).

Galleria mellonella has been successfully employed to investigate diverse intracellular bacterial pathogens, including Legionella pneumophila, Coxiella burnetii, and Brucella species, yielding important insights into immune evasion strategies and intracellular persistence mechanisms (Barquero-Calvo et al. 2007, Harding et al. 2013, Norville et al. 2014, Elizalde-Bielsa et al. 2023). These studies have demonstrated the model’s capacity to recapitulate key aspects of pathogen-host interactions observed in mammalian systems while providing unique insights into arthropod-specific immune responses.

To our knowledge, no previous studies have investigated E. ruminantium infection in G. mellonella or established this system as a model for heartwater research. As summarized in Fig. 1, this study aims to: (i) establish G. mellonella as a viable infection model for E. ruminantium; (ii) characterize bacterial localization, dissemination, and persistence within the larval host; (iii) document the macroscopic host response and establish the basis for future investigation of immune evasion; and (iv) evaluate the model’s potential for advancing our understanding of E. ruminantium biology and developing new therapeutic interventions.

Figure 1.

Experimental timeline for the 15-day longitudinal study of Galleria mellonella infection with Ehrlichia ruminantium. Larvae were inoculated via proleg injection with 10⁵ E. ruminantium elementary bodies per larva at day 0. Survival assessment was conducted at 15 time points throughout the study period. Samples for nucleic acid isolation and qPCR quantification were collected at six time points (days 1, 3, 5, 7, 9, and 15 post-infection). Bacterial cultivability was assessed at six time points (4 hours, and days 1, 2, 3, 4, and 5 post-infection). Visualization experiments using rhodamine and Hoechst staining were performed at seven time points (4 hours, and days 1, 2, 4, 5, 6, and 8 post-infection). DPI, days post-infection.

Methods

Bacterial strains and culture conditions

Bovine aortic endothelial cells (BAEC) were maintained in High Glucose Dulbecco’s modified Eagle’s medium (Gibco, ThermoFisher) supplemented with 10% fetal bovine serum (Gibco, ThermoFisher), 2 mM L-glutamine (Gibco, ThermoFisher), 100 IU/mL penicillin, and 100 μg/mL streptomycin (LifeTechnologies, ThermoFisher). Cells were cultured in 25 cm² tissue culture flasks as monolayers at 37°C in a humidified atmosphere containing 5% CO₂. BAEC were used up to passage 40 to maintain consistency.

Synchronous E. ruminantium infection was established using bacterial suspensions (80–90% infectivity) harvested at 120 hours post-infection (HPI) by mechanical disruption of infected cells through 18 G and 26 G needles, followed by reinfection at a 1:20 (v/v) ratio. Culture medium was completely replaced at 24 HPI, with half-medium changes performed at 96 HPI to maintain optimal growth conditions.

Galleria  mellonella infection and survival analysis

Elementary bodies were purified by differential centrifugation: initial centrifugation at 300 × g for 7 min to remove cellular debris, followed by bacterial concentration at 20 000 × g for 20 min. Bacterial pellets were resuspended in sterile phosphate-buffered saline (PBS), and enumeration was performed using flow cytometry (Accuri C6 Plus, BD Biosciences).

Galleria mellonella last instar larvae (200–250 mg weight) were obtained from MICALIS (INRAE, Jouy en Josas, France) and maintained on bees wax and pollen, at room temperature, until use. Groups of 10 larvae were injected with 10 μL containing 10⁵ purified elementary bodies per larva into the uppermost right proleg using a sterile Hamilton syringe. Larvae were incubated at 37°C in darkness without food, with survival monitored at 24-hour intervals for 15 days. Death was determined by absence of movement response to gentle mechanical stimulation. Appropriate controls included PBS-injected and untreated groups. Larvae were examined macroscopically at each 24 h assessment for cuticular darkening indicative of melanization, from 24 HPI to the end of the experiment; melanization was scored as present or absent, and no quantitative melanization index was determined. All experiments were carried out in triplicate (n = 30).

Bacterial visualization and tracking

For in vivo bacterial visualization, Ehrlichia ruminantium elementary bodies were labeled with rhodamine B-conjugated polymyxin B (Sigma–Aldrich) at 1.1 μg/mL final concentration, incubated for 30 min in darkness, and washed with PBS. Larvae were inoculated with 10⁵ labeled bacteria and observed from 4 HPI to 120 HPI using stereomicroscopy (M205FA, Leica). Individual larvae were immobilized on ice for 10 min before observation to prevent movement artifacts. Controls comprised PBS-injected larvae, larvae injected with unlabeled bacteria, and untreated larvae. A free-dye control (rhodamine B-polymyxin B without bacteria) was not performed; such a control would be required to exclude formally any contribution of unbound dye to the observed distribution.

Hemocyte analysis

For hemocyte analysis, larvae were infected with 10⁵ rhodamine-labeled bacteria as described above. Ten microliters hemolymph was collected by aseptically severing the bottom 2 mm of larvae and draining fluid into sterile microcentrifuge tubes. Samples were stained with 100 μL of 10 μg/mL Hoechst 33 342 (Sigma–Aldrich) for 15–30 min at room temperature before microscopic examination using a fluorescent cell imager (ZOE, BioRad). Observations were done directly on hemolymph without any anticoagulation medium and were conducted from 4 HPI to 120 HPI to track bacterial persistence. The selected images are representative of triplicated experiments.

Bacterial viability assessment

To confirm viability and infectivity of E. ruminantium during larval infection, 10 μL of hemolymph was collected at 4 HPI and at 1, 2, 3, 4 and 5 DPI from infected G. mellonella and directly inoculated onto BAEC monolayers in T25 flasks. Cultures were maintained under standard conditions with regular medium changes, and bacterial viability was assessed by monitoring morula formation and lysis synchrony characteristic of E. ruminantium infection. This assay samples total hemolymph and is qualitative: it does not establish that viable bacteria are those associated with hemocytes, nor does it quantify viable bacteria over time.

Bacterial quantification by qPCR

Bacterial loads onto larvae were quantified using real-time PCR targeting the pCS20 region (SOL-1). Relative quantification was calculated using the 2^(-ΔΔCt) method with eukaryotic 18S rRNA as the reference gene, normalizing bacterial DNA to total host DNA within each larva. Quantification is therefore of relative bacterial burden, not of absolute bacterial number per larva. TaqMan probes were labeled with FAM reporter dye (5') and BHQ-1 quencher (3'). Primer and probe sequences are provided in Table 1.

Table 1.

Primers and probes used for qPCR quantification of E. ruminantium.

Oligo name	5'—Oligonucleotide sequence–3'
18S Forward	TGTCTCAAAGATTAAGCCATGCA
18S Reverse	CCAAAGGAACCATAACTGATTTAATG
18S Probe	FAM-TTTCACTGTACCGGCCGTGCGTACTT-BHQ1
SOL-1 Forward	ACAAATCTGGYCCAGATCAC
SOL-1 Reverse	CAGCTTTCTGTTCAGCTAGT
SOL-1 Probe	FAM-ATCAATTCACATGAAACATTACATGCAACTGG-BHQ1

Nucleic acid extraction was performed using the IDEAL-32™ system with ID Gene^® Mag Fast Extraction Kit (ID Solutions). Individual larvae were surface-sterilized with ethanol, placed in 2 mL SafeLock tubes (Eppendorf) with 3 mm steel beads, and homogenized using TissueLyser II (Qiagen) at 30 Hz for 2 min. Following addition of 340 μL LysFast reagent and 2-min vortexing, samples were processed according to manufacturer specifications.

Real-time PCR reactions contained 2 μL template DNA, forward primer (250 nM), reverse primer (250 nM), probe (200 nM), and TaqMan Universal PCR Master Mix (ThermoFisher). Cycling conditions comprised initial denaturation at 50°C for 2 min, followed by 95°C for 10 min, then 40 cycles of 95°C for 15 s and 55°C for 1 min.

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 10 (GraphPad Software, San Diego, CA, USA). Kaplan–Meier survival curves were analyzed using the log-rank (Mantel–Cox) test. Quantitative PCR data were analyzed using one-way ANOVA with Tukey’s multiple comparison test. Results were considered statistically significant at P < 0.05.

Results

Galleria mellonella demonstrates susceptibility to E. ruminantium infection with characteristic delayed mortality

To establish G. mellonella as an infection model for E. ruminantium, larvae were challenged with the Gardel strain at 10⁵ bacteria per larva, a dose selected in preliminary experiments as giving ~50% survival at 15 DPI and thus permitting longitudinal sampling throughout the observation period, and monitored for 15 days post-infection (Fig. 1). Infected larvae exhibited a delayed and moderate mortality, with a final survival rate of 50%, whereas PBS-injected controls showed 100% survival (P = 0.047, log-rank (Mantel–Cox) test) (Fig. 2A). Monitoring was discontinued after the second death in the control group, and individuals that pupated during the experiment were censored from the survival analysis. Because pupation removes an animal from observation for reasons unrelated to death, and because infection could in principle alter the propensity or timing of pupation, this censoring may bias the survival estimate. This mortality pattern contrasted markedly with the rapid, high-level mortality typically observed with highly pathogenic bacteria such as Klebsiella pneumoniae. Virulence was assessed at a single inoculum, so this profile describes the outcome obtained at 10⁵ bacteria per larva rather than an intrinsic property of E. ruminantium.

Figure 2.

Ehrlichia ruminantium induces delayed mortality in G. mellonella with characteristic macroscopic changes. (A) Kaplan–Meier survival curves comparing larvae (n = 10 per group) inoculated with 10⁵ E. ruminantium bacteria per larva versus PBS-injected controls. Data represent means of three independent replicates. Statistical significance was determined using the log-rank (Mantel–Cox) test (*P < 0.05). (B) Representative macroscopic changes in G. mellonella larvae at 0 and 15 days post-infection, showing reduced size and altered morphology in infected larvae compared to healthy controls. Scale bar, 5 mm.

Macroscopic examination revealed progressive changes in infected larvae, including reduced motility and size reduction over the infection period (Fig. 2B), while control groups remained healthy throughout the study. No cuticular darkening indicative of melanization (color change from cream to dark gray/black) was recorded at any 24 h assessment, including during the first 48 h after inoculation, when this response is normally maximal; representative images at 0 and 15 DPI are shown in Fig. 2B. In the absence of a positive control confirming that these larvae were competent to melanize under identical conditions, and of a quantitative measurement of prophenoloxidase activation, this observation cannot be attributed to suppression of the melanization response by E. ruminantium.

Ehrlichia  ruminantium exhibits systemic dissemination and persistent localization within G. mellonella

To track bacterial localization and dissemination, E. ruminantium was labeled with rhodamine B-polymyxin B, achieving 97.9% labeling efficiency as confirmed by flow cytometry (Fig. 3A). This high labeling efficiency ensured reliable tracking of bacterial distribution throughout the infection period.

Figure 3.

Temporal localization and systemic dissemination of E. ruminantium in G. mellonella. (A) Flow cytometric validation of rhodamine labeling efficiency. Unlabeled E. ruminantium control showing minimal autofluorescence (0.2% positive events). Rhodamine-labeled E. ruminantium demonstrating 97.9% labeling with high intensity (high MFI). (B) Control groups showing absence of fluorescence in PBS-injected and unlabeled bacteria-injected larvae. (C) Time-course imaging of bacterial dissemination in larvae infected with rhodamine-labeled E. ruminantium (10⁷ bacteria/mL). Individual larvae were tracked by stereomicroscopy from 3 to 120 hours post-infection, showing progressive bacterial spread from the injection site to systemic segmental distribution throughout the larval body. Note the characteristic band-like pattern suggesting compartmentalized bacterial accumulation. HPI, hours post-infection.

Following injection of 10⁵ labeled bacteria into the proleg, fluorescence microscopy revealed dynamic bacterial dissemination patterns (Fig. 3B, C). Initially, fluorescence was localized exclusively at the injection site (3 HPI). By 24 HPI, bacterial fluorescence appeared in the dorsal region, indicating successful systemic dissemination. From 48 HPI onwards, the fluorescence adopted a repetitive, band-like distribution whose periodicity corresponded approximately to larval segmentation. This distribution intensified progressively through 96–120 HPI, with additional signal observed in the rectal region at the later time points.

This pattern was observed in all infected larvae examined up to 5 days post-infection. Its biological basis was not determined in this study, and we note that the whole-larva images report the distribution of the fluorescent label rather than that of viable bacteria.

Ehrlichia  ruminantium forms a long-lasting association with G. mellonella hemocytes

Fluorescence microscopy of hemolymph from infected larvae showed rhodamine signal spatially coincident with Hoechst-stained hemocyte nuclei, with hemocyte morphology unaltered relative to uninfected controls; no rhodamine signal was detected in uninfected controls (Fig. 4). This spatial coincidence is consistent with intracellular localization but does not establish it, since widefield imaging without optical sectioning cannot distinguish internalized bacteria from bacteria adherent to the hemocyte surface. We therefore refer throughout to hemocyte-associated bacteria.

Figure 4.

Association of E. ruminantium with G. mellonella hemocytes. Larvae were infected with rhodamine-labeled E. ruminantium (10⁷ bacteria/mL), and hemolymph was collected for fluorescence microscopy analysis. Hemocyte nuclei were stained with Hoechst 33 342 (blue), and bacteria appear as rhodamine fluorescence (red). Brightfield images show normal hemocyte morphology in both uninfected and infected samples. The merged images show rhodamine signal spatially coincident with hemocytes (white dashed box and inset, representative hemocyte-associated bacteria in red). Uninfected controls show no rhodamine signal. This spatial coincidence is consistent with intracellular localization; confocal z-stack acquisition would be required to confirm that and distinguish internalized from surface-adherent bacteria. Scale bar, 25 μm.

Hemocyte-associated bacteria remained detectable for at least 6 days post-infection (Fig. 5A), and whole-larva fluorescence persisted for at least 8 days (Fig. 5B), with no progressive loss of signal and no overt hemocyte destruction.

Figure 5.

Long-term persistence of E. ruminantium in G. mellonella and viability of recovered bacteria. (A) Fluorescence microscopy of hemolymph collected 6 days post-infection, showing sustained hemocyte-associated bacteria. Nuclei are stained with Hoechst (blue) and bacteria with rhodamine (red). Scale bar, 30 μm. (B) Stereomicroscopy demonstrating whole-larva bacterial fluorescence persistence at 8 days post-infection. Scale bar, 5 mm. (C, D) Viability confirmation through re-infection assay. Hemolymph collected from infected G. mellonella at 4 HPI and at 1–5 DPI was inoculated onto bovine aortic endothelial cell (BAEC) monolayers. (C) Brightfield image at 4 days post-inoculation showing characteristic morulae and synchronized lysis areas. (D) Corresponding fluorescence image showing rhodamine-labeled bacteria within BAEC. Together these confirm that bacteria recovered from the hemolymph remain viable and infectious. Scale bars, 100 μm. N, nuclei; DPI, days post-infection.

Bacterial viability and infectivity are maintained during G. mellonella infection

To determine whether bacteria recovered from infected larvae remained viable, hemolymph collected at 4 HPI and at 1, 2, 3, 4, and 5 DPI was inoculated onto BAEC monolayers. At every time point, characteristic morulae and synchronized lysis areas typical of active E. ruminantium infection were observed at 4 days post-inoculation (Fig. 5C), and rhodamine-labeled bacteria were detected within BAEC (Fig. 5D). Bacteria present in the hemolymph during the first 5 days of infection are therefore viable and infectious, and rhodamine B-polymyxin B labeling does not detectably impair infectivity. This assay samples total hemolymph and is qualitative: it does not establish that the viable bacteria are those associated with hemocytes, nor does it demonstrate replication within larval tissues.

Ehrlichia ruminantium establishes stable, persistent infection with minimal bacterial clearance

Bacterial DNA remained readily detectable throughout the 15-day observation period, with a variation of ~0.5 log and no significant difference between time points (P > 0.05, one-way ANOVA) (Fig. 6). Ehrlichia ruminantium therefore persists in G. mellonella without net clearance by the host.

Figure 6.

Ehrlichia ruminantium establishes stable, persistent infection in G. mellonella. Bacterial loads were quantified by the SOL1 qPCR targeting the pCS20 gene in larvae (n = 3 larvae per time point) inoculated with 10⁵ bacteria per larva. Samples were collected at indicated time points and bacterial enumeration assessed using the 2^−ΔΔCt method with eukaryotic 18S rRNA as reference gene, normalizing bacterial DNA to total host DNA within each larva; values therefore represent relative bacterial burden, not absolute bacterial number per larva. Violin plots show individual data points and distribution. Bacterial loads remained stable throughout the 15-day observation period with minimal variation (P > 0.05, one-way ANOVA), indicating persistence without net bacterial clearance. DPI, days post-infection.

A stable signal does not, however, distinguish between active intracellular replication balanced by clearance, slow-turnover persistence of non-replicating bacteria, and slow degradation of the inoculum. Bacterial loads were quantified in three larvae per time point, which is sufficient to establish that the load is maintained without significant variation but does not support finer comparisons between individual time points.

Discussion

This study represents the first successful establishment of G. mellonella as an infection model for E. ruminantium, providing a novel platform for investigating host–pathogen interactions in this economically important tick-borne disease. Our findings show that E. ruminantium infects G. mellonella, persists there for at least 2 weeks, remains viable and infectious, and associates with hemocytes without inducing a detectable melanization response. These are descriptive observations; the mechanisms underlying them remain to be established.

The observations reported here, namely bacterial persistence, hemocyte association, absence of visible melanization, and stable bacterial DNA levels, are compatible with active immune evasion, but they do not demonstrate it. The contribution of this study is to establish a system in which such hypotheses can be tested directly: through infection with T4SS-defective mutants, transcriptomic profiling of infected hemocytes, quantitative measurement of prophenoloxidase activation, and gentamicin-protection assays coupled to hemocyte isolation.

Advantages and applications of the G. mellonella model for E. ruminantium research

The G. mellonella model offers several significant advantages over traditional mammalian systems for E. ruminantium research. First, the ethical acceptability and cost-effectiveness of this invertebrate model enable high-throughput experimental approaches that would be prohibitive with mammalian models. Second, the ability to maintain larvae at 37°C ensures optimal expression of bacterial virulence factors that are temperature-regulated and critical for pathogenesis in mammalian hosts. Third, the experimental tractability allows controlled dose administration and precise temporal sampling that facilitates mechanistic studies of infection dynamics.

Perhaps most importantly, this model provides a bridge between mammalian and arthropod research contexts. While G. mellonella and Amblyomma ticks differ substantially in biology and ecology, both are arthropods with innate immune systems that share fundamental features including cellular immunity, humoral responses, and antimicrobial peptide production (Kavanagh and Reeves 2004, Ribeiro and Brehélin 2006). This model therefore offers opportunities to investigate arthropod-specific aspects of E. ruminantium biology that cannot be studied in mammalian systems, while providing insights that may be relevant to understanding some pathogen–tick interactions aspects.

Hypotheses for persistence and reduced immune activation

The delayed, moderate mortality observed in E. ruminantium-infected larvae contrasts sharply with the rapid, high-level mortality induced by typical Gram-negative pathogens such as Klebsiella pneumoniae (Elizalde-Bielsa et al. 2023). This pattern, together with the absence of visible melanization, is compatible with a “stealth” infection strategy of the kind described for Brucella species (Barquero-Calvo et al. 2007), although our data do not test it. Melanization is a central component of invertebrate immunity, involving prophenoloxidase activation and pathogen encapsulation (Cerenius et al. 2008); its apparent absence here is therefore worth explaining, but any explanation requires a positive control confirming melanization competence in these larvae and a quantitative assay of prophenoloxidase activation.

Several mechanisms may contribute to E. ruminantium immune evasion in G. mellonella. The type IV secretion system (T4SS) VirB/D4, which is conserved across Ehrlichia species, translocates bacterial effectors into host cell cytoplasm, subverting cellular pathways and immune responses (Celli et al. 2003, Rikihisa 2017). Studies with Legionella pneumophila have demonstrated that T4SS mutants show reduced virulence and impaired intracellular persistence in G. mellonella (Harding et al. 2012), suggesting that similar mechanisms may operate in E. ruminantium infections.

Additionally, the absence of canonical lipopolysaccharide (LPS), a major pathogen-associated molecular pattern in many Gram-negative bacteria, may contribute to immune evasion by reducing pathogen-associated molecular pattern (PAMP) recognition. In Ehrlichia and other members of the Anaplasmataceae, the lack of canonical LPS and peptidoglycan is thought to reduce innate immune detection and favor intracellular survival (McBride et al. 2011). In our model, the limited mortality and absence of visible melanization are consistent with this hypothesis, although the underlying mechanisms remain to be determined.

Bacterial dissemination and cellular targeting strategies

The segmental distribution of fluorescence is reproducible and its periodicity is suggestive, but our data do not identify its origin. Candidate explanations include the segmental arrangement of sessile hemocyte clusters, which organize along body-wall and dorsal-vessel landmarks in other insects (King and Hillyer 2013, Leitão and Sucena 2015); trafficking or entrapment along the tracheal system, itself segmentally organized; hemolymph circulation dynamics; and accumulation of the label in particular tissues independently of bacterial distribution. Distinguishing between these requires histological sectioning with hemocyte-specific labeling, confocal reconstruction of the fluorescence in situ, and comparison with the distribution of free dye. We report the pattern as a reproducible observation that warrants explanation, not as evidence of a dissemination strategy.

The association of E. ruminantium with G. mellonella hemocytes is notable in light of recent work in a natural vector: Ehrlichia chaffeensis co-opts phagocytic hemocytes for systemic dissemination in Amblyomma americanum (Adegoke et al. 2024). Hemocyte association may therefore be a recurrent feature of Ehrlichia biology in arthropods rather than a peculiarity of the G. mellonella system. Establishing whether hemocyte-associated bacteria are themselves viable, and whether they replicate, will require gentamicin protection to eliminate the extracellular population, followed by hemocyte isolation and quantification of recovered infectious bacteria over time. The same question remains open in the tick vector, where it is not known whether E. ruminantium replicates or merely survives between blood meals (Moumène and Meyer 2016).

Future research directions and applications

The established G. mellonella model opens numerous avenues for advancing E. ruminantium research. Transcriptomic and proteomic analyses of infected larvae could identify host factors influencing susceptibility and bacterial effectors involved in immune modulation. Screening of mutant bacterial strains, particularly those defective in T4SS function, could elucidate specific virulence mechanisms and their contributions to persistence.

The model’s compatibility with high-throughput approaches makes it particularly valuable for antimicrobial compound screening and vaccine development research. The ability to rapidly assess large numbers of potential therapeutic interventions in a cost-effective, ethically acceptable system could accelerate the development of new strategies for heartwater control. Furthermore, comparative studies examining E. ruminantium behavior in G. mellonella versus other invertebrate models or arthropod cell cultures could provide insights into conserved versus host-specific aspects of pathogenesis. Such studies may ultimately inform our understanding of E. ruminantium interactions with natural tick vectors.

Model limitations and validation considerations

While the G. mellonella model offers significant advantages, important limitations must be acknowledged. The substantial evolutionary distance between lepidopteran larvae and natural ruminant or tick hosts necessitates careful validation of findings in established mammalian models or tick systems. Additionally, the simplified immune system of G. mellonella lacks adaptive immunity components that may be crucial for certain aspects of heartwater pathogenesis.

The differences in life cycles, feeding habits, and environmental conditions between G. mellonella and Amblyomma ticks limit direct extrapolation of results to tick biology. However, the model’s value lies not in replacing existing systems but in complementing them by providing accessible experimental approaches for mechanistic studies that would be challenging or impossible in other systems.

Several controls that would strengthen the interpretations offered here were not performed. A heat-inactivated inoculum would establish formally that fluorescence persistence requires bacterial viability; we note that killed bacteria and cellular debris are cleared efficiently from the insect hemocoel by phagocytosis, nodulation and enzymatic degradation, and would therefore be expected to yield a progressively declining signal, whereas the signal observed here was stable to at least 8 DPI and the bacteria recovered from hemolymph remained infectious. A free-dye control would exclude any contribution of unbound rhodamine B-polymyxin B to the observed distribution. A strongly melanizing control organism would confirm that the larvae used were competent to mount the response whose absence we report. Higher-resolution imaging, in particular confocal acquisition with optical sectioning, would be required to determine whether hemocyte-associated bacteria are internalized.

Finally, virulence was assessed at a single inoculum; a dose-response series remains necessary to characterize it and to determine whether the persistence phenotype is maintained across inocula.

Conclusions

This study establishes G. mellonella as the first tractable invertebrate model for E. ruminantium infection, and documents persistence of viable bacteria in association with hemocytes together with a striking segmental distribution of the fluorescent label in an arthropod host. The model offers practical advantages for mechanistic studies, compound screening, and high-throughput experimental approaches.

The G. mellonella system represents a significant addition to the toolkit available for E. ruminantium research, offering opportunities to accelerate discovery of new therapeutic targets and interventions for heartwater disease control. While this model should complement rather than replace existing systems, its accessibility and experimental tractability position it as a valuable experimental platform for advancing our understanding of this important veterinary pathogen.

Author Contributions

Maëlle Bayet (Investigation, Visualization, Writing – original draft & editing), Christina Nielsen-Leroux (Methodology, Resources, Validation, Writing – review & editing), Valérie Rodrigues (Data curation, Investigation cytometry data acquisition, Validation, Writing – review & editing, Project administration), and Damien F. Meyer (Conceptualization, Methodology, Data curation, Validation, Visualization, Writing – review & editing, Supervision, Funding acquisition, Project administration).

Conflicts of interest

The authors declare no financial or commercial conflicts of interest related to this research.

Funding

This research was supported by the United States Department of Agriculture grant 58-3022-1-018-F (Risk of Arthropod-borne diseases in the Caribbean) and Région Guadeloupe grant FEDER Guadeloupe Une Santé.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.